mirror of
https://github.com/exo-explore/exo.git
synced 2026-02-12 07:01:06 -05:00
Compare commits
1 Commits
alexcheema
...
rust-explo
| Author | SHA1 | Date | |
|---|---|---|---|
|
49df704dcd |
171
Cargo.lock
generated
171
Cargo.lock
generated
@@ -141,12 +141,6 @@ version = "0.3.9"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "76a2e8124351fda1ef8aaaa3bbd7ebbcb486bbcd4225aca0aa0d84bb2db8fecb"
|
||||
|
||||
[[package]]
|
||||
name = "arrayvec"
|
||||
version = "0.7.6"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "7c02d123df017efcdfbd739ef81735b36c5ba83ec3c59c80a9d7ecc718f92e50"
|
||||
|
||||
[[package]]
|
||||
name = "asn1-rs"
|
||||
version = "0.7.1"
|
||||
@@ -304,19 +298,6 @@ version = "1.8.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "55248b47b0caf0546f7988906588779981c43bb1bc9d0c44087278f80cdb44ba"
|
||||
|
||||
[[package]]
|
||||
name = "bigdecimal"
|
||||
version = "0.4.9"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "560f42649de9fa436b73517378a147ec21f6c997a546581df4b4b31677828934"
|
||||
dependencies = [
|
||||
"autocfg",
|
||||
"libm",
|
||||
"num-bigint",
|
||||
"num-integer",
|
||||
"num-traits",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "bimap"
|
||||
version = "0.6.3"
|
||||
@@ -516,15 +497,6 @@ version = "0.4.3"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "2f421161cb492475f1661ddc9815a745a1c894592070661180fdec3d4872e9c3"
|
||||
|
||||
[[package]]
|
||||
name = "convert_case"
|
||||
version = "0.10.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "633458d4ef8c78b72454de2d54fd6ab2e60f9e02be22f3c6104cdc8a4e0fceb9"
|
||||
dependencies = [
|
||||
"unicode-segmentation",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "core-foundation"
|
||||
version = "0.9.4"
|
||||
@@ -701,17 +673,6 @@ dependencies = [
|
||||
"syn 2.0.111",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "delegate"
|
||||
version = "0.13.5"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "780eb241654bf097afb00fc5f054a09b687dad862e485fdcf8399bb056565370"
|
||||
dependencies = [
|
||||
"proc-macro2",
|
||||
"quote",
|
||||
"syn 2.0.111",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "der"
|
||||
version = "0.7.10"
|
||||
@@ -746,29 +707,6 @@ dependencies = [
|
||||
"powerfmt",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "derive_more"
|
||||
version = "2.1.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "10b768e943bed7bf2cab53df09f4bc34bfd217cdb57d971e769874c9a6710618"
|
||||
dependencies = [
|
||||
"derive_more-impl",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "derive_more-impl"
|
||||
version = "2.1.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "6d286bfdaf75e988b4a78e013ecd79c581e06399ab53fbacd2d916c2f904f30b"
|
||||
dependencies = [
|
||||
"convert_case",
|
||||
"proc-macro2",
|
||||
"quote",
|
||||
"rustc_version",
|
||||
"syn 2.0.111",
|
||||
"unicode-xid",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "digest"
|
||||
version = "0.10.7"
|
||||
@@ -938,37 +876,23 @@ dependencies = [
|
||||
name = "exo_pyo3_bindings"
|
||||
version = "0.0.1"
|
||||
dependencies = [
|
||||
"delegate",
|
||||
"derive_more",
|
||||
"env_logger",
|
||||
"extend",
|
||||
"futures",
|
||||
"impl-trait-for-tuples",
|
||||
"futures-lite",
|
||||
"libp2p",
|
||||
"log",
|
||||
"networking",
|
||||
"once_cell",
|
||||
"pin-project",
|
||||
"pyo3",
|
||||
"pyo3-async-runtimes",
|
||||
"pyo3-log",
|
||||
"pyo3-stub-gen",
|
||||
"thiserror 2.0.17",
|
||||
"thread_local",
|
||||
"tokio",
|
||||
"util",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "extend"
|
||||
version = "1.2.0"
|
||||
name = "fastrand"
|
||||
version = "2.3.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "311a6d2f1f9d60bff73d2c78a0af97ed27f79672f15c238192a5bbb64db56d00"
|
||||
dependencies = [
|
||||
"proc-macro2",
|
||||
"quote",
|
||||
"syn 2.0.111",
|
||||
]
|
||||
checksum = "37909eebbb50d72f9059c3b6d82c0463f2ff062c9e95845c43a6c9c0355411be"
|
||||
|
||||
[[package]]
|
||||
name = "ff"
|
||||
@@ -1078,7 +1002,10 @@ version = "2.6.1"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "f78e10609fe0e0b3f4157ffab1876319b5b0db102a2c60dc4626306dc46b44ad"
|
||||
dependencies = [
|
||||
"fastrand",
|
||||
"futures-core",
|
||||
"futures-io",
|
||||
"parking",
|
||||
"pin-project-lite",
|
||||
]
|
||||
|
||||
@@ -1640,17 +1567,6 @@ dependencies = [
|
||||
"xmltree",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "impl-trait-for-tuples"
|
||||
version = "0.2.3"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "a0eb5a3343abf848c0984fe4604b2b105da9539376e24fc0a3b0007411ae4fd9"
|
||||
dependencies = [
|
||||
"proc-macro2",
|
||||
"quote",
|
||||
"syn 2.0.111",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "indexmap"
|
||||
version = "2.12.1"
|
||||
@@ -1805,12 +1721,6 @@ dependencies = [
|
||||
"cpufeatures",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "keccak-const"
|
||||
version = "0.2.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "57d8d8ce877200136358e0bbff3a77965875db3af755a11e1fa6b1b3e2df13ea"
|
||||
|
||||
[[package]]
|
||||
name = "lalrpop-util"
|
||||
version = "0.20.2"
|
||||
@@ -1829,12 +1739,6 @@ version = "0.2.178"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "37c93d8daa9d8a012fd8ab92f088405fb202ea0b6ab73ee2482ae66af4f42091"
|
||||
|
||||
[[package]]
|
||||
name = "libm"
|
||||
version = "0.2.15"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "f9fbbcab51052fe104eb5e5d351cf728d30a5be1fe14d9be8a3b097481fb97de"
|
||||
|
||||
[[package]]
|
||||
name = "libp2p"
|
||||
version = "0.56.0"
|
||||
@@ -2823,20 +2727,10 @@ dependencies = [
|
||||
name = "networking"
|
||||
version = "0.0.1"
|
||||
dependencies = [
|
||||
"delegate",
|
||||
"derive_more",
|
||||
"either",
|
||||
"extend",
|
||||
"futures",
|
||||
"futures-timer",
|
||||
"impl-trait-for-tuples",
|
||||
"keccak-const",
|
||||
"libp2p",
|
||||
"log",
|
||||
"thiserror 2.0.17",
|
||||
"tokio",
|
||||
"tracing-subscriber",
|
||||
"util",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
@@ -2918,17 +2812,6 @@ dependencies = [
|
||||
"num-traits",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "num-rational"
|
||||
version = "0.4.2"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "f83d14da390562dca69fc84082e73e548e1ad308d24accdedd2720017cb37824"
|
||||
dependencies = [
|
||||
"num-bigint",
|
||||
"num-integer",
|
||||
"num-traits",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "num-traits"
|
||||
version = "0.2.19"
|
||||
@@ -3279,28 +3162,14 @@ version = "0.27.2"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "ab53c047fcd1a1d2a8820fe84f05d6be69e9526be40cb03b73f86b6b03e6d87d"
|
||||
dependencies = [
|
||||
"bigdecimal",
|
||||
"either",
|
||||
"hashbrown 0.16.1",
|
||||
"indexmap",
|
||||
"indoc",
|
||||
"inventory",
|
||||
"libc",
|
||||
"lock_api",
|
||||
"memoffset",
|
||||
"num-bigint",
|
||||
"num-complex",
|
||||
"num-rational",
|
||||
"num-traits",
|
||||
"once_cell",
|
||||
"ordered-float",
|
||||
"parking_lot",
|
||||
"portable-atomic",
|
||||
"pyo3-build-config",
|
||||
"pyo3-ffi",
|
||||
"pyo3-macros",
|
||||
"rust_decimal",
|
||||
"smallvec",
|
||||
"unindent",
|
||||
]
|
||||
|
||||
@@ -3741,16 +3610,6 @@ dependencies = [
|
||||
"tokio",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "rust_decimal"
|
||||
version = "1.39.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "35affe401787a9bd846712274d97654355d21b2a2c092a3139aabe31e9022282"
|
||||
dependencies = [
|
||||
"arrayvec",
|
||||
"num-traits",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "rustc-hash"
|
||||
version = "1.1.0"
|
||||
@@ -4615,24 +4474,12 @@ version = "1.0.22"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "9312f7c4f6ff9069b165498234ce8be658059c6728633667c526e27dc2cf1df5"
|
||||
|
||||
[[package]]
|
||||
name = "unicode-segmentation"
|
||||
version = "1.12.0"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "f6ccf251212114b54433ec949fd6a7841275f9ada20dddd2f29e9ceea4501493"
|
||||
|
||||
[[package]]
|
||||
name = "unicode-width"
|
||||
version = "0.2.2"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "b4ac048d71ede7ee76d585517add45da530660ef4390e49b098733c6e897f254"
|
||||
|
||||
[[package]]
|
||||
name = "unicode-xid"
|
||||
version = "0.2.6"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "ebc1c04c71510c7f702b52b7c350734c9ff1295c464a03335b00bb84fc54f853"
|
||||
|
||||
[[package]]
|
||||
name = "unicode_names2"
|
||||
version = "1.3.0"
|
||||
@@ -4713,10 +4560,6 @@ version = "0.2.2"
|
||||
source = "registry+https://github.com/rust-lang/crates.io-index"
|
||||
checksum = "06abde3611657adf66d383f00b093d7faecc7fa57071cce2578660c9f1010821"
|
||||
|
||||
[[package]]
|
||||
name = "util"
|
||||
version = "0.0.1"
|
||||
|
||||
[[package]]
|
||||
name = "uuid"
|
||||
version = "1.19.0"
|
||||
|
||||
108
Cargo.toml
108
Cargo.toml
@@ -3,7 +3,6 @@ resolver = "3"
|
||||
members = [
|
||||
"rust/networking",
|
||||
"rust/exo_pyo3_bindings",
|
||||
"rust/util",
|
||||
]
|
||||
|
||||
[workspace.package]
|
||||
@@ -24,62 +23,18 @@ opt-level = 3
|
||||
[workspace.dependencies]
|
||||
## Crate members as common dependencies
|
||||
networking = { path = "rust/networking" }
|
||||
util = { path = "rust/util" }
|
||||
|
||||
# Proc-macro authoring tools
|
||||
syn = "2.0"
|
||||
quote = "1.0"
|
||||
proc-macro2 = "1.0"
|
||||
darling = "0.20"
|
||||
|
||||
# Macro dependecies
|
||||
extend = "1.2"
|
||||
delegate = "0.13"
|
||||
impl-trait-for-tuples = "0.2"
|
||||
clap = "4.5"
|
||||
derive_more = { version = "2.0.1", features = ["display"] }
|
||||
pin-project = "1"
|
||||
|
||||
# Utility dependencies
|
||||
itertools = "0.14"
|
||||
thiserror = "2"
|
||||
internment = "0.8"
|
||||
recursion = "0.5"
|
||||
regex = "1.11"
|
||||
once_cell = "1.21"
|
||||
thread_local = "1.1"
|
||||
bon = "3.4"
|
||||
generativity = "1.1"
|
||||
anyhow = "1.0"
|
||||
keccak-const = "0.2"
|
||||
|
||||
# Functional generics/lenses frameworks
|
||||
frunk_core = "0.4"
|
||||
frunk = "0.4"
|
||||
frunk_utils = "0.2"
|
||||
frunk-enum-core = "0.3"
|
||||
|
||||
# Async dependencies
|
||||
tokio = "1.46"
|
||||
futures = "0.3"
|
||||
futures-util = "0.3"
|
||||
futures-timer = "3.0"
|
||||
|
||||
# Data structures
|
||||
either = "1.15"
|
||||
ordered-float = "5.0"
|
||||
ahash = "0.8"
|
||||
|
||||
# Tracing/logging
|
||||
log = "0.4"
|
||||
|
||||
# networking
|
||||
libp2p = "0.56"
|
||||
libp2p-tcp = "0.44"
|
||||
|
||||
[workspace.lints.rust]
|
||||
static_mut_refs = "warn" # Or use "warn" instead of deny
|
||||
incomplete_features = "allow"
|
||||
static_mut_refs = "warn"
|
||||
|
||||
# Clippy's lint category level configurations;
|
||||
# every member crate needs to inherit these by adding
|
||||
@@ -100,64 +55,3 @@ perf = { level = "warn", priority = -1 }
|
||||
pedantic = { level = "warn", priority = -1 }
|
||||
nursery = { level = "warn", priority = -1 }
|
||||
cargo = { level = "warn", priority = -1 }
|
||||
|
||||
# Individual Clippy lints from the `restriction` category
|
||||
arithmetic_side_effects = "warn"
|
||||
as_conversions = "warn"
|
||||
assertions_on_result_states = "warn"
|
||||
clone_on_ref_ptr = "warn"
|
||||
decimal_literal_representation = "warn"
|
||||
default_union_representation = "warn"
|
||||
deref_by_slicing = "warn"
|
||||
disallowed_script_idents = "deny"
|
||||
else_if_without_else = "warn"
|
||||
empty_enum_variants_with_brackets = "warn"
|
||||
empty_structs_with_brackets = "warn"
|
||||
error_impl_error = "warn"
|
||||
exit = "deny"
|
||||
expect_used = "warn"
|
||||
float_cmp_const = "warn"
|
||||
get_unwrap = "warn"
|
||||
if_then_some_else_none = "warn"
|
||||
impl_trait_in_params = "warn"
|
||||
indexing_slicing = "warn"
|
||||
infinite_loop = "warn"
|
||||
let_underscore_must_use = "warn"
|
||||
let_underscore_untyped = "warn"
|
||||
lossy_float_literal = "warn"
|
||||
mem_forget = "warn"
|
||||
missing_inline_in_public_items = "warn"
|
||||
multiple_inherent_impl = "warn"
|
||||
multiple_unsafe_ops_per_block = "warn"
|
||||
mutex_atomic = "warn"
|
||||
non_zero_suggestions = "warn"
|
||||
panic = "warn"
|
||||
partial_pub_fields = "warn"
|
||||
pattern_type_mismatch = "warn"
|
||||
pub_without_shorthand = "warn"
|
||||
rc_buffer = "warn"
|
||||
rc_mutex = "warn"
|
||||
redundant_type_annotations = "warn"
|
||||
renamed_function_params = "warn"
|
||||
rest_pat_in_fully_bound_structs = "warn"
|
||||
same_name_method = "warn"
|
||||
self_named_module_files = "deny"
|
||||
semicolon_inside_block = "warn"
|
||||
shadow_same = "warn"
|
||||
shadow_unrelated = "warn"
|
||||
str_to_string = "warn"
|
||||
string_add = "warn"
|
||||
string_lit_chars_any = "warn"
|
||||
string_to_string = "warn"
|
||||
tests_outside_test_module = "warn"
|
||||
todo = "warn"
|
||||
try_err = "warn"
|
||||
undocumented_unsafe_blocks = "warn"
|
||||
unnecessary_safety_comment = "warn"
|
||||
unnecessary_safety_doc = "warn"
|
||||
unneeded_field_pattern = "warn"
|
||||
unseparated_literal_suffix = "warn"
|
||||
unused_result_ok = "warn"
|
||||
unused_trait_names = "warn"
|
||||
unwrap_used = "warn"
|
||||
verbose_file_reads = "warn"
|
||||
|
||||
@@ -1,5 +1,5 @@
|
||||
# Missed things
|
||||
[X] Log EXO_LIBP2P_NAMESPACE on start in exo/main.py
|
||||
[X] Log namespace on start in exo/main.py
|
||||
[X] Ordering of warmup was changed, which is wrong. It was changed to rank < n-1, then rank=n-1. It should be rank!=0 then rank=0 (this matches the auto_parallel implementation. NOTE: we use a different convention to mlx-lm, our terminal rank is rank=n-1 whereas mlx-lm is rank=0 hence i can see why this was changed wrongly).
|
||||
[X] Downloads keying by model_id not shard_metadata (worker/plan.py, worker/main.py).
|
||||
[X] Fetching download status of all models on start
|
||||
|
||||
@@ -199,14 +199,14 @@ The app will ask for permission to modify system settings and install a new Netw
|
||||
|
||||
**Custom Namespace for Cluster Isolation:**
|
||||
|
||||
The macOS app includes a custom namespace feature that allows you to isolate your exo cluster from others on the same network. This is configured through the `EXO_LIBP2P_NAMESPACE` setting:
|
||||
The macOS app includes a custom namespace feature that allows you to isolate your exo cluster from others on the same network. This is configured through the `--namespace` cli arg:
|
||||
|
||||
- **Use cases**:
|
||||
- Running multiple separate exo clusters on the same network
|
||||
- Isolating development/testing clusters from production clusters
|
||||
- Preventing accidental cluster joining
|
||||
|
||||
- **Configuration**: Access this setting in the app's Advanced settings (or set the `EXO_LIBP2P_NAMESPACE` environment variable when running from source)
|
||||
- **Configuration**: Access this setting in the app's Advanced settings (or set the `--namespace` argument when running from source)
|
||||
|
||||
The namespace is logged on startup for debugging purposes.
|
||||
|
||||
@@ -418,4 +418,4 @@ On macOS, exo uses the GPU. On Linux, exo currently runs on CPU. We are working
|
||||
|
||||
## Contributing
|
||||
|
||||
See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines on how to contribute to exo.
|
||||
See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines on how to contribute to exo.
|
||||
|
||||
@@ -82,6 +82,7 @@ final class ExoProcessController: ObservableObject {
|
||||
|
||||
let child = Process()
|
||||
child.executableURL = executableURL
|
||||
child.arguments = ["--namespace", computeNamespace()]
|
||||
let exoHomeURL = Self.exoDirectoryURL
|
||||
try? FileManager.default.createDirectory(
|
||||
at: exoHomeURL, withIntermediateDirectories: true
|
||||
@@ -216,7 +217,6 @@ final class ExoProcessController: ObservableObject {
|
||||
private func makeEnvironment(for runtimeURL: URL) -> [String: String] {
|
||||
var environment = ProcessInfo.processInfo.environment
|
||||
environment["EXO_RUNTIME_DIR"] = runtimeURL.path
|
||||
environment["EXO_LIBP2P_NAMESPACE"] = computeNamespace()
|
||||
if !hfToken.isEmpty {
|
||||
environment["HF_TOKEN"] = hfToken
|
||||
}
|
||||
|
||||
@@ -185,7 +185,11 @@
|
||||
|
||||
let instanceType: string | null = null;
|
||||
if (instanceTag === "MlxRingInstance") instanceType = "MLX Ring";
|
||||
else if (instanceTag === "MlxJacclInstance") instanceType = "MLX RDMA";
|
||||
else if (
|
||||
instanceTag === "MlxIbvInstance" ||
|
||||
instanceTag === "MlxJacclInstance"
|
||||
)
|
||||
instanceType = "MLX RDMA";
|
||||
|
||||
let sharding: string | null = null;
|
||||
const inst = instance as {
|
||||
|
||||
@@ -21,7 +21,7 @@
|
||||
} | null;
|
||||
nodes?: Record<string, NodeInfo>;
|
||||
sharding?: "Pipeline" | "Tensor";
|
||||
runtime?: "MlxRing" | "MlxJaccl";
|
||||
runtime?: "MlxRing" | "MlxIbv" | "MlxJaccl";
|
||||
onLaunch?: () => void;
|
||||
tags?: string[];
|
||||
apiPreview?: PlacementPreview | null;
|
||||
@@ -348,7 +348,7 @@
|
||||
// Debug mode state
|
||||
const isDebugMode = $derived(debugMode());
|
||||
const topology = $derived(topologyData());
|
||||
const isRdma = $derived(runtime === "MlxJaccl");
|
||||
const isRdma = $derived(runtime === "MlxIbv" || runtime === "MlxJaccl");
|
||||
|
||||
// Get interface name for an IP from node data
|
||||
function getInterfaceForIp(nodeId: string, ip?: string): string | null {
|
||||
@@ -575,7 +575,7 @@
|
||||
>
|
||||
{runtime === "MlxRing"
|
||||
? "MLX Ring"
|
||||
: runtime === "MlxJaccl"
|
||||
: runtime === "MlxIbv" || runtime === "MlxJaccl"
|
||||
? "MLX RDMA"
|
||||
: runtime}
|
||||
</span>
|
||||
|
||||
@@ -168,7 +168,7 @@ export interface ModelDownloadStatus {
|
||||
export interface PlacementPreview {
|
||||
model_id: string;
|
||||
sharding: "Pipeline" | "Tensor";
|
||||
instance_meta: "MlxRing" | "MlxJaccl";
|
||||
instance_meta: "MlxRing" | "MlxIbv" | "MlxJaccl";
|
||||
instance: unknown | null;
|
||||
memory_delta_by_node: Record<string, number> | null;
|
||||
error: string | null;
|
||||
@@ -219,6 +219,7 @@ interface RawStateResponse {
|
||||
string,
|
||||
{
|
||||
MlxRingInstance?: Instance;
|
||||
MlxIbvInstance?: Instance;
|
||||
MlxJacclInstance?: Instance;
|
||||
}
|
||||
>;
|
||||
@@ -249,20 +250,6 @@ interface RawStateResponse {
|
||||
>;
|
||||
// Thunderbolt bridge cycles (nodes with bridge enabled forming loops)
|
||||
thunderboltBridgeCycles?: string[][];
|
||||
// MetaInstances (declarative instance constraints)
|
||||
metaInstances?: Record<string, MetaInstanceData>;
|
||||
}
|
||||
|
||||
export interface MetaInstanceData {
|
||||
metaInstanceId: string;
|
||||
modelId: string;
|
||||
sharding: string;
|
||||
instanceMeta: string;
|
||||
minNodes: number;
|
||||
nodeIds: string[] | null;
|
||||
placementError: string | null;
|
||||
consecutiveFailures: number;
|
||||
lastFailureError: string | null;
|
||||
}
|
||||
|
||||
export interface MessageAttachment {
|
||||
@@ -548,7 +535,6 @@ class AppStore {
|
||||
isLoadingPreviews = $state(false);
|
||||
previewNodeFilter = $state<Set<string>>(new Set());
|
||||
lastUpdate = $state<number | null>(null);
|
||||
metaInstances = $state<Record<string, MetaInstanceData>>({});
|
||||
nodeIdentities = $state<Record<string, RawNodeIdentity>>({});
|
||||
thunderboltBridgeCycles = $state<string[][]>([]);
|
||||
nodeThunderbolt = $state<
|
||||
@@ -905,7 +891,11 @@ class AppStore {
|
||||
|
||||
let instanceType: string | null = null;
|
||||
if (instanceTag === "MlxRingInstance") instanceType = "MLX Ring";
|
||||
else if (instanceTag === "MlxJacclInstance") instanceType = "MLX RDMA";
|
||||
else if (
|
||||
instanceTag === "MlxIbvInstance" ||
|
||||
instanceTag === "MlxJacclInstance"
|
||||
)
|
||||
instanceType = "MLX RDMA";
|
||||
|
||||
let sharding: string | null = null;
|
||||
const inst = instance as {
|
||||
@@ -1270,8 +1260,6 @@ class AppStore {
|
||||
if (data.downloads) {
|
||||
this.downloads = data.downloads;
|
||||
}
|
||||
// MetaInstances
|
||||
this.metaInstances = data.metaInstances ?? {};
|
||||
if (data.nodeDisk) {
|
||||
this.nodeDisk = data.nodeDisk;
|
||||
}
|
||||
@@ -3031,7 +3019,6 @@ export const tps = () => appStore.tps;
|
||||
export const totalTokens = () => appStore.totalTokens;
|
||||
export const topologyData = () => appStore.topologyData;
|
||||
export const instances = () => appStore.instances;
|
||||
export const metaInstances = () => appStore.metaInstances;
|
||||
export const runners = () => appStore.runners;
|
||||
export const downloads = () => appStore.downloads;
|
||||
export const nodeDisk = () => appStore.nodeDisk;
|
||||
|
||||
@@ -42,7 +42,6 @@
|
||||
toggleTopologyOnlyMode,
|
||||
chatSidebarVisible,
|
||||
toggleChatSidebarVisible,
|
||||
metaInstances,
|
||||
nodeThunderbolt,
|
||||
nodeRdmaCtl,
|
||||
thunderboltBridgeCycles,
|
||||
@@ -50,7 +49,6 @@
|
||||
nodeIdentities,
|
||||
type DownloadProgress,
|
||||
type PlacementPreview,
|
||||
type MetaInstanceData,
|
||||
} from "$lib/stores/app.svelte";
|
||||
import HeaderNav from "$lib/components/HeaderNav.svelte";
|
||||
import { fade, fly } from "svelte/transition";
|
||||
@@ -70,72 +68,7 @@
|
||||
const debugEnabled = $derived(debugMode());
|
||||
const topologyOnlyEnabled = $derived(topologyOnlyMode());
|
||||
const sidebarVisible = $derived(chatSidebarVisible());
|
||||
const metaInstancesData = $derived(metaInstances());
|
||||
const tbBridgeCycles = $derived(thunderboltBridgeCycles());
|
||||
|
||||
// Get status for a MetaInstance that has no backing instance yet
|
||||
function getMetaInstancePlacingStatus(metaInstanceId: string) {
|
||||
const meta = metaInstancesData[metaInstanceId];
|
||||
const placementError = meta?.placementError;
|
||||
const failures = meta?.consecutiveFailures ?? 0;
|
||||
const lastError = meta?.lastFailureError;
|
||||
|
||||
if (placementError) {
|
||||
return {
|
||||
statusText: "PLACEMENT FAILED",
|
||||
statusClass: "failed",
|
||||
isDownloading: false as const,
|
||||
isFailed: true,
|
||||
progress: null,
|
||||
perNode: [] as Array<{
|
||||
nodeId: string;
|
||||
nodeName: string;
|
||||
progress: DownloadProgress;
|
||||
}>,
|
||||
perNodeStatus: [] as PerNodeRunnerStatus[],
|
||||
errorMessage: placementError,
|
||||
};
|
||||
}
|
||||
|
||||
if (failures > 0) {
|
||||
const retryPosition = ((failures - 1) % 3) + 1;
|
||||
const isRecreated = failures % 3 === 0;
|
||||
return {
|
||||
statusText: isRecreated ? "PLACING" : `RETRYING (${retryPosition}/3)`,
|
||||
statusClass: "starting",
|
||||
isDownloading: false as const,
|
||||
isFailed: false,
|
||||
progress: null,
|
||||
perNode: [] as Array<{
|
||||
nodeId: string;
|
||||
nodeName: string;
|
||||
progress: DownloadProgress;
|
||||
}>,
|
||||
perNodeStatus: [] as PerNodeRunnerStatus[],
|
||||
errorMessage: isRecreated
|
||||
? `Instance re-created due to failure: ${lastError}`
|
||||
: lastError
|
||||
? `Previous failure: ${lastError}`
|
||||
: null,
|
||||
};
|
||||
}
|
||||
|
||||
return {
|
||||
statusText: "PLACING",
|
||||
statusClass: "starting",
|
||||
isDownloading: false as const,
|
||||
isFailed: false,
|
||||
progress: null,
|
||||
perNode: [] as Array<{
|
||||
nodeId: string;
|
||||
nodeName: string;
|
||||
progress: DownloadProgress;
|
||||
}>,
|
||||
perNodeStatus: [] as PerNodeRunnerStatus[],
|
||||
errorMessage: null,
|
||||
};
|
||||
}
|
||||
|
||||
const tbBridgeData = $derived(nodeThunderboltBridge());
|
||||
const identitiesData = $derived(nodeIdentities());
|
||||
const tbIdentifiers = $derived(nodeThunderbolt());
|
||||
@@ -181,17 +114,6 @@
|
||||
});
|
||||
let tb5InfoDismissed = $state(false);
|
||||
|
||||
// Detect [jaccl] RDMA driver errors from MetaInstance failure errors
|
||||
const jacclError = $derived.by(() => {
|
||||
for (const mi of Object.values(metaInstancesData)) {
|
||||
if (mi.lastFailureError?.includes("[jaccl]")) {
|
||||
return mi.lastFailureError;
|
||||
}
|
||||
}
|
||||
return null;
|
||||
});
|
||||
let jacclDismissedError = $state<string | null>(null);
|
||||
|
||||
// Helper to get friendly node name from node ID
|
||||
function getNodeName(nodeId: string): string {
|
||||
const node = data?.nodes?.[nodeId];
|
||||
@@ -289,7 +211,7 @@
|
||||
return model.tasks.includes("ImageToImage");
|
||||
}
|
||||
let selectedSharding = $state<"Pipeline" | "Tensor">("Pipeline");
|
||||
type InstanceMeta = "MlxRing" | "MlxJaccl";
|
||||
type InstanceMeta = "MlxRing" | "MlxIbv" | "MlxJaccl";
|
||||
|
||||
// Launch defaults persistence
|
||||
const LAUNCH_DEFAULTS_KEY = "exo-launch-defaults";
|
||||
@@ -546,7 +468,7 @@
|
||||
const matchesSelectedRuntime = (runtime: InstanceMeta): boolean =>
|
||||
selectedInstanceType === "MlxRing"
|
||||
? runtime === "MlxRing"
|
||||
: runtime === "MlxJaccl" || runtime === "MlxJaccl";
|
||||
: runtime === "MlxIbv" || runtime === "MlxJaccl";
|
||||
|
||||
// Helper to check if a model can be launched (has valid placement with >= minNodes)
|
||||
function canModelFit(modelId: string): boolean {
|
||||
@@ -762,30 +684,39 @@
|
||||
launchingModelId = modelId;
|
||||
|
||||
try {
|
||||
// Use the specific preview if provided, otherwise fall back to filtered preview
|
||||
const preview = specificPreview ?? filteredPreview();
|
||||
|
||||
// Extract node IDs from the preview the user is seeing
|
||||
const previewNodeIds = preview?.memory_delta_by_node
|
||||
? Object.keys(preview.memory_delta_by_node)
|
||||
: nodeFilter.size > 0
|
||||
? Array.from(nodeFilter)
|
||||
: undefined;
|
||||
let instanceData: unknown;
|
||||
|
||||
const response = await fetch("/meta_instance", {
|
||||
if (preview?.instance) {
|
||||
// Use the instance from the preview
|
||||
instanceData = preview.instance;
|
||||
} else {
|
||||
// Fallback: GET placement from API
|
||||
const placementResponse = await fetch(
|
||||
`/instance/placement?model_id=${encodeURIComponent(modelId)}&sharding=${selectedSharding}&instance_meta=${selectedInstanceType}&min_nodes=${selectedMinNodes}`,
|
||||
);
|
||||
|
||||
if (!placementResponse.ok) {
|
||||
const errorText = await placementResponse.text();
|
||||
console.error("Failed to get placement:", errorText);
|
||||
return;
|
||||
}
|
||||
|
||||
instanceData = await placementResponse.json();
|
||||
}
|
||||
|
||||
// POST the instance to create it
|
||||
const response = await fetch("/instance", {
|
||||
method: "POST",
|
||||
headers: { "Content-Type": "application/json" },
|
||||
body: JSON.stringify({
|
||||
model_id: modelId,
|
||||
sharding: preview?.sharding ?? selectedSharding,
|
||||
instance_meta: preview?.instance_meta ?? selectedInstanceType,
|
||||
min_nodes: selectedMinNodes,
|
||||
node_ids: previewNodeIds,
|
||||
}),
|
||||
body: JSON.stringify({ instance: instanceData }),
|
||||
});
|
||||
|
||||
if (!response.ok) {
|
||||
const errorText = await response.text();
|
||||
console.error("Failed to create meta instance:", errorText);
|
||||
console.error("Failed to launch instance:", errorText);
|
||||
} else {
|
||||
// Always auto-select the newly launched model so the user chats to what they just launched
|
||||
setSelectedChatModel(modelId);
|
||||
@@ -808,7 +739,7 @@
|
||||
setTimeout(scrollToBottom, 1000);
|
||||
}
|
||||
} catch (error) {
|
||||
console.error("Error creating meta instance:", error);
|
||||
console.error("Error launching instance:", error);
|
||||
} finally {
|
||||
launchingModelId = null;
|
||||
}
|
||||
@@ -1010,18 +941,15 @@
|
||||
nodeName: string;
|
||||
progress: DownloadProgress;
|
||||
}>;
|
||||
perNodeStatus: PerNodeRunnerStatus[];
|
||||
} {
|
||||
if (!downloadsData || Object.keys(downloadsData).length === 0) {
|
||||
const statusInfo = deriveInstanceStatus(instanceWrapped);
|
||||
return {
|
||||
isDownloading: false,
|
||||
isFailed: statusInfo.statusText === "FAILED",
|
||||
errorMessage: statusInfo.errorMessage,
|
||||
isFailed: false,
|
||||
errorMessage: null,
|
||||
progress: null,
|
||||
statusText: statusInfo.statusText,
|
||||
statusText: "RUNNING",
|
||||
perNode: [],
|
||||
perNodeStatus: statusInfo.perNodeStatus,
|
||||
};
|
||||
}
|
||||
|
||||
@@ -1035,7 +963,6 @@
|
||||
progress: null,
|
||||
statusText: "PREPARING",
|
||||
perNode: [],
|
||||
perNodeStatus: [],
|
||||
};
|
||||
}
|
||||
|
||||
@@ -1104,7 +1031,6 @@
|
||||
progress: null,
|
||||
statusText: "FAILED",
|
||||
perNode: [],
|
||||
perNodeStatus: [],
|
||||
};
|
||||
}
|
||||
}
|
||||
@@ -1145,11 +1071,10 @@
|
||||
return {
|
||||
isDownloading: false,
|
||||
isFailed: statusInfo.statusText === "FAILED",
|
||||
errorMessage: statusInfo.errorMessage,
|
||||
errorMessage: null,
|
||||
progress: null,
|
||||
statusText: statusInfo.statusText,
|
||||
perNode: [],
|
||||
perNodeStatus: statusInfo.perNodeStatus,
|
||||
};
|
||||
}
|
||||
|
||||
@@ -1173,172 +1098,92 @@
|
||||
},
|
||||
statusText: "DOWNLOADING",
|
||||
perNode,
|
||||
perNodeStatus: [],
|
||||
};
|
||||
}
|
||||
|
||||
// Derive instance status from runners
|
||||
// Get color class for a status
|
||||
function getStatusColor(statusText: string): string {
|
||||
if (statusText === "FAILED" || statusText === "PLACEMENT FAILED")
|
||||
return "text-red-400";
|
||||
if (statusText.startsWith("RETRYING")) return "text-orange-400";
|
||||
if (statusText === "SHUTDOWN") return "text-gray-400";
|
||||
if (statusText === "DOWNLOADING") return "text-blue-400";
|
||||
if (
|
||||
statusText.startsWith("LOADING") ||
|
||||
statusText.startsWith("WARMING UP") ||
|
||||
statusText === "WAITING" ||
|
||||
statusText === "INITIALIZING"
|
||||
)
|
||||
return "text-yellow-400";
|
||||
if (statusText === "RUNNING") return "text-teal-400";
|
||||
if (statusText === "READY" || statusText === "LOADED")
|
||||
return "text-green-400";
|
||||
return "text-exo-light-gray";
|
||||
}
|
||||
|
||||
const RUNNER_STATUS_MAP: Record<string, string> = {
|
||||
RunnerWaitingForInitialization: "WaitingForInitialization",
|
||||
RunnerInitializingBackend: "InitializingBackend",
|
||||
RunnerWaitingForModel: "WaitingForModel",
|
||||
RunnerLoading: "Loading",
|
||||
RunnerLoaded: "Loaded",
|
||||
RunnerWarmingUp: "WarmingUp",
|
||||
RunnerReady: "Ready",
|
||||
RunnerRunning: "Running",
|
||||
RunnerShutdown: "Shutdown",
|
||||
RunnerFailed: "Failed",
|
||||
};
|
||||
|
||||
// Friendly labels for display
|
||||
const RUNNER_STATUS_DISPLAY: Record<string, string> = {
|
||||
WaitingForInitialization: "Initializing",
|
||||
InitializingBackend: "Initializing",
|
||||
WaitingForModel: "Waiting",
|
||||
Loading: "Loading",
|
||||
Loaded: "Loaded",
|
||||
WarmingUp: "Warming Up",
|
||||
Ready: "Ready",
|
||||
Running: "Running",
|
||||
Shutdown: "Shutdown",
|
||||
Failed: "Failed",
|
||||
};
|
||||
|
||||
interface PerNodeRunnerStatus {
|
||||
nodeId: string;
|
||||
nodeName: string;
|
||||
status: string; // friendly display status
|
||||
switch (statusText) {
|
||||
case "FAILED":
|
||||
return "text-red-400";
|
||||
case "SHUTDOWN":
|
||||
return "text-gray-400";
|
||||
case "DOWNLOADING":
|
||||
return "text-blue-400";
|
||||
case "LOADING":
|
||||
case "WARMING UP":
|
||||
case "WAITING":
|
||||
case "INITIALIZING":
|
||||
return "text-yellow-400";
|
||||
case "RUNNING":
|
||||
return "text-teal-400";
|
||||
case "READY":
|
||||
case "LOADED":
|
||||
return "text-green-400";
|
||||
default:
|
||||
return "text-exo-light-gray";
|
||||
}
|
||||
}
|
||||
|
||||
function deriveInstanceStatus(instanceWrapped: unknown): {
|
||||
statusText: string;
|
||||
statusClass: string;
|
||||
perNodeStatus: PerNodeRunnerStatus[];
|
||||
errorMessage: string | null;
|
||||
} {
|
||||
const [, instance] = getTagged(instanceWrapped);
|
||||
if (!instance || typeof instance !== "object") {
|
||||
return {
|
||||
statusText: "PREPARING",
|
||||
statusClass: "inactive",
|
||||
perNodeStatus: [],
|
||||
errorMessage: null,
|
||||
};
|
||||
return { statusText: "PREPARING", statusClass: "inactive" };
|
||||
}
|
||||
|
||||
const inst = instance as {
|
||||
shardAssignments?: {
|
||||
runnerToShard?: Record<string, unknown>;
|
||||
nodeToRunner?: Record<string, string>;
|
||||
};
|
||||
shardAssignments?: { runnerToShard?: Record<string, unknown> };
|
||||
};
|
||||
const nodeToRunner = inst.shardAssignments?.nodeToRunner || {};
|
||||
const runnerIds = Object.keys(inst.shardAssignments?.runnerToShard || {});
|
||||
const totalNodes = runnerIds.length;
|
||||
|
||||
// Build per-node status and extract error messages from RunnerFailed
|
||||
const perNodeStatus: PerNodeRunnerStatus[] = [];
|
||||
const statuses: string[] = [];
|
||||
const failedErrors: string[] = [];
|
||||
for (const [nodeId, runnerId] of Object.entries(nodeToRunner)) {
|
||||
const r = runnersData[runnerId];
|
||||
let status: string | null = null;
|
||||
if (r) {
|
||||
const [kind, runnerData] = getTagged(r);
|
||||
status = kind ? RUNNER_STATUS_MAP[kind] || null : null;
|
||||
// Extract error message from RunnerFailed
|
||||
if (
|
||||
kind === "RunnerFailed" &&
|
||||
runnerData &&
|
||||
typeof runnerData === "object"
|
||||
) {
|
||||
const rd = runnerData as { errorMessage?: string };
|
||||
if (rd.errorMessage) failedErrors.push(`${getNodeName(nodeId)}: ${rd.errorMessage}`);
|
||||
}
|
||||
}
|
||||
if (status) {
|
||||
statuses.push(status);
|
||||
perNodeStatus.push({
|
||||
nodeId,
|
||||
nodeName: getNodeName(nodeId),
|
||||
status: RUNNER_STATUS_DISPLAY[status] || status,
|
||||
});
|
||||
}
|
||||
}
|
||||
const statuses = runnerIds
|
||||
.map((rid) => {
|
||||
const r = runnersData[rid];
|
||||
if (!r) return null;
|
||||
const [kind] = getTagged(r);
|
||||
const statusMap: Record<string, string> = {
|
||||
RunnerWaitingForInitialization: "WaitingForInitialization",
|
||||
RunnerInitializingBackend: "InitializingBackend",
|
||||
RunnerWaitingForModel: "WaitingForModel",
|
||||
RunnerLoading: "Loading",
|
||||
RunnerLoaded: "Loaded",
|
||||
RunnerWarmingUp: "WarmingUp",
|
||||
RunnerReady: "Ready",
|
||||
RunnerRunning: "Running",
|
||||
RunnerShutdown: "Shutdown",
|
||||
RunnerFailed: "Failed",
|
||||
};
|
||||
return kind ? statusMap[kind] || null : null;
|
||||
})
|
||||
.filter((s): s is string => s !== null);
|
||||
|
||||
const has = (s: string) => statuses.includes(s);
|
||||
const count = (s: string) => statuses.filter((v) => v === s).length;
|
||||
|
||||
if (statuses.length === 0)
|
||||
return {
|
||||
statusText: "PREPARING",
|
||||
statusClass: "inactive",
|
||||
perNodeStatus,
|
||||
errorMessage: null,
|
||||
};
|
||||
if (has("Failed"))
|
||||
return {
|
||||
statusText: "FAILED",
|
||||
statusClass: "failed",
|
||||
perNodeStatus,
|
||||
errorMessage: failedErrors.length > 0 ? failedErrors.join("; ") : null,
|
||||
};
|
||||
return { statusText: "PREPARING", statusClass: "inactive" };
|
||||
if (has("Failed")) return { statusText: "FAILED", statusClass: "failed" };
|
||||
if (has("Shutdown"))
|
||||
return { statusText: "SHUTDOWN", statusClass: "inactive", perNodeStatus, errorMessage: null };
|
||||
|
||||
// For loading/warming states, show node progress when multi-node
|
||||
if (has("Loading")) {
|
||||
const readyCount = count("Ready") + count("Running") + count("Loaded");
|
||||
const statusText =
|
||||
totalNodes > 1
|
||||
? `LOADING (${readyCount}/${totalNodes} nodes ready)`
|
||||
: "LOADING";
|
||||
return { statusText, statusClass: "starting", perNodeStatus, errorMessage: null };
|
||||
}
|
||||
if (has("WarmingUp")) {
|
||||
const readyCount = count("Ready") + count("Running");
|
||||
const statusText =
|
||||
totalNodes > 1
|
||||
? `WARMING UP (${readyCount}/${totalNodes} nodes ready)`
|
||||
: "WARMING UP";
|
||||
return { statusText, statusClass: "starting", perNodeStatus, errorMessage: null };
|
||||
}
|
||||
|
||||
return { statusText: "SHUTDOWN", statusClass: "inactive" };
|
||||
if (has("Loading"))
|
||||
return { statusText: "LOADING", statusClass: "starting" };
|
||||
if (has("WarmingUp"))
|
||||
return { statusText: "WARMING UP", statusClass: "starting" };
|
||||
if (has("Running"))
|
||||
return { statusText: "RUNNING", statusClass: "running", perNodeStatus, errorMessage: null };
|
||||
if (has("Ready"))
|
||||
return { statusText: "READY", statusClass: "loaded", perNodeStatus, errorMessage: null };
|
||||
if (has("Loaded"))
|
||||
return { statusText: "LOADED", statusClass: "loaded", perNodeStatus, errorMessage: null };
|
||||
return { statusText: "RUNNING", statusClass: "running" };
|
||||
if (has("Ready")) return { statusText: "READY", statusClass: "loaded" };
|
||||
if (has("Loaded")) return { statusText: "LOADED", statusClass: "loaded" };
|
||||
if (has("WaitingForModel"))
|
||||
return { statusText: "WAITING", statusClass: "starting", perNodeStatus, errorMessage: null };
|
||||
return { statusText: "WAITING", statusClass: "starting" };
|
||||
if (has("InitializingBackend"))
|
||||
return { statusText: "INITIALIZING", statusClass: "starting", perNodeStatus, errorMessage: null };
|
||||
return { statusText: "INITIALIZING", statusClass: "starting" };
|
||||
if (has("WaitingForInitialization"))
|
||||
return { statusText: "INITIALIZING", statusClass: "starting", perNodeStatus, errorMessage: null };
|
||||
return { statusText: "INITIALIZING", statusClass: "starting" };
|
||||
|
||||
return { statusText: "RUNNING", statusClass: "active", perNodeStatus, errorMessage: null };
|
||||
return { statusText: "RUNNING", statusClass: "active" };
|
||||
}
|
||||
|
||||
function getBytes(value: unknown): number {
|
||||
@@ -1397,75 +1242,6 @@
|
||||
}
|
||||
}
|
||||
|
||||
async function deleteMetaInstance(metaInstanceId: string) {
|
||||
const meta = metaInstancesData[metaInstanceId];
|
||||
const modelId = meta?.modelId ?? "unknown";
|
||||
if (!confirm(`Delete model ${modelId}?`)) return;
|
||||
|
||||
const wasSelected = selectedChatModel() === modelId;
|
||||
|
||||
try {
|
||||
const response = await fetch(`/meta_instance/${metaInstanceId}`, {
|
||||
method: "DELETE",
|
||||
headers: { "Content-Type": "application/json" },
|
||||
});
|
||||
|
||||
if (!response.ok) {
|
||||
console.error("Failed to delete meta instance:", response.status);
|
||||
} else if (wasSelected) {
|
||||
// Switch to another available model or clear selection
|
||||
const remainingInstances = Object.entries(instanceData).filter(
|
||||
([id]) => id !== getBackingInstanceId(metaInstanceId),
|
||||
);
|
||||
if (remainingInstances.length > 0) {
|
||||
const [, lastInstance] =
|
||||
remainingInstances[remainingInstances.length - 1];
|
||||
const newModelId = getInstanceModelId(lastInstance);
|
||||
if (
|
||||
newModelId &&
|
||||
newModelId !== "Unknown" &&
|
||||
newModelId !== "Unknown Model"
|
||||
) {
|
||||
setSelectedChatModel(newModelId);
|
||||
} else {
|
||||
setSelectedChatModel("");
|
||||
}
|
||||
} else {
|
||||
setSelectedChatModel("");
|
||||
}
|
||||
}
|
||||
} catch (error) {
|
||||
console.error("Error deleting meta instance:", error);
|
||||
}
|
||||
}
|
||||
|
||||
// Find the backing Instance ID for a MetaInstance by scanning instances
|
||||
function getBackingInstanceId(metaInstanceId: string): string | null {
|
||||
for (const [id, inst] of Object.entries(instanceData)) {
|
||||
const [, inner] = getTagged(inst);
|
||||
if (
|
||||
inner &&
|
||||
typeof inner === "object" &&
|
||||
(inner as Record<string, unknown>).metaInstanceId === metaInstanceId
|
||||
) {
|
||||
return id;
|
||||
}
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
// Get orphan Instance IDs (not backing any MetaInstance)
|
||||
function getOrphanInstanceIds(): string[] {
|
||||
return Object.keys(instanceData).filter((id) => {
|
||||
const [, inner] = getTagged(instanceData[id]);
|
||||
return (
|
||||
!inner ||
|
||||
typeof inner !== "object" ||
|
||||
!(inner as Record<string, unknown>).metaInstanceId
|
||||
);
|
||||
});
|
||||
}
|
||||
|
||||
// Helper to unwrap tagged unions like { MlxRingInstance: {...} }
|
||||
function getTagged(obj: unknown): [string | null, unknown] {
|
||||
if (!obj || typeof obj !== "object") return [null, null];
|
||||
@@ -1506,7 +1282,11 @@
|
||||
// Instance type from tag
|
||||
let instanceType = "Unknown";
|
||||
if (instanceTag === "MlxRingInstance") instanceType = "MLX Ring";
|
||||
else if (instanceTag === "MlxJacclInstance") instanceType = "MLX RDMA";
|
||||
else if (
|
||||
instanceTag === "MlxIbvInstance" ||
|
||||
instanceTag === "MlxJacclInstance"
|
||||
)
|
||||
instanceType = "MLX RDMA";
|
||||
|
||||
const inst = instance as {
|
||||
shardAssignments?: {
|
||||
@@ -1854,51 +1634,7 @@
|
||||
}
|
||||
|
||||
const nodeCount = $derived(data ? Object.keys(data.nodes).length : 0);
|
||||
const metaInstanceCount = $derived(Object.keys(metaInstancesData).length);
|
||||
const orphanInstanceIds = $derived(getOrphanInstanceIds());
|
||||
const instanceCount = $derived(metaInstanceCount + orphanInstanceIds.length);
|
||||
|
||||
// Unified display items: MetaInstances first, then orphan Instances
|
||||
interface DisplayItem {
|
||||
id: string; // MetaInstance ID or Instance ID (used as key and displayed)
|
||||
modelId: string;
|
||||
instance: unknown | null; // The backing/orphan instance (tagged union) or null if placing
|
||||
instanceId: string | null; // The actual Instance ID (for topology hover)
|
||||
isMetaInstance: boolean;
|
||||
sharding: string | null; // From MetaInstance constraints (used when instance is null)
|
||||
instanceMeta: string | null; // From MetaInstance constraints (used when instance is null)
|
||||
}
|
||||
|
||||
const unifiedDisplayItems = $derived((): DisplayItem[] => {
|
||||
const items: DisplayItem[] = [];
|
||||
// MetaInstances
|
||||
for (const [metaId, meta] of Object.entries(metaInstancesData)) {
|
||||
const backingId = getBackingInstanceId(metaId);
|
||||
items.push({
|
||||
id: metaId,
|
||||
modelId: meta.modelId,
|
||||
instance: backingId ? instanceData[backingId] : null,
|
||||
instanceId: backingId,
|
||||
isMetaInstance: true,
|
||||
sharding: meta.sharding,
|
||||
instanceMeta: meta.instanceMeta,
|
||||
});
|
||||
}
|
||||
// Orphan Instances
|
||||
for (const orphanId of getOrphanInstanceIds()) {
|
||||
const inst = instanceData[orphanId];
|
||||
items.push({
|
||||
id: orphanId,
|
||||
modelId: getInstanceModelId(inst),
|
||||
instance: inst,
|
||||
instanceId: orphanId,
|
||||
isMetaInstance: false,
|
||||
sharding: null,
|
||||
instanceMeta: null,
|
||||
});
|
||||
}
|
||||
return items;
|
||||
});
|
||||
const instanceCount = $derived(Object.keys(instanceData).length);
|
||||
|
||||
// Helper to get the number of nodes in a placement preview
|
||||
function getPreviewNodeCount(preview: PlacementPreview): number {
|
||||
@@ -2016,71 +1752,8 @@
|
||||
</script>
|
||||
|
||||
{#snippet clusterWarnings()}
|
||||
{#if tbBridgeCycles.length > 0 || macosVersionMismatch || (tb5WithoutRdma && !tb5InfoDismissed) || (jacclError && jacclError !== jacclDismissedError)}
|
||||
{#if tbBridgeCycles.length > 0 || macosVersionMismatch || (tb5WithoutRdma && !tb5InfoDismissed)}
|
||||
<div class="absolute top-4 left-4 flex flex-col gap-2 z-40">
|
||||
{#if jacclError && jacclError !== jacclDismissedError}
|
||||
<div class="group relative" role="alert">
|
||||
<div
|
||||
class="flex items-center gap-2 px-3 py-2 rounded border border-red-500/50 bg-red-500/10 backdrop-blur-sm cursor-help"
|
||||
>
|
||||
<svg
|
||||
class="w-5 h-5 text-red-400 flex-shrink-0"
|
||||
fill="none"
|
||||
viewBox="0 0 24 24"
|
||||
stroke="currentColor"
|
||||
stroke-width="2"
|
||||
>
|
||||
<path
|
||||
stroke-linecap="round"
|
||||
stroke-linejoin="round"
|
||||
d={warningIconPath}
|
||||
/>
|
||||
</svg>
|
||||
<span class="text-sm font-mono text-red-200">
|
||||
JACCL RDMA ERROR
|
||||
</span>
|
||||
<button
|
||||
type="button"
|
||||
onclick={() => (jacclDismissedError = jacclError)}
|
||||
class="ml-1 text-red-300/60 hover:text-red-200 transition-colors cursor-pointer"
|
||||
title="Dismiss"
|
||||
>
|
||||
<svg
|
||||
class="w-4 h-4"
|
||||
fill="none"
|
||||
viewBox="0 0 24 24"
|
||||
stroke="currentColor"
|
||||
stroke-width="2"
|
||||
>
|
||||
<path
|
||||
stroke-linecap="round"
|
||||
stroke-linejoin="round"
|
||||
d="M6 18L18 6M6 6l12 12"
|
||||
/>
|
||||
</svg>
|
||||
</button>
|
||||
</div>
|
||||
|
||||
<!-- Tooltip on hover -->
|
||||
<div
|
||||
class="absolute top-full left-0 mt-2 w-80 p-3 rounded border border-red-500/30 bg-exo-dark-gray/95 backdrop-blur-sm opacity-0 invisible group-hover:opacity-100 group-hover:visible transition-all duration-200 z-50 shadow-lg"
|
||||
>
|
||||
<p class="text-xs text-white/80 mb-2">
|
||||
A macOS RDMA driver error was detected. This is a known issue
|
||||
with the experimental RDMA driver in macOS.
|
||||
</p>
|
||||
<p class="text-xs text-white/60 mb-2">
|
||||
<span class="text-red-300">Error:</span>
|
||||
{jacclError}
|
||||
</p>
|
||||
<p class="text-xs text-white/60">
|
||||
<span class="text-red-300">To fix:</span> Restart the affected machine.
|
||||
There is currently no other workaround for this issue.
|
||||
</p>
|
||||
</div>
|
||||
</div>
|
||||
{/if}
|
||||
|
||||
{#if tbBridgeCycles.length > 0}
|
||||
{@const cycle = tbBridgeCycles[0]}
|
||||
{@const serviceName = getTbBridgeServiceName(cycle)}
|
||||
@@ -2249,29 +1922,8 @@
|
||||
{/snippet}
|
||||
|
||||
{#snippet clusterWarningsCompact()}
|
||||
{#if tbBridgeCycles.length > 0 || macosVersionMismatch || (tb5WithoutRdma && !tb5InfoDismissed) || (jacclError && jacclError !== jacclDismissedError)}
|
||||
{#if tbBridgeCycles.length > 0 || macosVersionMismatch || (tb5WithoutRdma && !tb5InfoDismissed)}
|
||||
<div class="absolute top-2 left-2 flex flex-col gap-1">
|
||||
{#if jacclError && jacclError !== jacclDismissedError}
|
||||
<div
|
||||
class="flex items-center gap-1.5 px-2 py-1 rounded border border-red-500/50 bg-red-500/10 backdrop-blur-sm"
|
||||
title="JACCL RDMA driver error — restart affected machine"
|
||||
>
|
||||
<svg
|
||||
class="w-3.5 h-3.5 text-red-400"
|
||||
fill="none"
|
||||
viewBox="0 0 24 24"
|
||||
stroke="currentColor"
|
||||
stroke-width="2"
|
||||
>
|
||||
<path
|
||||
stroke-linecap="round"
|
||||
stroke-linejoin="round"
|
||||
d={warningIconPath}
|
||||
/>
|
||||
</svg>
|
||||
<span class="text-[10px] font-mono text-red-200">JACCL ERROR</span>
|
||||
</div>
|
||||
{/if}
|
||||
{#if tbBridgeCycles.length > 0}
|
||||
<div
|
||||
class="flex items-center gap-1.5 px-2 py-1 rounded border border-yellow-500/50 bg-yellow-500/10 backdrop-blur-sm"
|
||||
@@ -2649,57 +2301,31 @@
|
||||
bind:this={instancesContainerRef}
|
||||
class="max-h-72 xl:max-h-96 space-y-3 overflow-y-auto overflow-x-hidden py-px"
|
||||
>
|
||||
{#each unifiedDisplayItems() as item (item.id)}
|
||||
{@const id = item.id}
|
||||
{@const instance = item.instance}
|
||||
{@const downloadInfo = instance
|
||||
? getInstanceDownloadStatus(item.instanceId ?? id, instance)
|
||||
: getMetaInstancePlacingStatus(id)}
|
||||
{@const metaData = item.isMetaInstance ? metaInstancesData[id] : null}
|
||||
{@const retryError = metaData?.lastFailureError && !downloadInfo.isFailed
|
||||
? metaData.consecutiveFailures > 0
|
||||
? `(${((metaData.consecutiveFailures - 1) % 3) + 1}/3) ${metaData.lastFailureError}`
|
||||
: metaData.lastFailureError
|
||||
: null}
|
||||
{#each Object.entries(instanceData) as [id, instance]}
|
||||
{@const downloadInfo = getInstanceDownloadStatus(
|
||||
id,
|
||||
instance,
|
||||
)}
|
||||
{@const statusText = downloadInfo.statusText}
|
||||
{@const isDownloading = downloadInfo.isDownloading}
|
||||
{@const isFailed =
|
||||
statusText === "FAILED" ||
|
||||
statusText === "PLACEMENT FAILED"}
|
||||
{@const isFailed = statusText === "FAILED"}
|
||||
{@const isLoading =
|
||||
statusText.startsWith("LOADING") ||
|
||||
statusText.startsWith("WARMING UP") ||
|
||||
statusText === "WAITING" ||
|
||||
statusText === "PLACING" ||
|
||||
statusText.startsWith("RETRYING")}
|
||||
statusText === "LOADING" ||
|
||||
statusText === "WARMING UP" ||
|
||||
statusText === "WAITING"}
|
||||
{@const isReady =
|
||||
statusText === "READY" || statusText === "LOADED"}
|
||||
{@const isRunning = statusText === "RUNNING"}
|
||||
<!-- Instance Card -->
|
||||
{@const instanceModelId = item.modelId}
|
||||
{@const instanceInfo = instance
|
||||
? getInstanceInfo(instance)
|
||||
: {
|
||||
instanceType:
|
||||
item.instanceMeta === "MlxRing"
|
||||
? "MLX Ring"
|
||||
: item.instanceMeta === "MlxJaccl"
|
||||
? "MLX RDMA"
|
||||
: "Unknown",
|
||||
sharding: item.sharding ?? "Unknown",
|
||||
nodeNames: [] as string[],
|
||||
nodeIds: [] as string[],
|
||||
nodeCount: 0,
|
||||
}}
|
||||
{@const instanceConnections = instance
|
||||
? getInstanceConnections(instance)
|
||||
: []}
|
||||
{@const instanceModelId = getInstanceModelId(instance)}
|
||||
{@const instanceInfo = getInstanceInfo(instance)}
|
||||
{@const instanceConnections =
|
||||
getInstanceConnections(instance)}
|
||||
<div
|
||||
class="relative group cursor-pointer"
|
||||
role="button"
|
||||
tabindex="0"
|
||||
onmouseenter={() =>
|
||||
(hoveredInstanceId = item.instanceId ?? id)}
|
||||
onmouseenter={() => (hoveredInstanceId = id)}
|
||||
onmouseleave={() => (hoveredInstanceId = null)}
|
||||
onclick={() => {
|
||||
if (
|
||||
@@ -2798,10 +2424,7 @@
|
||||
>
|
||||
</div>
|
||||
<button
|
||||
onclick={() =>
|
||||
item.isMetaInstance
|
||||
? deleteMetaInstance(id)
|
||||
: deleteInstance(id)}
|
||||
onclick={() => deleteInstance(id)}
|
||||
class="text-xs px-2 py-1 font-mono tracking-wider uppercase border border-red-500/30 text-red-400 hover:bg-red-500/20 hover:text-red-400 hover:border-red-500/50 transition-all duration-200 cursor-pointer"
|
||||
>
|
||||
DELETE
|
||||
@@ -2811,7 +2434,7 @@
|
||||
<div
|
||||
class="text-exo-yellow text-xs font-mono tracking-wide truncate"
|
||||
>
|
||||
{instanceModelId}
|
||||
{getInstanceModelId(instance)}
|
||||
</div>
|
||||
<div class="text-white/60 text-xs font-mono">
|
||||
Strategy: <span class="text-white/80"
|
||||
@@ -3079,30 +2702,6 @@
|
||||
>
|
||||
{downloadInfo.errorMessage}
|
||||
</div>
|
||||
{:else if retryError}
|
||||
<div
|
||||
class="text-xs text-orange-400/80 font-mono mt-1 break-words"
|
||||
>
|
||||
Retrying after error: {retryError}
|
||||
</div>
|
||||
{/if}
|
||||
{#if downloadInfo.perNodeStatus.length > 1 && (statusText.startsWith("LOADING") || statusText.startsWith("WARMING UP") || statusText === "WAITING" || statusText === "INITIALIZING")}
|
||||
<div class="mt-1.5 space-y-0.5">
|
||||
{#each downloadInfo.perNodeStatus as node}
|
||||
<div
|
||||
class="flex items-center justify-between text-[10px] font-mono"
|
||||
>
|
||||
<span class="text-white/60 truncate pr-2"
|
||||
>{node.nodeName}</span
|
||||
>
|
||||
<span
|
||||
class={getStatusColor(
|
||||
node.status.toUpperCase(),
|
||||
)}>{node.status}</span
|
||||
>
|
||||
</div>
|
||||
{/each}
|
||||
</div>
|
||||
{/if}
|
||||
{/if}
|
||||
</div>
|
||||
@@ -3271,21 +2870,21 @@
|
||||
</button>
|
||||
<button
|
||||
onclick={() => {
|
||||
selectedInstanceType = "MlxJaccl";
|
||||
selectedInstanceType = "MlxIbv";
|
||||
saveLaunchDefaults();
|
||||
}}
|
||||
class="flex items-center gap-2 py-2 px-4 text-sm font-mono border rounded transition-all duration-200 cursor-pointer {selectedInstanceType ===
|
||||
'MlxJaccl'
|
||||
'MlxIbv'
|
||||
? 'bg-transparent text-exo-yellow border-exo-yellow'
|
||||
: 'bg-transparent text-white/70 border-exo-medium-gray/50 hover:border-exo-yellow/50'}"
|
||||
>
|
||||
<span
|
||||
class="w-4 h-4 rounded-full border-2 flex items-center justify-center {selectedInstanceType ===
|
||||
'MlxJaccl'
|
||||
'MlxIbv'
|
||||
? 'border-exo-yellow'
|
||||
: 'border-exo-medium-gray'}"
|
||||
>
|
||||
{#if selectedInstanceType === "MlxJaccl"}
|
||||
{#if selectedInstanceType === "MlxIbv"}
|
||||
<span class="w-2 h-2 rounded-full bg-exo-yellow"></span>
|
||||
{/if}
|
||||
</span>
|
||||
@@ -3514,60 +3113,31 @@
|
||||
<div
|
||||
class="space-y-3 max-h-72 xl:max-h-96 overflow-y-auto overflow-x-hidden py-px pr-1"
|
||||
>
|
||||
{#each unifiedDisplayItems() as item (item.id)}
|
||||
{@const id = item.id}
|
||||
{@const instance = item.instance}
|
||||
{@const downloadInfo = instance
|
||||
? getInstanceDownloadStatus(
|
||||
item.instanceId ?? id,
|
||||
instance,
|
||||
)
|
||||
: getMetaInstancePlacingStatus(id)}
|
||||
{@const metaData = item.isMetaInstance ? metaInstancesData[id] : null}
|
||||
{@const retryError = metaData?.lastFailureError && !downloadInfo.isFailed
|
||||
? metaData.consecutiveFailures > 0
|
||||
? `(${((metaData.consecutiveFailures - 1) % 3) + 1}/3) ${metaData.lastFailureError}`
|
||||
: metaData.lastFailureError
|
||||
: null}
|
||||
{#each Object.entries(instanceData) as [id, instance]}
|
||||
{@const downloadInfo = getInstanceDownloadStatus(
|
||||
id,
|
||||
instance,
|
||||
)}
|
||||
{@const statusText = downloadInfo.statusText}
|
||||
{@const isDownloading = downloadInfo.isDownloading}
|
||||
{@const isFailed =
|
||||
statusText === "FAILED" ||
|
||||
statusText === "PLACEMENT FAILED"}
|
||||
{@const isFailed = statusText === "FAILED"}
|
||||
{@const isLoading =
|
||||
statusText.startsWith("LOADING") ||
|
||||
statusText.startsWith("WARMING UP") ||
|
||||
statusText === "WAITING" ||
|
||||
statusText === "PLACING" ||
|
||||
statusText.startsWith("RETRYING")}
|
||||
statusText === "LOADING" ||
|
||||
statusText === "WARMING UP" ||
|
||||
statusText === "WAITING"}
|
||||
{@const isReady =
|
||||
statusText === "READY" || statusText === "LOADED"}
|
||||
{@const isRunning = statusText === "RUNNING"}
|
||||
<!-- Instance Card -->
|
||||
{@const instanceModelId = item.modelId}
|
||||
{@const instanceInfo = instance
|
||||
? getInstanceInfo(instance)
|
||||
: {
|
||||
instanceType:
|
||||
item.instanceMeta === "MlxRing"
|
||||
? "MLX Ring"
|
||||
: item.instanceMeta === "MlxJaccl"
|
||||
? "MLX RDMA"
|
||||
: "Unknown",
|
||||
sharding: item.sharding ?? "Unknown",
|
||||
nodeNames: [] as string[],
|
||||
nodeIds: [] as string[],
|
||||
nodeCount: 0,
|
||||
}}
|
||||
{@const instanceConnections = instance
|
||||
? getInstanceConnections(instance)
|
||||
: []}
|
||||
{@const instanceModelId = getInstanceModelId(instance)}
|
||||
{@const instanceInfo = getInstanceInfo(instance)}
|
||||
{@const instanceConnections =
|
||||
getInstanceConnections(instance)}
|
||||
<div
|
||||
class="relative group cursor-pointer"
|
||||
role="button"
|
||||
tabindex="0"
|
||||
onmouseenter={() =>
|
||||
(hoveredInstanceId = item.instanceId ?? id)}
|
||||
onmouseenter={() => (hoveredInstanceId = id)}
|
||||
onmouseleave={() => (hoveredInstanceId = null)}
|
||||
onclick={() => {
|
||||
if (
|
||||
@@ -3666,10 +3236,7 @@
|
||||
>
|
||||
</div>
|
||||
<button
|
||||
onclick={() =>
|
||||
item.isMetaInstance
|
||||
? deleteMetaInstance(id)
|
||||
: deleteInstance(id)}
|
||||
onclick={() => deleteInstance(id)}
|
||||
class="text-xs px-2 py-1 font-mono tracking-wider uppercase border border-red-500/30 text-red-400 hover:bg-red-500/20 hover:text-red-400 hover:border-red-500/50 transition-all duration-200 cursor-pointer"
|
||||
>
|
||||
DELETE
|
||||
@@ -3679,7 +3246,7 @@
|
||||
<div
|
||||
class="text-exo-yellow text-xs font-mono tracking-wide truncate"
|
||||
>
|
||||
{instanceModelId}
|
||||
{getInstanceModelId(instance)}
|
||||
</div>
|
||||
<div class="text-white/60 text-xs font-mono">
|
||||
Strategy: <span class="text-white/80"
|
||||
@@ -3957,30 +3524,6 @@
|
||||
>
|
||||
{downloadInfo.errorMessage}
|
||||
</div>
|
||||
{:else if retryError}
|
||||
<div
|
||||
class="text-xs text-orange-400/80 font-mono mt-1 break-words"
|
||||
>
|
||||
Retrying after error: {retryError}
|
||||
</div>
|
||||
{/if}
|
||||
{#if downloadInfo.perNodeStatus.length > 1 && (statusText.startsWith("LOADING") || statusText.startsWith("WARMING UP") || statusText === "WAITING" || statusText === "INITIALIZING")}
|
||||
<div class="mt-1.5 space-y-0.5">
|
||||
{#each downloadInfo.perNodeStatus as node}
|
||||
<div
|
||||
class="flex items-center justify-between text-[10px] font-mono"
|
||||
>
|
||||
<span class="text-white/60 truncate pr-2"
|
||||
>{node.nodeName}</span
|
||||
>
|
||||
<span
|
||||
class={getStatusColor(
|
||||
node.status.toUpperCase(),
|
||||
)}>{node.status}</span
|
||||
>
|
||||
</div>
|
||||
{/each}
|
||||
</div>
|
||||
{/if}
|
||||
{/if}
|
||||
</div>
|
||||
|
||||
@@ -1,2 +0,0 @@
|
||||
# we can manually exclude false-positive lint errors for dual packages (if in dependencies)
|
||||
#allowed-duplicate-crates = ["hashbrown"]
|
||||
@@ -25,44 +25,26 @@ workspace = true
|
||||
networking = { workspace = true }
|
||||
|
||||
# interop
|
||||
pyo3 = { version = "0.27.1", features = [
|
||||
# "abi3-py311", # tells pyo3 (and maturin) to build using the stable ABI with minimum Python version 3.11
|
||||
"nightly", # enables better-supported GIL integration
|
||||
"experimental-async", # async support in #[pyfunction] & #[pymethods]
|
||||
#"experimental-inspect", # inspection of generated binary => easier to automate type-hint generation
|
||||
#"py-clone", # adding Clone-ing of `Py<T>` without GIL (may cause panics - remove if panics happen)
|
||||
"multiple-pymethods", # allows multiple #[pymethods] sections per class
|
||||
pyo3 = { version = "0.27.2", features = [
|
||||
"abi3-py313", # tells pyo3 (and maturin) to build using the stable ABI with minimum Python version 3.13
|
||||
# "nightly", # enables better-supported GIL integration
|
||||
"experimental-async" # async support in #[pyfunction] & #[pymethods]
|
||||
# "experimental-inspect", # inspection of generated binary => easier to automate type-hint generation
|
||||
# "py-clone", # adding Clone-ing of `Py<T>` without GIL (may cause panics - remove if panics happen)
|
||||
# "multiple-pymethods", # allows multiple #[pymethods] sections per class
|
||||
|
||||
# integrations with other libraries
|
||||
"arc_lock", "bigdecimal", "either", "hashbrown", "indexmap", "num-bigint", "num-complex", "num-rational",
|
||||
"ordered-float", "rust_decimal", "smallvec",
|
||||
# "arc_lock", "bigdecimal", "either", "hashbrown", "indexmap", "num-bigint", "num-complex", "num-rational",
|
||||
# "ordered-float", "rust_decimal", "smallvec",
|
||||
# "anyhow", "chrono", "chrono-local", "chrono-tz", "eyre", "jiff-02", "lock_api", "parking-lot", "time", "serde",
|
||||
] }
|
||||
pyo3-stub-gen = { version = "0.17.2" }
|
||||
pyo3-async-runtimes = { version = "0.27.0", features = ["attributes", "tokio-runtime", "testing"] }
|
||||
pyo3-log = "0.13.2"
|
||||
|
||||
# macro dependencies
|
||||
extend = { workspace = true }
|
||||
delegate = { workspace = true }
|
||||
impl-trait-for-tuples = { workspace = true }
|
||||
derive_more = { workspace = true }
|
||||
pin-project = { workspace = true }
|
||||
|
||||
# async runtime
|
||||
tokio = { workspace = true, features = ["full", "tracing"] }
|
||||
futures = { workspace = true }
|
||||
|
||||
# utility dependencies
|
||||
once_cell = "1.21.3"
|
||||
thread_local = "1.1.9"
|
||||
util = { workspace = true }
|
||||
thiserror = { workspace = true }
|
||||
#internment = { workspace = true }
|
||||
#recursion = { workspace = true }
|
||||
#generativity = { workspace = true }
|
||||
#itertools = { workspace = true }
|
||||
|
||||
futures-lite = "2.6.1"
|
||||
|
||||
# Tracing
|
||||
#tracing = "0.1"
|
||||
|
||||
@@ -2,220 +2,39 @@
|
||||
# ruff: noqa: E501, F401
|
||||
|
||||
import builtins
|
||||
import enum
|
||||
import typing
|
||||
|
||||
@typing.final
|
||||
class AllQueuesFullError(builtins.Exception):
|
||||
def __new__(cls, *args: typing.Any) -> AllQueuesFullError: ...
|
||||
def __repr__(self) -> builtins.str: ...
|
||||
def __str__(self) -> builtins.str: ...
|
||||
|
||||
@typing.final
|
||||
class ConnectionUpdate:
|
||||
@property
|
||||
def update_type(self) -> ConnectionUpdateType:
|
||||
r"""
|
||||
Whether this is a connection or disconnection event
|
||||
"""
|
||||
@property
|
||||
def peer_id(self) -> PeerId:
|
||||
r"""
|
||||
Identity of the peer that we have connected to or disconnected from.
|
||||
"""
|
||||
@property
|
||||
def remote_ipv4(self) -> builtins.str:
|
||||
r"""
|
||||
Remote connection's IPv4 address.
|
||||
"""
|
||||
@property
|
||||
def remote_tcp_port(self) -> builtins.int:
|
||||
r"""
|
||||
Remote connection's TCP port.
|
||||
"""
|
||||
|
||||
@typing.final
|
||||
class Keypair:
|
||||
r"""
|
||||
Identity keypair of a node.
|
||||
"""
|
||||
@staticmethod
|
||||
def generate_ed25519() -> Keypair:
|
||||
def generate() -> Keypair:
|
||||
r"""
|
||||
Generate a new Ed25519 keypair.
|
||||
"""
|
||||
@staticmethod
|
||||
def generate_ecdsa() -> Keypair:
|
||||
r"""
|
||||
Generate a new ECDSA keypair.
|
||||
"""
|
||||
@staticmethod
|
||||
def generate_secp256k1() -> Keypair:
|
||||
r"""
|
||||
Generate a new Secp256k1 keypair.
|
||||
Generate a new ed25519 keypair
|
||||
"""
|
||||
@staticmethod
|
||||
def from_protobuf_encoding(bytes: bytes) -> Keypair:
|
||||
r"""
|
||||
Decode a private key from a protobuf structure and parse it as a `Keypair`.
|
||||
"""
|
||||
@staticmethod
|
||||
def rsa_from_pkcs8(bytes: bytes) -> Keypair:
|
||||
r"""
|
||||
Decode an keypair from a DER-encoded secret key in PKCS#8 `PrivateKeyInfo`
|
||||
format (i.e. unencrypted) as defined in [RFC5208].
|
||||
|
||||
[RFC5208]: https://tools.ietf.org/html/rfc5208#section-5
|
||||
"""
|
||||
@staticmethod
|
||||
def secp256k1_from_der(bytes: bytes) -> Keypair:
|
||||
r"""
|
||||
Decode a keypair from a DER-encoded Secp256k1 secret key in an `ECPrivateKey`
|
||||
structure as defined in [RFC5915].
|
||||
|
||||
[RFC5915]: https://tools.ietf.org/html/rfc5915
|
||||
"""
|
||||
@staticmethod
|
||||
def ed25519_from_bytes(bytes: bytes) -> Keypair: ...
|
||||
def to_protobuf_encoding(self) -> bytes:
|
||||
r"""
|
||||
Encode a private key as protobuf structure.
|
||||
"""
|
||||
def to_peer_id(self) -> PeerId:
|
||||
r"""
|
||||
Convert the `Keypair` into the corresponding `PeerId`.
|
||||
Encode a private key to a protobuf structure.
|
||||
"""
|
||||
def to_string(self) -> builtins.str: ...
|
||||
|
||||
@typing.final
|
||||
class Multiaddr:
|
||||
r"""
|
||||
Representation of a Multiaddr.
|
||||
"""
|
||||
class PyPeer:
|
||||
@staticmethod
|
||||
def empty() -> Multiaddr:
|
||||
r"""
|
||||
Create a new, empty multiaddress.
|
||||
"""
|
||||
@staticmethod
|
||||
def with_capacity(n: builtins.int) -> Multiaddr:
|
||||
r"""
|
||||
Create a new, empty multiaddress with the given capacity.
|
||||
"""
|
||||
@staticmethod
|
||||
def from_bytes(bytes: bytes) -> Multiaddr:
|
||||
r"""
|
||||
Parse a `Multiaddr` value from its byte slice representation.
|
||||
"""
|
||||
@staticmethod
|
||||
def from_string(string: builtins.str) -> Multiaddr:
|
||||
r"""
|
||||
Parse a `Multiaddr` value from its string representation.
|
||||
"""
|
||||
def len(self) -> builtins.int:
|
||||
r"""
|
||||
Return the length in bytes of this multiaddress.
|
||||
"""
|
||||
def is_empty(self) -> builtins.bool:
|
||||
r"""
|
||||
Returns true if the length of this multiaddress is 0.
|
||||
"""
|
||||
def to_bytes(self) -> bytes:
|
||||
r"""
|
||||
Return a copy of this [`Multiaddr`]'s byte representation.
|
||||
"""
|
||||
def to_string(self) -> builtins.str:
|
||||
r"""
|
||||
Convert a Multiaddr to a string.
|
||||
"""
|
||||
def new(kp: Keypair, namespace: builtins.str) -> PyPeer: ...
|
||||
async def subscribe(self, topic: builtins.str) -> None: ...
|
||||
async def unsubscribe(self, topic: builtins.str) -> None: ...
|
||||
async def send(self, topic: builtins.str, payload: bytes) -> None: ...
|
||||
async def run(self) -> None: ...
|
||||
async def recv(self) -> PySwarmEvent: ...
|
||||
|
||||
@typing.final
|
||||
class NetworkingHandle:
|
||||
def __new__(cls, identity: Keypair) -> NetworkingHandle: ...
|
||||
async def connection_update_recv(self) -> ConnectionUpdate:
|
||||
r"""
|
||||
Receives the next `ConnectionUpdate` from networking.
|
||||
"""
|
||||
async def connection_update_recv_many(self, limit: builtins.int) -> builtins.list[ConnectionUpdate]:
|
||||
r"""
|
||||
Receives at most `limit` `ConnectionUpdate`s from networking and returns them.
|
||||
|
||||
For `limit = 0`, an empty collection of `ConnectionUpdate`s will be returned immediately.
|
||||
For `limit > 0`, if there are no `ConnectionUpdate`s in the channel's queue this method
|
||||
will sleep until a `ConnectionUpdate`s is sent.
|
||||
"""
|
||||
async def gossipsub_subscribe(self, topic: builtins.str) -> builtins.bool:
|
||||
r"""
|
||||
Subscribe to a `GossipSub` topic.
|
||||
|
||||
Returns `True` if the subscription worked. Returns `False` if we were already subscribed.
|
||||
"""
|
||||
async def gossipsub_unsubscribe(self, topic: builtins.str) -> builtins.bool:
|
||||
r"""
|
||||
Unsubscribes from a `GossipSub` topic.
|
||||
|
||||
Returns `True` if we were subscribed to this topic. Returns `False` if we were not subscribed.
|
||||
"""
|
||||
async def gossipsub_publish(self, topic: builtins.str, data: bytes) -> None:
|
||||
r"""
|
||||
Publishes a message with multiple topics to the `GossipSub` network.
|
||||
|
||||
If no peers are found that subscribe to this topic, throws `NoPeersSubscribedToTopicError` exception.
|
||||
"""
|
||||
async def gossipsub_recv(self) -> tuple[builtins.str, bytes]:
|
||||
r"""
|
||||
Receives the next message from the `GossipSub` network.
|
||||
"""
|
||||
async def gossipsub_recv_many(self, limit: builtins.int) -> builtins.list[tuple[builtins.str, bytes]]:
|
||||
r"""
|
||||
Receives at most `limit` messages from the `GossipSub` network and returns them.
|
||||
|
||||
For `limit = 0`, an empty collection of messages will be returned immediately.
|
||||
For `limit > 0`, if there are no messages in the channel's queue this method
|
||||
will sleep until a message is sent.
|
||||
"""
|
||||
|
||||
@typing.final
|
||||
class NoPeersSubscribedToTopicError(builtins.Exception):
|
||||
def __new__(cls, *args: typing.Any) -> NoPeersSubscribedToTopicError: ...
|
||||
def __repr__(self) -> builtins.str: ...
|
||||
def __str__(self) -> builtins.str: ...
|
||||
|
||||
@typing.final
|
||||
class PeerId:
|
||||
r"""
|
||||
Identifier of a peer of the network.
|
||||
|
||||
The data is a `CIDv0` compatible multihash of the protobuf encoded public key of the peer
|
||||
as specified in [specs/peer-ids](https://github.com/libp2p/specs/blob/master/peer-ids/peer-ids.md).
|
||||
"""
|
||||
@staticmethod
|
||||
def random() -> PeerId:
|
||||
r"""
|
||||
Generates a random peer ID from a cryptographically secure PRNG.
|
||||
|
||||
This is useful for randomly walking on a DHT, or for testing purposes.
|
||||
"""
|
||||
@staticmethod
|
||||
def from_bytes(bytes: bytes) -> PeerId:
|
||||
r"""
|
||||
Parses a `PeerId` from bytes.
|
||||
"""
|
||||
def to_bytes(self) -> bytes:
|
||||
r"""
|
||||
Returns a raw bytes representation of this `PeerId`.
|
||||
"""
|
||||
def to_base58(self) -> builtins.str:
|
||||
r"""
|
||||
Returns a base-58 encoded string of this `PeerId`.
|
||||
"""
|
||||
def __repr__(self) -> builtins.str: ...
|
||||
def __str__(self) -> builtins.str: ...
|
||||
|
||||
@typing.final
|
||||
class ConnectionUpdateType(enum.Enum):
|
||||
r"""
|
||||
Connection or disconnection event discriminant type.
|
||||
"""
|
||||
Connected = ...
|
||||
Disconnected = ...
|
||||
class PySwarmEvent:
|
||||
def downcast_discovered(self) -> typing.Optional[builtins.str]: ...
|
||||
def downcast_expired(self) -> typing.Optional[builtins.str]: ...
|
||||
def downcast_message(self) -> typing.Optional[tuple[builtins.str, builtins.str, bytes]]: ...
|
||||
|
||||
|
||||
@@ -1,8 +1,4 @@
|
||||
//! SEE: https://pyo3.rs/v0.26.0/async-await.html#detaching-from-the-interpreter-across-await
|
||||
//!
|
||||
|
||||
use pin_project::pin_project;
|
||||
use pyo3::marker::Ungil;
|
||||
//! See: <https://pyo3.rs/v0.27.2/async-await.html#detaching-from-the-interpreter-across-await>
|
||||
use pyo3::prelude::*;
|
||||
use std::{
|
||||
future::Future,
|
||||
@@ -10,31 +6,17 @@ use std::{
|
||||
task::{Context, Poll},
|
||||
};
|
||||
|
||||
/// SEE: https://pyo3.rs/v0.26.0/async-await.html#detaching-from-the-interpreter-across-await
|
||||
#[pin_project]
|
||||
#[repr(transparent)]
|
||||
pub(crate) struct AllowThreads<F>(#[pin] F);
|
||||
|
||||
impl<F> AllowThreads<F>
|
||||
where
|
||||
Self: Future,
|
||||
{
|
||||
pub fn new(f: F) -> Self {
|
||||
Self(f)
|
||||
}
|
||||
}
|
||||
pub struct AllowThreads<F>(pub(crate) F);
|
||||
|
||||
impl<F> Future for AllowThreads<F>
|
||||
where
|
||||
F: Future + Ungil,
|
||||
F::Output: Ungil,
|
||||
F: Future + Unpin + Send,
|
||||
F::Output: Send,
|
||||
{
|
||||
type Output = F::Output;
|
||||
|
||||
fn poll(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Self::Output> {
|
||||
fn poll(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Self::Output> {
|
||||
let waker = cx.waker();
|
||||
Python::with_gil(|py| {
|
||||
py.allow_threads(|| self.project().0.poll(&mut Context::from_waker(waker)))
|
||||
})
|
||||
Python::attach(|py| py.detach(|| pin!(&mut self.0).poll(&mut Context::from_waker(waker))))
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,240 +0,0 @@
|
||||
//! This module exists to hold examples of some pyo3 patterns that may be too complex to
|
||||
//! re-create from scratch, but too inhomogenous to create an abstraction/wrapper around.
|
||||
//!
|
||||
//! Pattern examples include:
|
||||
//! - Async task handles: with GC-integrated cleanup
|
||||
//! - Sync/async callbacks from python: with propper eventloop handling
|
||||
//!
|
||||
//! Mutability pattern: https://pyo3.rs/v0.26.0/async-await.html#send--static-constraint
|
||||
//! - Store mutable fields in tokio's `Mutex<T>`
|
||||
//! - For async code: take `&self` and `.lock().await`
|
||||
//! - For sync code: take `&mut self` and `.get_mut()`
|
||||
|
||||
use crate::ext::{PyResultExt as _, ResultExt as _, TokioRuntimeExt as _};
|
||||
use futures::FutureExt as _;
|
||||
use futures::future::BoxFuture;
|
||||
use pyo3::exceptions::PyRuntimeError;
|
||||
use pyo3::prelude::{PyModule, PyModuleMethods as _};
|
||||
use pyo3::{
|
||||
Bound, Py, PyAny, PyErr, PyResult, PyTraverseError, PyVisit, Python, pyclass, pymethods,
|
||||
};
|
||||
use std::time::Duration;
|
||||
use tokio::sync::mpsc;
|
||||
use tokio::sync::mpsc::error::TryRecvError;
|
||||
|
||||
fn needs_tokio_runtime() {
|
||||
tokio::runtime::Handle::current();
|
||||
}
|
||||
|
||||
type SyncCallback = Box<dyn Fn() + Send + Sync>;
|
||||
type AsyncCallback = Box<dyn Fn() -> BoxFuture<'static, ()> + Send + Sync>;
|
||||
|
||||
enum AsyncTaskMessage {
|
||||
SyncCallback(SyncCallback),
|
||||
AsyncCallback(AsyncCallback),
|
||||
}
|
||||
|
||||
async fn async_task(
|
||||
sender: mpsc::UnboundedSender<()>,
|
||||
mut receiver: mpsc::UnboundedReceiver<AsyncTaskMessage>,
|
||||
) {
|
||||
log::info!("RUST: async task started");
|
||||
|
||||
// task state
|
||||
let mut interval = tokio::time::interval(Duration::from_secs(1));
|
||||
|
||||
let mut sync_cbs: Vec<SyncCallback> = vec![];
|
||||
let mut async_cbs: Vec<AsyncCallback> = vec![];
|
||||
|
||||
loop {
|
||||
tokio::select! {
|
||||
// handle incoming messages from task-handle
|
||||
message = receiver.recv() => {
|
||||
// handle closed channel by exiting
|
||||
let Some(message) = message else {
|
||||
log::info!("RUST: channel closed");
|
||||
break;
|
||||
};
|
||||
|
||||
// dispatch incoming event
|
||||
match message {
|
||||
AsyncTaskMessage::SyncCallback(cb) => {
|
||||
sync_cbs.push(cb);
|
||||
}
|
||||
AsyncTaskMessage::AsyncCallback(cb) => {
|
||||
async_cbs.push(cb);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// handle all other events
|
||||
_ = interval.tick() => {
|
||||
log::info!("RUST: async task tick");
|
||||
|
||||
// call back all sync callbacks
|
||||
for cb in &sync_cbs {
|
||||
cb();
|
||||
}
|
||||
|
||||
// call back all async callbacks
|
||||
for cb in &async_cbs {
|
||||
cb().await;
|
||||
}
|
||||
|
||||
// send event on unbounded channel
|
||||
sender.send(()).expect("handle receiver cannot be closed/dropped");
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
log::info!("RUST: async task stopped");
|
||||
}
|
||||
|
||||
// #[gen_stub_pyclass]
|
||||
#[pyclass(name = "AsyncTaskHandle")]
|
||||
#[derive(Debug)]
|
||||
struct PyAsyncTaskHandle {
|
||||
sender: Option<mpsc::UnboundedSender<AsyncTaskMessage>>,
|
||||
receiver: mpsc::UnboundedReceiver<()>,
|
||||
}
|
||||
|
||||
#[allow(clippy::expect_used)]
|
||||
impl PyAsyncTaskHandle {
|
||||
const fn sender(&self) -> &mpsc::UnboundedSender<AsyncTaskMessage> {
|
||||
self.sender
|
||||
.as_ref()
|
||||
.expect("The sender should only be None after de-initialization.")
|
||||
}
|
||||
|
||||
const fn sender_mut(&mut self) -> &mpsc::UnboundedSender<AsyncTaskMessage> {
|
||||
self.sender
|
||||
.as_mut()
|
||||
.expect("The sender should only be None after de-initialization.")
|
||||
}
|
||||
|
||||
const fn new(
|
||||
sender: mpsc::UnboundedSender<AsyncTaskMessage>,
|
||||
receiver: mpsc::UnboundedReceiver<()>,
|
||||
) -> Self {
|
||||
Self {
|
||||
sender: Some(sender),
|
||||
receiver,
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// #[gen_stub_pymethods]
|
||||
#[pymethods]
|
||||
impl PyAsyncTaskHandle {
|
||||
#[new]
|
||||
fn py_new(py: Python<'_>) -> PyResult<Self> {
|
||||
use pyo3_async_runtimes::tokio::get_runtime;
|
||||
|
||||
// create communication channel TOWARDS our task
|
||||
let (h_sender, t_receiver) = mpsc::unbounded_channel::<AsyncTaskMessage>();
|
||||
|
||||
// create communication channel FROM our task
|
||||
let (t_sender, h_receiver) = mpsc::unbounded_channel::<()>();
|
||||
|
||||
// perform necessary setup within tokio context - or it crashes
|
||||
let () = get_runtime().block_on(async { needs_tokio_runtime() });
|
||||
|
||||
// spawn tokio task with this thread's task-locals - without this, async callbacks on the new threads will not work!!
|
||||
_ = get_runtime().spawn_with_scope(py, async move {
|
||||
async_task(t_sender, t_receiver).await;
|
||||
});
|
||||
Ok(Self::new(h_sender, h_receiver))
|
||||
}
|
||||
|
||||
/// NOTE: exceptions in callbacks are silently ignored until end of execution
|
||||
fn add_sync_callback(
|
||||
&self,
|
||||
// #[gen_stub(override_type(
|
||||
// type_repr="collections.abc.Callable[[], None]",
|
||||
// imports=("collections.abc")
|
||||
// ))]
|
||||
callback: Py<PyAny>,
|
||||
) -> PyResult<()> {
|
||||
// blocking call to async method -> can do non-blocking if needed
|
||||
self.sender()
|
||||
.send(AsyncTaskMessage::SyncCallback(Box::new(move || {
|
||||
_ = Python::with_gil(|py| callback.call0(py).write_unraisable_with(py));
|
||||
})))
|
||||
.pyerr()?;
|
||||
Ok(())
|
||||
}
|
||||
|
||||
/// NOTE: exceptions in callbacks are silently ignored until end of execution
|
||||
fn add_async_callback(
|
||||
&self,
|
||||
// #[gen_stub(override_type(
|
||||
// type_repr="collections.abc.Callable[[], collections.abc.Awaitable[None]]",
|
||||
// imports=("collections.abc")
|
||||
// ))]
|
||||
callback: Py<PyAny>,
|
||||
) -> PyResult<()> {
|
||||
// blocking call to async method -> can do non-blocking if needed
|
||||
self.sender()
|
||||
.send(AsyncTaskMessage::AsyncCallback(Box::new(move || {
|
||||
let c = Python::with_gil(|py| callback.clone_ref(py));
|
||||
async move {
|
||||
if let Some(f) = Python::with_gil(|py| {
|
||||
let coroutine = c.call0(py).write_unraisable_with(py)?;
|
||||
pyo3_async_runtimes::tokio::into_future(coroutine.into_bound(py))
|
||||
.write_unraisable_with(py)
|
||||
}) {
|
||||
_ = f.await.write_unraisable();
|
||||
}
|
||||
}
|
||||
.boxed()
|
||||
})))
|
||||
.pyerr()?;
|
||||
Ok(())
|
||||
}
|
||||
|
||||
async fn receive_unit(&mut self) -> PyResult<()> {
|
||||
self.receiver
|
||||
.recv()
|
||||
.await
|
||||
.ok_or(PyErr::new::<PyRuntimeError, _>(
|
||||
"cannot receive unit on closed channel",
|
||||
))
|
||||
}
|
||||
|
||||
fn drain_units(&mut self) -> PyResult<i32> {
|
||||
let mut cnt = 0;
|
||||
loop {
|
||||
match self.receiver.try_recv() {
|
||||
Err(TryRecvError::Disconnected) => {
|
||||
return Err(PyErr::new::<PyRuntimeError, _>(
|
||||
"cannot receive unit on closed channel",
|
||||
));
|
||||
}
|
||||
Err(TryRecvError::Empty) => return Ok(cnt),
|
||||
Ok(()) => {
|
||||
cnt += 1;
|
||||
continue;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// #[gen_stub(skip)]
|
||||
const fn __traverse__(&self, _visit: PyVisit<'_>) -> Result<(), PyTraverseError> {
|
||||
Ok(()) // This is needed purely so `__clear__` can work
|
||||
}
|
||||
|
||||
// #[gen_stub(skip)]
|
||||
fn __clear__(&mut self) {
|
||||
// TODO: may or may not need to await a "kill-signal" oneshot channel message,
|
||||
// to ensure that the networking task is done BEFORE exiting the clear function...
|
||||
// but this may require GIL?? and it may not be safe to call GIL here??
|
||||
self.sender = None; // Using Option<T> as a trick to force `sender` channel to be dropped
|
||||
}
|
||||
}
|
||||
|
||||
pub fn examples_submodule(m: &Bound<'_, PyModule>) -> PyResult<()> {
|
||||
m.add_class::<PyAsyncTaskHandle>()?;
|
||||
|
||||
Ok(())
|
||||
}
|
||||
@@ -1,216 +1,42 @@
|
||||
//! TODO: crate documentation
|
||||
//!
|
||||
//! this is here as a placeholder documentation
|
||||
//!
|
||||
//!
|
||||
pub(crate) mod allow_threading;
|
||||
|
||||
// enable Rust-unstable features for convenience
|
||||
#![feature(trait_alias)]
|
||||
#![feature(tuple_trait)]
|
||||
#![feature(unboxed_closures)]
|
||||
// #![feature(stmt_expr_attributes)]
|
||||
// #![feature(assert_matches)]
|
||||
// #![feature(async_fn_in_dyn_trait)]
|
||||
// #![feature(async_for_loop)]
|
||||
// #![feature(auto_traits)]
|
||||
// #![feature(negative_impls)]
|
||||
|
||||
extern crate core;
|
||||
mod allow_threading;
|
||||
mod examples;
|
||||
pub(crate) mod networking;
|
||||
pub(crate) mod pylibp2p;
|
||||
pub(crate) mod take_once {
|
||||
use std::sync::Mutex;
|
||||
|
||||
pub struct TakeOnce<T>(Mutex<Option<T>>);
|
||||
impl<T> TakeOnce<T> {
|
||||
pub fn new(t: T) -> Self {
|
||||
Self(Mutex::new(Some(t)))
|
||||
}
|
||||
pub fn take(&self) -> Option<T> {
|
||||
match self.0.try_lock() {
|
||||
Ok(mut o) => o.take(),
|
||||
Err(_) => None,
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
use crate::networking::networking_submodule;
|
||||
use crate::pylibp2p::ident::ident_submodule;
|
||||
use crate::pylibp2p::multiaddr::multiaddr_submodule;
|
||||
use pyo3::prelude::PyModule;
|
||||
use pyo3::prelude::*;
|
||||
use pyo3::{Bound, PyResult, pyclass, pymodule};
|
||||
|
||||
use pyo3_stub_gen::define_stub_info_gatherer;
|
||||
|
||||
/// Namespace for all the constants used by this crate.
|
||||
pub(crate) mod r#const {
|
||||
pub const MPSC_CHANNEL_SIZE: usize = 1024;
|
||||
}
|
||||
|
||||
/// Namespace for all the type/trait aliases used by this crate.
|
||||
pub(crate) mod alias {
|
||||
use std::error::Error;
|
||||
use std::marker::Tuple;
|
||||
|
||||
pub trait SendFn<Args: Tuple + Send + 'static, Output> =
|
||||
Fn<Args, Output = Output> + Send + 'static;
|
||||
|
||||
pub type AnyError = Box<dyn Error + Send + Sync + 'static>;
|
||||
pub type AnyResult<T> = Result<T, AnyError>;
|
||||
}
|
||||
|
||||
/// Namespace for crate-wide extension traits/methods
|
||||
pub(crate) mod ext {
|
||||
use crate::allow_threading::AllowThreads;
|
||||
use extend::ext;
|
||||
use pyo3::exceptions::{PyConnectionError, PyRuntimeError};
|
||||
use pyo3::marker::Ungil;
|
||||
use pyo3::types::PyBytes;
|
||||
use pyo3::{Py, PyErr, PyResult, Python};
|
||||
use tokio::runtime::Runtime;
|
||||
use tokio::sync::mpsc;
|
||||
use tokio::sync::mpsc::error::TryRecvError;
|
||||
use tokio::task::JoinHandle;
|
||||
|
||||
#[ext(pub, name = ByteArrayExt)]
|
||||
impl [u8] {
|
||||
fn pybytes(&self) -> Py<PyBytes> {
|
||||
Python::with_gil(|py| PyBytes::new(py, self).unbind())
|
||||
}
|
||||
}
|
||||
|
||||
#[ext(pub, name = ResultExt)]
|
||||
impl<T, E> Result<T, E>
|
||||
where
|
||||
E: ToString,
|
||||
{
|
||||
fn pyerr(self) -> PyResult<T> {
|
||||
self.map_err(|e| PyRuntimeError::new_err(e.to_string()))
|
||||
}
|
||||
}
|
||||
|
||||
pub trait FutureExt: Future + Sized {
|
||||
/// SEE: https://pyo3.rs/v0.26.0/async-await.html#detaching-from-the-interpreter-across-await
|
||||
fn allow_threads_py(self) -> AllowThreads<Self>
|
||||
where
|
||||
AllowThreads<Self>: Future,
|
||||
{
|
||||
AllowThreads::new(self)
|
||||
}
|
||||
}
|
||||
|
||||
impl<T: Future> FutureExt for T {}
|
||||
|
||||
#[ext(pub, name = PyErrExt)]
|
||||
impl PyErr {
|
||||
fn receiver_channel_closed() -> Self {
|
||||
PyConnectionError::new_err("Receiver channel closed unexpectedly")
|
||||
}
|
||||
}
|
||||
|
||||
#[ext(pub, name = PyResultExt)]
|
||||
impl<T> PyResult<T> {
|
||||
fn write_unraisable(self) -> Option<T> {
|
||||
Python::with_gil(|py| self.write_unraisable_with(py))
|
||||
}
|
||||
|
||||
fn write_unraisable_with(self, py: Python<'_>) -> Option<T> {
|
||||
match self {
|
||||
Ok(v) => Some(v),
|
||||
Err(e) => {
|
||||
// write error back to python
|
||||
e.write_unraisable(py, None);
|
||||
None
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#[ext(pub, name = TokioRuntimeExt)]
|
||||
impl Runtime {
|
||||
fn spawn_with_scope<F>(&self, py: Python<'_>, future: F) -> PyResult<JoinHandle<F::Output>>
|
||||
where
|
||||
F: Future + Send + 'static,
|
||||
F::Output: Send + 'static,
|
||||
{
|
||||
let locals = pyo3_async_runtimes::tokio::get_current_locals(py)?;
|
||||
Ok(self.spawn(pyo3_async_runtimes::tokio::scope(locals, future)))
|
||||
}
|
||||
}
|
||||
|
||||
#[ext(pub, name = TokioMpscSenderExt)]
|
||||
impl<T> mpsc::Sender<T> {
|
||||
/// Sends a value, waiting until there is capacity.
|
||||
///
|
||||
/// A successful send occurs when it is determined that the other end of the
|
||||
/// channel has not hung up already. An unsuccessful send would be one where
|
||||
/// the corresponding receiver has already been closed.
|
||||
async fn send_py(&self, value: T) -> PyResult<()> {
|
||||
self.send(value)
|
||||
.await
|
||||
.map_err(|_| PyErr::receiver_channel_closed())
|
||||
}
|
||||
}
|
||||
|
||||
#[ext(pub, name = TokioMpscReceiverExt)]
|
||||
impl<T> mpsc::Receiver<T> {
|
||||
/// Receives the next value for this receiver.
|
||||
async fn recv_py(&mut self) -> PyResult<T> {
|
||||
self.recv().await.ok_or_else(PyErr::receiver_channel_closed)
|
||||
}
|
||||
|
||||
/// Receives at most `limit` values for this receiver and returns them.
|
||||
///
|
||||
/// For `limit = 0`, an empty collection of messages will be returned immediately.
|
||||
/// For `limit > 0`, if there are no messages in the channel's queue this method
|
||||
/// will sleep until a message is sent.
|
||||
async fn recv_many_py(&mut self, limit: usize) -> PyResult<Vec<T>> {
|
||||
// get updates from receiver channel
|
||||
let mut updates = Vec::with_capacity(limit);
|
||||
let received = self.recv_many(&mut updates, limit).await;
|
||||
|
||||
// if we received zero items, then the channel was unexpectedly closed
|
||||
if limit != 0 && received == 0 {
|
||||
return Err(PyErr::receiver_channel_closed());
|
||||
}
|
||||
|
||||
Ok(updates)
|
||||
}
|
||||
|
||||
/// Tries to receive the next value for this receiver.
|
||||
fn try_recv_py(&mut self) -> PyResult<Option<T>> {
|
||||
match self.try_recv() {
|
||||
Ok(v) => Ok(Some(v)),
|
||||
Err(TryRecvError::Empty) => Ok(None),
|
||||
Err(TryRecvError::Disconnected) => Err(PyErr::receiver_channel_closed()),
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
pub(crate) mod private {
|
||||
use std::marker::Sized;
|
||||
|
||||
/// Sealed traits support
|
||||
pub trait Sealed {}
|
||||
impl<T: ?Sized> Sealed for T {}
|
||||
}
|
||||
|
||||
/// A wrapper around [`Py`] that implements [`Clone`] using [`Python::with_gil`].
|
||||
#[repr(transparent)]
|
||||
pub(crate) struct ClonePy<T>(pub Py<T>);
|
||||
|
||||
impl<T> Clone for ClonePy<T> {
|
||||
fn clone(&self) -> Self {
|
||||
Python::with_gil(|py| Self(self.0.clone_ref(py)))
|
||||
}
|
||||
}
|
||||
|
||||
/// A Python module implemented in Rust. The name of this function must match
|
||||
/// the `lib.name` setting in the `Cargo.toml`, else Python will not be able to
|
||||
/// import the module.
|
||||
#[pymodule(name = "exo_pyo3_bindings")]
|
||||
fn main_module(m: &Bound<'_, PyModule>) -> PyResult<()> {
|
||||
pub fn networking_module(m: &Bound<'_, PyModule>) -> PyResult<()> {
|
||||
// install logger
|
||||
pyo3_log::init();
|
||||
// setup runtime
|
||||
let mut builder = tokio::runtime::Builder::new_multi_thread();
|
||||
builder.enable_all();
|
||||
pyo3_async_runtimes::tokio::init(builder);
|
||||
|
||||
// TODO: for now this is all NOT a submodule, but figure out how to make the submodule system
|
||||
// work with maturin, where the types generate correctly, in the right folder, without
|
||||
// too many importing issues...
|
||||
ident_submodule(m)?;
|
||||
multiaddr_submodule(m)?;
|
||||
networking_submodule(m)?;
|
||||
|
||||
// top-level constructs
|
||||
// TODO: ...
|
||||
|
||||
m.add_class::<networking::PyPeer>()?;
|
||||
m.add_class::<networking::PyKeypair>()?;
|
||||
Ok(())
|
||||
}
|
||||
|
||||
|
||||
@@ -1,572 +1,214 @@
|
||||
#![allow(
|
||||
clippy::multiple_inherent_impl,
|
||||
clippy::unnecessary_wraps,
|
||||
clippy::unused_self,
|
||||
clippy::needless_pass_by_value
|
||||
)]
|
||||
use crate::allow_threading::AllowThreads;
|
||||
use crate::take_once::TakeOnce;
|
||||
|
||||
use crate::r#const::MPSC_CHANNEL_SIZE;
|
||||
use crate::ext::{ByteArrayExt as _, FutureExt, PyErrExt as _};
|
||||
use crate::ext::{ResultExt as _, TokioMpscReceiverExt as _, TokioMpscSenderExt as _};
|
||||
use crate::pyclass;
|
||||
use crate::pylibp2p::ident::{PyKeypair, PyPeerId};
|
||||
use libp2p::futures::StreamExt as _;
|
||||
use libp2p::gossipsub::{IdentTopic, Message, MessageId, PublishError};
|
||||
use libp2p::swarm::SwarmEvent;
|
||||
use libp2p::{gossipsub, mdns};
|
||||
use networking::discovery;
|
||||
use networking::swarm::create_swarm;
|
||||
use pyo3::prelude::{PyModule, PyModuleMethods as _};
|
||||
use pyo3::types::PyBytes;
|
||||
use pyo3::{Bound, Py, PyErr, PyResult, PyTraverseError, PyVisit, Python, pymethods};
|
||||
use pyo3_stub_gen::derive::{gen_stub_pyclass, gen_stub_pyclass_enum, gen_stub_pymethods};
|
||||
use std::net::IpAddr;
|
||||
use tokio::sync::{Mutex, mpsc, oneshot};
|
||||
use std::pin::pin;
|
||||
|
||||
mod exception {
|
||||
use pyo3::types::PyTuple;
|
||||
use pyo3::{PyErrArguments, exceptions::PyException, prelude::*};
|
||||
use pyo3_stub_gen::derive::*;
|
||||
|
||||
#[gen_stub_pyclass]
|
||||
#[pyclass(frozen, extends=PyException, name="NoPeersSubscribedToTopicError")]
|
||||
pub struct PyNoPeersSubscribedToTopicError {}
|
||||
|
||||
impl PyNoPeersSubscribedToTopicError {
|
||||
const MSG: &'static str = "\
|
||||
No peers are currently subscribed to receive messages on this topic. \
|
||||
Wait for peers to subscribe or check your network connectivity.";
|
||||
|
||||
/// Creates a new [ `PyErr` ] of this type.
|
||||
///
|
||||
/// [`PyErr`] : https://docs.rs/pyo3/latest/pyo3/struct.PyErr.html "PyErr in pyo3"
|
||||
pub(crate) fn new_err() -> PyErr {
|
||||
PyErr::new::<Self, _>(()) // TODO: check if this needs to be replaced???
|
||||
}
|
||||
}
|
||||
|
||||
#[gen_stub_pymethods]
|
||||
#[pymethods]
|
||||
impl PyNoPeersSubscribedToTopicError {
|
||||
#[new]
|
||||
#[pyo3(signature = (*args))]
|
||||
#[allow(unused_variables)]
|
||||
pub(crate) fn new(args: &Bound<'_, PyTuple>) -> Self {
|
||||
Self {}
|
||||
}
|
||||
|
||||
fn __repr__(&self) -> String {
|
||||
format!("PeerId(\"{}\")", Self::MSG)
|
||||
}
|
||||
|
||||
fn __str__(&self) -> String {
|
||||
Self::MSG.to_string()
|
||||
}
|
||||
}
|
||||
|
||||
#[gen_stub_pyclass]
|
||||
#[pyclass(frozen, extends=PyException, name="AllQueuesFullError")]
|
||||
pub struct PyAllQueuesFullError {}
|
||||
|
||||
impl PyAllQueuesFullError {
|
||||
const MSG: &'static str =
|
||||
"All libp2p peers are unresponsive, resend the message or reconnect.";
|
||||
|
||||
/// Creates a new [ `PyErr` ] of this type.
|
||||
///
|
||||
/// [`PyErr`] : https://docs.rs/pyo3/latest/pyo3/struct.PyErr.html "PyErr in pyo3"
|
||||
pub(crate) fn new_err() -> PyErr {
|
||||
PyErr::new::<Self, _>(()) // TODO: check if this needs to be replaced???
|
||||
}
|
||||
}
|
||||
|
||||
#[gen_stub_pymethods]
|
||||
#[pymethods]
|
||||
impl PyAllQueuesFullError {
|
||||
#[new]
|
||||
#[pyo3(signature = (*args))]
|
||||
#[allow(unused_variables)]
|
||||
pub(crate) fn new(args: &Bound<'_, PyTuple>) -> Self {
|
||||
Self {}
|
||||
}
|
||||
|
||||
fn __repr__(&self) -> String {
|
||||
format!("PeerId(\"{}\")", Self::MSG)
|
||||
}
|
||||
|
||||
fn __str__(&self) -> String {
|
||||
Self::MSG.to_string()
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// Connection or disconnection event discriminant type.
|
||||
#[gen_stub_pyclass_enum]
|
||||
#[pyclass(eq, eq_int, name = "ConnectionUpdateType")]
|
||||
#[derive(Debug, Clone, PartialEq)]
|
||||
enum PyConnectionUpdateType {
|
||||
Connected = 0,
|
||||
Disconnected,
|
||||
}
|
||||
use futures_lite::FutureExt;
|
||||
use libp2p::{gossipsub::PublishError, identity::Keypair};
|
||||
use networking::{FromSwarm, Peer, ToSwarm};
|
||||
use pyo3::{
|
||||
coroutine::CancelHandle,
|
||||
exceptions::{PyConnectionError, PyRuntimeError, PyValueError},
|
||||
prelude::*,
|
||||
types::PyBytes,
|
||||
};
|
||||
use pyo3_stub_gen::{
|
||||
derive::{gen_methods_from_python, gen_stub_pyclass, gen_stub_pymethods},
|
||||
inventory::submit,
|
||||
};
|
||||
use tokio::sync::{Mutex, mpsc};
|
||||
|
||||
#[gen_stub_pyclass]
|
||||
#[pyclass(frozen, name = "ConnectionUpdate")]
|
||||
#[derive(Debug, Clone)]
|
||||
struct PyConnectionUpdate {
|
||||
/// Whether this is a connection or disconnection event
|
||||
#[pyo3(get)]
|
||||
update_type: PyConnectionUpdateType,
|
||||
#[pyclass(name = "Keypair", frozen)]
|
||||
#[derive(Clone)]
|
||||
pub struct PyKeypair(Keypair);
|
||||
|
||||
/// Identity of the peer that we have connected to or disconnected from.
|
||||
#[pyo3(get)]
|
||||
peer_id: PyPeerId,
|
||||
#[gen_stub_pymethods]
|
||||
#[pymethods]
|
||||
impl PyKeypair {
|
||||
/// Generate a new ed25519 keypair
|
||||
#[staticmethod]
|
||||
fn generate() -> Self {
|
||||
Self(Keypair::generate_ed25519())
|
||||
}
|
||||
|
||||
/// Remote connection's IPv4 address.
|
||||
#[pyo3(get)]
|
||||
remote_ipv4: String,
|
||||
/// Decode a private key from a protobuf structure and parse it as a `Keypair`.
|
||||
#[staticmethod]
|
||||
fn from_protobuf_encoding(bytes: &Bound<'_, PyBytes>) -> Self {
|
||||
let bytes = Vec::from(bytes.as_bytes());
|
||||
Self(Keypair::from_protobuf_encoding(&bytes).expect("todo"))
|
||||
}
|
||||
|
||||
/// Remote connection's TCP port.
|
||||
#[pyo3(get)]
|
||||
remote_tcp_port: u16,
|
||||
}
|
||||
|
||||
enum ToTask {
|
||||
GossipsubSubscribe {
|
||||
topic: String,
|
||||
result_tx: oneshot::Sender<PyResult<bool>>,
|
||||
},
|
||||
GossipsubUnsubscribe {
|
||||
topic: String,
|
||||
result_tx: oneshot::Sender<bool>,
|
||||
},
|
||||
GossipsubPublish {
|
||||
topic: String,
|
||||
data: Vec<u8>,
|
||||
result_tx: oneshot::Sender<PyResult<MessageId>>,
|
||||
},
|
||||
}
|
||||
|
||||
#[allow(clippy::enum_glob_use)]
|
||||
async fn networking_task(
|
||||
mut swarm: networking::swarm::Swarm,
|
||||
mut to_task_rx: mpsc::Receiver<ToTask>,
|
||||
connection_update_tx: mpsc::Sender<PyConnectionUpdate>,
|
||||
gossipsub_message_tx: mpsc::Sender<(String, Vec<u8>)>,
|
||||
) {
|
||||
use SwarmEvent::*;
|
||||
use ToTask::*;
|
||||
use mdns::Event::*;
|
||||
use networking::swarm::BehaviourEvent::*;
|
||||
|
||||
log::info!("RUST: networking task started");
|
||||
|
||||
loop {
|
||||
tokio::select! {
|
||||
message = to_task_rx.recv() => {
|
||||
// handle closed channel
|
||||
let Some(message) = message else {
|
||||
log::info!("RUST: channel closed");
|
||||
break;
|
||||
};
|
||||
|
||||
// dispatch incoming messages
|
||||
match message {
|
||||
GossipsubSubscribe { topic, result_tx } => {
|
||||
// try to subscribe
|
||||
let result = swarm.behaviour_mut()
|
||||
.gossipsub.subscribe(&IdentTopic::new(topic));
|
||||
|
||||
// send response oneshot
|
||||
if let Err(e) = result_tx.send(result.pyerr()) {
|
||||
log::error!("RUST: could not subscribe to gossipsub topic since channel already closed: {e:?}");
|
||||
continue;
|
||||
}
|
||||
}
|
||||
GossipsubUnsubscribe { topic, result_tx } => {
|
||||
// try to unsubscribe from the topic
|
||||
let result = swarm.behaviour_mut()
|
||||
.gossipsub.unsubscribe(&IdentTopic::new(topic));
|
||||
|
||||
// send response oneshot (or exit if connection closed)
|
||||
if let Err(e) = result_tx.send(result) {
|
||||
log::error!("RUST: could not unsubscribe from gossipsub topic since channel already closed: {e:?}");
|
||||
continue;
|
||||
}
|
||||
}
|
||||
GossipsubPublish { topic, data, result_tx } => {
|
||||
// try to publish the data -> catch NoPeersSubscribedToTopic error & convert to correct exception
|
||||
let result = swarm.behaviour_mut().gossipsub.publish(
|
||||
IdentTopic::new(topic), data);
|
||||
let pyresult: PyResult<MessageId> = if let Err(PublishError::NoPeersSubscribedToTopic) = result {
|
||||
Err(exception::PyNoPeersSubscribedToTopicError::new_err())
|
||||
} else if let Err(PublishError::AllQueuesFull(_)) = result {
|
||||
Err(exception::PyAllQueuesFullError::new_err())
|
||||
} else {
|
||||
result.pyerr()
|
||||
};
|
||||
|
||||
// send response oneshot (or exit if connection closed)
|
||||
if let Err(e) = result_tx.send(pyresult) {
|
||||
log::error!("RUST: could not publish gossipsub message since channel already closed: {e:?}");
|
||||
continue;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// architectural solution to this problem:
|
||||
// create keep_alive behavior who's job it is to dial peers discovered by mDNS (and drop when expired)
|
||||
// -> it will emmit TRUE connected/disconnected events consumable elsewhere
|
||||
//
|
||||
// gossipsub will feed off-of dial attempts created by networking, and that will bootstrap its' peers list
|
||||
// then for actual communication it will dial those peers if need-be
|
||||
swarm_event = swarm.select_next_some() => {
|
||||
match swarm_event {
|
||||
Behaviour(Gossipsub(gossipsub::Event::Message {
|
||||
message: Message {
|
||||
topic,
|
||||
data,
|
||||
..
|
||||
},
|
||||
..
|
||||
})) => {
|
||||
// topic-ID is just the topic hash!!! (since we used identity hasher)
|
||||
let message = (topic.into_string(), data);
|
||||
|
||||
// send incoming message to channel (or exit if connection closed)
|
||||
if let Err(e) = gossipsub_message_tx.send(message).await {
|
||||
log::error!("RUST: could not send incoming gossipsub message since channel already closed: {e}");
|
||||
continue;
|
||||
}
|
||||
},
|
||||
Behaviour(Discovery(discovery::Event::ConnectionEstablished { peer_id, remote_ip, remote_tcp_port, .. })) => {
|
||||
// grab IPv4 string
|
||||
let remote_ipv4 = match remote_ip {
|
||||
IpAddr::V4(ip) => ip.to_string(),
|
||||
IpAddr::V6(ip) => {
|
||||
log::warn!("RUST: ignoring connection to IPv6 address: {ip}");
|
||||
continue;
|
||||
}
|
||||
};
|
||||
|
||||
// send connection event to channel (or exit if connection closed)
|
||||
if let Err(e) = connection_update_tx.send(PyConnectionUpdate {
|
||||
update_type: PyConnectionUpdateType::Connected,
|
||||
peer_id: PyPeerId(peer_id),
|
||||
remote_ipv4,
|
||||
remote_tcp_port,
|
||||
}).await {
|
||||
log::error!("RUST: could not send connection update since channel already closed: {e}");
|
||||
continue;
|
||||
}
|
||||
},
|
||||
Behaviour(Discovery(discovery::Event::ConnectionClosed { peer_id, remote_ip, remote_tcp_port, .. })) => {
|
||||
// grab IPv4 string
|
||||
let remote_ipv4 = match remote_ip {
|
||||
IpAddr::V4(ip) => ip.to_string(),
|
||||
IpAddr::V6(ip) => {
|
||||
log::warn!("RUST: ignoring disconnection from IPv6 address: {ip}");
|
||||
continue;
|
||||
}
|
||||
};
|
||||
|
||||
// send disconnection event to channel (or exit if connection closed)
|
||||
if let Err(e) = connection_update_tx.send(PyConnectionUpdate {
|
||||
update_type: PyConnectionUpdateType::Disconnected,
|
||||
peer_id: PyPeerId(peer_id),
|
||||
remote_ipv4,
|
||||
remote_tcp_port,
|
||||
}).await {
|
||||
log::error!("RUST: could not send connection update since channel already closed: {e}");
|
||||
continue;
|
||||
}
|
||||
},
|
||||
e => {
|
||||
log::info!("RUST: other event {e:?}");
|
||||
}
|
||||
}
|
||||
}
|
||||
/// Encode a private key to a protobuf structure.
|
||||
fn to_protobuf_encoding<'py>(&self, py: Python<'py>) -> PyResult<Bound<'py, PyBytes>> {
|
||||
match self.0.to_protobuf_encoding() {
|
||||
Ok(bytes) => Ok(PyBytes::new(py, &bytes)),
|
||||
Err(e) => Err(PyValueError::new_err(e.to_string())),
|
||||
}
|
||||
}
|
||||
|
||||
log::info!("RUST: networking task stopped");
|
||||
fn to_string(&self) -> String {
|
||||
self.0.public().to_peer_id().to_base58()
|
||||
}
|
||||
}
|
||||
|
||||
struct PeerBuilder(
|
||||
String,
|
||||
Keypair,
|
||||
mpsc::Sender<FromSwarm>,
|
||||
mpsc::Receiver<ToSwarm>,
|
||||
);
|
||||
|
||||
#[gen_stub_pyclass]
|
||||
#[pyclass(name = "NetworkingHandle")]
|
||||
#[derive(Debug)]
|
||||
struct PyNetworkingHandle {
|
||||
// channels
|
||||
to_task_tx: Option<mpsc::Sender<ToTask>>,
|
||||
connection_update_rx: Mutex<mpsc::Receiver<PyConnectionUpdate>>,
|
||||
gossipsub_message_rx: Mutex<mpsc::Receiver<(String, Vec<u8>)>>,
|
||||
}
|
||||
|
||||
impl Drop for PyNetworkingHandle {
|
||||
fn drop(&mut self) {
|
||||
// TODO: may or may not need to await a "kill-signal" oneshot channel message,
|
||||
// to ensure that the networking task is done BEFORE exiting the clear function...
|
||||
// but this may require GIL?? and it may not be safe to call GIL here??
|
||||
self.to_task_tx = None; // Using Option<T> as a trick to force channel to be dropped
|
||||
}
|
||||
}
|
||||
|
||||
#[allow(clippy::expect_used)]
|
||||
impl PyNetworkingHandle {
|
||||
fn new(
|
||||
to_task_tx: mpsc::Sender<ToTask>,
|
||||
connection_update_rx: mpsc::Receiver<PyConnectionUpdate>,
|
||||
gossipsub_message_rx: mpsc::Receiver<(String, Vec<u8>)>,
|
||||
) -> Self {
|
||||
Self {
|
||||
to_task_tx: Some(to_task_tx),
|
||||
connection_update_rx: Mutex::new(connection_update_rx),
|
||||
gossipsub_message_rx: Mutex::new(gossipsub_message_rx),
|
||||
}
|
||||
}
|
||||
|
||||
const fn to_task_tx(&self) -> &mpsc::Sender<ToTask> {
|
||||
self.to_task_tx
|
||||
.as_ref()
|
||||
.expect("The sender should only be None after de-initialization.")
|
||||
}
|
||||
#[pyclass]
|
||||
pub struct PyPeer {
|
||||
peer: TakeOnce<PeerBuilder>,
|
||||
to_swarm: mpsc::Sender<ToSwarm>,
|
||||
from_swarm: Mutex<mpsc::Receiver<FromSwarm>>,
|
||||
}
|
||||
|
||||
#[gen_stub_pymethods]
|
||||
#[pymethods]
|
||||
impl PyNetworkingHandle {
|
||||
// NOTE: `async fn`s here that use `.await` will wrap the future in `.allow_threads_py()`
|
||||
// immediately beforehand to release the interpreter.
|
||||
// SEE: https://pyo3.rs/v0.26.0/async-await.html#detaching-from-the-interpreter-across-await
|
||||
|
||||
// ---- Lifecycle management methods ----
|
||||
|
||||
#[new]
|
||||
fn py_new(identity: Bound<'_, PyKeypair>) -> PyResult<Self> {
|
||||
use pyo3_async_runtimes::tokio::get_runtime;
|
||||
|
||||
// create communication channels
|
||||
let (to_task_tx, to_task_rx) = mpsc::channel(MPSC_CHANNEL_SIZE);
|
||||
let (connection_update_tx, connection_update_rx) = mpsc::channel(MPSC_CHANNEL_SIZE);
|
||||
let (gossipsub_message_tx, gossipsub_message_rx) = mpsc::channel(MPSC_CHANNEL_SIZE);
|
||||
|
||||
// get identity
|
||||
let identity = identity.borrow().0.clone();
|
||||
|
||||
// create networking swarm (within tokio context!! or it crashes)
|
||||
let swarm = get_runtime()
|
||||
.block_on(async { create_swarm(identity) })
|
||||
.pyerr()?;
|
||||
|
||||
// spawn tokio task running the networking logic
|
||||
get_runtime().spawn(async move {
|
||||
networking_task(
|
||||
swarm,
|
||||
to_task_rx,
|
||||
connection_update_tx,
|
||||
gossipsub_message_tx,
|
||||
)
|
||||
.await;
|
||||
});
|
||||
Ok(Self::new(
|
||||
to_task_tx,
|
||||
connection_update_rx,
|
||||
gossipsub_message_rx,
|
||||
))
|
||||
impl PyPeer {
|
||||
#[staticmethod]
|
||||
fn new(kp: PyKeypair, namespace: String) -> PyResult<Self> {
|
||||
let (to_client, from_swarm) = mpsc::channel(1024);
|
||||
let (to_swarm, from_client) = mpsc::channel(1024);
|
||||
Ok(Self {
|
||||
peer: TakeOnce::new(PeerBuilder(namespace, kp.0, to_client, from_client)),
|
||||
to_swarm,
|
||||
from_swarm: Mutex::new(from_swarm),
|
||||
})
|
||||
}
|
||||
|
||||
#[gen_stub(skip)]
|
||||
const fn __traverse__(&self, _visit: PyVisit<'_>) -> Result<(), PyTraverseError> {
|
||||
Ok(()) // This is needed purely so `__clear__` can work
|
||||
async fn run(&self, #[pyo3(cancel_handle)] mut cancel: CancelHandle) -> PyResult<()> {
|
||||
let builder = self
|
||||
.peer
|
||||
.take()
|
||||
.ok_or_else(|| PyRuntimeError::new_err("tried to run peer twice"))?;
|
||||
let jh = pyo3_async_runtimes::tokio::get_runtime()
|
||||
.spawn(async move {
|
||||
let mut peer =
|
||||
Peer::new(builder.0, builder.1, builder.2, builder.3).map_err(|_| {
|
||||
PyConnectionError::new_err("peer failed to listen on default address")
|
||||
})?;
|
||||
peer.run()
|
||||
.await
|
||||
.map_err(|()| PyConnectionError::new_err("peer communication closed"))
|
||||
})
|
||||
.or(async {
|
||||
cancel.cancelled().await;
|
||||
Ok(Ok(()))
|
||||
});
|
||||
match AllowThreads(pin!(jh)).await {
|
||||
Err(e) if e.is_cancelled() => Ok(()),
|
||||
Err(e) if e.is_panic() => Err(PyRuntimeError::new_err(format!("tokio panic {e}"))),
|
||||
Err(_) => unreachable!(),
|
||||
Ok(res) => res,
|
||||
}
|
||||
}
|
||||
|
||||
async fn subscribe(&self, topic: String) -> PyResult<()> {
|
||||
self.to_swarm
|
||||
.send(ToSwarm::Subscribe(topic))
|
||||
.await
|
||||
.map_err(|_| PyRuntimeError::new_err("swarm communication closed"))
|
||||
}
|
||||
async fn unsubscribe(&self, topic: String) -> PyResult<()> {
|
||||
self.to_swarm
|
||||
.send(ToSwarm::Unsubscribe(topic))
|
||||
.await
|
||||
.map_err(|_| PyRuntimeError::new_err("swarm communication closed"))
|
||||
}
|
||||
async fn send(&self, topic: String, payload: Py<PyBytes>) -> PyResult<()> {
|
||||
// this function attaches to the python interpreter synchronously to avoid holding the GIL
|
||||
let bytes = Python::attach(|py| Vec::from(payload.bind(py).as_bytes()));
|
||||
self.to_swarm
|
||||
.send(ToSwarm::Message(topic, bytes))
|
||||
.await
|
||||
.map_err(|_| PyRuntimeError::new_err("swarm communication closed"))
|
||||
}
|
||||
|
||||
#[gen_stub(skip)]
|
||||
fn __clear__(&mut self) {
|
||||
// TODO: may or may not need to await a "kill-signal" oneshot channel message,
|
||||
// to ensure that the networking task is done BEFORE exiting the clear function...
|
||||
// but this may require GIL?? and it may not be safe to call GIL here??
|
||||
self.to_task_tx = None; // Using Option<T> as a trick to force channel to be dropped
|
||||
async fn recv(
|
||||
&self,
|
||||
#[pyo3(cancel_handle)] mut cancel: CancelHandle,
|
||||
) -> PyResult<PySwarmEvent> {
|
||||
loop {
|
||||
return match AllowThreads(pin!(
|
||||
self.from_swarm
|
||||
.try_lock()
|
||||
.map_err(|_| PyRuntimeError::new_err("tried to recv twice"))?
|
||||
.recv()
|
||||
.or(async {
|
||||
cancel.cancelled().await;
|
||||
None
|
||||
})
|
||||
))
|
||||
.await
|
||||
{
|
||||
Some(FromSwarm::PublishError(p)) => match p {
|
||||
PublishError::AllQueuesFull(_) => {
|
||||
Err(PyConnectionError::new_err("swarm overloaded"))
|
||||
}
|
||||
PublishError::MessageTooLarge => {
|
||||
Err(PyValueError::new_err("message too large"))
|
||||
}
|
||||
PublishError::NoPeersSubscribedToTopic => {
|
||||
continue;
|
||||
}
|
||||
// TODO(evan): logs here
|
||||
_ => continue,
|
||||
},
|
||||
None => Err(PyRuntimeError::new_err("swarm communication closed")),
|
||||
Some(fs) => Ok(PySwarmEvent(fs)),
|
||||
};
|
||||
}
|
||||
}
|
||||
|
||||
// ---- Connection update receiver methods ----
|
||||
|
||||
/// Receives the next `ConnectionUpdate` from networking.
|
||||
async fn connection_update_recv(&self) -> PyResult<PyConnectionUpdate> {
|
||||
self.connection_update_rx
|
||||
.lock()
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await
|
||||
.recv_py()
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await
|
||||
}
|
||||
|
||||
/// Receives at most `limit` `ConnectionUpdate`s from networking and returns them.
|
||||
///
|
||||
/// For `limit = 0`, an empty collection of `ConnectionUpdate`s will be returned immediately.
|
||||
/// For `limit > 0`, if there are no `ConnectionUpdate`s in the channel's queue this method
|
||||
/// will sleep until a `ConnectionUpdate`s is sent.
|
||||
async fn connection_update_recv_many(&self, limit: usize) -> PyResult<Vec<PyConnectionUpdate>> {
|
||||
self.connection_update_rx
|
||||
.lock()
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await
|
||||
.recv_many_py(limit)
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await
|
||||
}
|
||||
|
||||
// TODO: rn this blocks main thread if anything else is awaiting the channel (bc its a mutex)
|
||||
// so its too dangerous to expose just yet. figure out a better semantics for handling this,
|
||||
// so things don't randomly block
|
||||
// /// Tries to receive the next `ConnectionUpdate` from networking.
|
||||
// fn connection_update_try_recv(&self) -> PyResult<Option<PyConnectionUpdate>> {
|
||||
// self.connection_update_rx.blocking_lock().try_recv_py()
|
||||
// }
|
||||
//
|
||||
// /// Checks if the `ConnectionUpdate` channel is empty.
|
||||
// fn connection_update_is_empty(&self) -> bool {
|
||||
// self.connection_update_rx.blocking_lock().is_empty()
|
||||
// }
|
||||
//
|
||||
// /// Returns the number of `ConnectionUpdate`s in the channel.
|
||||
// fn connection_update_len(&self) -> usize {
|
||||
// self.connection_update_rx.blocking_lock().len()
|
||||
// }
|
||||
|
||||
// ---- Gossipsub management methods ----
|
||||
|
||||
/// Subscribe to a `GossipSub` topic.
|
||||
///
|
||||
/// Returns `True` if the subscription worked. Returns `False` if we were already subscribed.
|
||||
async fn gossipsub_subscribe(&self, topic: String) -> PyResult<bool> {
|
||||
let (tx, rx) = oneshot::channel();
|
||||
|
||||
// send off request to subscribe
|
||||
self.to_task_tx()
|
||||
.send_py(ToTask::GossipsubSubscribe {
|
||||
topic,
|
||||
result_tx: tx,
|
||||
})
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await?;
|
||||
|
||||
// wait for response & return any errors
|
||||
rx.allow_threads_py() // allow-threads-aware async call
|
||||
.await
|
||||
.map_err(|_| PyErr::receiver_channel_closed())?
|
||||
}
|
||||
|
||||
/// Unsubscribes from a `GossipSub` topic.
|
||||
///
|
||||
/// Returns `True` if we were subscribed to this topic. Returns `False` if we were not subscribed.
|
||||
async fn gossipsub_unsubscribe(&self, topic: String) -> PyResult<bool> {
|
||||
let (tx, rx) = oneshot::channel();
|
||||
|
||||
// send off request to unsubscribe
|
||||
self.to_task_tx()
|
||||
.send_py(ToTask::GossipsubUnsubscribe {
|
||||
topic,
|
||||
result_tx: tx,
|
||||
})
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await?;
|
||||
|
||||
// wait for response & convert any errors
|
||||
rx.allow_threads_py() // allow-threads-aware async call
|
||||
.await
|
||||
.map_err(|_| PyErr::receiver_channel_closed())
|
||||
}
|
||||
|
||||
/// Publishes a message with multiple topics to the `GossipSub` network.
|
||||
///
|
||||
/// If no peers are found that subscribe to this topic, throws `NoPeersSubscribedToTopicError` exception.
|
||||
async fn gossipsub_publish(&self, topic: String, data: Py<PyBytes>) -> PyResult<()> {
|
||||
let (tx, rx) = oneshot::channel();
|
||||
|
||||
// send off request to subscribe
|
||||
let data = Python::with_gil(|py| Vec::from(data.as_bytes(py)));
|
||||
self.to_task_tx()
|
||||
.send_py(ToTask::GossipsubPublish {
|
||||
topic,
|
||||
data,
|
||||
result_tx: tx,
|
||||
})
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await?;
|
||||
|
||||
// wait for response & return any errors => ignore messageID for now!!!
|
||||
let _ = rx
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await
|
||||
.map_err(|_| PyErr::receiver_channel_closed())??;
|
||||
Ok(())
|
||||
}
|
||||
|
||||
// ---- Gossipsub message receiver methods ----
|
||||
|
||||
/// Receives the next message from the `GossipSub` network.
|
||||
async fn gossipsub_recv(&self) -> PyResult<(String, Py<PyBytes>)> {
|
||||
self.gossipsub_message_rx
|
||||
.lock()
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await
|
||||
.recv_py()
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await
|
||||
.map(|(t, d)| (t, d.pybytes()))
|
||||
}
|
||||
|
||||
/// Receives at most `limit` messages from the `GossipSub` network and returns them.
|
||||
///
|
||||
/// For `limit = 0`, an empty collection of messages will be returned immediately.
|
||||
/// For `limit > 0`, if there are no messages in the channel's queue this method
|
||||
/// will sleep until a message is sent.
|
||||
async fn gossipsub_recv_many(&self, limit: usize) -> PyResult<Vec<(String, Py<PyBytes>)>> {
|
||||
Ok(self
|
||||
.gossipsub_message_rx
|
||||
.lock()
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await
|
||||
.recv_many_py(limit)
|
||||
.allow_threads_py() // allow-threads-aware async call
|
||||
.await?
|
||||
.into_iter()
|
||||
.map(|(t, d)| (t, d.pybytes()))
|
||||
.collect())
|
||||
}
|
||||
|
||||
// TODO: rn this blocks main thread if anything else is awaiting the channel (bc its a mutex)
|
||||
// so its too dangerous to expose just yet. figure out a better semantics for handling this,
|
||||
// so things don't randomly block
|
||||
// /// Tries to receive the next message from the `GossipSub` network.
|
||||
// fn gossipsub_try_recv(&self) -> PyResult<Option<(String, Py<PyBytes>)>> {
|
||||
// Ok(self
|
||||
// .gossipsub_message_rx
|
||||
// .blocking_lock()
|
||||
// .try_recv_py()?
|
||||
// .map(|(t, d)| (t, d.pybytes())))
|
||||
// }
|
||||
//
|
||||
// /// Checks if the `GossipSub` message channel is empty.
|
||||
// fn gossipsub_is_empty(&self) -> bool {
|
||||
// self.gossipsub_message_rx.blocking_lock().is_empty()
|
||||
// }
|
||||
//
|
||||
// /// Returns the number of `GossipSub` messages in the channel.
|
||||
// fn gossipsub_len(&self) -> usize {
|
||||
// self.gossipsub_message_rx.blocking_lock().len()
|
||||
// }
|
||||
}
|
||||
|
||||
pub fn networking_submodule(m: &Bound<'_, PyModule>) -> PyResult<()> {
|
||||
m.add_class::<exception::PyNoPeersSubscribedToTopicError>()?;
|
||||
m.add_class::<exception::PyAllQueuesFullError>()?;
|
||||
|
||||
m.add_class::<PyConnectionUpdateType>()?;
|
||||
m.add_class::<PyConnectionUpdate>()?;
|
||||
m.add_class::<PyConnectionUpdateType>()?;
|
||||
m.add_class::<PyNetworkingHandle>()?;
|
||||
|
||||
Ok(())
|
||||
// Manually submit the run()/recv() stub because the cancelhandle is poorly understood
|
||||
submit! {
|
||||
gen_methods_from_python! {
|
||||
r#"
|
||||
class PyPeer:
|
||||
async def run(self): ...
|
||||
async def recv(self) -> PySwarmEvent: ...
|
||||
"#
|
||||
}
|
||||
}
|
||||
|
||||
#[gen_stub_pyclass]
|
||||
#[pyclass]
|
||||
pub struct PySwarmEvent(FromSwarm);
|
||||
|
||||
#[gen_stub_pymethods]
|
||||
#[pymethods]
|
||||
impl PySwarmEvent {
|
||||
// probably a better way to do this, but...
|
||||
fn downcast_discovered(&self) -> Option<String> {
|
||||
if let FromSwarm::Discovered(peer_id) = self.0 {
|
||||
Some(peer_id.to_base58())
|
||||
} else {
|
||||
None
|
||||
}
|
||||
}
|
||||
fn downcast_expired(&self) -> Option<String> {
|
||||
if let FromSwarm::Expired(peer_id) = self.0 {
|
||||
Some(peer_id.to_base58())
|
||||
} else {
|
||||
None
|
||||
}
|
||||
}
|
||||
fn downcast_message<'py>(
|
||||
&self,
|
||||
py: Python<'py>,
|
||||
) -> Option<(String, String, Bound<'py, PyBytes>)> {
|
||||
if let FromSwarm::Message(peer_id, topic, data) = &self.0 {
|
||||
Some((peer_id.to_base58(), topic.clone(), PyBytes::new(py, data)))
|
||||
} else {
|
||||
None
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,159 +0,0 @@
|
||||
use crate::ext::ResultExt as _;
|
||||
use libp2p::PeerId;
|
||||
use libp2p::identity::Keypair;
|
||||
use pyo3::prelude::{PyBytesMethods as _, PyModule, PyModuleMethods as _};
|
||||
use pyo3::types::PyBytes;
|
||||
use pyo3::{Bound, PyResult, Python, pyclass, pymethods};
|
||||
use pyo3_stub_gen::derive::{gen_stub_pyclass, gen_stub_pymethods};
|
||||
|
||||
/// Identity keypair of a node.
|
||||
#[gen_stub_pyclass]
|
||||
#[pyclass(name = "Keypair", frozen)]
|
||||
#[repr(transparent)]
|
||||
pub struct PyKeypair(pub Keypair);
|
||||
|
||||
#[gen_stub_pymethods]
|
||||
#[pymethods]
|
||||
#[allow(clippy::needless_pass_by_value)]
|
||||
impl PyKeypair {
|
||||
/// Generate a new Ed25519 keypair.
|
||||
#[staticmethod]
|
||||
fn generate_ed25519() -> Self {
|
||||
Self(Keypair::generate_ed25519())
|
||||
}
|
||||
|
||||
/// Generate a new ECDSA keypair.
|
||||
#[staticmethod]
|
||||
fn generate_ecdsa() -> Self {
|
||||
Self(Keypair::generate_ecdsa())
|
||||
}
|
||||
|
||||
/// Generate a new Secp256k1 keypair.
|
||||
#[staticmethod]
|
||||
fn generate_secp256k1() -> Self {
|
||||
Self(Keypair::generate_secp256k1())
|
||||
}
|
||||
|
||||
/// Decode a private key from a protobuf structure and parse it as a `Keypair`.
|
||||
#[staticmethod]
|
||||
fn from_protobuf_encoding(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
|
||||
let bytes = Vec::from(bytes.as_bytes());
|
||||
Ok(Self(Keypair::from_protobuf_encoding(&bytes).pyerr()?))
|
||||
}
|
||||
|
||||
/// Decode an keypair from a DER-encoded secret key in PKCS#8 `PrivateKeyInfo`
|
||||
/// format (i.e. unencrypted) as defined in [RFC5208].
|
||||
///
|
||||
/// [RFC5208]: https://tools.ietf.org/html/rfc5208#section-5
|
||||
#[staticmethod]
|
||||
fn rsa_from_pkcs8(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
|
||||
let mut bytes = Vec::from(bytes.as_bytes());
|
||||
Ok(Self(Keypair::rsa_from_pkcs8(&mut bytes).pyerr()?))
|
||||
}
|
||||
|
||||
/// Decode a keypair from a DER-encoded Secp256k1 secret key in an `ECPrivateKey`
|
||||
/// structure as defined in [RFC5915].
|
||||
///
|
||||
/// [RFC5915]: https://tools.ietf.org/html/rfc5915
|
||||
#[staticmethod]
|
||||
fn secp256k1_from_der(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
|
||||
let mut bytes = Vec::from(bytes.as_bytes());
|
||||
Ok(Self(Keypair::secp256k1_from_der(&mut bytes).pyerr()?))
|
||||
}
|
||||
|
||||
#[staticmethod]
|
||||
fn ed25519_from_bytes(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
|
||||
let mut bytes = Vec::from(bytes.as_bytes());
|
||||
Ok(Self(Keypair::ed25519_from_bytes(&mut bytes).pyerr()?))
|
||||
}
|
||||
|
||||
/// Encode a private key as protobuf structure.
|
||||
fn to_protobuf_encoding<'py>(&self, py: Python<'py>) -> PyResult<Bound<'py, PyBytes>> {
|
||||
let bytes = self.0.to_protobuf_encoding().pyerr()?;
|
||||
Ok(PyBytes::new(py, &bytes))
|
||||
}
|
||||
|
||||
/// Convert the `Keypair` into the corresponding `PeerId`.
|
||||
fn to_peer_id(&self) -> PyPeerId {
|
||||
PyPeerId(self.0.public().to_peer_id())
|
||||
}
|
||||
|
||||
// /// Hidden constructor for pickling support. TODO: figure out how to do pickling...
|
||||
// #[gen_stub(skip)]
|
||||
// #[new]
|
||||
// fn py_new(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
|
||||
// Self::from_protobuf_encoding(bytes)
|
||||
// }
|
||||
//
|
||||
// #[gen_stub(skip)]
|
||||
// fn __setstate__(&mut self, state: Bound<'_, PyBytes>) -> PyResult<()> {
|
||||
// *self = Self::from_protobuf_encoding(state)?;
|
||||
// Ok(())
|
||||
// }
|
||||
//
|
||||
// #[gen_stub(skip)]
|
||||
// fn __getstate__<'py>(&self, py: Python<'py>) -> PyResult<Bound<'py, PyBytes>> {
|
||||
// self.to_protobuf_encoding(py)
|
||||
// }
|
||||
//
|
||||
// #[gen_stub(skip)]
|
||||
// pub fn __getnewargs__<'py>(&self, py: Python<'py>) -> PyResult<(Bound<'py, PyBytes>,)> {
|
||||
// Ok((self.to_protobuf_encoding(py)?,))
|
||||
// }
|
||||
}
|
||||
|
||||
/// Identifier of a peer of the network.
|
||||
///
|
||||
/// The data is a `CIDv0` compatible multihash of the protobuf encoded public key of the peer
|
||||
/// as specified in [specs/peer-ids](https://github.com/libp2p/specs/blob/master/peer-ids/peer-ids.md).
|
||||
#[gen_stub_pyclass]
|
||||
#[pyclass(name = "PeerId", frozen)]
|
||||
#[derive(Debug, Clone)]
|
||||
#[repr(transparent)]
|
||||
pub struct PyPeerId(pub PeerId);
|
||||
|
||||
#[gen_stub_pymethods]
|
||||
#[pymethods]
|
||||
#[allow(clippy::needless_pass_by_value)]
|
||||
impl PyPeerId {
|
||||
/// Generates a random peer ID from a cryptographically secure PRNG.
|
||||
///
|
||||
/// This is useful for randomly walking on a DHT, or for testing purposes.
|
||||
#[staticmethod]
|
||||
fn random() -> Self {
|
||||
Self(PeerId::random())
|
||||
}
|
||||
|
||||
/// Parses a `PeerId` from bytes.
|
||||
#[staticmethod]
|
||||
fn from_bytes(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
|
||||
let bytes = Vec::from(bytes.as_bytes());
|
||||
Ok(Self(PeerId::from_bytes(&bytes).pyerr()?))
|
||||
}
|
||||
|
||||
/// Returns a raw bytes representation of this `PeerId`.
|
||||
fn to_bytes<'py>(&self, py: Python<'py>) -> Bound<'py, PyBytes> {
|
||||
let bytes = self.0.to_bytes();
|
||||
PyBytes::new(py, &bytes)
|
||||
}
|
||||
|
||||
/// Returns a base-58 encoded string of this `PeerId`.
|
||||
fn to_base58(&self) -> String {
|
||||
self.0.to_base58()
|
||||
}
|
||||
|
||||
fn __repr__(&self) -> String {
|
||||
format!("PeerId({})", self.to_base58())
|
||||
}
|
||||
|
||||
fn __str__(&self) -> String {
|
||||
self.to_base58()
|
||||
}
|
||||
}
|
||||
|
||||
pub fn ident_submodule(m: &Bound<'_, PyModule>) -> PyResult<()> {
|
||||
m.add_class::<PyKeypair>()?;
|
||||
m.add_class::<PyPeerId>()?;
|
||||
|
||||
Ok(())
|
||||
}
|
||||
@@ -1,8 +0,0 @@
|
||||
//! A module for exposing Rust's libp2p datatypes over Pyo3
|
||||
//!
|
||||
//! TODO: right now we are coupled to libp2p's identity, but eventually we want to create our own
|
||||
//! independent identity type of some kind or another. This may require handshaking.
|
||||
//!
|
||||
|
||||
pub mod ident;
|
||||
pub mod multiaddr;
|
||||
@@ -1,81 +0,0 @@
|
||||
use crate::ext::ResultExt as _;
|
||||
use libp2p::Multiaddr;
|
||||
use pyo3::prelude::{PyBytesMethods as _, PyModule, PyModuleMethods as _};
|
||||
use pyo3::types::PyBytes;
|
||||
use pyo3::{Bound, PyResult, Python, pyclass, pymethods};
|
||||
use pyo3_stub_gen::derive::{gen_stub_pyclass, gen_stub_pymethods};
|
||||
use std::str::FromStr as _;
|
||||
|
||||
/// Representation of a Multiaddr.
|
||||
#[gen_stub_pyclass]
|
||||
#[pyclass(name = "Multiaddr", frozen)]
|
||||
#[derive(Debug, Clone)]
|
||||
#[repr(transparent)]
|
||||
pub struct PyMultiaddr(pub Multiaddr);
|
||||
|
||||
#[gen_stub_pymethods]
|
||||
#[pymethods]
|
||||
#[allow(clippy::needless_pass_by_value)]
|
||||
impl PyMultiaddr {
|
||||
/// Create a new, empty multiaddress.
|
||||
#[staticmethod]
|
||||
fn empty() -> Self {
|
||||
Self(Multiaddr::empty())
|
||||
}
|
||||
|
||||
/// Create a new, empty multiaddress with the given capacity.
|
||||
#[staticmethod]
|
||||
fn with_capacity(n: usize) -> Self {
|
||||
Self(Multiaddr::with_capacity(n))
|
||||
}
|
||||
|
||||
/// Parse a `Multiaddr` value from its byte slice representation.
|
||||
#[staticmethod]
|
||||
fn from_bytes(bytes: Bound<'_, PyBytes>) -> PyResult<Self> {
|
||||
let bytes = Vec::from(bytes.as_bytes());
|
||||
Ok(Self(Multiaddr::try_from(bytes).pyerr()?))
|
||||
}
|
||||
|
||||
/// Parse a `Multiaddr` value from its string representation.
|
||||
#[staticmethod]
|
||||
fn from_string(string: String) -> PyResult<Self> {
|
||||
Ok(Self(Multiaddr::from_str(&string).pyerr()?))
|
||||
}
|
||||
|
||||
/// Return the length in bytes of this multiaddress.
|
||||
fn len(&self) -> usize {
|
||||
self.0.len()
|
||||
}
|
||||
|
||||
/// Returns true if the length of this multiaddress is 0.
|
||||
fn is_empty(&self) -> bool {
|
||||
self.0.is_empty()
|
||||
}
|
||||
|
||||
/// Return a copy of this [`Multiaddr`]'s byte representation.
|
||||
fn to_bytes<'py>(&self, py: Python<'py>) -> Bound<'py, PyBytes> {
|
||||
let bytes = self.0.to_vec();
|
||||
PyBytes::new(py, &bytes)
|
||||
}
|
||||
|
||||
/// Convert a Multiaddr to a string.
|
||||
fn to_string(&self) -> String {
|
||||
self.0.to_string()
|
||||
}
|
||||
|
||||
#[gen_stub(skip)]
|
||||
fn __repr__(&self) -> String {
|
||||
format!("Multiaddr({})", self.0)
|
||||
}
|
||||
|
||||
#[gen_stub(skip)]
|
||||
fn __str__(&self) -> String {
|
||||
self.to_string()
|
||||
}
|
||||
}
|
||||
|
||||
pub fn multiaddr_submodule(m: &Bound<'_, PyModule>) -> PyResult<()> {
|
||||
m.add_class::<PyMultiaddr>()?;
|
||||
|
||||
Ok(())
|
||||
}
|
||||
@@ -1,54 +0,0 @@
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use core::mem::drop;
|
||||
use core::option::Option::Some;
|
||||
use core::time::Duration;
|
||||
use tokio;
|
||||
use tokio::sync::mpsc;
|
||||
|
||||
#[tokio::test]
|
||||
async fn test_drop_channel() {
|
||||
struct Ping;
|
||||
|
||||
let (tx, mut rx) = mpsc::channel::<Ping>(10);
|
||||
|
||||
let _ = tokio::spawn(async move {
|
||||
println!("TASK: entered");
|
||||
|
||||
loop {
|
||||
tokio::select! {
|
||||
result = rx.recv() => {
|
||||
match result {
|
||||
Some(_) => {
|
||||
println!("TASK: pinged");
|
||||
}
|
||||
None => {
|
||||
println!("TASK: closing channel");
|
||||
break;
|
||||
}
|
||||
}
|
||||
}
|
||||
_ = tokio::time::sleep(Duration::from_secs_f32(0.1)) => {
|
||||
println!("TASK: heartbeat");
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
println!("TASK: exited");
|
||||
});
|
||||
|
||||
let tx2 = tx.clone();
|
||||
|
||||
tokio::time::sleep(Duration::from_secs_f32(0.11)).await;
|
||||
|
||||
tx.send(Ping).await.expect("Should not fail");
|
||||
drop(tx);
|
||||
|
||||
tokio::time::sleep(Duration::from_secs_f32(0.11)).await;
|
||||
|
||||
tx2.send(Ping).await.expect("Should not fail");
|
||||
drop(tx2);
|
||||
|
||||
tokio::time::sleep(Duration::from_secs_f32(0.11)).await;
|
||||
}
|
||||
}
|
||||
@@ -13,32 +13,14 @@ path = "src/lib.rs"
|
||||
workspace = true
|
||||
|
||||
[dependencies]
|
||||
# datastructures
|
||||
either = { workspace = true }
|
||||
|
||||
# macro dependencies
|
||||
extend = { workspace = true }
|
||||
delegate = { workspace = true }
|
||||
impl-trait-for-tuples = { workspace = true }
|
||||
derive_more = { workspace = true }
|
||||
|
||||
# async
|
||||
tokio = { workspace = true, features = ["full"] }
|
||||
futures = { workspace = true }
|
||||
futures-timer = { workspace = true }
|
||||
|
||||
# utility dependencies
|
||||
util = { workspace = true }
|
||||
thiserror = { workspace = true }
|
||||
#internment = { workspace = true }
|
||||
#recursion = { workspace = true }
|
||||
#generativity = { workspace = true }
|
||||
#itertools = { workspace = true }
|
||||
tracing-subscriber = { version = "0.3.19", features = ["default", "env-filter"] }
|
||||
keccak-const = { workspace = true }
|
||||
|
||||
# tracing/logging
|
||||
log = { workspace = true }
|
||||
|
||||
# networking
|
||||
libp2p = { workspace = true, features = ["full"] }
|
||||
libp2p = { workspace = true, features = ["full"] }
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
use futures::stream::StreamExt as _;
|
||||
use libp2p::{gossipsub, identity, swarm::SwarmEvent};
|
||||
use networking::{discovery, swarm};
|
||||
use libp2p::identity;
|
||||
use networking::{self, FromSwarm, ToSwarm};
|
||||
use tokio::sync::mpsc;
|
||||
use tokio::{io, io::AsyncBufReadExt as _, select};
|
||||
use tracing_subscriber::EnvFilter;
|
||||
use tracing_subscriber::filter::LevelFilter;
|
||||
@@ -12,63 +12,51 @@ async fn main() {
|
||||
.try_init();
|
||||
|
||||
// Configure swarm
|
||||
let mut swarm =
|
||||
swarm::create_swarm(identity::Keypair::generate_ed25519()).expect("Swarm creation failed");
|
||||
let (to_client, mut from_swarm) = mpsc::channel(20);
|
||||
let (to_swarm, from_client) = mpsc::channel(20);
|
||||
let mut peer = networking::Peer::new(
|
||||
"chatroom!".to_string(),
|
||||
identity::Keypair::generate_ed25519(),
|
||||
to_client,
|
||||
from_client,
|
||||
)
|
||||
.expect("listen error");
|
||||
|
||||
// Create a Gossipsub topic & subscribe
|
||||
let topic = gossipsub::IdentTopic::new("test-net");
|
||||
swarm
|
||||
.behaviour_mut()
|
||||
.gossipsub
|
||||
.subscribe(&topic)
|
||||
.expect("Subscribing to topic failed");
|
||||
|
||||
// Read full lines from stdin
|
||||
let mut stdin = io::BufReader::new(io::stdin()).lines();
|
||||
println!("Enter messages via STDIN and they will be sent to connected peers using Gossipsub");
|
||||
|
||||
let jh = tokio::spawn(async move { peer.run().await });
|
||||
_ = to_swarm
|
||||
.send(ToSwarm::Subscribe("chatting".to_string()))
|
||||
.await;
|
||||
|
||||
// Kick it off
|
||||
loop {
|
||||
select! {
|
||||
// on gossipsub outgoing
|
||||
Ok(Some(line)) = stdin.next_line() => {
|
||||
if let Err(e) = swarm
|
||||
.behaviour_mut().gossipsub
|
||||
.publish(topic.clone(), line.as_bytes()) {
|
||||
println!("Publish error: {e:?}");
|
||||
}
|
||||
_ = to_swarm.send(ToSwarm::Message("chatting".to_string(), line.into_bytes())).await;
|
||||
}
|
||||
event = swarm.select_next_some() => match event {
|
||||
event = from_swarm.recv() => match event {
|
||||
// on gossipsub incoming
|
||||
SwarmEvent::Behaviour(swarm::BehaviourEvent::Gossipsub(gossipsub::Event::Message {
|
||||
propagation_source: peer_id,
|
||||
message_id: id,
|
||||
message,
|
||||
})) => println!(
|
||||
"\n\nGot message: '{}' with id: {id} from peer: {peer_id}\n\n",
|
||||
String::from_utf8_lossy(&message.data),
|
||||
Some(FromSwarm::Message(peer_id,_, data)) => println!(
|
||||
"\n\nGot message: '{}' from peer: {peer_id}\n\n",
|
||||
String::from_utf8_lossy(&data),
|
||||
),
|
||||
|
||||
// on discovery
|
||||
SwarmEvent::Behaviour(swarm::BehaviourEvent::Discovery(e)) => match e {
|
||||
discovery::Event::ConnectionEstablished {
|
||||
peer_id, connection_id, remote_ip, remote_tcp_port
|
||||
} => {
|
||||
println!("\n\nConnected to: {peer_id}; connection ID: {connection_id}; remote IP: {remote_ip}; remote TCP port: {remote_tcp_port}\n\n");
|
||||
}
|
||||
discovery::Event::ConnectionClosed {
|
||||
peer_id, connection_id, remote_ip, remote_tcp_port
|
||||
} => {
|
||||
eprintln!("\n\nDisconnected from: {peer_id}; connection ID: {connection_id}; remote IP: {remote_ip}; remote TCP port: {remote_tcp_port}\n\n");
|
||||
}
|
||||
Some(FromSwarm::Discovered(peer_id)) => {
|
||||
println!("\n\nConnected to: {peer_id}\n\n");
|
||||
}
|
||||
|
||||
// ignore outgoing errors: those are normal
|
||||
e@SwarmEvent::OutgoingConnectionError { .. } => { log::debug!("Outgoing connection error: {e:?}"); }
|
||||
|
||||
// otherwise log any other event
|
||||
e => { log::info!("Other event {e:?}"); }
|
||||
Some(FromSwarm::Expired(peer_id)) => {
|
||||
println!("\n\nDisconnected from: {peer_id}\n\n");
|
||||
}
|
||||
Some(FromSwarm::PublishError(e)) => eprintln!("\n\nError {e:?}\n\n"),
|
||||
None => break,
|
||||
}
|
||||
}
|
||||
}
|
||||
_ = jh.await;
|
||||
}
|
||||
|
||||
@@ -1,127 +0,0 @@
|
||||
// Copyright 2018 Parity Technologies (UK) Ltd.
|
||||
//
|
||||
// Permission is hereby granted, free of charge, to any person obtaining a
|
||||
// copy of this software and associated documentation files (the "Software"),
|
||||
// to deal in the Software without restriction, including without limitation
|
||||
// the rights to use, copy, modify, merge, publish, distribute, sublicense,
|
||||
// and/or sell copies of the Software, and to permit persons to whom the
|
||||
// Software is furnished to do so, subject to the following conditions:
|
||||
//
|
||||
// The above copyright notice and this permission notice shall be included in
|
||||
// all copies or substantial portions of the Software.
|
||||
//
|
||||
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
|
||||
// OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
||||
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
||||
// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
||||
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
|
||||
// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
|
||||
// DEALINGS IN THE SOFTWARE.
|
||||
|
||||
use futures::stream::StreamExt;
|
||||
use libp2p::{
|
||||
gossipsub, mdns, noise,
|
||||
swarm::{NetworkBehaviour, SwarmEvent},
|
||||
tcp, yamux,
|
||||
};
|
||||
use std::time::Duration;
|
||||
use std::{error::Error, hash::Hash};
|
||||
use tokio::{io, io::AsyncBufReadExt, select};
|
||||
use tracing_subscriber::EnvFilter;
|
||||
|
||||
// We create a custom network behaviour that combines Gossipsub and Mdns.
|
||||
#[derive(NetworkBehaviour)]
|
||||
struct MyBehaviour {
|
||||
gossipsub: gossipsub::Behaviour,
|
||||
mdns: mdns::tokio::Behaviour,
|
||||
}
|
||||
|
||||
#[tokio::main]
|
||||
async fn main() -> Result<(), Box<dyn Error>> {
|
||||
let _ = tracing_subscriber::fmt()
|
||||
.with_env_filter(EnvFilter::from_default_env())
|
||||
.try_init();
|
||||
|
||||
let mut swarm = libp2p::SwarmBuilder::with_new_identity()
|
||||
.with_tokio()
|
||||
.with_tcp(
|
||||
tcp::Config::default(),
|
||||
noise::Config::new,
|
||||
yamux::Config::default,
|
||||
)?
|
||||
.with_behaviour(|key| {
|
||||
// Set a custom gossipsub configuration
|
||||
let gossipsub_config = gossipsub::ConfigBuilder::default()
|
||||
.heartbeat_interval(Duration::from_secs(10))
|
||||
.validation_mode(gossipsub::ValidationMode::Strict) // This sets the kind of message validation. The default is Strict (enforce message signing)
|
||||
.build()
|
||||
.map_err(io::Error::other)?; // Temporary hack because `build` does not return a proper `std::error::Error`.
|
||||
|
||||
// build a gossipsub network behaviour
|
||||
let gossipsub = gossipsub::Behaviour::new(
|
||||
gossipsub::MessageAuthenticity::Signed(key.clone()),
|
||||
gossipsub_config,
|
||||
)?;
|
||||
|
||||
let mdns =
|
||||
mdns::tokio::Behaviour::new(mdns::Config::default(), key.public().to_peer_id())?;
|
||||
Ok(MyBehaviour { gossipsub, mdns })
|
||||
})?
|
||||
.build();
|
||||
|
||||
println!("Running swarm with identity {}", swarm.local_peer_id());
|
||||
|
||||
// Create a Gossipsub topic
|
||||
let topic = gossipsub::IdentTopic::new("test-net");
|
||||
// subscribes to our topic
|
||||
swarm.behaviour_mut().gossipsub.subscribe(&topic)?;
|
||||
|
||||
// Read full lines from stdin
|
||||
let mut stdin = io::BufReader::new(io::stdin()).lines();
|
||||
|
||||
// Listen on all interfaces and whatever port the OS assigns
|
||||
swarm.listen_on("/ip4/0.0.0.0/tcp/0".parse()?)?;
|
||||
|
||||
println!("Enter messages via STDIN and they will be sent to connected peers using Gossipsub");
|
||||
|
||||
// Kick it off
|
||||
loop {
|
||||
select! {
|
||||
Ok(Some(line)) = stdin.next_line() => {
|
||||
if let Err(e) = swarm
|
||||
.behaviour_mut().gossipsub
|
||||
.publish(topic.clone(), line.as_bytes()) {
|
||||
println!("Publish error: {e:?}");
|
||||
}
|
||||
}
|
||||
event = swarm.select_next_some() => match event {
|
||||
SwarmEvent::Behaviour(MyBehaviourEvent::Mdns(mdns::Event::Discovered(list))) => {
|
||||
for (peer_id, multiaddr) in list {
|
||||
println!("mDNS discovered a new peer: {peer_id} on {multiaddr}");
|
||||
swarm.behaviour_mut().gossipsub.add_explicit_peer(&peer_id);
|
||||
}
|
||||
},
|
||||
SwarmEvent::Behaviour(MyBehaviourEvent::Mdns(mdns::Event::Expired(list))) => {
|
||||
for (peer_id, multiaddr) in list {
|
||||
println!("mDNS discover peer has expired: {peer_id} on {multiaddr}");
|
||||
swarm.behaviour_mut().gossipsub.remove_explicit_peer(&peer_id);
|
||||
}
|
||||
},
|
||||
SwarmEvent::Behaviour(MyBehaviourEvent::Gossipsub(gossipsub::Event::Message {
|
||||
propagation_source: peer_id,
|
||||
message_id: id,
|
||||
message,
|
||||
})) => println!(
|
||||
"Got message: '{}' with id: {id} from peer: {peer_id}",
|
||||
String::from_utf8_lossy(&message.data),
|
||||
),
|
||||
SwarmEvent::NewListenAddr { address, .. } => {
|
||||
println!("Local node is listening on {address}");
|
||||
}
|
||||
e => {
|
||||
println!("Other swarm event: {:?}", e);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -1,44 +0,0 @@
|
||||
https://github.com/ml-explore/mlx/commit/3fe98bacc7640d857acf3539f1d21b47a32e5609
|
||||
^raw sockets distributed -> `<net/ndrv.h>` -> https://newosxbook.com/code/xnu-3247.1.106/bsd/net/ndrv.h.auto.html
|
||||
--> header file for a networking component found in the macOS kernel (XNU) that defines structures for network device driver registration, specifically the ndrv_demux_desc and ndrv_protocol_desc structures used for demultiplexing protocol data at the network interface level. It specifies how to describe protocol data, such as an Ethernet type or a SNAP header, and how to associate these descriptions with a specific protocol family to receive matching packets.
|
||||
--> Used to bind an NDRV socket so that packets that match given protocol demux descriptions can be received.
|
||||
--> An NDRV socket is a special kind of socket in the Darwin/macOS operating system's XNU kernel, used for low-level network packet manipulation and binding to specific protocols for packet processing. It allows user-space applications or drivers to directly write Layer 2 (L2) network packets or interact with the network stack at a lower level, often by binding to protocol descriptors like the ndrv_protocol_desc. This type of socket is used for functions such as capturing and injecting packets, especially in network infrastructure software like routers or for kernel-level network monitoring and security tools.
|
||||
--> also called PF_NDRV sockets --> https://newosxbook.com/bonus/vol1ch16.html
|
||||
----> they are conceptually similar to https://scapy.disruptivelabs.in/networking/socket-interface PF_RAW or PF_PACKET
|
||||
|
||||
https://stackoverflow.com/questions/17169298/af-packet-on-osx
|
||||
^AF_PACKET duplicates the packets as soon as it receives them from the physical layer (for incoming packets) or just before sending them out to the physical layer (for outgoing packets). -> this is on Linux only
|
||||
^it doesn't exist on OS X so you can use /dev/bpfX (Berkeley Packet Filter) for sniffing
|
||||
|
||||
https://www.unix.com/man_page/mojave/4/ip/
|
||||
^OS X manpages for IP
|
||||
|
||||
https://developer.apple.com/documentation/kernel/implementing_drivers_system_extensions_and_kexts
|
||||
^driver kit, system extensions & kexts for macOS
|
||||
|
||||
----
|
||||
|
||||
To set up a Linux system to use a Thunderbolt connection as a network device, connect the two computers with a Thunderbolt cable, load the thunderbolt-net kernel module (usually automatic but modprobe is an option for manual loading), and then the operating system will create virtual Ethernet interfaces (e.g., thunderbolt0) for networking. You can then use standard tools like ifconfig or your desktop environment's network manager to configure these new interfaces for a link-local network.
|
||||
--> https://gist.github.com/geosp/80fbd39e617b7d1d9421683df4ea224a
|
||||
----> here is a guide on how to set up thunderbolt-ethernet on linux
|
||||
----> I may be able to steal the thunderbolt-net code ideas to implement a kernel module for MacOS
|
||||
|
||||
https://chatgpt.com/s/t_68af8e41a8548191993281a014f846a7
|
||||
^GPT discussion about making socket interface
|
||||
|
||||
https://chatgpt.com/s/t_68afb798a85c8191973c02a0fa7a48a3 --> link-local address,,??
|
||||
https://chatgpt.com/s/t_68afb02987e08191b2b0044d3667ece2
|
||||
^GPT discussion about accessing TB on MacOS low level interactions
|
||||
|
||||
--------------------------------
|
||||
|
||||
https://www.intel.com/content/www/us/en/support/articles/000098893/software.html
|
||||
^Thunderbolt Share & Thunderbolt Networking Mode => intel's equivalent of thunderbolt bridge
|
||||
|
||||
|
||||
---------------------------------
|
||||
|
||||
https://www.zerotier.com/blog/how-zerotier-eliminated-kernel-extensions-on-macos/
|
||||
-->fake ethernet devices on MacOS -> omg??? we can detect thunderbolt bridge, then bind to it, then re-expose it as fake ethernet??
|
||||
-->ps: https://chatgpt.com/s/t_68afb2b25fb881919526763fb5d7359c, AF/PF_NDRV are one and the same!!!
|
||||
-->https://github.com/zerotier/ZeroTierOne/blob/dev/osdep/MacEthernetTapAgent.c
|
||||
@@ -1,383 +0,0 @@
|
||||
use crate::ext::MultiaddrExt;
|
||||
use crate::keep_alive;
|
||||
use delegate::delegate;
|
||||
use either::Either;
|
||||
use futures::FutureExt;
|
||||
use futures_timer::Delay;
|
||||
use libp2p::core::transport::PortUse;
|
||||
use libp2p::core::{ConnectedPoint, Endpoint};
|
||||
use libp2p::swarm::behaviour::ConnectionEstablished;
|
||||
use libp2p::swarm::dial_opts::DialOpts;
|
||||
use libp2p::swarm::{
|
||||
CloseConnection, ConnectionClosed, ConnectionDenied, ConnectionHandler,
|
||||
ConnectionHandlerSelect, ConnectionId, FromSwarm, NetworkBehaviour, THandler, THandlerInEvent,
|
||||
THandlerOutEvent, ToSwarm, dummy,
|
||||
};
|
||||
use libp2p::{Multiaddr, PeerId, identity, mdns};
|
||||
use std::collections::{BTreeSet, HashMap};
|
||||
use std::convert::Infallible;
|
||||
use std::io;
|
||||
use std::net::IpAddr;
|
||||
use std::task::{Context, Poll};
|
||||
use std::time::Duration;
|
||||
use util::wakerdeque::WakerDeque;
|
||||
|
||||
const RETRY_CONNECT_INTERVAL: Duration = Duration::from_secs(5);
|
||||
|
||||
mod managed {
|
||||
use libp2p::swarm::NetworkBehaviour;
|
||||
use libp2p::{identity, mdns, ping};
|
||||
use std::io;
|
||||
use std::time::Duration;
|
||||
|
||||
const MDNS_RECORD_TTL: Duration = Duration::from_secs(2_500);
|
||||
const MDNS_QUERY_INTERVAL: Duration = Duration::from_secs(1_500);
|
||||
const PING_TIMEOUT: Duration = Duration::from_millis(2_500);
|
||||
const PING_INTERVAL: Duration = Duration::from_millis(2_500);
|
||||
|
||||
#[derive(NetworkBehaviour)]
|
||||
pub struct Behaviour {
|
||||
mdns: mdns::tokio::Behaviour,
|
||||
ping: ping::Behaviour,
|
||||
}
|
||||
|
||||
impl Behaviour {
|
||||
pub fn new(keypair: &identity::Keypair) -> io::Result<Self> {
|
||||
Ok(Self {
|
||||
mdns: mdns_behaviour(keypair)?,
|
||||
ping: ping_behaviour(),
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
fn mdns_behaviour(keypair: &identity::Keypair) -> io::Result<mdns::tokio::Behaviour> {
|
||||
use mdns::{Config, tokio};
|
||||
|
||||
// mDNS config => enable IPv6
|
||||
let mdns_config = Config {
|
||||
ttl: MDNS_RECORD_TTL,
|
||||
query_interval: MDNS_QUERY_INTERVAL,
|
||||
|
||||
// enable_ipv6: true, // TODO: for some reason, TCP+mDNS don't work well with ipv6?? figure out how to make work
|
||||
..Default::default()
|
||||
};
|
||||
|
||||
let mdns_behaviour = tokio::Behaviour::new(mdns_config, keypair.public().to_peer_id());
|
||||
Ok(mdns_behaviour?)
|
||||
}
|
||||
|
||||
fn ping_behaviour() -> ping::Behaviour {
|
||||
ping::Behaviour::new(
|
||||
ping::Config::new()
|
||||
.with_timeout(PING_TIMEOUT)
|
||||
.with_interval(PING_INTERVAL),
|
||||
)
|
||||
}
|
||||
}
|
||||
|
||||
/// Events for when a listening connection is truly established and truly closed.
|
||||
#[derive(Debug, Clone)]
|
||||
pub enum Event {
|
||||
ConnectionEstablished {
|
||||
peer_id: PeerId,
|
||||
connection_id: ConnectionId,
|
||||
remote_ip: IpAddr,
|
||||
remote_tcp_port: u16,
|
||||
},
|
||||
ConnectionClosed {
|
||||
peer_id: PeerId,
|
||||
connection_id: ConnectionId,
|
||||
remote_ip: IpAddr,
|
||||
remote_tcp_port: u16,
|
||||
},
|
||||
}
|
||||
|
||||
/// Discovery behavior that wraps mDNS to produce truly discovered durable peer-connections.
|
||||
///
|
||||
/// The behaviour operates as such:
|
||||
/// 1) All true (listening) connections/disconnections are tracked, emitting corresponding events
|
||||
/// to the swarm.
|
||||
/// 1) mDNS discovered/expired peers are tracked; discovered but not connected peers are dialed
|
||||
/// immediately, and expired but connected peers are disconnected from immediately.
|
||||
/// 2) Every fixed interval: discovered but not connected peers are dialed, and expired but
|
||||
/// connected peers are disconnected from.
|
||||
pub struct Behaviour {
|
||||
// state-tracking for managed behaviors & mDNS-discovered peers
|
||||
managed: managed::Behaviour,
|
||||
mdns_discovered: HashMap<PeerId, BTreeSet<Multiaddr>>,
|
||||
|
||||
retry_delay: Delay, // retry interval
|
||||
|
||||
// pending events to emmit => waker-backed Deque to control polling
|
||||
pending_events: WakerDeque<ToSwarm<Event, Infallible>>,
|
||||
}
|
||||
|
||||
impl Behaviour {
|
||||
pub fn new(keypair: &identity::Keypair) -> io::Result<Self> {
|
||||
Ok(Self {
|
||||
managed: managed::Behaviour::new(keypair)?,
|
||||
mdns_discovered: HashMap::new(),
|
||||
retry_delay: Delay::new(RETRY_CONNECT_INTERVAL),
|
||||
pending_events: WakerDeque::new(),
|
||||
})
|
||||
}
|
||||
|
||||
fn dial(&mut self, peer_id: PeerId, addr: Multiaddr) {
|
||||
self.pending_events.push_back(ToSwarm::Dial {
|
||||
opts: DialOpts::peer_id(peer_id).addresses(vec![addr]).build(),
|
||||
})
|
||||
}
|
||||
|
||||
fn close_connection(&mut self, peer_id: PeerId, connection: ConnectionId) {
|
||||
// push front to make this IMMEDIATE
|
||||
self.pending_events.push_front(ToSwarm::CloseConnection {
|
||||
peer_id,
|
||||
connection: CloseConnection::One(connection),
|
||||
})
|
||||
}
|
||||
|
||||
fn handle_mdns_discovered(&mut self, peers: Vec<(PeerId, Multiaddr)>) {
|
||||
for (p, ma) in peers {
|
||||
self.dial(p, ma.clone()); // always connect
|
||||
|
||||
// get peer's multi-addresses or insert if missing
|
||||
let Some(mas) = self.mdns_discovered.get_mut(&p) else {
|
||||
self.mdns_discovered.insert(p, BTreeSet::from([ma]));
|
||||
continue;
|
||||
};
|
||||
|
||||
// multiaddress should never already be present - else something has gone wrong
|
||||
let is_new_addr = mas.insert(ma);
|
||||
assert!(is_new_addr, "cannot discover a discovered peer");
|
||||
}
|
||||
}
|
||||
|
||||
fn handle_mdns_expired(&mut self, peers: Vec<(PeerId, Multiaddr)>) {
|
||||
for (p, ma) in peers {
|
||||
// at this point, we *must* have the peer
|
||||
let mas = self
|
||||
.mdns_discovered
|
||||
.get_mut(&p)
|
||||
.expect("nonexistent peer cannot expire");
|
||||
|
||||
// at this point, we *must* have the multiaddress
|
||||
let was_present = mas.remove(&ma);
|
||||
assert!(was_present, "nonexistent multiaddress cannot expire");
|
||||
|
||||
// if empty, remove the peer-id entirely
|
||||
if mas.is_empty() {
|
||||
self.mdns_discovered.remove(&p);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
fn on_connection_established(
|
||||
&mut self,
|
||||
peer_id: PeerId,
|
||||
connection_id: ConnectionId,
|
||||
remote_ip: IpAddr,
|
||||
remote_tcp_port: u16,
|
||||
) {
|
||||
// send out connected event
|
||||
self.pending_events
|
||||
.push_back(ToSwarm::GenerateEvent(Event::ConnectionEstablished {
|
||||
peer_id,
|
||||
connection_id,
|
||||
remote_ip,
|
||||
remote_tcp_port,
|
||||
}));
|
||||
}
|
||||
|
||||
fn on_connection_closed(
|
||||
&mut self,
|
||||
peer_id: PeerId,
|
||||
connection_id: ConnectionId,
|
||||
remote_ip: IpAddr,
|
||||
remote_tcp_port: u16,
|
||||
) {
|
||||
// send out disconnected event
|
||||
self.pending_events
|
||||
.push_back(ToSwarm::GenerateEvent(Event::ConnectionClosed {
|
||||
peer_id,
|
||||
connection_id,
|
||||
remote_ip,
|
||||
remote_tcp_port,
|
||||
}));
|
||||
}
|
||||
}
|
||||
|
||||
impl NetworkBehaviour for Behaviour {
|
||||
type ConnectionHandler =
|
||||
ConnectionHandlerSelect<dummy::ConnectionHandler, THandler<managed::Behaviour>>;
|
||||
type ToSwarm = Event;
|
||||
|
||||
// simply delegate to underlying mDNS behaviour
|
||||
|
||||
delegate! {
|
||||
to self.managed {
|
||||
fn handle_pending_inbound_connection(&mut self, connection_id: ConnectionId, local_addr: &Multiaddr, remote_addr: &Multiaddr) -> Result<(), ConnectionDenied>;
|
||||
fn handle_pending_outbound_connection(&mut self, connection_id: ConnectionId, maybe_peer: Option<PeerId>, addresses: &[Multiaddr], effective_role: Endpoint) -> Result<Vec<Multiaddr>, ConnectionDenied>;
|
||||
}
|
||||
}
|
||||
|
||||
fn handle_established_inbound_connection(
|
||||
&mut self,
|
||||
connection_id: ConnectionId,
|
||||
peer: PeerId,
|
||||
local_addr: &Multiaddr,
|
||||
remote_addr: &Multiaddr,
|
||||
) -> Result<THandler<Self>, ConnectionDenied> {
|
||||
Ok(ConnectionHandler::select(
|
||||
dummy::ConnectionHandler,
|
||||
self.managed.handle_established_inbound_connection(
|
||||
connection_id,
|
||||
peer,
|
||||
local_addr,
|
||||
remote_addr,
|
||||
)?,
|
||||
))
|
||||
}
|
||||
|
||||
#[allow(clippy::needless_question_mark)]
|
||||
fn handle_established_outbound_connection(
|
||||
&mut self,
|
||||
connection_id: ConnectionId,
|
||||
peer: PeerId,
|
||||
addr: &Multiaddr,
|
||||
role_override: Endpoint,
|
||||
port_use: PortUse,
|
||||
) -> Result<THandler<Self>, ConnectionDenied> {
|
||||
Ok(ConnectionHandler::select(
|
||||
dummy::ConnectionHandler,
|
||||
self.managed.handle_established_outbound_connection(
|
||||
connection_id,
|
||||
peer,
|
||||
addr,
|
||||
role_override,
|
||||
port_use,
|
||||
)?,
|
||||
))
|
||||
}
|
||||
|
||||
fn on_connection_handler_event(
|
||||
&mut self,
|
||||
peer_id: PeerId,
|
||||
connection_id: ConnectionId,
|
||||
event: THandlerOutEvent<Self>,
|
||||
) {
|
||||
match event {
|
||||
Either::Left(ev) => libp2p::core::util::unreachable(ev),
|
||||
Either::Right(ev) => {
|
||||
self.managed
|
||||
.on_connection_handler_event(peer_id, connection_id, ev)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// hook into these methods to drive behavior
|
||||
|
||||
fn on_swarm_event(&mut self, event: FromSwarm) {
|
||||
self.managed.on_swarm_event(event); // let mDNS handle swarm events
|
||||
|
||||
// handle swarm events to update internal state:
|
||||
match event {
|
||||
FromSwarm::ConnectionEstablished(ConnectionEstablished {
|
||||
peer_id,
|
||||
connection_id,
|
||||
endpoint,
|
||||
..
|
||||
}) => {
|
||||
let remote_address = match endpoint {
|
||||
ConnectedPoint::Dialer { address, .. } => address,
|
||||
ConnectedPoint::Listener { send_back_addr, .. } => send_back_addr,
|
||||
};
|
||||
|
||||
if let Some((ip, port)) = remote_address.try_to_tcp_addr() {
|
||||
// handle connection established event which is filtered correctly
|
||||
self.on_connection_established(peer_id, connection_id, ip, port)
|
||||
}
|
||||
}
|
||||
FromSwarm::ConnectionClosed(ConnectionClosed {
|
||||
peer_id,
|
||||
connection_id,
|
||||
endpoint,
|
||||
..
|
||||
}) => {
|
||||
let remote_address = match endpoint {
|
||||
ConnectedPoint::Dialer { address, .. } => address,
|
||||
ConnectedPoint::Listener { send_back_addr, .. } => send_back_addr,
|
||||
};
|
||||
|
||||
if let Some((ip, port)) = remote_address.try_to_tcp_addr() {
|
||||
// handle connection closed event which is filtered correctly
|
||||
self.on_connection_closed(peer_id, connection_id, ip, port)
|
||||
}
|
||||
}
|
||||
|
||||
// since we are running TCP/IP transport layer, we are assuming that
|
||||
// no address changes can occur, hence encountering one is a fatal error
|
||||
FromSwarm::AddressChange(a) => {
|
||||
unreachable!("unhandlable: address change encountered: {:?}", a)
|
||||
}
|
||||
_ => {}
|
||||
}
|
||||
}
|
||||
|
||||
fn poll(&mut self, cx: &mut Context) -> Poll<ToSwarm<Self::ToSwarm, THandlerInEvent<Self>>> {
|
||||
// delegate to managed behaviors for any behaviors they need to perform
|
||||
match self.managed.poll(cx) {
|
||||
Poll::Ready(ToSwarm::GenerateEvent(e)) => {
|
||||
match e {
|
||||
// handle discovered and expired events from mDNS
|
||||
managed::BehaviourEvent::Mdns(e) => match e.clone() {
|
||||
mdns::Event::Discovered(peers) => {
|
||||
self.handle_mdns_discovered(peers);
|
||||
}
|
||||
mdns::Event::Expired(peers) => {
|
||||
self.handle_mdns_expired(peers);
|
||||
}
|
||||
},
|
||||
|
||||
// handle ping events => if error then disconnect
|
||||
managed::BehaviourEvent::Ping(e) => {
|
||||
if let Err(_) = e.result {
|
||||
self.close_connection(e.peer, e.connection.clone())
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// since we just consumed an event, we should immediately wake just in case
|
||||
// there are more events to come where that came from
|
||||
cx.waker().wake_by_ref();
|
||||
}
|
||||
|
||||
// forward any other mDNS event to the swarm or its connection handler(s)
|
||||
Poll::Ready(e) => {
|
||||
return Poll::Ready(
|
||||
e.map_out(|_| unreachable!("events returning to swarm already handled"))
|
||||
.map_in(Either::Right),
|
||||
);
|
||||
}
|
||||
|
||||
Poll::Pending => {}
|
||||
}
|
||||
|
||||
// retry connecting to all mDNS peers periodically (fails safely if already connected)
|
||||
if self.retry_delay.poll_unpin(cx).is_ready() {
|
||||
for (p, mas) in self.mdns_discovered.clone() {
|
||||
for ma in mas {
|
||||
self.dial(p, ma)
|
||||
}
|
||||
}
|
||||
self.retry_delay.reset(RETRY_CONNECT_INTERVAL) // reset timeout
|
||||
}
|
||||
|
||||
// send out any pending events from our own service
|
||||
if let Some(e) = self.pending_events.pop_front(cx) {
|
||||
return Poll::Ready(e.map_in(Either::Left));
|
||||
}
|
||||
|
||||
// wait for pending events
|
||||
Poll::Pending
|
||||
}
|
||||
}
|
||||
@@ -1,44 +0,0 @@
|
||||
use delegate::delegate;
|
||||
use libp2p::swarm::handler::ConnectionEvent;
|
||||
use libp2p::swarm::{ConnectionHandlerEvent, SubstreamProtocol, dummy, handler};
|
||||
use std::task::{Context, Poll};
|
||||
|
||||
/// An implementation of [`ConnectionHandler`] that doesn't handle any protocols, but it keeps
|
||||
/// the connection alive.
|
||||
#[derive(Clone)]
|
||||
#[repr(transparent)]
|
||||
pub struct ConnectionHandler(dummy::ConnectionHandler);
|
||||
|
||||
impl ConnectionHandler {
|
||||
pub fn new() -> Self {
|
||||
ConnectionHandler(dummy::ConnectionHandler)
|
||||
}
|
||||
}
|
||||
|
||||
impl handler::ConnectionHandler for ConnectionHandler {
|
||||
// delegate types and implementation mostly to dummy handler
|
||||
type FromBehaviour = <dummy::ConnectionHandler as handler::ConnectionHandler>::FromBehaviour;
|
||||
type ToBehaviour = <dummy::ConnectionHandler as handler::ConnectionHandler>::ToBehaviour;
|
||||
type InboundProtocol =
|
||||
<dummy::ConnectionHandler as handler::ConnectionHandler>::InboundProtocol;
|
||||
type OutboundProtocol =
|
||||
<dummy::ConnectionHandler as handler::ConnectionHandler>::OutboundProtocol;
|
||||
type InboundOpenInfo =
|
||||
<dummy::ConnectionHandler as handler::ConnectionHandler>::InboundOpenInfo;
|
||||
type OutboundOpenInfo =
|
||||
<dummy::ConnectionHandler as handler::ConnectionHandler>::OutboundOpenInfo;
|
||||
|
||||
delegate! {
|
||||
to self.0 {
|
||||
fn listen_protocol(&self) -> SubstreamProtocol<Self::InboundProtocol, Self::InboundOpenInfo>;
|
||||
fn poll(&mut self, cx: &mut Context<'_>) -> Poll<ConnectionHandlerEvent<Self::OutboundProtocol, Self::OutboundOpenInfo, Self::ToBehaviour>>;
|
||||
fn on_behaviour_event(&mut self, event: Self::FromBehaviour);
|
||||
fn on_connection_event(&mut self, event: ConnectionEvent<Self::InboundProtocol, Self::OutboundProtocol, Self::InboundOpenInfo, Self::OutboundOpenInfo>);
|
||||
}
|
||||
}
|
||||
|
||||
// specifically override this to force connection to stay alive
|
||||
fn connection_keep_alive(&self) -> bool {
|
||||
true
|
||||
}
|
||||
}
|
||||
@@ -1,64 +1,299 @@
|
||||
//! TODO: crate documentation
|
||||
//!
|
||||
//! this is here as a placeholder documentation
|
||||
//!
|
||||
//!
|
||||
use libp2p::{
|
||||
Multiaddr, PeerId,
|
||||
futures::StreamExt,
|
||||
gossipsub::{self, TopicHash},
|
||||
identify,
|
||||
identity::Keypair,
|
||||
mdns,
|
||||
swarm::{NetworkBehaviour, SwarmEvent, dial_opts::DialOpts},
|
||||
};
|
||||
use std::collections::HashMap;
|
||||
use tokio::sync::mpsc;
|
||||
|
||||
// enable Rust-unstable features for convenience
|
||||
#![feature(trait_alias)]
|
||||
// #![feature(stmt_expr_attributes)]
|
||||
// #![feature(unboxed_closures)]
|
||||
// #![feature(assert_matches)]
|
||||
// #![feature(async_fn_in_dyn_trait)]
|
||||
// #![feature(async_for_loop)]
|
||||
// #![feature(auto_traits)]
|
||||
// #![feature(negative_impls)]
|
||||
#[derive(Debug)]
|
||||
pub struct ListenError;
|
||||
|
||||
pub mod discovery;
|
||||
pub mod keep_alive;
|
||||
pub mod swarm;
|
||||
|
||||
/// Namespace for all the type/trait aliases used by this crate.
|
||||
pub(crate) mod alias {
|
||||
use std::error::Error;
|
||||
|
||||
pub type AnyError = Box<dyn Error + Send + Sync + 'static>;
|
||||
pub type AnyResult<T> = Result<T, AnyError>;
|
||||
pub enum FromSwarm {
|
||||
PublishError(gossipsub::PublishError),
|
||||
Discovered(PeerId),
|
||||
Expired(PeerId),
|
||||
Message(PeerId, String, Vec<u8>),
|
||||
}
|
||||
pub enum ToSwarm {
|
||||
Message(String, Vec<u8>),
|
||||
Subscribe(String),
|
||||
Unsubscribe(String),
|
||||
}
|
||||
|
||||
/// Namespace for crate-wide extension traits/methods
|
||||
pub(crate) mod ext {
|
||||
use extend::ext;
|
||||
use libp2p::Multiaddr;
|
||||
use libp2p::multiaddr::Protocol;
|
||||
use std::net::IpAddr;
|
||||
pub struct Peer {
|
||||
pub swarm: libp2p::Swarm<Behaviour>,
|
||||
to_client: mpsc::Sender<FromSwarm>,
|
||||
from_client: mpsc::Receiver<ToSwarm>,
|
||||
namespace: String,
|
||||
known_peers: HashMap<PeerId, Vec<Multiaddr>>,
|
||||
}
|
||||
impl Peer {
|
||||
pub fn new(
|
||||
namespace: String,
|
||||
kp: Keypair,
|
||||
to_client: mpsc::Sender<FromSwarm>,
|
||||
from_client: mpsc::Receiver<ToSwarm>,
|
||||
) -> Result<Self, ListenError> {
|
||||
let mut swarm = libp2p::SwarmBuilder::with_existing_identity(kp)
|
||||
.with_tokio()
|
||||
.with_quic()
|
||||
// TODO(evan) .with_bandwidth_metrics()
|
||||
.with_behaviour(|kp| Behaviour::new(namespace.clone(), kp))
|
||||
.expect("invalid swarm behaviour")
|
||||
.build();
|
||||
|
||||
#[ext(pub, name = MultiaddrExt)]
|
||||
impl Multiaddr {
|
||||
/// If the multiaddress corresponds to a TCP address, extracts it
|
||||
fn try_to_tcp_addr(&self) -> Option<(IpAddr, u16)> {
|
||||
let mut ps = self.into_iter();
|
||||
let ip = if let Some(p) = ps.next() {
|
||||
match p {
|
||||
Protocol::Ip4(ip) => IpAddr::V4(ip),
|
||||
Protocol::Ip6(ip) => IpAddr::V6(ip),
|
||||
_ => return None,
|
||||
swarm
|
||||
.listen_on("/ip6/::/udp/0/quic-v1".parse().expect("invalid multiaddr"))
|
||||
.map_err(|_| ListenError)?;
|
||||
swarm
|
||||
.listen_on(
|
||||
"/ip4/0.0.0.0/udp/0/quic-v1"
|
||||
.parse()
|
||||
.expect("invalid multiaddr"),
|
||||
)
|
||||
.map_err(|_| ListenError)?;
|
||||
Ok(Self {
|
||||
swarm,
|
||||
to_client,
|
||||
from_client,
|
||||
namespace,
|
||||
known_peers: HashMap::default(),
|
||||
})
|
||||
}
|
||||
pub async fn run(&mut self) -> Result<(), ()> {
|
||||
loop {
|
||||
tokio::select! {
|
||||
event = self.swarm.next() => self.handle_event(event.ok_or(())?).await?,
|
||||
msg = self.from_client.recv() => self.handle_message(msg.ok_or(())?).await?,
|
||||
}
|
||||
}
|
||||
}
|
||||
async fn handle_message(&mut self, message: ToSwarm) -> Result<(), ()> {
|
||||
match message {
|
||||
ToSwarm::Message(topic, data) => {
|
||||
if let Err(e) = self
|
||||
.swarm
|
||||
.behaviour_mut()
|
||||
.gossipsub
|
||||
.publish(TopicHash::from_raw(topic), data)
|
||||
{
|
||||
self.to_client
|
||||
.send(FromSwarm::PublishError(e))
|
||||
.await
|
||||
.map_err(|_| ())?;
|
||||
}
|
||||
} else {
|
||||
return None;
|
||||
};
|
||||
let Some(Protocol::Tcp(port)) = ps.next() else {
|
||||
return None;
|
||||
};
|
||||
Some((ip, port))
|
||||
}
|
||||
ToSwarm::Subscribe(topic) => {
|
||||
match self
|
||||
.swarm
|
||||
.behaviour_mut()
|
||||
.gossipsub
|
||||
.subscribe(&gossipsub::IdentTopic::new(topic))
|
||||
{
|
||||
Ok(_) => {}
|
||||
Err(gossipsub::SubscriptionError::NotAllowed) => {
|
||||
unreachable!("subscription filter hit")
|
||||
}
|
||||
Err(gossipsub::SubscriptionError::PublishError(e)) => self
|
||||
.to_client
|
||||
.send(FromSwarm::PublishError(e))
|
||||
.await
|
||||
.map_err(|_| ())?,
|
||||
}
|
||||
}
|
||||
ToSwarm::Unsubscribe(topic) => {
|
||||
self.swarm
|
||||
.behaviour_mut()
|
||||
.gossipsub
|
||||
.unsubscribe(&gossipsub::IdentTopic::new(topic));
|
||||
}
|
||||
}
|
||||
Ok(())
|
||||
}
|
||||
async fn handle_event(&mut self, event: SwarmEvent<BehaviourEvent>) -> Result<(), ()> {
|
||||
let SwarmEvent::Behaviour(event) = event else {
|
||||
return Ok(());
|
||||
};
|
||||
match event {
|
||||
BehaviourEvent::Gossipsub(gossipsub::Event::Message { message, .. }) => {
|
||||
if let Some(source) = message.source {
|
||||
self.to_client
|
||||
.send(FromSwarm::Message(
|
||||
source,
|
||||
message.topic.into_string(),
|
||||
message.data,
|
||||
))
|
||||
.await
|
||||
.map_err(|_| ())?;
|
||||
}
|
||||
}
|
||||
BehaviourEvent::Identify(identify::Event::Received { peer_id, info, .. }) => {
|
||||
log::debug!(
|
||||
"identify from {peer_id}: protocol_version='{}' agent_version='{}' (local namespace='{}')",
|
||||
info.protocol_version,
|
||||
info.agent_version,
|
||||
self.namespace
|
||||
);
|
||||
if info.protocol_version == self.namespace {
|
||||
self.passed_namespace(peer_id);
|
||||
self.to_client
|
||||
.send(FromSwarm::Discovered(peer_id))
|
||||
.await
|
||||
.map_err(|_| ())?;
|
||||
} else {
|
||||
self.failed_namespace(peer_id);
|
||||
}
|
||||
}
|
||||
BehaviourEvent::Mdns(mdns::Event::Discovered(v)) => {
|
||||
for (peer_id, addr) in v {
|
||||
self.known_peers.entry(peer_id).or_default().push(addr);
|
||||
}
|
||||
for (peer_id, addrs) in &self.known_peers {
|
||||
// dialopts handles rate limiting, we should check errors if we want to blacklist earlier
|
||||
let _ = self
|
||||
.swarm
|
||||
.dial(DialOpts::peer_id(*peer_id).addresses(addrs.clone()).build());
|
||||
}
|
||||
}
|
||||
BehaviourEvent::Mdns(mdns::Event::Expired(v)) => {
|
||||
for (peer_id, addr) in v {
|
||||
let addrs = self.known_peers.entry(peer_id).or_default();
|
||||
addrs.retain(|a| *a != addr);
|
||||
if addrs.is_empty() {
|
||||
self.known_peers.remove(&peer_id);
|
||||
self.swarm
|
||||
.behaviour_mut()
|
||||
.gossipsub
|
||||
.remove_explicit_peer(&peer_id);
|
||||
self.to_client
|
||||
.send(FromSwarm::Expired(peer_id))
|
||||
.await
|
||||
.map_err(|_| ())?;
|
||||
}
|
||||
}
|
||||
}
|
||||
_ => {}
|
||||
}
|
||||
Ok(())
|
||||
}
|
||||
fn passed_namespace(&mut self, peer_id: PeerId) {
|
||||
self.swarm
|
||||
.behaviour_mut()
|
||||
.gossipsub
|
||||
.remove_blacklisted_peer(&peer_id);
|
||||
self.swarm
|
||||
.behaviour_mut()
|
||||
.gossipsub
|
||||
.add_explicit_peer(&peer_id);
|
||||
}
|
||||
fn failed_namespace(&mut self, peer_id: PeerId) {
|
||||
self.swarm
|
||||
.behaviour_mut()
|
||||
.gossipsub
|
||||
.blacklist_peer(&peer_id);
|
||||
self.swarm
|
||||
.behaviour_mut()
|
||||
.gossipsub
|
||||
.remove_explicit_peer(&peer_id);
|
||||
}
|
||||
}
|
||||
|
||||
#[derive(NetworkBehaviour)]
|
||||
pub struct Behaviour {
|
||||
gossipsub: gossipsub::Behaviour,
|
||||
mdns: mdns::tokio::Behaviour,
|
||||
identify: identify::Behaviour,
|
||||
}
|
||||
|
||||
impl Behaviour {
|
||||
fn new(namespace: String, kp: &Keypair) -> Self {
|
||||
let mdns = mdns::Behaviour::new(mdns::Config::default(), kp.public().to_peer_id())
|
||||
.expect("mdns behaviour failed to build");
|
||||
|
||||
let identify =
|
||||
identify::Behaviour::new(identify::Config::new_with_signed_peer_record(namespace, kp));
|
||||
|
||||
let gossipsub = gossipsub::Behaviour::new(
|
||||
gossipsub::MessageAuthenticity::Signed(kp.clone()),
|
||||
gossipsub::ConfigBuilder::default()
|
||||
.max_transmit_size(1024 * 1024)
|
||||
.validation_mode(gossipsub::ValidationMode::Strict)
|
||||
.build()
|
||||
.expect("invalid gossipsub configuration"),
|
||||
)
|
||||
.expect("gossipsub behaviour failed ot build");
|
||||
|
||||
Self {
|
||||
gossipsub,
|
||||
mdns,
|
||||
identify,
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
pub(crate) mod private {
|
||||
#![allow(dead_code)]
|
||||
// TODO: more tests
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
use tokio::time::{Duration, timeout};
|
||||
|
||||
/// Sealed traits support
|
||||
pub trait Sealed {}
|
||||
impl<T: ?Sized> Sealed for T {}
|
||||
fn make_peer(namespace: &str) -> (Peer, mpsc::Receiver<FromSwarm>, mpsc::Sender<ToSwarm>) {
|
||||
let kp = Keypair::generate_ed25519();
|
||||
|
||||
let (to_client_tx, to_client_rx) = mpsc::channel(64);
|
||||
let (to_peer_tx, to_peer_rx) = mpsc::channel(64);
|
||||
|
||||
let peer = Peer::new(namespace.to_string(), kp, to_client_tx, to_peer_rx)
|
||||
.expect("Peer::new should succeed in tests");
|
||||
|
||||
(peer, to_client_rx, to_peer_tx)
|
||||
}
|
||||
|
||||
async fn next_listen_addr(peer: &mut Peer) -> Multiaddr {
|
||||
loop {
|
||||
match peer.swarm.next().await {
|
||||
Some(SwarmEvent::NewListenAddr { address, .. }) => return address,
|
||||
Some(_) => {}
|
||||
None => panic!("swarm stream ended unexpectedly"),
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#[tokio::test]
|
||||
async fn subscribe_and_unsubscribe_do_not_error() {
|
||||
let (mut peer, mut events_rx, commands_tx) = make_peer("ns-test");
|
||||
|
||||
// Drive the swarm just enough to get at least one listen address event,
|
||||
// so the background run loop has something initialized.
|
||||
let _addr = next_listen_addr(&mut peer).await;
|
||||
|
||||
// Run the peer loop in the background.
|
||||
let handle = tokio::spawn(async move {
|
||||
let _ = peer.run().await;
|
||||
});
|
||||
|
||||
commands_tx
|
||||
.send(ToSwarm::Subscribe("topic-a".to_string()))
|
||||
.await
|
||||
.unwrap();
|
||||
|
||||
commands_tx
|
||||
.send(ToSwarm::Unsubscribe("topic-a".to_string()))
|
||||
.await
|
||||
.unwrap();
|
||||
|
||||
// We don't *require* any FromSwarm events here; this is mainly a
|
||||
// smoke test that the message-handling path doesn't panic/hang.
|
||||
// Still, poll briefly to ensure the task is alive.
|
||||
let _ = timeout(Duration::from_millis(200), events_rx.recv()).await;
|
||||
|
||||
// Shut down: dropping the command sender closes the channel, causing run() to return Err.
|
||||
drop(commands_tx);
|
||||
let _ = handle.await;
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,143 +0,0 @@
|
||||
use crate::alias;
|
||||
use crate::swarm::transport::tcp_transport;
|
||||
pub use behaviour::{Behaviour, BehaviourEvent};
|
||||
use libp2p::{SwarmBuilder, identity};
|
||||
|
||||
pub type Swarm = libp2p::Swarm<Behaviour>;
|
||||
|
||||
/// The current version of the network: this prevents devices running different versions of the
|
||||
/// software from interacting with each other.
|
||||
///
|
||||
/// TODO: right now this is a hardcoded constant; figure out what the versioning semantics should
|
||||
/// even be, and how to inject the right version into this config/initialization. E.g. should
|
||||
/// this be passed in as a parameter? What about rapidly changing versions in debug builds?
|
||||
/// this is all VERY very hard to figure out and needs to be mulled over as a team.
|
||||
pub const NETWORK_VERSION: &[u8] = b"v0.0.1";
|
||||
pub const OVERRIDE_VERSION_ENV_VAR: &str = "EXO_LIBP2P_NAMESPACE";
|
||||
|
||||
/// Create and configure a swarm which listens to all ports on OS
|
||||
pub fn create_swarm(keypair: identity::Keypair) -> alias::AnyResult<Swarm> {
|
||||
let mut swarm = SwarmBuilder::with_existing_identity(keypair)
|
||||
.with_tokio()
|
||||
.with_other_transport(tcp_transport)?
|
||||
.with_behaviour(Behaviour::new)?
|
||||
.build();
|
||||
|
||||
// Listen on all interfaces and whatever port the OS assigns
|
||||
swarm.listen_on("/ip4/0.0.0.0/tcp/0".parse()?)?;
|
||||
Ok(swarm)
|
||||
}
|
||||
|
||||
mod transport {
|
||||
use crate::alias;
|
||||
use crate::swarm::{NETWORK_VERSION, OVERRIDE_VERSION_ENV_VAR};
|
||||
use futures::{AsyncRead, AsyncWrite};
|
||||
use keccak_const::Sha3_256;
|
||||
use libp2p::core::muxing;
|
||||
use libp2p::core::transport::Boxed;
|
||||
use libp2p::pnet::{PnetError, PnetOutput};
|
||||
use libp2p::{PeerId, Transport, identity, noise, pnet, yamux};
|
||||
use std::{env, sync::LazyLock};
|
||||
|
||||
/// Key used for networking's private network; parametrized on the [`NETWORK_VERSION`].
|
||||
/// See [`pnet_upgrade`] for more.
|
||||
static PNET_PRESHARED_KEY: LazyLock<[u8; 32]> = LazyLock::new(|| {
|
||||
let builder = Sha3_256::new().update(b"exo_discovery_network");
|
||||
|
||||
if let Ok(var) = env::var(OVERRIDE_VERSION_ENV_VAR) {
|
||||
let bytes = var.into_bytes();
|
||||
builder.update(&bytes)
|
||||
} else {
|
||||
builder.update(NETWORK_VERSION)
|
||||
}
|
||||
.finalize()
|
||||
});
|
||||
|
||||
/// Make the Swarm run on a private network, as to not clash with public libp2p nodes and
|
||||
/// also different-versioned instances of this same network.
|
||||
/// This is implemented as an additional "upgrade" ontop of existing [`libp2p::Transport`] layers.
|
||||
async fn pnet_upgrade<TSocket>(
|
||||
socket: TSocket,
|
||||
_: impl Sized,
|
||||
) -> Result<PnetOutput<TSocket>, PnetError>
|
||||
where
|
||||
TSocket: AsyncRead + AsyncWrite + Send + Unpin + 'static,
|
||||
{
|
||||
use pnet::{PnetConfig, PreSharedKey};
|
||||
PnetConfig::new(PreSharedKey::new(*PNET_PRESHARED_KEY))
|
||||
.handshake(socket)
|
||||
.await
|
||||
}
|
||||
|
||||
/// TCP/IP transport layer configuration.
|
||||
pub fn tcp_transport(
|
||||
keypair: &identity::Keypair,
|
||||
) -> alias::AnyResult<Boxed<(PeerId, muxing::StreamMuxerBox)>> {
|
||||
use libp2p::{
|
||||
core::upgrade::Version,
|
||||
tcp::{Config, tokio},
|
||||
};
|
||||
|
||||
// `TCP_NODELAY` enabled => avoid latency
|
||||
let tcp_config = Config::default().nodelay(true);
|
||||
|
||||
// V1 + lazy flushing => 0-RTT negotiation
|
||||
let upgrade_version = Version::V1Lazy;
|
||||
|
||||
// Noise is faster than TLS + we don't care much for security
|
||||
let noise_config = noise::Config::new(keypair)?;
|
||||
|
||||
// Use default Yamux config for multiplexing
|
||||
let yamux_config = yamux::Config::default();
|
||||
|
||||
// Create new Tokio-driven TCP/IP transport layer
|
||||
let base_transport = tokio::Transport::new(tcp_config)
|
||||
.and_then(pnet_upgrade)
|
||||
.upgrade(upgrade_version)
|
||||
.authenticate(noise_config)
|
||||
.multiplex(yamux_config);
|
||||
|
||||
// Return boxed transport (to flatten complex type)
|
||||
Ok(base_transport.boxed())
|
||||
}
|
||||
}
|
||||
|
||||
mod behaviour {
|
||||
use crate::{alias, discovery};
|
||||
use libp2p::swarm::NetworkBehaviour;
|
||||
use libp2p::{gossipsub, identity};
|
||||
|
||||
/// Behavior of the Swarm which composes all desired behaviors:
|
||||
/// Right now its just [`discovery::Behaviour`] and [`gossipsub::Behaviour`].
|
||||
#[derive(NetworkBehaviour)]
|
||||
pub struct Behaviour {
|
||||
pub discovery: discovery::Behaviour,
|
||||
pub gossipsub: gossipsub::Behaviour,
|
||||
}
|
||||
|
||||
impl Behaviour {
|
||||
pub fn new(keypair: &identity::Keypair) -> alias::AnyResult<Self> {
|
||||
Ok(Self {
|
||||
discovery: discovery::Behaviour::new(keypair)?,
|
||||
gossipsub: gossipsub_behaviour(keypair),
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
fn gossipsub_behaviour(keypair: &identity::Keypair) -> gossipsub::Behaviour {
|
||||
use gossipsub::{ConfigBuilder, MessageAuthenticity, ValidationMode};
|
||||
|
||||
// build a gossipsub network behaviour
|
||||
// => signed message authenticity + strict validation mode means the message-ID is
|
||||
// automatically provided by gossipsub w/out needing to provide custom message-ID function
|
||||
gossipsub::Behaviour::new(
|
||||
MessageAuthenticity::Signed(keypair.clone()),
|
||||
ConfigBuilder::default()
|
||||
.max_transmit_size(1024 * 1024)
|
||||
.validation_mode(ValidationMode::Strict)
|
||||
.build()
|
||||
.expect("the configuration should always be valid"),
|
||||
)
|
||||
.expect("creating gossipsub behavior should always work")
|
||||
}
|
||||
}
|
||||
@@ -1,7 +0,0 @@
|
||||
// maybe this will hold test in the future...??
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
#[test]
|
||||
fn does_nothing() {}
|
||||
}
|
||||
@@ -1,2 +0,0 @@
|
||||
[toolchain]
|
||||
channel = "nightly"
|
||||
@@ -1,15 +0,0 @@
|
||||
[package]
|
||||
name = "util"
|
||||
version = { workspace = true }
|
||||
edition = { workspace = true }
|
||||
publish = false
|
||||
|
||||
[lib]
|
||||
doctest = false
|
||||
name = "util"
|
||||
path = "src/lib.rs"
|
||||
|
||||
[lints]
|
||||
workspace = true
|
||||
|
||||
[dependencies]
|
||||
@@ -1 +0,0 @@
|
||||
pub mod wakerdeque;
|
||||
@@ -1,55 +0,0 @@
|
||||
use std::collections::VecDeque;
|
||||
use std::fmt::{Debug, Formatter};
|
||||
use std::task::{Context, Waker};
|
||||
|
||||
/// A wrapper around [`VecDeque`] which wakes (if it can) on any `push_*` methods,
|
||||
/// and updates the internally stored waker by consuming [`Context`] on any `pop_*` methods.
|
||||
pub struct WakerDeque<T> {
|
||||
waker: Option<Waker>,
|
||||
deque: VecDeque<T>,
|
||||
}
|
||||
|
||||
impl<T: Debug> Debug for WakerDeque<T> {
|
||||
fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
|
||||
self.deque.fmt(f)
|
||||
}
|
||||
}
|
||||
|
||||
impl<T> WakerDeque<T> {
|
||||
pub fn new() -> Self {
|
||||
Self {
|
||||
waker: None,
|
||||
deque: VecDeque::new(),
|
||||
}
|
||||
}
|
||||
|
||||
fn update(&mut self, cx: &mut Context<'_>) {
|
||||
self.waker = Some(cx.waker().clone());
|
||||
}
|
||||
|
||||
fn wake(&mut self) {
|
||||
let Some(ref mut w) = self.waker else { return };
|
||||
w.wake_by_ref();
|
||||
self.waker = None;
|
||||
}
|
||||
|
||||
pub fn pop_front(&mut self, cx: &mut Context<'_>) -> Option<T> {
|
||||
self.update(cx);
|
||||
self.deque.pop_front()
|
||||
}
|
||||
|
||||
pub fn pop_back(&mut self, cx: &mut Context<'_>) -> Option<T> {
|
||||
self.update(cx);
|
||||
self.deque.pop_back()
|
||||
}
|
||||
|
||||
pub fn push_front(&mut self, value: T) {
|
||||
self.wake();
|
||||
self.deque.push_front(value);
|
||||
}
|
||||
|
||||
pub fn push_back(&mut self, value: T) {
|
||||
self.wake();
|
||||
self.deque.push_back(value);
|
||||
}
|
||||
}
|
||||
@@ -1,4 +1,5 @@
|
||||
import argparse
|
||||
import importlib.metadata
|
||||
import itertools
|
||||
import multiprocessing as mp
|
||||
import os
|
||||
@@ -44,9 +45,9 @@ class Node:
|
||||
@classmethod
|
||||
async def create(cls, args: "Args") -> "Self":
|
||||
keypair = get_node_id_keypair()
|
||||
node_id = NodeId(keypair.to_peer_id().to_base58())
|
||||
node_id = NodeId(keypair.to_string())
|
||||
session_id = SessionId(master_node_id=node_id, election_clock=0)
|
||||
router = Router.create(keypair)
|
||||
router = Router.create(keypair, namespace=args.namespace)
|
||||
await router.register_topic(topics.GLOBAL_EVENTS)
|
||||
await router.register_topic(topics.LOCAL_EVENTS)
|
||||
await router.register_topic(topics.COMMANDS)
|
||||
@@ -72,7 +73,7 @@ class Node:
|
||||
else:
|
||||
download_coordinator = None
|
||||
|
||||
if args.spawn_api:
|
||||
if not args.no_api:
|
||||
api = API(
|
||||
node_id,
|
||||
session_id,
|
||||
@@ -258,7 +259,7 @@ def main():
|
||||
# TODO: Refactor the current verbosity system
|
||||
logger_setup(EXO_LOG, args.verbosity)
|
||||
logger.info("Starting EXO")
|
||||
logger.info(f"EXO_LIBP2P_NAMESPACE: {os.getenv('EXO_LIBP2P_NAMESPACE')}")
|
||||
logger.info(f"Namespace: {args.namespace}")
|
||||
|
||||
# Set FAST_SYNCH override env var for runner subprocesses
|
||||
if args.fast_synch is True:
|
||||
@@ -275,13 +276,13 @@ def main():
|
||||
|
||||
|
||||
class Args(CamelCaseModel):
|
||||
verbosity: int = 0
|
||||
force_master: bool = False
|
||||
spawn_api: bool = False
|
||||
api_port: PositiveInt = 52415
|
||||
tb_only: bool = False
|
||||
verbosity: int
|
||||
force_master: bool
|
||||
no_api: bool
|
||||
api_port: PositiveInt
|
||||
no_worker: bool = False
|
||||
no_downloads: bool = False
|
||||
namespace: str
|
||||
fast_synch: bool | None = None # None = auto, True = force on, False = force off
|
||||
|
||||
@classmethod
|
||||
@@ -311,14 +312,15 @@ class Args(CamelCaseModel):
|
||||
)
|
||||
parser.add_argument(
|
||||
"--no-api",
|
||||
action="store_false",
|
||||
dest="spawn_api",
|
||||
action="store_true",
|
||||
help="Disable the API server for this node",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--api-port",
|
||||
type=int,
|
||||
dest="api_port",
|
||||
default=52415,
|
||||
help="Which port the API server will be available on",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--no-worker",
|
||||
@@ -329,6 +331,11 @@ class Args(CamelCaseModel):
|
||||
action="store_true",
|
||||
help="Disable the download coordinator (node won't download models)",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--namespace",
|
||||
default=importlib.metadata.version("exo"),
|
||||
help="Set the EXO namespace to run multiple isolated clusters",
|
||||
)
|
||||
fast_synch_group = parser.add_mutually_exclusive_group()
|
||||
fast_synch_group.add_argument(
|
||||
"--fast-synch",
|
||||
|
||||
@@ -71,11 +71,8 @@ from exo.shared.types.api import (
|
||||
ChatCompletionResponse,
|
||||
CreateInstanceParams,
|
||||
CreateInstanceResponse,
|
||||
CreateMetaInstanceParams,
|
||||
CreateMetaInstanceResponse,
|
||||
DeleteDownloadResponse,
|
||||
DeleteInstanceResponse,
|
||||
DeleteMetaInstanceResponse,
|
||||
ErrorInfo,
|
||||
ErrorResponse,
|
||||
FinishReason,
|
||||
@@ -118,10 +115,8 @@ from exo.shared.types.claude_api import (
|
||||
from exo.shared.types.commands import (
|
||||
Command,
|
||||
CreateInstance,
|
||||
CreateMetaInstance,
|
||||
DeleteDownload,
|
||||
DeleteInstance,
|
||||
DeleteMetaInstance,
|
||||
DownloadCommand,
|
||||
ForwarderCommand,
|
||||
ForwarderDownloadCommand,
|
||||
@@ -133,7 +128,7 @@ from exo.shared.types.commands import (
|
||||
TaskFinished,
|
||||
TextGeneration,
|
||||
)
|
||||
from exo.shared.types.common import CommandId, Id, MetaInstanceId, NodeId, SessionId
|
||||
from exo.shared.types.common import CommandId, Id, NodeId, SessionId
|
||||
from exo.shared.types.events import (
|
||||
ChunkGenerated,
|
||||
Event,
|
||||
@@ -142,7 +137,6 @@ from exo.shared.types.events import (
|
||||
TracesMerged,
|
||||
)
|
||||
from exo.shared.types.memory import Memory
|
||||
from exo.shared.types.meta_instance import MetaInstance
|
||||
from exo.shared.types.openai_responses import (
|
||||
ResponsesRequest,
|
||||
ResponsesResponse,
|
||||
@@ -281,8 +275,6 @@ class API:
|
||||
self.app.get("/instance/previews")(self.get_placement_previews)
|
||||
self.app.get("/instance/{instance_id}")(self.get_instance)
|
||||
self.app.delete("/instance/{instance_id}")(self.delete_instance)
|
||||
self.app.post("/meta_instance")(self.create_meta_instance)
|
||||
self.app.delete("/meta_instance/{meta_instance_id}")(self.delete_meta_instance)
|
||||
self.app.get("/models")(self.get_models)
|
||||
self.app.get("/v1/models")(self.get_models)
|
||||
self.app.post("/models/add")(self.add_custom_model)
|
||||
@@ -529,46 +521,6 @@ class API:
|
||||
instance_id=instance_id,
|
||||
)
|
||||
|
||||
async def create_meta_instance(
|
||||
self, payload: CreateMetaInstanceParams
|
||||
) -> CreateMetaInstanceResponse:
|
||||
meta_instance = MetaInstance(
|
||||
model_id=payload.model_id,
|
||||
sharding=payload.sharding,
|
||||
instance_meta=payload.instance_meta,
|
||||
min_nodes=payload.min_nodes,
|
||||
node_ids=payload.node_ids,
|
||||
)
|
||||
command = CreateMetaInstance(meta_instance=meta_instance)
|
||||
await self._send(command)
|
||||
return CreateMetaInstanceResponse(
|
||||
message="Command received.",
|
||||
command_id=command.command_id,
|
||||
meta_instance_id=meta_instance.meta_instance_id,
|
||||
)
|
||||
|
||||
async def delete_meta_instance(
|
||||
self, meta_instance_id: MetaInstanceId
|
||||
) -> DeleteMetaInstanceResponse:
|
||||
meta = self.state.meta_instances.get(meta_instance_id)
|
||||
if not meta:
|
||||
raise HTTPException(status_code=404, detail="MetaInstance not found")
|
||||
|
||||
# Delete MetaInstance first to prevent reconciler from re-placing
|
||||
command = DeleteMetaInstance(meta_instance_id=meta_instance_id)
|
||||
await self._send(command)
|
||||
|
||||
# Then cascade-delete any backing instances
|
||||
for instance_id, instance in self.state.instances.items():
|
||||
if instance.meta_instance_id == meta_instance_id:
|
||||
await self._send(DeleteInstance(instance_id=instance_id))
|
||||
|
||||
return DeleteMetaInstanceResponse(
|
||||
message="Command received.",
|
||||
command_id=command.command_id,
|
||||
meta_instance_id=meta_instance_id,
|
||||
)
|
||||
|
||||
async def _token_chunk_stream(
|
||||
self, command_id: CommandId
|
||||
) -> AsyncGenerator[ErrorChunk | ToolCallChunk | TokenChunk, None]:
|
||||
|
||||
@@ -1,5 +1,4 @@
|
||||
from collections.abc import Sequence
|
||||
from datetime import datetime, timezone
|
||||
from datetime import datetime, timedelta, timezone
|
||||
|
||||
import anyio
|
||||
from anyio.abc import TaskGroup
|
||||
@@ -13,19 +12,11 @@ from exo.master.placement import (
|
||||
get_transition_events,
|
||||
place_instance,
|
||||
)
|
||||
from exo.master.process_managers import ProcessManager
|
||||
from exo.master.process_managers.instance_health import InstanceHealthReconciler
|
||||
from exo.master.process_managers.meta_instance import MetaInstanceReconciler
|
||||
from exo.master.process_managers.node_timeout import NodeTimeoutReconciler
|
||||
from exo.master.reconcile import try_place_for_meta_instance
|
||||
from exo.shared.apply import apply
|
||||
from exo.shared.constants import EXO_EVENT_LOG_DIR, EXO_TRACING_ENABLED
|
||||
from exo.shared.models.model_cards import ModelCard
|
||||
from exo.shared.types.commands import (
|
||||
CreateInstance,
|
||||
CreateMetaInstance,
|
||||
DeleteInstance,
|
||||
DeleteMetaInstance,
|
||||
ForwarderCommand,
|
||||
ForwarderDownloadCommand,
|
||||
ImageEdits,
|
||||
@@ -43,9 +34,9 @@ from exo.shared.types.events import (
|
||||
ForwarderEvent,
|
||||
IndexedEvent,
|
||||
InputChunkReceived,
|
||||
MetaInstanceCreated,
|
||||
MetaInstanceDeleted,
|
||||
InstanceDeleted,
|
||||
NodeGatheredInfo,
|
||||
NodeTimedOut,
|
||||
TaskCreated,
|
||||
TaskDeleted,
|
||||
TraceEventData,
|
||||
@@ -67,7 +58,7 @@ from exo.shared.types.tasks import (
|
||||
TextGeneration as TextGenerationTask,
|
||||
)
|
||||
from exo.shared.types.worker.instances import InstanceId
|
||||
from exo.utils.channels import Receiver, Sender
|
||||
from exo.utils.channels import Receiver, Sender, channel
|
||||
from exo.utils.event_buffer import MultiSourceBuffer
|
||||
|
||||
|
||||
@@ -91,15 +82,16 @@ class Master:
|
||||
self.local_event_receiver = local_event_receiver
|
||||
self.global_event_sender = global_event_sender
|
||||
self.download_command_sender = download_command_sender
|
||||
send, recv = channel[Event]()
|
||||
self.event_sender: Sender[Event] = send
|
||||
self._loopback_event_receiver: Receiver[Event] = recv
|
||||
self._loopback_event_sender: Sender[ForwarderEvent] = (
|
||||
local_event_receiver.clone_sender()
|
||||
)
|
||||
self._multi_buffer = MultiSourceBuffer[NodeId, Event]()
|
||||
self._event_log = DiskEventLog(EXO_EVENT_LOG_DIR / "master")
|
||||
self._pending_traces: dict[TaskId, dict[int, list[TraceEventData]]] = {}
|
||||
self._expected_ranks: dict[TaskId, set[int]] = {}
|
||||
self._process_managers: Sequence[ProcessManager] = [
|
||||
InstanceHealthReconciler(),
|
||||
NodeTimeoutReconciler(),
|
||||
MetaInstanceReconciler(),
|
||||
]
|
||||
|
||||
async def run(self):
|
||||
logger.info("Starting Master")
|
||||
@@ -108,12 +100,15 @@ class Master:
|
||||
async with self._tg as tg:
|
||||
tg.start_soon(self._event_processor)
|
||||
tg.start_soon(self._command_processor)
|
||||
tg.start_soon(self._reconcile)
|
||||
tg.start_soon(self._loopback_processor)
|
||||
tg.start_soon(self._plan)
|
||||
finally:
|
||||
self._event_log.close()
|
||||
self.global_event_sender.close()
|
||||
self.local_event_receiver.close()
|
||||
self.command_receiver.close()
|
||||
self._loopback_event_sender.close()
|
||||
self._loopback_event_receiver.close()
|
||||
|
||||
async def shutdown(self):
|
||||
logger.info("Stopping Master")
|
||||
@@ -295,29 +290,6 @@ class Master:
|
||||
)
|
||||
)
|
||||
generated_events.extend(transition_events)
|
||||
case CreateMetaInstance():
|
||||
generated_events.append(
|
||||
MetaInstanceCreated(meta_instance=command.meta_instance)
|
||||
)
|
||||
# Immediate placement attempt for responsiveness
|
||||
model_card = await ModelCard.load(
|
||||
command.meta_instance.model_id
|
||||
)
|
||||
result = try_place_for_meta_instance(
|
||||
command.meta_instance,
|
||||
model_card,
|
||||
self.state.topology,
|
||||
self.state.instances,
|
||||
self.state.node_memory,
|
||||
self.state.node_network,
|
||||
)
|
||||
generated_events.extend(result.events)
|
||||
case DeleteMetaInstance():
|
||||
generated_events.append(
|
||||
MetaInstanceDeleted(
|
||||
meta_instance_id=command.meta_instance_id
|
||||
)
|
||||
)
|
||||
case PlaceInstance():
|
||||
placement = place_instance(
|
||||
command,
|
||||
@@ -369,32 +341,31 @@ class Master:
|
||||
):
|
||||
await self._send_event(IndexedEvent(idx=i, event=event))
|
||||
for event in generated_events:
|
||||
await self._apply_and_broadcast(event)
|
||||
await self.event_sender.send(event)
|
||||
except ValueError as e:
|
||||
logger.opt(exception=e).warning("Error in command processor")
|
||||
|
||||
async def _apply_and_broadcast(self, event: Event) -> None:
|
||||
"""Apply event to state, persist to disk, and broadcast to workers.
|
||||
|
||||
State is updated synchronously (before any await), so callers can
|
||||
rely on ``self.state`` reflecting this event immediately after the
|
||||
call. Python's cooperative scheduling guarantees no interleaving
|
||||
between the state read and write.
|
||||
"""
|
||||
logger.debug(f"Master indexing event: {str(event)[:100]}")
|
||||
indexed = IndexedEvent(event=event, idx=len(self._event_log))
|
||||
self.state = apply(self.state, indexed)
|
||||
event._master_time_stamp = datetime.now(tz=timezone.utc) # pyright: ignore[reportPrivateUsage]
|
||||
self._event_log.append(event)
|
||||
await self._send_event(indexed)
|
||||
|
||||
async def _reconcile(self) -> None:
|
||||
# These plan loops are the cracks showing in our event sourcing architecture - more things could be commands
|
||||
async def _plan(self) -> None:
|
||||
while True:
|
||||
for pm in self._process_managers:
|
||||
events = await pm.reconcile(self.state)
|
||||
for event in events:
|
||||
await self._apply_and_broadcast(event)
|
||||
await anyio.sleep(1)
|
||||
# kill broken instances
|
||||
connected_node_ids = set(self.state.topology.list_nodes())
|
||||
for instance_id, instance in self.state.instances.items():
|
||||
for node_id in instance.shard_assignments.node_to_runner:
|
||||
if node_id not in connected_node_ids:
|
||||
await self.event_sender.send(
|
||||
InstanceDeleted(instance_id=instance_id)
|
||||
)
|
||||
break
|
||||
|
||||
# time out dead nodes
|
||||
for node_id, time in self.state.last_seen.items():
|
||||
now = datetime.now(tz=timezone.utc)
|
||||
if now - time > timedelta(seconds=30):
|
||||
logger.info(f"Manually removing node {node_id} due to inactivity")
|
||||
await self.event_sender.send(NodeTimedOut(node_id=node_id))
|
||||
|
||||
await anyio.sleep(10)
|
||||
|
||||
async def _event_processor(self) -> None:
|
||||
with self.local_event_receiver as local_events:
|
||||
@@ -412,10 +383,32 @@ class Master:
|
||||
await self._handle_traces_collected(event)
|
||||
continue
|
||||
|
||||
logger.trace(f"Master indexing event: {str(event)[:100]}")
|
||||
indexed = IndexedEvent(event=event, idx=len(self._event_log))
|
||||
self.state = apply(self.state, indexed)
|
||||
|
||||
event._master_time_stamp = datetime.now(tz=timezone.utc) # pyright: ignore[reportPrivateUsage]
|
||||
if isinstance(event, NodeGatheredInfo):
|
||||
event.when = str(datetime.now(tz=timezone.utc))
|
||||
|
||||
await self._apply_and_broadcast(event)
|
||||
self._event_log.append(event)
|
||||
await self._send_event(indexed)
|
||||
|
||||
async def _loopback_processor(self) -> None:
|
||||
# this would ideally not be necessary.
|
||||
# this is WAY less hacky than how I was working around this before
|
||||
local_index = 0
|
||||
with self._loopback_event_receiver as events:
|
||||
async for event in events:
|
||||
await self._loopback_event_sender.send(
|
||||
ForwarderEvent(
|
||||
origin=NodeId(f"master_{self.node_id}"),
|
||||
origin_idx=local_index,
|
||||
session=self.session_id,
|
||||
event=event,
|
||||
)
|
||||
)
|
||||
local_index += 1
|
||||
|
||||
# This function is re-entrant, take care!
|
||||
async def _send_event(self, event: IndexedEvent):
|
||||
@@ -447,7 +440,7 @@ class Master:
|
||||
for trace_data in self._pending_traces[task_id].values():
|
||||
all_trace_data.extend(trace_data)
|
||||
|
||||
await self._apply_and_broadcast(
|
||||
await self.event_sender.send(
|
||||
TracesMerged(task_id=task_id, traces=all_trace_data)
|
||||
)
|
||||
|
||||
|
||||
@@ -63,9 +63,7 @@ def place_instance(
|
||||
required_nodes: set[NodeId] | None = None,
|
||||
) -> dict[InstanceId, Instance]:
|
||||
cycles = topology.get_cycles()
|
||||
candidate_cycles = list(
|
||||
filter(lambda it: len(it) >= command.min_nodes, cycles)
|
||||
)
|
||||
candidate_cycles = list(filter(lambda it: len(it) >= command.min_nodes, cycles))
|
||||
|
||||
# Filter to cycles containing all required nodes (subset matching)
|
||||
if required_nodes:
|
||||
@@ -108,11 +106,7 @@ def place_instance(
|
||||
cycle for cycle in smallest_cycles if topology.is_rdma_cycle(cycle)
|
||||
]
|
||||
|
||||
if command.instance_meta == InstanceMeta.MlxJaccl:
|
||||
if not smallest_rdma_cycles:
|
||||
raise ValueError(
|
||||
"Requested RDMA (MlxJaccl) but no RDMA-connected cycles available"
|
||||
)
|
||||
if command.instance_meta == InstanceMeta.MlxJaccl and smallest_rdma_cycles != []:
|
||||
smallest_cycles = smallest_rdma_cycles
|
||||
|
||||
cycles_with_leaf_nodes: list[Cycle] = [
|
||||
|
||||
@@ -1,12 +0,0 @@
|
||||
from collections.abc import Sequence
|
||||
from typing import Protocol, runtime_checkable
|
||||
|
||||
from exo.shared.types.events import Event
|
||||
from exo.shared.types.state import State
|
||||
|
||||
|
||||
@runtime_checkable
|
||||
class ProcessManager(Protocol):
|
||||
"""A reconciliation step that examines state and returns corrective events."""
|
||||
|
||||
async def reconcile(self, state: State) -> Sequence[Event]: ...
|
||||
@@ -1,49 +0,0 @@
|
||||
from collections.abc import Sequence
|
||||
from typing import final
|
||||
|
||||
from exo.master.reconcile import instance_connections_healthy, instance_runners_failed
|
||||
from exo.shared.types.events import Event, InstanceDeleted, InstanceRetrying
|
||||
from exo.shared.types.state import State
|
||||
|
||||
MAX_INSTANCE_RETRIES = 3
|
||||
|
||||
|
||||
@final
|
||||
class InstanceHealthReconciler:
|
||||
"""Delete instances whose network connections are broken or whose runners have all failed."""
|
||||
|
||||
async def reconcile(self, state: State) -> Sequence[Event]:
|
||||
events: list[Event] = []
|
||||
for instance_id, instance in state.instances.items():
|
||||
if not instance_connections_healthy(instance, state.topology):
|
||||
events.append(
|
||||
InstanceDeleted(
|
||||
instance_id=instance_id,
|
||||
failure_error="Network connection lost",
|
||||
)
|
||||
)
|
||||
continue
|
||||
|
||||
is_failed, error_message = instance_runners_failed(
|
||||
instance, state.runners, state.node_identities
|
||||
)
|
||||
if is_failed:
|
||||
# Retry within the same instance if backed by a MetaInstance
|
||||
mid = instance.meta_instance_id
|
||||
mi = state.meta_instances.get(mid) if mid else None
|
||||
if mid and mi and mi.consecutive_failures < MAX_INSTANCE_RETRIES:
|
||||
events.append(
|
||||
InstanceRetrying(
|
||||
instance_id=instance_id,
|
||||
meta_instance_id=mid,
|
||||
failure_error=error_message or "Runner failed",
|
||||
)
|
||||
)
|
||||
else:
|
||||
events.append(
|
||||
InstanceDeleted(
|
||||
instance_id=instance_id,
|
||||
failure_error=error_message,
|
||||
)
|
||||
)
|
||||
return events
|
||||
@@ -1,53 +0,0 @@
|
||||
from collections.abc import Sequence
|
||||
from typing import final
|
||||
|
||||
from exo.master.reconcile import (
|
||||
find_unsatisfied_meta_instances,
|
||||
try_place_for_meta_instance,
|
||||
)
|
||||
from exo.shared.models.model_cards import ModelCard
|
||||
from exo.shared.types.events import Event, InstanceCreated, MetaInstancePlacementFailed
|
||||
from exo.shared.types.state import State
|
||||
from exo.shared.types.worker.instances import Instance, InstanceId
|
||||
|
||||
|
||||
@final
|
||||
class MetaInstanceReconciler:
|
||||
"""Place instances for unsatisfied MetaInstances."""
|
||||
|
||||
async def reconcile(self, state: State) -> Sequence[Event]:
|
||||
all_events: list[Event] = []
|
||||
# Local copy for intermediate tracking — so placement of B
|
||||
# sees A's instance and doesn't double-place on same resources.
|
||||
current_instances: dict[InstanceId, Instance] = dict(state.instances)
|
||||
|
||||
unsatisfied = find_unsatisfied_meta_instances(
|
||||
state.meta_instances,
|
||||
current_instances,
|
||||
state.topology,
|
||||
)
|
||||
for meta_instance in unsatisfied:
|
||||
model_card = await ModelCard.load(meta_instance.model_id)
|
||||
result = try_place_for_meta_instance(
|
||||
meta_instance,
|
||||
model_card,
|
||||
state.topology,
|
||||
current_instances,
|
||||
state.node_memory,
|
||||
state.node_network,
|
||||
)
|
||||
# Update local instance map so next placement sees this one
|
||||
for event in result.events:
|
||||
if isinstance(event, InstanceCreated):
|
||||
current_instances[event.instance.instance_id] = event.instance
|
||||
all_events.extend(result.events)
|
||||
|
||||
# Emit placement failure if error differs from what's already in state
|
||||
if result.error is not None and meta_instance.placement_error != result.error:
|
||||
all_events.append(
|
||||
MetaInstancePlacementFailed(
|
||||
meta_instance_id=meta_instance.meta_instance_id,
|
||||
reason=result.error,
|
||||
)
|
||||
)
|
||||
return all_events
|
||||
@@ -1,27 +0,0 @@
|
||||
from collections.abc import Sequence
|
||||
from datetime import datetime, timedelta, timezone
|
||||
from typing import final
|
||||
|
||||
from loguru import logger
|
||||
|
||||
from exo.shared.types.events import Event, NodeTimedOut
|
||||
from exo.shared.types.state import State
|
||||
|
||||
_DEFAULT_TIMEOUT = timedelta(seconds=30)
|
||||
|
||||
|
||||
@final
|
||||
class NodeTimeoutReconciler:
|
||||
"""Time out nodes that haven't been seen recently."""
|
||||
|
||||
def __init__(self, timeout: timedelta = _DEFAULT_TIMEOUT) -> None:
|
||||
self.timeout = timeout
|
||||
|
||||
async def reconcile(self, state: State) -> Sequence[Event]:
|
||||
now = datetime.now(tz=timezone.utc)
|
||||
events: list[Event] = []
|
||||
for node_id, last_seen in state.last_seen.items():
|
||||
if now - last_seen > self.timeout:
|
||||
logger.info(f"Removing node {node_id} due to inactivity")
|
||||
events.append(NodeTimedOut(node_id=node_id))
|
||||
return events
|
||||
@@ -1,236 +0,0 @@
|
||||
from collections.abc import Mapping, Sequence
|
||||
from typing import NamedTuple
|
||||
|
||||
from loguru import logger
|
||||
|
||||
from exo.master.placement import get_transition_events, place_instance
|
||||
from exo.shared.models.model_cards import ModelCard
|
||||
from exo.shared.topology import Topology
|
||||
from exo.shared.types.commands import PlaceInstance
|
||||
from exo.shared.types.common import MetaInstanceId, NodeId
|
||||
from exo.shared.types.events import Event
|
||||
from exo.shared.types.meta_instance import MetaInstance
|
||||
from exo.shared.types.profiling import MemoryUsage, NodeIdentity, NodeNetworkInfo
|
||||
from exo.shared.types.topology import RDMAConnection, SocketConnection
|
||||
from exo.shared.types.worker.instances import (
|
||||
BaseInstance,
|
||||
Instance,
|
||||
InstanceId,
|
||||
MlxJacclInstance,
|
||||
MlxRingInstance,
|
||||
)
|
||||
from exo.shared.types.worker.runners import (
|
||||
RunnerFailed,
|
||||
RunnerId,
|
||||
RunnerShutdown,
|
||||
RunnerStatus,
|
||||
)
|
||||
|
||||
|
||||
class PlacementResult(NamedTuple):
|
||||
"""Result of a placement attempt: events to apply and optional error reason."""
|
||||
|
||||
events: Sequence[Event]
|
||||
error: str | None
|
||||
|
||||
|
||||
def _get_ring_order(instance: BaseInstance) -> list[NodeId]:
|
||||
"""Reconstruct ring order from shard device_rank."""
|
||||
node_ranks: list[tuple[NodeId, int]] = []
|
||||
for node_id, runner_id in instance.shard_assignments.node_to_runner.items():
|
||||
shard = instance.shard_assignments.runner_to_shard[runner_id]
|
||||
node_ranks.append((node_id, shard.device_rank))
|
||||
node_ranks.sort(key=lambda x: x[1])
|
||||
return [node_id for node_id, _ in node_ranks]
|
||||
|
||||
|
||||
def _ring_connections_healthy(instance: MlxRingInstance, topology: Topology) -> bool:
|
||||
"""Check that the specific IPs used by a ring instance still exist in the topology."""
|
||||
ring = _get_ring_order(instance)
|
||||
n = len(ring)
|
||||
for node in ring:
|
||||
hosts = instance.hosts_by_node[node]
|
||||
for idx in range(n):
|
||||
host = hosts[idx]
|
||||
if host.ip in ("0.0.0.0", "198.51.100.1"):
|
||||
continue # self or placeholder
|
||||
# Real connection: node → ring[idx]. Check specific IP.
|
||||
connections = topology.get_all_connections_between(node, ring[idx])
|
||||
if not any(
|
||||
isinstance(c, SocketConnection)
|
||||
and c.sink_multiaddr.ip_address == host.ip
|
||||
for c in connections
|
||||
):
|
||||
return False
|
||||
return True
|
||||
|
||||
|
||||
def _jaccl_connections_healthy(instance: MlxJacclInstance, topology: Topology) -> bool:
|
||||
"""Check that the specific RDMA interfaces used by a JACCL instance still exist."""
|
||||
ring = _get_ring_order(instance)
|
||||
n = len(ring)
|
||||
for i in range(n):
|
||||
for j in range(n):
|
||||
iface = instance.jaccl_devices[i][j]
|
||||
if iface is None:
|
||||
continue
|
||||
connections = topology.get_all_connections_between(ring[i], ring[j])
|
||||
if not any(
|
||||
isinstance(c, RDMAConnection) and c.source_rdma_iface == iface
|
||||
for c in connections
|
||||
):
|
||||
return False
|
||||
return True
|
||||
|
||||
|
||||
def instance_connections_healthy(instance: Instance, topology: Topology) -> bool:
|
||||
"""Check that an instance's nodes and specific connections are still in the topology."""
|
||||
instance_nodes = set(instance.shard_assignments.node_to_runner.keys())
|
||||
if not all(topology.contains_node(n) for n in instance_nodes):
|
||||
return False
|
||||
if len(instance_nodes) <= 1:
|
||||
return True
|
||||
match instance:
|
||||
case MlxRingInstance():
|
||||
return _ring_connections_healthy(instance, topology)
|
||||
case MlxJacclInstance():
|
||||
return _jaccl_connections_healthy(instance, topology)
|
||||
|
||||
|
||||
def instance_runners_failed(
|
||||
instance: Instance,
|
||||
runners: Mapping[RunnerId, RunnerStatus],
|
||||
node_identities: Mapping[NodeId, NodeIdentity],
|
||||
) -> tuple[bool, str | None]:
|
||||
"""Check if an instance's runners have all reached terminal failure states.
|
||||
|
||||
Returns ``(True, error_message)`` when ALL runners are terminal
|
||||
(``RunnerFailed`` or ``RunnerShutdown``) and at least one is ``RunnerFailed``.
|
||||
|
||||
Returns ``(False, None)`` when runners are still active, haven't reported
|
||||
yet, or all gracefully shut down (no ``RunnerFailed``).
|
||||
"""
|
||||
instance_runner_ids = set(instance.shard_assignments.node_to_runner.values())
|
||||
|
||||
if not instance_runner_ids:
|
||||
return False, None
|
||||
|
||||
# Build reverse mapping: runner_id -> node_id
|
||||
runner_to_node: dict[RunnerId, NodeId] = {
|
||||
runner_id: node_id
|
||||
for node_id, runner_id in instance.shard_assignments.node_to_runner.items()
|
||||
}
|
||||
|
||||
has_any_failed = False
|
||||
error_messages: list[str] = []
|
||||
|
||||
for runner_id in instance_runner_ids:
|
||||
status = runners.get(runner_id)
|
||||
if status is None:
|
||||
# Runner hasn't reported yet — instance is still starting
|
||||
return False, None
|
||||
if isinstance(status, RunnerFailed):
|
||||
has_any_failed = True
|
||||
if status.error_message:
|
||||
node_id = runner_to_node.get(runner_id)
|
||||
name = node_identities[node_id].friendly_name if node_id and node_id in node_identities else node_id or "unknown"
|
||||
error_messages.append(f"{name}: {status.error_message}")
|
||||
elif isinstance(status, RunnerShutdown):
|
||||
pass # Terminal but not a failure indicator on its own
|
||||
else:
|
||||
# Runner is still active (connecting, loading, running, etc.)
|
||||
return False, None
|
||||
|
||||
if has_any_failed:
|
||||
return True, "; ".join(error_messages) if error_messages else "Runner failed"
|
||||
|
||||
# All runners are Shutdown but none Failed — graceful shutdown, not a failure
|
||||
return False, None
|
||||
|
||||
|
||||
def instance_satisfies_meta_instance(
|
||||
meta_instance: MetaInstance,
|
||||
instance: Instance,
|
||||
) -> bool:
|
||||
"""Check if a single instance satisfies a meta-instance's constraints.
|
||||
|
||||
This is a pure constraint check (model, min_nodes, node_ids).
|
||||
Use ``instance_connections_healthy`` separately for topology health.
|
||||
"""
|
||||
if instance.shard_assignments.model_id != meta_instance.model_id:
|
||||
return False
|
||||
|
||||
instance_nodes = set(instance.shard_assignments.node_to_runner.keys())
|
||||
|
||||
if len(instance_nodes) < meta_instance.min_nodes:
|
||||
return False
|
||||
|
||||
return meta_instance.node_ids is None or set(meta_instance.node_ids).issubset(
|
||||
instance_nodes
|
||||
)
|
||||
|
||||
|
||||
def find_unsatisfied_meta_instances(
|
||||
meta_instances: Mapping[MetaInstanceId, MetaInstance],
|
||||
instances: Mapping[InstanceId, Instance],
|
||||
topology: Topology,
|
||||
) -> Sequence[MetaInstance]:
|
||||
"""Return meta-instances that have no healthy backing instance."""
|
||||
unsatisfied: list[MetaInstance] = []
|
||||
for meta_id, meta_instance in meta_instances.items():
|
||||
has_healthy_backing = any(
|
||||
instance.meta_instance_id == meta_id
|
||||
and instance_connections_healthy(instance, topology)
|
||||
for instance in instances.values()
|
||||
)
|
||||
if not has_healthy_backing:
|
||||
unsatisfied.append(meta_instance)
|
||||
return unsatisfied
|
||||
|
||||
|
||||
def try_place_for_meta_instance(
|
||||
meta_instance: MetaInstance,
|
||||
model_card: ModelCard,
|
||||
topology: Topology,
|
||||
current_instances: Mapping[InstanceId, Instance],
|
||||
node_memory: Mapping[NodeId, MemoryUsage],
|
||||
node_network: Mapping[NodeId, NodeNetworkInfo],
|
||||
) -> PlacementResult:
|
||||
"""Try to place an instance satisfying the meta-instance constraints.
|
||||
|
||||
Returns a :class:`PlacementResult` with events on success, or an error
|
||||
reason on failure.
|
||||
"""
|
||||
command = PlaceInstance(
|
||||
model_card=model_card,
|
||||
sharding=meta_instance.sharding,
|
||||
instance_meta=meta_instance.instance_meta,
|
||||
min_nodes=meta_instance.min_nodes,
|
||||
)
|
||||
try:
|
||||
target_instances = place_instance(
|
||||
command,
|
||||
topology,
|
||||
current_instances,
|
||||
node_memory,
|
||||
node_network,
|
||||
required_nodes=(
|
||||
set(meta_instance.node_ids) if meta_instance.node_ids else None
|
||||
),
|
||||
)
|
||||
# Tag the new instance with meta_instance_id
|
||||
new_instance_ids = set(target_instances.keys()) - set(current_instances.keys())
|
||||
if new_instance_ids:
|
||||
new_id = next(iter(new_instance_ids))
|
||||
target_instances[new_id] = target_instances[new_id].model_copy(
|
||||
update={"meta_instance_id": meta_instance.meta_instance_id}
|
||||
)
|
||||
return PlacementResult(
|
||||
events=list(get_transition_events(current_instances, target_instances)),
|
||||
error=None,
|
||||
)
|
||||
except ValueError as e:
|
||||
logger.debug(
|
||||
f"MetaInstance placement not possible for {meta_instance.model_id}: {e}"
|
||||
)
|
||||
return PlacementResult(events=[], error=str(e))
|
||||
@@ -42,7 +42,7 @@ from exo.utils.channels import channel
|
||||
@pytest.mark.asyncio
|
||||
async def test_master():
|
||||
keypair = get_node_id_keypair()
|
||||
node_id = NodeId(keypair.to_peer_id().to_base58())
|
||||
node_id = NodeId(keypair.to_string())
|
||||
session_id = SessionId(master_node_id=node_id, election_clock=0)
|
||||
|
||||
ge_sender, global_event_receiver = channel[ForwarderEvent]()
|
||||
@@ -75,7 +75,7 @@ async def test_master():
|
||||
async with anyio.create_task_group() as tg:
|
||||
tg.start_soon(master.run)
|
||||
|
||||
sender_node_id = NodeId(f"{keypair.to_peer_id().to_base58()}_sender")
|
||||
sender_node_id = NodeId(f"{keypair.to_string()}_sender")
|
||||
# inject a NodeGatheredInfo event
|
||||
logger.info("inject a NodeGatheredInfo event")
|
||||
await local_event_sender.send(
|
||||
|
||||
@@ -1,750 +0,0 @@
|
||||
from exo.master.process_managers.instance_health import InstanceHealthReconciler
|
||||
from exo.master.reconcile import (
|
||||
find_unsatisfied_meta_instances,
|
||||
instance_connections_healthy,
|
||||
instance_runners_failed,
|
||||
instance_satisfies_meta_instance,
|
||||
)
|
||||
from exo.shared.apply import apply
|
||||
from exo.shared.models.model_cards import ModelCard, ModelId, ModelTask
|
||||
from exo.shared.topology import Topology
|
||||
from exo.shared.types.common import Host, MetaInstanceId, NodeId
|
||||
from exo.shared.types.events import (
|
||||
IndexedEvent,
|
||||
InstanceCreated,
|
||||
InstanceDeleted,
|
||||
InstanceRetrying,
|
||||
MetaInstanceCreated,
|
||||
MetaInstanceDeleted,
|
||||
)
|
||||
from exo.shared.types.memory import Memory
|
||||
from exo.shared.types.meta_instance import MetaInstance
|
||||
from exo.shared.types.multiaddr import Multiaddr
|
||||
from exo.shared.types.state import State
|
||||
from exo.shared.types.topology import Connection, SocketConnection
|
||||
from exo.shared.types.worker.instances import (
|
||||
InstanceId,
|
||||
MlxRingInstance,
|
||||
)
|
||||
from exo.shared.types.worker.runners import (
|
||||
RunnerFailed,
|
||||
RunnerId,
|
||||
RunnerLoading,
|
||||
RunnerReady,
|
||||
RunnerShutdown,
|
||||
ShardAssignments,
|
||||
)
|
||||
from exo.shared.types.worker.shards import PipelineShardMetadata
|
||||
|
||||
|
||||
def _model_card(model_id: str = "test-org/test-model") -> ModelCard:
|
||||
return ModelCard(
|
||||
model_id=ModelId(model_id),
|
||||
storage_size=Memory.from_kb(1000),
|
||||
n_layers=10,
|
||||
hidden_size=30,
|
||||
supports_tensor=True,
|
||||
tasks=[ModelTask.TextGeneration],
|
||||
)
|
||||
|
||||
|
||||
def _topology(*node_ids: str, connect: bool = True) -> Topology:
|
||||
"""Build a topology with nodes connected in a bidirectional ring with unique IPs.
|
||||
|
||||
Node at index ``i`` gets IP ``10.0.0.{i+1}``. Edges go in both directions
|
||||
between consecutive nodes (including wrap-around).
|
||||
"""
|
||||
t = Topology()
|
||||
nodes = [NodeId(n) for n in node_ids]
|
||||
for n in nodes:
|
||||
t.add_node(n)
|
||||
if connect and len(nodes) > 1:
|
||||
for i in range(len(nodes)):
|
||||
j = (i + 1) % len(nodes)
|
||||
t.add_connection(
|
||||
Connection(
|
||||
source=nodes[i],
|
||||
sink=nodes[j],
|
||||
edge=SocketConnection(
|
||||
sink_multiaddr=Multiaddr(
|
||||
address=f"/ip4/10.0.0.{j + 1}/tcp/50000"
|
||||
)
|
||||
),
|
||||
)
|
||||
)
|
||||
t.add_connection(
|
||||
Connection(
|
||||
source=nodes[j],
|
||||
sink=nodes[i],
|
||||
edge=SocketConnection(
|
||||
sink_multiaddr=Multiaddr(
|
||||
address=f"/ip4/10.0.0.{i + 1}/tcp/50000"
|
||||
)
|
||||
),
|
||||
)
|
||||
)
|
||||
return t
|
||||
|
||||
|
||||
def _meta_instance(
|
||||
model_id: str = "test-org/test-model",
|
||||
*,
|
||||
min_nodes: int = 1,
|
||||
node_ids: list[NodeId] | None = None,
|
||||
meta_instance_id: MetaInstanceId | None = None,
|
||||
) -> MetaInstance:
|
||||
return MetaInstance(
|
||||
meta_instance_id=meta_instance_id or MetaInstanceId(),
|
||||
model_id=ModelId(model_id),
|
||||
min_nodes=min_nodes,
|
||||
node_ids=node_ids,
|
||||
)
|
||||
|
||||
|
||||
def _instance(
|
||||
model_id: str = "test-org/test-model",
|
||||
node_ids: list[str] | None = None,
|
||||
instance_id: InstanceId | None = None,
|
||||
meta_instance_id: MetaInstanceId | None = None,
|
||||
) -> tuple[InstanceId, MlxRingInstance]:
|
||||
"""Create a test instance with hosts_by_node matching ``_topology()`` IPs."""
|
||||
iid = instance_id or InstanceId()
|
||||
nodes = node_ids or ["node-a"]
|
||||
n = len(nodes)
|
||||
mc = _model_card(model_id)
|
||||
ephemeral_port = 50000
|
||||
node_to_runner = {NodeId(nd): RunnerId() for nd in nodes}
|
||||
runner_to_shard = {
|
||||
runner_id: PipelineShardMetadata(
|
||||
model_card=mc,
|
||||
device_rank=i,
|
||||
world_size=n,
|
||||
start_layer=0,
|
||||
end_layer=mc.n_layers,
|
||||
n_layers=mc.n_layers,
|
||||
)
|
||||
for i, runner_id in enumerate(node_to_runner.values())
|
||||
}
|
||||
# Build hosts_by_node with IPs matching _topology() convention:
|
||||
# node at index idx has IP 10.0.0.{idx+1}
|
||||
hosts_by_node: dict[NodeId, list[Host]] = {}
|
||||
for r, node_str in enumerate(nodes):
|
||||
hosts: list[Host] = []
|
||||
for idx in range(n):
|
||||
if idx == r:
|
||||
hosts.append(Host(ip="0.0.0.0", port=ephemeral_port))
|
||||
elif n > 1 and idx in ((r - 1) % n, (r + 1) % n):
|
||||
hosts.append(Host(ip=f"10.0.0.{idx + 1}", port=ephemeral_port))
|
||||
else:
|
||||
hosts.append(Host(ip="198.51.100.1", port=0))
|
||||
hosts_by_node[NodeId(node_str)] = hosts
|
||||
return iid, MlxRingInstance(
|
||||
instance_id=iid,
|
||||
shard_assignments=ShardAssignments(
|
||||
model_id=ModelId(model_id),
|
||||
runner_to_shard=runner_to_shard,
|
||||
node_to_runner=node_to_runner,
|
||||
),
|
||||
hosts_by_node=hosts_by_node,
|
||||
ephemeral_port=ephemeral_port,
|
||||
meta_instance_id=meta_instance_id,
|
||||
)
|
||||
|
||||
|
||||
# --- instance_satisfies_meta_instance (pure constraint matching) ---
|
||||
|
||||
|
||||
def test_satisfies_matching_model():
|
||||
meta = _meta_instance()
|
||||
_, inst = _instance(node_ids=["node-a"])
|
||||
assert instance_satisfies_meta_instance(meta, inst) is True
|
||||
|
||||
|
||||
def test_not_satisfies_wrong_model():
|
||||
meta = _meta_instance("test-org/model-a")
|
||||
_, inst = _instance("test-org/model-b")
|
||||
assert instance_satisfies_meta_instance(meta, inst) is False
|
||||
|
||||
|
||||
def test_not_satisfies_missing_required_node():
|
||||
meta = _meta_instance(node_ids=[NodeId("node-c")])
|
||||
_, inst = _instance(node_ids=["node-a", "node-b"])
|
||||
assert instance_satisfies_meta_instance(meta, inst) is False
|
||||
|
||||
|
||||
def test_not_satisfies_fewer_than_min_nodes():
|
||||
meta = _meta_instance(min_nodes=3)
|
||||
_, inst = _instance(node_ids=["node-a", "node-b"])
|
||||
assert instance_satisfies_meta_instance(meta, inst) is False
|
||||
|
||||
|
||||
def test_satisfies_with_node_ids_specified():
|
||||
meta = _meta_instance(
|
||||
node_ids=[NodeId("node-a"), NodeId("node-b")], min_nodes=2
|
||||
)
|
||||
_, inst = _instance(node_ids=["node-a", "node-b", "node-c"])
|
||||
assert instance_satisfies_meta_instance(meta, inst) is True
|
||||
|
||||
|
||||
# --- instance_connections_healthy ---
|
||||
|
||||
|
||||
def test_healthy_single_node_present():
|
||||
_, inst = _instance(node_ids=["node-a"])
|
||||
topology = _topology("node-a")
|
||||
assert instance_connections_healthy(inst, topology) is True
|
||||
|
||||
|
||||
def test_unhealthy_single_node_missing():
|
||||
_, inst = _instance(node_ids=["node-a"])
|
||||
topology = Topology() # empty
|
||||
assert instance_connections_healthy(inst, topology) is False
|
||||
|
||||
|
||||
def test_healthy_two_node_ring():
|
||||
_, inst = _instance(node_ids=["node-a", "node-b"])
|
||||
topology = _topology("node-a", "node-b")
|
||||
assert instance_connections_healthy(inst, topology) is True
|
||||
|
||||
|
||||
def test_unhealthy_two_node_edge_removed():
|
||||
"""Nodes present but edge removed — ring broken."""
|
||||
_, inst = _instance(node_ids=["node-a", "node-b"])
|
||||
topology = _topology("node-a", "node-b", connect=False)
|
||||
assert instance_connections_healthy(inst, topology) is False
|
||||
|
||||
|
||||
def test_unhealthy_two_node_ip_changed():
|
||||
"""Edge exists but with a different IP than instance was configured with."""
|
||||
_, inst = _instance(node_ids=["node-a", "node-b"])
|
||||
# Build topology with different IPs than _instance() expects
|
||||
topology = Topology()
|
||||
topology.add_node(NodeId("node-a"))
|
||||
topology.add_node(NodeId("node-b"))
|
||||
topology.add_connection(
|
||||
Connection(
|
||||
source=NodeId("node-a"),
|
||||
sink=NodeId("node-b"),
|
||||
edge=SocketConnection(
|
||||
sink_multiaddr=Multiaddr(address="/ip4/192.168.99.99/tcp/50000")
|
||||
),
|
||||
)
|
||||
)
|
||||
topology.add_connection(
|
||||
Connection(
|
||||
source=NodeId("node-b"),
|
||||
sink=NodeId("node-a"),
|
||||
edge=SocketConnection(
|
||||
sink_multiaddr=Multiaddr(address="/ip4/192.168.99.98/tcp/50000")
|
||||
),
|
||||
)
|
||||
)
|
||||
assert instance_connections_healthy(inst, topology) is False
|
||||
|
||||
|
||||
def test_healthy_three_node_ring():
|
||||
_, inst = _instance(node_ids=["node-a", "node-b", "node-c"])
|
||||
topology = _topology("node-a", "node-b", "node-c")
|
||||
assert instance_connections_healthy(inst, topology) is True
|
||||
|
||||
|
||||
def test_unhealthy_three_node_one_edge_removed():
|
||||
"""Remove one edge from a three-node ring — instance unhealthy."""
|
||||
_, inst = _instance(node_ids=["node-a", "node-b", "node-c"])
|
||||
# Build topology with one direction of one edge missing
|
||||
topology = Topology()
|
||||
nodes = [NodeId("node-a"), NodeId("node-b"), NodeId("node-c")]
|
||||
for n in nodes:
|
||||
topology.add_node(n)
|
||||
# Add all edges except node-a → node-b
|
||||
topology.add_connection(
|
||||
Connection(
|
||||
source=nodes[1],
|
||||
sink=nodes[0],
|
||||
edge=SocketConnection(
|
||||
sink_multiaddr=Multiaddr(address="/ip4/10.0.0.1/tcp/50000")
|
||||
),
|
||||
)
|
||||
)
|
||||
topology.add_connection(
|
||||
Connection(
|
||||
source=nodes[1],
|
||||
sink=nodes[2],
|
||||
edge=SocketConnection(
|
||||
sink_multiaddr=Multiaddr(address="/ip4/10.0.0.3/tcp/50000")
|
||||
),
|
||||
)
|
||||
)
|
||||
topology.add_connection(
|
||||
Connection(
|
||||
source=nodes[2],
|
||||
sink=nodes[1],
|
||||
edge=SocketConnection(
|
||||
sink_multiaddr=Multiaddr(address="/ip4/10.0.0.2/tcp/50000")
|
||||
),
|
||||
)
|
||||
)
|
||||
topology.add_connection(
|
||||
Connection(
|
||||
source=nodes[2],
|
||||
sink=nodes[0],
|
||||
edge=SocketConnection(
|
||||
sink_multiaddr=Multiaddr(address="/ip4/10.0.0.1/tcp/50000")
|
||||
),
|
||||
)
|
||||
)
|
||||
topology.add_connection(
|
||||
Connection(
|
||||
source=nodes[0],
|
||||
sink=nodes[2],
|
||||
edge=SocketConnection(
|
||||
sink_multiaddr=Multiaddr(address="/ip4/10.0.0.3/tcp/50000")
|
||||
),
|
||||
)
|
||||
)
|
||||
# Missing: node-a → node-b (ip 10.0.0.2)
|
||||
assert instance_connections_healthy(inst, topology) is False
|
||||
|
||||
|
||||
def test_unhealthy_node_missing_from_topology():
|
||||
"""Instance has a node that's not in the topology at all."""
|
||||
_, inst = _instance(node_ids=["node-a", "node-b"])
|
||||
topology = _topology("node-a") # node-b not present
|
||||
assert instance_connections_healthy(inst, topology) is False
|
||||
|
||||
|
||||
def test_healthy_extra_nodes_in_topology():
|
||||
"""Extra nodes in topology don't affect instance health."""
|
||||
_, inst = _instance(node_ids=["node-a", "node-b"])
|
||||
topology = _topology("node-a", "node-b", "node-c")
|
||||
assert instance_connections_healthy(inst, topology) is True
|
||||
|
||||
|
||||
# --- find_unsatisfied_meta_instances ---
|
||||
|
||||
|
||||
def test_unsatisfied_no_meta_instances():
|
||||
result = find_unsatisfied_meta_instances({}, {}, Topology())
|
||||
assert list(result) == []
|
||||
|
||||
|
||||
def test_unsatisfied_one_satisfied():
|
||||
meta = _meta_instance()
|
||||
id_a, inst_a = _instance(meta_instance_id=meta.meta_instance_id)
|
||||
topology = _topology("node-a")
|
||||
result = find_unsatisfied_meta_instances(
|
||||
{meta.meta_instance_id: meta},
|
||||
{id_a: inst_a},
|
||||
topology,
|
||||
)
|
||||
assert list(result) == []
|
||||
|
||||
|
||||
def test_unsatisfied_one_not_satisfied():
|
||||
meta = _meta_instance("test-org/model-x")
|
||||
id_a, inst_a = _instance("test-org/model-y")
|
||||
topology = _topology("node-a")
|
||||
result = find_unsatisfied_meta_instances(
|
||||
{meta.meta_instance_id: meta}, {id_a: inst_a}, topology
|
||||
)
|
||||
assert list(result) == [meta]
|
||||
|
||||
|
||||
def test_unsatisfied_mix():
|
||||
meta_satisfied = _meta_instance("test-org/model-a")
|
||||
meta_unsatisfied = _meta_instance("test-org/model-b")
|
||||
id_a, inst_a = _instance(
|
||||
"test-org/model-a", meta_instance_id=meta_satisfied.meta_instance_id
|
||||
)
|
||||
topology = _topology("node-a")
|
||||
result = find_unsatisfied_meta_instances(
|
||||
{
|
||||
meta_satisfied.meta_instance_id: meta_satisfied,
|
||||
meta_unsatisfied.meta_instance_id: meta_unsatisfied,
|
||||
},
|
||||
{id_a: inst_a},
|
||||
topology,
|
||||
)
|
||||
assert list(result) == [meta_unsatisfied]
|
||||
|
||||
|
||||
def test_unsatisfied_node_disconnect():
|
||||
meta = _meta_instance()
|
||||
id_a, inst_a = _instance(
|
||||
node_ids=["node-a", "node-b"], meta_instance_id=meta.meta_instance_id
|
||||
)
|
||||
topology = _topology("node-a") # node-b disconnected
|
||||
result = find_unsatisfied_meta_instances(
|
||||
{meta.meta_instance_id: meta},
|
||||
{id_a: inst_a},
|
||||
topology,
|
||||
)
|
||||
assert list(result) == [meta]
|
||||
|
||||
|
||||
def test_unsatisfied_edge_break():
|
||||
"""Instance exists but its connections broke — meta-instance becomes unsatisfied."""
|
||||
meta = _meta_instance()
|
||||
id_a, inst_a = _instance(
|
||||
node_ids=["node-a", "node-b"], meta_instance_id=meta.meta_instance_id
|
||||
)
|
||||
topology = _topology("node-a", "node-b", connect=False) # nodes present, no edges
|
||||
result = find_unsatisfied_meta_instances(
|
||||
{meta.meta_instance_id: meta},
|
||||
{id_a: inst_a},
|
||||
topology,
|
||||
)
|
||||
assert list(result) == [meta]
|
||||
|
||||
|
||||
def test_unsatisfied_idempotent():
|
||||
meta = _meta_instance("test-org/model-x")
|
||||
topology = _topology("node-a")
|
||||
meta_instances = {meta.meta_instance_id: meta}
|
||||
instances: dict[InstanceId, MlxRingInstance] = {}
|
||||
result_1 = list(
|
||||
find_unsatisfied_meta_instances(meta_instances, instances, topology)
|
||||
)
|
||||
result_2 = list(
|
||||
find_unsatisfied_meta_instances(meta_instances, instances, topology)
|
||||
)
|
||||
assert result_1 == result_2
|
||||
|
||||
|
||||
def test_unsatisfied_exclusive_binding():
|
||||
"""Two MetaInstances for the same model: one is bound via meta_instance_id, the other is unsatisfied."""
|
||||
meta_a = _meta_instance("test-org/model-x")
|
||||
meta_b = _meta_instance("test-org/model-x")
|
||||
id_inst, inst = _instance(
|
||||
"test-org/model-x", meta_instance_id=meta_a.meta_instance_id
|
||||
)
|
||||
topology = _topology("node-a")
|
||||
result = find_unsatisfied_meta_instances(
|
||||
{
|
||||
meta_a.meta_instance_id: meta_a,
|
||||
meta_b.meta_instance_id: meta_b,
|
||||
},
|
||||
{id_inst: inst},
|
||||
topology,
|
||||
)
|
||||
assert list(result) == [meta_b]
|
||||
|
||||
|
||||
# --- apply handlers ---
|
||||
|
||||
|
||||
def test_apply_meta_instance_created():
|
||||
state = State()
|
||||
meta = _meta_instance()
|
||||
event = MetaInstanceCreated(meta_instance=meta)
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
assert meta.meta_instance_id in new_state.meta_instances
|
||||
assert new_state.meta_instances[meta.meta_instance_id] == meta
|
||||
|
||||
|
||||
def test_apply_meta_instance_deleted():
|
||||
meta = _meta_instance()
|
||||
state = State(meta_instances={meta.meta_instance_id: meta})
|
||||
event = MetaInstanceDeleted(meta_instance_id=meta.meta_instance_id)
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
assert meta.meta_instance_id not in new_state.meta_instances
|
||||
|
||||
|
||||
def test_apply_meta_instance_deleted_clears_failure_info():
|
||||
meta = _meta_instance().model_copy(
|
||||
update={"consecutive_failures": 2, "last_failure_error": "OOM"}
|
||||
)
|
||||
state = State(meta_instances={meta.meta_instance_id: meta})
|
||||
event = MetaInstanceDeleted(meta_instance_id=meta.meta_instance_id)
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
assert meta.meta_instance_id not in new_state.meta_instances
|
||||
|
||||
|
||||
# --- instance_runners_failed ---
|
||||
|
||||
|
||||
def test_runners_failed_all_failed():
|
||||
"""All runners in RunnerFailed -> instance is failed."""
|
||||
_, inst = _instance(node_ids=["node-a", "node-b"])
|
||||
runners = {
|
||||
rid: RunnerFailed(error_message="OOM")
|
||||
for rid in inst.shard_assignments.node_to_runner.values()
|
||||
}
|
||||
is_failed, error = instance_runners_failed(inst, runners, {})
|
||||
assert is_failed is True
|
||||
assert error is not None
|
||||
assert "OOM" in error
|
||||
|
||||
|
||||
def test_runners_failed_mixed_failed_shutdown():
|
||||
"""One Failed + one Shutdown = failed."""
|
||||
_, inst = _instance(node_ids=["node-a", "node-b"])
|
||||
runner_ids = list(inst.shard_assignments.node_to_runner.values())
|
||||
runners = {
|
||||
runner_ids[0]: RunnerFailed(error_message="crash"),
|
||||
runner_ids[1]: RunnerShutdown(),
|
||||
}
|
||||
is_failed, error = instance_runners_failed(inst, runners, {})
|
||||
assert is_failed is True
|
||||
assert error is not None
|
||||
assert "crash" in error
|
||||
|
||||
|
||||
def test_runners_not_failed_all_shutdown():
|
||||
"""All Shutdown (graceful) = not a failure."""
|
||||
_, inst = _instance(node_ids=["node-a"])
|
||||
runners = {
|
||||
rid: RunnerShutdown()
|
||||
for rid in inst.shard_assignments.node_to_runner.values()
|
||||
}
|
||||
is_failed, _ = instance_runners_failed(inst, runners, {})
|
||||
assert is_failed is False
|
||||
|
||||
|
||||
def test_runners_not_failed_still_active():
|
||||
"""Some runners still active = not failed yet."""
|
||||
_, inst = _instance(node_ids=["node-a", "node-b"])
|
||||
runner_ids = list(inst.shard_assignments.node_to_runner.values())
|
||||
runners = {
|
||||
runner_ids[0]: RunnerFailed(error_message="OOM"),
|
||||
runner_ids[1]: RunnerLoading(),
|
||||
}
|
||||
is_failed, _ = instance_runners_failed(inst, runners, {})
|
||||
assert is_failed is False
|
||||
|
||||
|
||||
def test_runners_not_failed_no_status():
|
||||
"""Runner not yet reported = not failed."""
|
||||
_, inst = _instance(node_ids=["node-a"])
|
||||
is_failed, _ = instance_runners_failed(inst, {}, {})
|
||||
assert is_failed is False
|
||||
|
||||
|
||||
def test_runners_not_failed_healthy():
|
||||
"""Runners in Ready state = not failed."""
|
||||
_, inst = _instance(node_ids=["node-a"])
|
||||
runners = {
|
||||
rid: RunnerReady()
|
||||
for rid in inst.shard_assignments.node_to_runner.values()
|
||||
}
|
||||
is_failed, _ = instance_runners_failed(inst, runners, {})
|
||||
assert is_failed is False
|
||||
|
||||
|
||||
# --- failure tracking in apply_instance_deleted ---
|
||||
|
||||
|
||||
def test_apply_instance_deleted_tracks_failure():
|
||||
"""InstanceDeleted with failure_error increments meta instance failure count."""
|
||||
meta = _meta_instance()
|
||||
iid, inst = _instance(
|
||||
node_ids=["node-a"], meta_instance_id=meta.meta_instance_id
|
||||
)
|
||||
state = State(
|
||||
meta_instances={meta.meta_instance_id: meta},
|
||||
instances={iid: inst},
|
||||
)
|
||||
event = InstanceDeleted(instance_id=iid, failure_error="Runner OOM")
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
mi = new_state.meta_instances[meta.meta_instance_id]
|
||||
assert mi.consecutive_failures == 1
|
||||
assert mi.last_failure_error == "Runner OOM"
|
||||
|
||||
|
||||
def test_apply_instance_deleted_increments_failure():
|
||||
"""Subsequent failures increment the counter."""
|
||||
meta = _meta_instance().model_copy(
|
||||
update={"consecutive_failures": 2, "last_failure_error": "previous error"}
|
||||
)
|
||||
iid, inst = _instance(
|
||||
node_ids=["node-a"], meta_instance_id=meta.meta_instance_id
|
||||
)
|
||||
state = State(
|
||||
meta_instances={meta.meta_instance_id: meta},
|
||||
instances={iid: inst},
|
||||
)
|
||||
event = InstanceDeleted(instance_id=iid, failure_error="new error")
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
mi = new_state.meta_instances[meta.meta_instance_id]
|
||||
assert mi.consecutive_failures == 3
|
||||
assert mi.last_failure_error == "new error"
|
||||
|
||||
|
||||
def test_apply_instance_deleted_no_failure_no_tracking():
|
||||
"""InstanceDeleted without failure_error does not track."""
|
||||
meta = _meta_instance()
|
||||
iid, inst = _instance(
|
||||
node_ids=["node-a"], meta_instance_id=meta.meta_instance_id
|
||||
)
|
||||
state = State(
|
||||
meta_instances={meta.meta_instance_id: meta},
|
||||
instances={iid: inst},
|
||||
)
|
||||
event = InstanceDeleted(instance_id=iid)
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
mi = new_state.meta_instances[meta.meta_instance_id]
|
||||
assert mi.consecutive_failures == 0
|
||||
|
||||
|
||||
def test_apply_instance_deleted_orphan_no_tracking():
|
||||
"""InstanceDeleted for orphan instance (no meta_instance_id) does not track."""
|
||||
iid, inst = _instance(node_ids=["node-a"])
|
||||
state = State(instances={iid: inst})
|
||||
event = InstanceDeleted(instance_id=iid, failure_error="crash")
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
assert len(new_state.meta_instances) == 0
|
||||
|
||||
|
||||
# --- InstanceRetrying ---
|
||||
|
||||
|
||||
def test_apply_instance_retrying_removes_runners():
|
||||
"""InstanceRetrying removes the instance's runners from state but keeps the instance."""
|
||||
meta = _meta_instance()
|
||||
iid, inst = _instance(node_ids=["node-a", "node-b"], meta_instance_id=meta.meta_instance_id)
|
||||
runner_ids = list(inst.shard_assignments.node_to_runner.values())
|
||||
runners = {
|
||||
runner_ids[0]: RunnerFailed(error_message="OOM"),
|
||||
runner_ids[1]: RunnerShutdown(),
|
||||
}
|
||||
state = State(
|
||||
meta_instances={meta.meta_instance_id: meta},
|
||||
instances={iid: inst},
|
||||
runners=runners,
|
||||
)
|
||||
event = InstanceRetrying(
|
||||
instance_id=iid,
|
||||
meta_instance_id=meta.meta_instance_id,
|
||||
failure_error="OOM",
|
||||
)
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
# Instance still exists
|
||||
assert iid in new_state.instances
|
||||
# Runners removed
|
||||
assert runner_ids[0] not in new_state.runners
|
||||
assert runner_ids[1] not in new_state.runners
|
||||
|
||||
|
||||
def test_apply_instance_retrying_increments_failure():
|
||||
"""InstanceRetrying increments consecutive_failures on the MetaInstance."""
|
||||
meta = _meta_instance()
|
||||
iid, inst = _instance(node_ids=["node-a"], meta_instance_id=meta.meta_instance_id)
|
||||
state = State(
|
||||
meta_instances={meta.meta_instance_id: meta},
|
||||
instances={iid: inst},
|
||||
)
|
||||
event = InstanceRetrying(
|
||||
instance_id=iid,
|
||||
meta_instance_id=meta.meta_instance_id,
|
||||
failure_error="crash",
|
||||
)
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
mi = new_state.meta_instances[meta.meta_instance_id]
|
||||
assert mi.consecutive_failures == 1
|
||||
assert mi.last_failure_error == "crash"
|
||||
|
||||
|
||||
def test_apply_instance_retrying_skips_missing_runners():
|
||||
"""InstanceRetrying doesn't assert if runners haven't reported yet."""
|
||||
meta = _meta_instance()
|
||||
iid, inst = _instance(node_ids=["node-a"], meta_instance_id=meta.meta_instance_id)
|
||||
# No runners in state at all
|
||||
state = State(
|
||||
meta_instances={meta.meta_instance_id: meta},
|
||||
instances={iid: inst},
|
||||
)
|
||||
event = InstanceRetrying(
|
||||
instance_id=iid,
|
||||
meta_instance_id=meta.meta_instance_id,
|
||||
failure_error="crash",
|
||||
)
|
||||
# Should not raise
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
assert iid in new_state.instances
|
||||
|
||||
|
||||
def test_apply_instance_created_resets_failure_counter():
|
||||
"""InstanceCreated resets consecutive_failures but preserves last_failure_error."""
|
||||
meta = _meta_instance().model_copy(
|
||||
update={"consecutive_failures": 3, "last_failure_error": "old error"}
|
||||
)
|
||||
_, inst = _instance(node_ids=["node-a"], meta_instance_id=meta.meta_instance_id)
|
||||
state = State(meta_instances={meta.meta_instance_id: meta})
|
||||
event = InstanceCreated(instance=inst)
|
||||
new_state = apply(state, IndexedEvent(idx=0, event=event))
|
||||
mi = new_state.meta_instances[meta.meta_instance_id]
|
||||
assert mi.consecutive_failures == 0
|
||||
assert mi.last_failure_error == "old error"
|
||||
assert mi.placement_error is None
|
||||
|
||||
|
||||
# --- InstanceHealthReconciler retry-vs-delete ---
|
||||
|
||||
|
||||
async def test_health_reconciler_retries_when_under_limit():
|
||||
"""InstanceHealthReconciler emits InstanceRetrying when consecutive_failures < 3."""
|
||||
meta = _meta_instance()
|
||||
iid, inst = _instance(node_ids=["node-a"], meta_instance_id=meta.meta_instance_id)
|
||||
runner_ids = list(inst.shard_assignments.node_to_runner.values())
|
||||
state = State(
|
||||
meta_instances={meta.meta_instance_id: meta},
|
||||
instances={iid: inst},
|
||||
runners={runner_ids[0]: RunnerFailed(error_message="OOM")},
|
||||
topology=_topology("node-a"),
|
||||
)
|
||||
reconciler = InstanceHealthReconciler()
|
||||
events = await reconciler.reconcile(state)
|
||||
assert len(events) == 1
|
||||
assert isinstance(events[0], InstanceRetrying)
|
||||
assert events[0].instance_id == iid
|
||||
assert events[0].meta_instance_id == meta.meta_instance_id
|
||||
|
||||
|
||||
async def test_health_reconciler_deletes_when_limit_reached():
|
||||
"""InstanceHealthReconciler emits InstanceDeleted when consecutive_failures >= 3."""
|
||||
meta = _meta_instance().model_copy(update={"consecutive_failures": 3})
|
||||
iid, inst = _instance(node_ids=["node-a"], meta_instance_id=meta.meta_instance_id)
|
||||
runner_ids = list(inst.shard_assignments.node_to_runner.values())
|
||||
state = State(
|
||||
meta_instances={meta.meta_instance_id: meta},
|
||||
instances={iid: inst},
|
||||
runners={runner_ids[0]: RunnerFailed(error_message="OOM")},
|
||||
topology=_topology("node-a"),
|
||||
)
|
||||
reconciler = InstanceHealthReconciler()
|
||||
events = await reconciler.reconcile(state)
|
||||
assert len(events) == 1
|
||||
assert isinstance(events[0], InstanceDeleted)
|
||||
|
||||
|
||||
async def test_health_reconciler_deletes_without_meta_instance():
|
||||
"""Instances without a MetaInstance are deleted immediately on runner failure."""
|
||||
iid, inst = _instance(node_ids=["node-a"])
|
||||
runner_ids = list(inst.shard_assignments.node_to_runner.values())
|
||||
state = State(
|
||||
instances={iid: inst},
|
||||
runners={runner_ids[0]: RunnerFailed(error_message="crash")},
|
||||
topology=_topology("node-a"),
|
||||
)
|
||||
reconciler = InstanceHealthReconciler()
|
||||
events = await reconciler.reconcile(state)
|
||||
assert len(events) == 1
|
||||
assert isinstance(events[0], InstanceDeleted)
|
||||
|
||||
|
||||
async def test_health_reconciler_network_failure_always_deletes():
|
||||
"""Network failure always triggers InstanceDeleted regardless of retry count."""
|
||||
meta = _meta_instance()
|
||||
iid, inst = _instance(
|
||||
node_ids=["node-a", "node-b"], meta_instance_id=meta.meta_instance_id
|
||||
)
|
||||
state = State(
|
||||
meta_instances={meta.meta_instance_id: meta},
|
||||
instances={iid: inst},
|
||||
topology=_topology("node-a"), # node-b missing
|
||||
)
|
||||
reconciler = InstanceHealthReconciler()
|
||||
events = await reconciler.reconcile(state)
|
||||
assert len(events) == 1
|
||||
assert isinstance(events[0], InstanceDeleted)
|
||||
assert events[0].failure_error == "Network connection lost"
|
||||
@@ -1,37 +1,9 @@
|
||||
from enum import Enum
|
||||
|
||||
from exo_pyo3_bindings import ConnectionUpdate, ConnectionUpdateType
|
||||
|
||||
from exo.shared.types.common import NodeId
|
||||
from exo.utils.pydantic_ext import CamelCaseModel
|
||||
|
||||
"""Serialisable types for Connection Updates/Messages"""
|
||||
|
||||
|
||||
class ConnectionMessageType(Enum):
|
||||
Connected = 0
|
||||
Disconnected = 1
|
||||
|
||||
@staticmethod
|
||||
def from_update_type(update_type: ConnectionUpdateType):
|
||||
match update_type:
|
||||
case ConnectionUpdateType.Connected:
|
||||
return ConnectionMessageType.Connected
|
||||
case ConnectionUpdateType.Disconnected:
|
||||
return ConnectionMessageType.Disconnected
|
||||
|
||||
|
||||
class ConnectionMessage(CamelCaseModel):
|
||||
node_id: NodeId
|
||||
connection_type: ConnectionMessageType
|
||||
remote_ipv4: str
|
||||
remote_tcp_port: int
|
||||
|
||||
@classmethod
|
||||
def from_update(cls, update: ConnectionUpdate) -> "ConnectionMessage":
|
||||
return cls(
|
||||
node_id=NodeId(update.peer_id.to_base58()),
|
||||
connection_type=ConnectionMessageType.from_update_type(update.update_type),
|
||||
remote_ipv4=update.remote_ipv4,
|
||||
remote_tcp_port=update.remote_tcp_port,
|
||||
)
|
||||
expired: bool
|
||||
|
||||
@@ -1,5 +1,5 @@
|
||||
from copy import copy
|
||||
from itertools import count
|
||||
from dataclasses import dataclass, field
|
||||
from math import inf
|
||||
from os import PathLike
|
||||
from pathlib import Path
|
||||
@@ -14,15 +14,14 @@ from anyio import (
|
||||
)
|
||||
from anyio.abc import TaskGroup
|
||||
from exo_pyo3_bindings import (
|
||||
AllQueuesFullError,
|
||||
Keypair,
|
||||
NetworkingHandle,
|
||||
NoPeersSubscribedToTopicError,
|
||||
PyPeer,
|
||||
)
|
||||
from filelock import FileLock
|
||||
from loguru import logger
|
||||
|
||||
from exo.shared.constants import EXO_NODE_ID_KEYPAIR
|
||||
from exo.shared.types.common import NodeId
|
||||
from exo.utils.channels import Receiver, Sender, channel
|
||||
from exo.utils.pydantic_ext import CamelCaseModel
|
||||
|
||||
@@ -99,28 +98,32 @@ class TopicRouter[T: CamelCaseModel]:
|
||||
)
|
||||
|
||||
|
||||
@dataclass
|
||||
class Router:
|
||||
@classmethod
|
||||
def create(cls, identity: Keypair) -> "Router":
|
||||
return cls(handle=NetworkingHandle(identity))
|
||||
_peer: PyPeer
|
||||
topic_routers: dict[str, TopicRouter[CamelCaseModel]] = field(
|
||||
init=False, default_factory=dict
|
||||
)
|
||||
networking_receiver: Receiver[tuple[str, bytes]] = field(init=False)
|
||||
_tmp_networking_sender: Sender[tuple[str, bytes]] | None = field(init=False)
|
||||
_tg: TaskGroup | None = None
|
||||
|
||||
def __init__(self, handle: NetworkingHandle):
|
||||
self.topic_routers: dict[str, TopicRouter[CamelCaseModel]] = {}
|
||||
send, recv = channel[tuple[str, bytes]]()
|
||||
self.networking_receiver: Receiver[tuple[str, bytes]] = recv
|
||||
self._net: NetworkingHandle = handle
|
||||
self._tmp_networking_sender: Sender[tuple[str, bytes]] | None = send
|
||||
self._id_count = count()
|
||||
self._tg: TaskGroup | None = None
|
||||
def __post_init__(self):
|
||||
self._tmp_networking_sender, self.networking_receiver = channel()
|
||||
|
||||
@classmethod
|
||||
def create(cls, identity: Keypair, namespace: str) -> "Router":
|
||||
return cls(_peer=PyPeer.new(identity, namespace))
|
||||
|
||||
async def register_topic[T: CamelCaseModel](self, topic: TypedTopic[T]):
|
||||
assert self._tg is None, "Attempted to register topic after setup time"
|
||||
send = self._tmp_networking_sender
|
||||
if send:
|
||||
self._tmp_networking_sender = None
|
||||
else:
|
||||
send = self.networking_receiver.clone_sender()
|
||||
router = TopicRouter[T](topic, send)
|
||||
if self._tg is not None:
|
||||
self._tg.start_soon(router.run)
|
||||
self.topic_routers[topic.topic] = cast(TopicRouter[CamelCaseModel], router)
|
||||
await self._networking_subscribe(str(topic.topic))
|
||||
|
||||
@@ -148,14 +151,18 @@ class Router:
|
||||
async def run(self):
|
||||
logger.debug("Starting Router")
|
||||
try:
|
||||
|
||||
async def _peer_run():
|
||||
await self._peer.run()
|
||||
|
||||
async with create_task_group() as tg:
|
||||
self._tg = tg
|
||||
for topic in self.topic_routers:
|
||||
router = self.topic_routers[topic]
|
||||
tg.start_soon(router.run)
|
||||
tg.start_soon(self._networking_recv)
|
||||
tg.start_soon(self._networking_recv_connection_messages)
|
||||
tg.start_soon(self._networking_publish)
|
||||
tg.start_soon(_peer_run)
|
||||
# Router only shuts down if you cancel it.
|
||||
await sleep_forever()
|
||||
finally:
|
||||
@@ -170,47 +177,58 @@ class Router:
|
||||
self._tg.cancel_scope.cancel()
|
||||
|
||||
async def _networking_subscribe(self, topic: str):
|
||||
await self._net.gossipsub_subscribe(topic)
|
||||
await self._peer.subscribe(topic)
|
||||
logger.info(f"Subscribed to {topic}")
|
||||
|
||||
async def _networking_unsubscribe(self, topic: str):
|
||||
await self._net.gossipsub_unsubscribe(topic)
|
||||
await self._peer.unsubscribe(topic)
|
||||
logger.info(f"Unsubscribed from {topic}")
|
||||
|
||||
async def _networking_recv(self):
|
||||
while True:
|
||||
topic, data = await self._net.gossipsub_recv()
|
||||
logger.trace(f"Received message on {topic} with payload {data}")
|
||||
try:
|
||||
swarm_event = await self._peer.recv()
|
||||
except ValueError:
|
||||
logger.error("Message too large for gossipsub, dropped")
|
||||
continue
|
||||
except ConnectionError:
|
||||
logger.error("All peer queues full, network overloaded")
|
||||
continue
|
||||
except RuntimeError:
|
||||
break
|
||||
|
||||
cm = None
|
||||
if (peer_id := swarm_event.downcast_discovered()) is not None:
|
||||
cm = ConnectionMessage(node_id=NodeId(peer_id), expired=False)
|
||||
if (peer_id := swarm_event.downcast_expired()) is not None:
|
||||
cm = ConnectionMessage(node_id=NodeId(peer_id), expired=True)
|
||||
|
||||
if cm is not None:
|
||||
if CONNECTION_MESSAGES.topic in self.topic_routers:
|
||||
router = self.topic_routers[CONNECTION_MESSAGES.topic]
|
||||
assert router.topic.model_type == ConnectionMessage
|
||||
router = cast(TopicRouter[ConnectionMessage], router)
|
||||
await router.publish(cm)
|
||||
continue
|
||||
|
||||
assert (msg := swarm_event.downcast_message()) is not None
|
||||
_origin, topic, payload = msg
|
||||
logger.debug(f"Received message on {topic} with payload {payload}")
|
||||
if topic not in self.topic_routers:
|
||||
logger.warning(f"Received message on unknown or inactive topic {topic}")
|
||||
continue
|
||||
|
||||
router = self.topic_routers[topic]
|
||||
await router.publish_bytes(data)
|
||||
|
||||
async def _networking_recv_connection_messages(self):
|
||||
while True:
|
||||
update = await self._net.connection_update_recv()
|
||||
message = ConnectionMessage.from_update(update)
|
||||
logger.trace(
|
||||
f"Received message on connection_messages with payload {message}"
|
||||
)
|
||||
if CONNECTION_MESSAGES.topic in self.topic_routers:
|
||||
router = self.topic_routers[CONNECTION_MESSAGES.topic]
|
||||
assert router.topic.model_type == ConnectionMessage
|
||||
router = cast(TopicRouter[ConnectionMessage], router)
|
||||
await router.publish(message)
|
||||
await router.publish_bytes(payload)
|
||||
|
||||
async def _networking_publish(self):
|
||||
with self.networking_receiver as networked_items:
|
||||
async for topic, data in networked_items:
|
||||
try:
|
||||
logger.trace(f"Sending message on {topic} with payload {data}")
|
||||
await self._net.gossipsub_publish(topic, data)
|
||||
except NoPeersSubscribedToTopicError:
|
||||
pass
|
||||
except AllQueuesFullError:
|
||||
logger.warning(f"All peer queues full, dropping message on {topic}")
|
||||
await self._peer.send(topic, data)
|
||||
except RuntimeError:
|
||||
break
|
||||
|
||||
|
||||
def get_node_id_keypair(
|
||||
@@ -221,7 +239,7 @@ def get_node_id_keypair(
|
||||
Obtain the :class:`PeerId` by from it.
|
||||
"""
|
||||
# TODO(evan): bring back node id persistence once we figure out how to deal with duplicates
|
||||
return Keypair.generate_ed25519()
|
||||
return Keypair.generate()
|
||||
|
||||
def lock_path(path: str | bytes | PathLike[str] | PathLike[bytes]) -> Path:
|
||||
return Path(str(path) + ".lock")
|
||||
|
||||
@@ -4,7 +4,7 @@ from datetime import datetime
|
||||
|
||||
from loguru import logger
|
||||
|
||||
from exo.shared.types.common import MetaInstanceId, NodeId
|
||||
from exo.shared.types.common import NodeId
|
||||
from exo.shared.types.events import (
|
||||
ChunkGenerated,
|
||||
Event,
|
||||
@@ -12,10 +12,6 @@ from exo.shared.types.events import (
|
||||
InputChunkReceived,
|
||||
InstanceCreated,
|
||||
InstanceDeleted,
|
||||
InstanceRetrying,
|
||||
MetaInstanceCreated,
|
||||
MetaInstanceDeleted,
|
||||
MetaInstancePlacementFailed,
|
||||
NodeDownloadProgress,
|
||||
NodeGatheredInfo,
|
||||
NodeTimedOut,
|
||||
@@ -32,7 +28,6 @@ from exo.shared.types.events import (
|
||||
TracesCollected,
|
||||
TracesMerged,
|
||||
)
|
||||
from exo.shared.types.meta_instance import MetaInstance
|
||||
from exo.shared.types.profiling import (
|
||||
NodeIdentity,
|
||||
NodeNetworkInfo,
|
||||
@@ -77,14 +72,6 @@ def event_apply(event: Event, state: State) -> State:
|
||||
return apply_instance_created(event, state)
|
||||
case InstanceDeleted():
|
||||
return apply_instance_deleted(event, state)
|
||||
case InstanceRetrying():
|
||||
return apply_instance_retrying(event, state)
|
||||
case MetaInstanceCreated():
|
||||
return apply_meta_instance_created(event, state)
|
||||
case MetaInstanceDeleted():
|
||||
return apply_meta_instance_deleted(event, state)
|
||||
case MetaInstancePlacementFailed():
|
||||
return apply_meta_instance_placement_failed(event, state)
|
||||
case NodeTimedOut():
|
||||
return apply_node_timed_out(event, state)
|
||||
case NodeDownloadProgress():
|
||||
@@ -187,119 +174,20 @@ def apply_task_failed(event: TaskFailed, state: State) -> State:
|
||||
return state.model_copy(update={"tasks": new_tasks})
|
||||
|
||||
|
||||
def _update_meta_instance(
|
||||
state: State, mid: MetaInstanceId, **fields: object
|
||||
) -> Mapping[MetaInstanceId, MetaInstance]:
|
||||
mi = state.meta_instances[mid]
|
||||
return {**state.meta_instances, mid: mi.model_copy(update=fields)}
|
||||
|
||||
|
||||
def apply_instance_created(event: InstanceCreated, state: State) -> State:
|
||||
instance = event.instance
|
||||
new_instances: Mapping[InstanceId, Instance] = {
|
||||
**state.instances,
|
||||
instance.instance_id: instance,
|
||||
}
|
||||
update: dict[str, object] = {"instances": new_instances}
|
||||
# Reset failure tracking when a new instance is created for a meta-instance
|
||||
if instance.meta_instance_id and instance.meta_instance_id in state.meta_instances:
|
||||
mi = state.meta_instances[instance.meta_instance_id]
|
||||
if mi.placement_error is not None or mi.consecutive_failures > 0:
|
||||
update["meta_instances"] = _update_meta_instance(
|
||||
state,
|
||||
instance.meta_instance_id,
|
||||
placement_error=None,
|
||||
consecutive_failures=0,
|
||||
)
|
||||
return state.model_copy(update=update)
|
||||
return state.model_copy(update={"instances": new_instances})
|
||||
|
||||
|
||||
def apply_instance_deleted(event: InstanceDeleted, state: State) -> State:
|
||||
deleted_instance = state.instances.get(event.instance_id)
|
||||
new_instances: Mapping[InstanceId, Instance] = {
|
||||
iid: inst for iid, inst in state.instances.items() if iid != event.instance_id
|
||||
}
|
||||
update: dict[str, object] = {"instances": new_instances}
|
||||
|
||||
# Track failure on the MetaInstance itself
|
||||
if (
|
||||
event.failure_error
|
||||
and deleted_instance
|
||||
and deleted_instance.meta_instance_id
|
||||
and deleted_instance.meta_instance_id in state.meta_instances
|
||||
):
|
||||
mid = deleted_instance.meta_instance_id
|
||||
mi = state.meta_instances[mid]
|
||||
update["meta_instances"] = {
|
||||
**state.meta_instances,
|
||||
mid: mi.model_copy(
|
||||
update={
|
||||
"consecutive_failures": mi.consecutive_failures + 1,
|
||||
"last_failure_error": event.failure_error,
|
||||
}
|
||||
),
|
||||
}
|
||||
|
||||
return state.model_copy(update=update)
|
||||
|
||||
|
||||
def apply_instance_retrying(event: InstanceRetrying, state: State) -> State:
|
||||
"""Runners failed but retry limit not reached — remove runners, keep instance."""
|
||||
instance = state.instances.get(event.instance_id)
|
||||
if instance is None:
|
||||
return state
|
||||
|
||||
# Remove all runners belonging to this instance from state
|
||||
runner_ids_to_remove = set(instance.shard_assignments.node_to_runner.values())
|
||||
new_runners: Mapping[RunnerId, RunnerStatus] = {
|
||||
rid: rs
|
||||
for rid, rs in state.runners.items()
|
||||
if rid not in runner_ids_to_remove
|
||||
}
|
||||
|
||||
update: dict[str, object] = {"runners": new_runners}
|
||||
|
||||
# Increment failure count on the MetaInstance
|
||||
if event.meta_instance_id in state.meta_instances:
|
||||
update["meta_instances"] = _update_meta_instance(
|
||||
state,
|
||||
event.meta_instance_id,
|
||||
consecutive_failures=state.meta_instances[event.meta_instance_id].consecutive_failures + 1,
|
||||
last_failure_error=event.failure_error,
|
||||
)
|
||||
|
||||
return state.model_copy(update=update)
|
||||
|
||||
|
||||
def apply_meta_instance_created(event: MetaInstanceCreated, state: State) -> State:
|
||||
new_meta: Mapping[MetaInstanceId, MetaInstance] = {
|
||||
**state.meta_instances,
|
||||
event.meta_instance.meta_instance_id: event.meta_instance,
|
||||
}
|
||||
return state.model_copy(update={"meta_instances": new_meta})
|
||||
|
||||
|
||||
def apply_meta_instance_deleted(event: MetaInstanceDeleted, state: State) -> State:
|
||||
new_meta: Mapping[MetaInstanceId, MetaInstance] = {
|
||||
mid: mi
|
||||
for mid, mi in state.meta_instances.items()
|
||||
if mid != event.meta_instance_id
|
||||
}
|
||||
return state.model_copy(update={"meta_instances": new_meta})
|
||||
|
||||
|
||||
def apply_meta_instance_placement_failed(
|
||||
event: MetaInstancePlacementFailed, state: State
|
||||
) -> State:
|
||||
if event.meta_instance_id not in state.meta_instances:
|
||||
return state
|
||||
return state.model_copy(
|
||||
update={
|
||||
"meta_instances": _update_meta_instance(
|
||||
state, event.meta_instance_id, placement_error=event.reason
|
||||
)
|
||||
}
|
||||
)
|
||||
return state.model_copy(update={"instances": new_instances})
|
||||
|
||||
|
||||
def apply_runner_status_updated(event: RunnerStatusUpdated, state: State) -> State:
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
import pytest
|
||||
from anyio import create_task_group, fail_after, move_on_after
|
||||
|
||||
from exo.routing.connection_message import ConnectionMessage, ConnectionMessageType
|
||||
from exo.routing.connection_message import ConnectionMessage
|
||||
from exo.shared.election import Election, ElectionMessage, ElectionResult
|
||||
from exo.shared.types.commands import ForwarderCommand, TestCommand
|
||||
from exo.shared.types.common import NodeId, SessionId
|
||||
@@ -330,9 +330,7 @@ async def test_connection_message_triggers_new_round_broadcast() -> None:
|
||||
await cm_tx.send(
|
||||
ConnectionMessage(
|
||||
node_id=NodeId(),
|
||||
connection_type=ConnectionMessageType.Connected,
|
||||
remote_ipv4="",
|
||||
remote_tcp_port=0,
|
||||
expired=False,
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
@@ -3,10 +3,11 @@ from collections.abc import Generator
|
||||
from typing import Annotated, Any, Literal
|
||||
from uuid import uuid4
|
||||
|
||||
from pydantic import BaseModel, Field
|
||||
from pydantic import BaseModel, Field, field_validator
|
||||
from pydantic_core import PydanticUseDefault
|
||||
|
||||
from exo.shared.models.model_cards import ModelCard, ModelId
|
||||
from exo.shared.types.common import CommandId, MetaInstanceId, NodeId
|
||||
from exo.shared.types.common import CommandId, NodeId
|
||||
from exo.shared.types.memory import Memory
|
||||
from exo.shared.types.worker.instances import Instance, InstanceId, InstanceMeta
|
||||
from exo.shared.types.worker.shards import Sharding, ShardMetadata
|
||||
@@ -226,6 +227,13 @@ class PlaceInstanceParams(BaseModel):
|
||||
instance_meta: InstanceMeta = InstanceMeta.MlxRing
|
||||
min_nodes: int = 1
|
||||
|
||||
@field_validator("sharding", "instance_meta", mode="plain")
|
||||
@classmethod
|
||||
def use_default(cls, v: object):
|
||||
if not v or not isinstance(v, (Sharding, InstanceMeta)):
|
||||
raise PydanticUseDefault()
|
||||
return v
|
||||
|
||||
|
||||
class CreateInstanceParams(BaseModel):
|
||||
instance: Instance
|
||||
@@ -261,26 +269,6 @@ class DeleteInstanceResponse(BaseModel):
|
||||
instance_id: InstanceId
|
||||
|
||||
|
||||
class CreateMetaInstanceParams(BaseModel):
|
||||
model_id: ModelId
|
||||
sharding: Sharding = Sharding.Pipeline
|
||||
instance_meta: InstanceMeta = InstanceMeta.MlxRing
|
||||
min_nodes: int = 1
|
||||
node_ids: list[NodeId] | None = None
|
||||
|
||||
|
||||
class CreateMetaInstanceResponse(BaseModel):
|
||||
message: str
|
||||
command_id: CommandId
|
||||
meta_instance_id: MetaInstanceId
|
||||
|
||||
|
||||
class DeleteMetaInstanceResponse(BaseModel):
|
||||
message: str
|
||||
command_id: CommandId
|
||||
meta_instance_id: MetaInstanceId
|
||||
|
||||
|
||||
class AdvancedImageParams(BaseModel):
|
||||
seed: Annotated[int, Field(ge=0)] | None = None
|
||||
num_inference_steps: Annotated[int, Field(ge=1, le=100)] | None = None
|
||||
|
||||
@@ -6,8 +6,7 @@ from exo.shared.types.api import (
|
||||
ImageGenerationTaskParams,
|
||||
)
|
||||
from exo.shared.types.chunks import InputImageChunk
|
||||
from exo.shared.types.common import CommandId, MetaInstanceId, NodeId
|
||||
from exo.shared.types.meta_instance import MetaInstance
|
||||
from exo.shared.types.common import CommandId, NodeId
|
||||
from exo.shared.types.text_generation import TextGenerationTaskParams
|
||||
from exo.shared.types.worker.instances import Instance, InstanceId, InstanceMeta
|
||||
from exo.shared.types.worker.shards import Sharding, ShardMetadata
|
||||
@@ -49,14 +48,6 @@ class DeleteInstance(BaseCommand):
|
||||
instance_id: InstanceId
|
||||
|
||||
|
||||
class CreateMetaInstance(BaseCommand):
|
||||
meta_instance: MetaInstance
|
||||
|
||||
|
||||
class DeleteMetaInstance(BaseCommand):
|
||||
meta_instance_id: MetaInstanceId
|
||||
|
||||
|
||||
class TaskFinished(BaseCommand):
|
||||
finished_command_id: CommandId
|
||||
|
||||
@@ -98,8 +89,6 @@ Command = (
|
||||
| PlaceInstance
|
||||
| CreateInstance
|
||||
| DeleteInstance
|
||||
| CreateMetaInstance
|
||||
| DeleteMetaInstance
|
||||
| TaskFinished
|
||||
| SendInputChunk
|
||||
)
|
||||
|
||||
@@ -42,10 +42,6 @@ class CommandId(Id):
|
||||
pass
|
||||
|
||||
|
||||
class MetaInstanceId(Id):
|
||||
"""Identifier for a MetaInstance."""
|
||||
|
||||
|
||||
class Host(CamelCaseModel):
|
||||
ip: str
|
||||
port: int
|
||||
|
||||
@@ -5,8 +5,7 @@ from pydantic import Field
|
||||
|
||||
from exo.shared.topology import Connection
|
||||
from exo.shared.types.chunks import GenerationChunk, InputImageChunk
|
||||
from exo.shared.types.common import CommandId, Id, MetaInstanceId, NodeId, SessionId
|
||||
from exo.shared.types.meta_instance import MetaInstance
|
||||
from exo.shared.types.common import CommandId, Id, NodeId, SessionId
|
||||
from exo.shared.types.tasks import Task, TaskId, TaskStatus
|
||||
from exo.shared.types.worker.downloads import DownloadProgress
|
||||
from exo.shared.types.worker.instances import Instance, InstanceId
|
||||
@@ -67,30 +66,6 @@ class InstanceCreated(BaseEvent):
|
||||
|
||||
class InstanceDeleted(BaseEvent):
|
||||
instance_id: InstanceId
|
||||
failure_error: str | None = None
|
||||
|
||||
|
||||
class MetaInstanceCreated(BaseEvent):
|
||||
meta_instance: MetaInstance
|
||||
|
||||
|
||||
class MetaInstanceDeleted(BaseEvent):
|
||||
meta_instance_id: MetaInstanceId
|
||||
|
||||
|
||||
@final
|
||||
class MetaInstancePlacementFailed(BaseEvent):
|
||||
meta_instance_id: MetaInstanceId
|
||||
reason: str
|
||||
|
||||
|
||||
@final
|
||||
class InstanceRetrying(BaseEvent):
|
||||
"""Runners failed but retry count is below the limit — restart runners, keep instance."""
|
||||
|
||||
instance_id: InstanceId
|
||||
meta_instance_id: MetaInstanceId
|
||||
failure_error: str
|
||||
|
||||
|
||||
class RunnerStatusUpdated(BaseEvent):
|
||||
@@ -166,10 +141,6 @@ Event = (
|
||||
| TaskAcknowledged
|
||||
| InstanceCreated
|
||||
| InstanceDeleted
|
||||
| InstanceRetrying
|
||||
| MetaInstanceCreated
|
||||
| MetaInstanceDeleted
|
||||
| MetaInstancePlacementFailed
|
||||
| RunnerStatusUpdated
|
||||
| RunnerDeleted
|
||||
| NodeTimedOut
|
||||
|
||||
@@ -1,25 +0,0 @@
|
||||
from typing import final
|
||||
|
||||
from pydantic import Field
|
||||
|
||||
from exo.shared.models.model_cards import ModelId
|
||||
from exo.shared.types.common import MetaInstanceId, NodeId
|
||||
from exo.shared.types.worker.instances import InstanceMeta
|
||||
from exo.shared.types.worker.shards import Sharding
|
||||
from exo.utils.pydantic_ext import FrozenModel
|
||||
|
||||
|
||||
@final
|
||||
class MetaInstance(FrozenModel):
|
||||
"""Declarative constraint: ensure an instance matching these parameters always exists."""
|
||||
|
||||
meta_instance_id: MetaInstanceId = Field(default_factory=MetaInstanceId)
|
||||
model_id: ModelId
|
||||
sharding: Sharding = Sharding.Pipeline
|
||||
instance_meta: InstanceMeta = InstanceMeta.MlxRing
|
||||
min_nodes: int = 1
|
||||
node_ids: list[NodeId] | None = None
|
||||
# Failure tracking
|
||||
placement_error: str | None = None
|
||||
consecutive_failures: int = 0
|
||||
last_failure_error: str | None = None
|
||||
@@ -6,8 +6,7 @@ from pydantic import ConfigDict, Field, field_serializer, field_validator
|
||||
from pydantic.alias_generators import to_camel
|
||||
|
||||
from exo.shared.topology import Topology, TopologySnapshot
|
||||
from exo.shared.types.common import MetaInstanceId, NodeId
|
||||
from exo.shared.types.meta_instance import MetaInstance
|
||||
from exo.shared.types.common import NodeId
|
||||
from exo.shared.types.profiling import (
|
||||
DiskUsage,
|
||||
MemoryUsage,
|
||||
@@ -42,7 +41,6 @@ class State(CamelCaseModel):
|
||||
arbitrary_types_allowed=True,
|
||||
)
|
||||
instances: Mapping[InstanceId, Instance] = {}
|
||||
meta_instances: Mapping[MetaInstanceId, MetaInstance] = {}
|
||||
runners: Mapping[RunnerId, RunnerStatus] = {}
|
||||
downloads: Mapping[NodeId, Sequence[DownloadProgress]] = {}
|
||||
tasks: Mapping[TaskId, Task] = {}
|
||||
|
||||
@@ -2,7 +2,7 @@ from enum import Enum
|
||||
|
||||
from pydantic import model_validator
|
||||
|
||||
from exo.shared.types.common import Host, Id, MetaInstanceId, NodeId
|
||||
from exo.shared.types.common import Host, Id, NodeId
|
||||
from exo.shared.types.worker.runners import RunnerId, ShardAssignments, ShardMetadata
|
||||
from exo.utils.pydantic_ext import CamelCaseModel, TaggedModel
|
||||
|
||||
@@ -19,7 +19,6 @@ class InstanceMeta(str, Enum):
|
||||
class BaseInstance(TaggedModel):
|
||||
instance_id: InstanceId
|
||||
shard_assignments: ShardAssignments
|
||||
meta_instance_id: MetaInstanceId | None = None
|
||||
|
||||
def shard(self, runner_id: RunnerId) -> ShardMetadata | None:
|
||||
return self.shard_assignments.runner_to_shard.get(runner_id, None)
|
||||
|
||||
@@ -342,7 +342,7 @@ class Worker:
|
||||
session=self.session_id,
|
||||
event=event,
|
||||
)
|
||||
logger.debug(f"Worker published event {idx}: {str(event)[:100]}")
|
||||
logger.trace(f"Worker published event {idx}: {str(event)[:100]}")
|
||||
await self.local_event_sender.send(fe)
|
||||
self.out_for_delivery[event.event_id] = fe
|
||||
|
||||
|
||||
@@ -34,7 +34,6 @@ from exo.shared.types.worker.runners import (
|
||||
RunnerLoading,
|
||||
RunnerReady,
|
||||
RunnerRunning,
|
||||
RunnerShutdown,
|
||||
RunnerStatus,
|
||||
RunnerWarmingUp,
|
||||
)
|
||||
@@ -55,7 +54,7 @@ def plan(
|
||||
# Python short circuiting OR logic should evaluate these sequentially.
|
||||
return (
|
||||
_kill_runner(runners, all_runners, instances)
|
||||
or _create_runner(node_id, runners, instances, all_runners)
|
||||
or _create_runner(node_id, runners, instances)
|
||||
or _model_needs_download(node_id, runners, global_download_status)
|
||||
or _init_distributed_backend(runners, all_runners)
|
||||
or _load_model(runners, all_runners, global_download_status)
|
||||
@@ -74,12 +73,6 @@ def _kill_runner(
|
||||
if (instance_id := runner.bound_instance.instance.instance_id) not in instances:
|
||||
return Shutdown(instance_id=instance_id, runner_id=runner_id)
|
||||
|
||||
# Master removed our runner from state (retry signal) and process is dead
|
||||
if runner_id not in all_runners and isinstance(
|
||||
runner.status, (RunnerFailed, RunnerShutdown)
|
||||
):
|
||||
return Shutdown(instance_id=instance_id, runner_id=runner_id)
|
||||
|
||||
for (
|
||||
global_runner_id
|
||||
) in runner.bound_instance.instance.shard_assignments.node_to_runner.values():
|
||||
@@ -97,7 +90,6 @@ def _create_runner(
|
||||
node_id: NodeId,
|
||||
runners: Mapping[RunnerId, RunnerSupervisor],
|
||||
instances: Mapping[InstanceId, Instance],
|
||||
all_runners: Mapping[RunnerId, RunnerStatus],
|
||||
) -> CreateRunner | None:
|
||||
for instance in instances.values():
|
||||
runner_id = instance.shard_assignments.node_to_runner.get(node_id, None)
|
||||
@@ -107,16 +99,6 @@ def _create_runner(
|
||||
if runner_id in runners:
|
||||
continue
|
||||
|
||||
# Don't create while any peer runner is in a terminal state — wait for
|
||||
# the master to emit InstanceRetrying which removes them from state.
|
||||
has_terminal_peer = any(
|
||||
isinstance(all_runners.get(peer_rid), (RunnerFailed, RunnerShutdown))
|
||||
for peer_rid in instance.shard_assignments.node_to_runner.values()
|
||||
if peer_rid != runner_id
|
||||
)
|
||||
if has_terminal_peer:
|
||||
continue
|
||||
|
||||
shard = instance.shard(runner_id)
|
||||
assert shard is not None
|
||||
|
||||
|
||||
@@ -28,12 +28,12 @@ trap 'cleanup' EXIT INT TERM
|
||||
|
||||
for host; do
|
||||
ssh -T -o BatchMode=yes -o ServerAliveInterval=30 "$host@$host" \
|
||||
"EXO_LIBP2P_NAMESPACE=$commit /nix/var/nix/profiles/default/bin/nix build github:exo-explore/exo/$commit" &
|
||||
"/nix/var/nix/profiles/default/bin/nix build github:exo-explore/exo/$commit" &
|
||||
done
|
||||
wait
|
||||
for host; do
|
||||
ssh -T -o BatchMode=yes -o ServerAliveInterval=30 "$host@$host" \
|
||||
"EXO_LIBP2P_NAMESPACE=$commit /nix/var/nix/profiles/default/bin/nix run github:exo-explore/exo/$commit" &>/dev/null &
|
||||
"/nix/var/nix/profiles/default/bin/nix run github:exo-explore/exo/$commit -- --namespace $commit" &>/dev/null &
|
||||
done
|
||||
|
||||
for host; do
|
||||
|
||||
@@ -35,7 +35,7 @@ i=0
|
||||
for host; do
|
||||
colour=${colours[i++ % 4]}
|
||||
ssh -T -o BatchMode=yes -o ServerAliveInterval=30 "$host@$host" \
|
||||
"EXO_LIBP2P_NAMESPACE=$commit /nix/var/nix/profiles/default/bin/nix run github:exo-explore/exo/$commit" |&
|
||||
"/nix/var/nix/profiles/default/bin/nix run github:exo-explore/exo/$commit -- --namespace $commit" |&
|
||||
awk -v p="${colour}[${host}]${reset}" '{ print p $0; fflush() }' &
|
||||
done
|
||||
|
||||
|
||||
12
uv.lock
generated
12
uv.lock
generated
@@ -193,20 +193,14 @@ sdist = { url = "https://files.pythonhosted.org/packages/eb/56/b1ba7935a17738ae8
|
||||
wheels = [
|
||||
{ url = "https://files.pythonhosted.org/packages/b0/1e/d22cc63332bd59b06481ceaac49d6c507598642e2230f201649058a7e704/cffi-2.0.0-cp313-cp313-manylinux1_i686.manylinux2014_i686.manylinux_2_17_i686.manylinux_2_5_i686.whl", hash = "sha256:07b271772c100085dd28b74fa0cd81c8fb1a3ba18b21e03d7c27f3436a10606b", size = 212446, upload-time = "2025-09-08T23:23:03.472Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/a9/f5/a2c23eb03b61a0b8747f211eb716446c826ad66818ddc7810cc2cc19b3f2/cffi-2.0.0-cp313-cp313-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:d48a880098c96020b02d5a1f7d9251308510ce8858940e6fa99ece33f610838b", size = 220101, upload-time = "2025-09-08T23:23:04.792Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/f2/7f/e6647792fc5850d634695bc0e6ab4111ae88e89981d35ac269956605feba/cffi-2.0.0-cp313-cp313-manylinux2014_ppc64le.manylinux_2_17_ppc64le.whl", hash = "sha256:f93fd8e5c8c0a4aa1f424d6173f14a892044054871c771f8566e4008eaa359d2", size = 207948, upload-time = "2025-09-08T23:23:06.127Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/cb/1e/a5a1bd6f1fb30f22573f76533de12a00bf274abcdc55c8edab639078abb6/cffi-2.0.0-cp313-cp313-manylinux2014_s390x.manylinux_2_17_s390x.whl", hash = "sha256:dd4f05f54a52fb558f1ba9f528228066954fee3ebe629fc1660d874d040ae5a3", size = 206422, upload-time = "2025-09-08T23:23:07.753Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/98/df/0a1755e750013a2081e863e7cd37e0cdd02664372c754e5560099eb7aa44/cffi-2.0.0-cp313-cp313-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:c8d3b5532fc71b7a77c09192b4a5a200ea992702734a2e9279a37f2478236f26", size = 219499, upload-time = "2025-09-08T23:23:09.648Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/50/e1/a969e687fcf9ea58e6e2a928ad5e2dd88cc12f6f0ab477e9971f2309b57c/cffi-2.0.0-cp313-cp313-musllinux_1_2_aarch64.whl", hash = "sha256:d9b29c1f0ae438d5ee9acb31cadee00a58c46cc9c0b2f9038c6b0b3470877a8c", size = 222928, upload-time = "2025-09-08T23:23:10.928Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/36/54/0362578dd2c9e557a28ac77698ed67323ed5b9775ca9d3fe73fe191bb5d8/cffi-2.0.0-cp313-cp313-musllinux_1_2_x86_64.whl", hash = "sha256:6d50360be4546678fc1b79ffe7a66265e28667840010348dd69a314145807a1b", size = 221302, upload-time = "2025-09-08T23:23:12.42Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/d6/43/0e822876f87ea8a4ef95442c3d766a06a51fc5298823f884ef87aaad168c/cffi-2.0.0-cp314-cp314-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:24b6f81f1983e6df8db3adc38562c83f7d4a0c36162885ec7f7b77c7dcbec97b", size = 220049, upload-time = "2025-09-08T23:23:20.853Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/b4/89/76799151d9c2d2d1ead63c2429da9ea9d7aac304603de0c6e8764e6e8e70/cffi-2.0.0-cp314-cp314-manylinux2014_ppc64le.manylinux_2_17_ppc64le.whl", hash = "sha256:12873ca6cb9b0f0d3a0da705d6086fe911591737a59f28b7936bdfed27c0d47c", size = 207793, upload-time = "2025-09-08T23:23:22.08Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/bb/dd/3465b14bb9e24ee24cb88c9e3730f6de63111fffe513492bf8c808a3547e/cffi-2.0.0-cp314-cp314-manylinux2014_s390x.manylinux_2_17_s390x.whl", hash = "sha256:d9b97165e8aed9272a6bb17c01e3cc5871a594a446ebedc996e2397a1c1ea8ef", size = 206300, upload-time = "2025-09-08T23:23:23.314Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/47/d9/d83e293854571c877a92da46fdec39158f8d7e68da75bf73581225d28e90/cffi-2.0.0-cp314-cp314-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:afb8db5439b81cf9c9d0c80404b60c3cc9c3add93e114dcae767f1477cb53775", size = 219244, upload-time = "2025-09-08T23:23:24.541Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/2b/0f/1f177e3683aead2bb00f7679a16451d302c436b5cbf2505f0ea8146ef59e/cffi-2.0.0-cp314-cp314-musllinux_1_2_aarch64.whl", hash = "sha256:737fe7d37e1a1bffe70bd5754ea763a62a066dc5913ca57e957824b72a85e205", size = 222828, upload-time = "2025-09-08T23:23:26.143Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/c6/0f/cafacebd4b040e3119dcb32fed8bdef8dfe94da653155f9d0b9dc660166e/cffi-2.0.0-cp314-cp314-musllinux_1_2_x86_64.whl", hash = "sha256:38100abb9d1b1435bc4cc340bb4489635dc2f0da7456590877030c9b3d40b0c1", size = 220926, upload-time = "2025-09-08T23:23:27.873Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/be/b4/c56878d0d1755cf9caa54ba71e5d049479c52f9e4afc230f06822162ab2f/cffi-2.0.0-cp314-cp314t-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:7cc09976e8b56f8cebd752f7113ad07752461f48a58cbba644139015ac24954c", size = 221593, upload-time = "2025-09-08T23:23:31.91Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/e0/0d/eb704606dfe8033e7128df5e90fee946bbcb64a04fcdaa97321309004000/cffi-2.0.0-cp314-cp314t-manylinux2014_ppc64le.manylinux_2_17_ppc64le.whl", hash = "sha256:92b68146a71df78564e4ef48af17551a5ddd142e5190cdf2c5624d0c3ff5b2e8", size = 209354, upload-time = "2025-09-08T23:23:33.214Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/d8/19/3c435d727b368ca475fb8742ab97c9cb13a0de600ce86f62eab7fa3eea60/cffi-2.0.0-cp314-cp314t-manylinux2014_s390x.manylinux_2_17_s390x.whl", hash = "sha256:b1e74d11748e7e98e2f426ab176d4ed720a64412b6a15054378afdb71e0f37dc", size = 208480, upload-time = "2025-09-08T23:23:34.495Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/d0/44/681604464ed9541673e486521497406fadcc15b5217c3e326b061696899a/cffi-2.0.0-cp314-cp314t-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:28a3a209b96630bca57cce802da70c266eb08c6e97e5afd61a75611ee6c64592", size = 221584, upload-time = "2025-09-08T23:23:36.096Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/25/8e/342a504ff018a2825d395d44d63a767dd8ebc927ebda557fecdaca3ac33a/cffi-2.0.0-cp314-cp314t-musllinux_1_2_aarch64.whl", hash = "sha256:7553fb2090d71822f02c629afe6042c299edf91ba1bf94951165613553984512", size = 224443, upload-time = "2025-09-08T23:23:37.328Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/e1/5e/b666bacbbc60fbf415ba9988324a132c9a7a0448a9a8f125074671c0f2c3/cffi-2.0.0-cp314-cp314t-musllinux_1_2_x86_64.whl", hash = "sha256:6c6c373cfc5c83a975506110d17457138c8c63016b563cc9ed6e056a82f13ce4", size = 223437, upload-time = "2025-09-08T23:23:38.945Z" },
|
||||
@@ -312,10 +306,8 @@ wheels = [
|
||||
{ url = "https://files.pythonhosted.org/packages/5c/49/498c86566a1d80e978b42f0d702795f69887005548c041636df6ae1ca64c/cryptography-46.0.3-cp311-abi3-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:01ca9ff2885f3acc98c29f1860552e37f6d7c7d013d7334ff2a9de43a449315d", size = 4450807, upload-time = "2025-10-15T23:16:56.414Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/4b/0a/863a3604112174c8624a2ac3c038662d9e59970c7f926acdcfaed8d61142/cryptography-46.0.3-cp311-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:6eae65d4c3d33da080cff9c4ab1f711b15c1d9760809dad6ea763f3812d254cb", size = 4299615, upload-time = "2025-10-15T23:16:58.442Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/64/02/b73a533f6b64a69f3cd3872acb6ebc12aef924d8d103133bb3ea750dc703/cryptography-46.0.3-cp311-abi3-manylinux_2_28_armv7l.manylinux_2_31_armv7l.whl", hash = "sha256:e5bf0ed4490068a2e72ac03d786693adeb909981cc596425d09032d372bcc849", size = 4016800, upload-time = "2025-10-15T23:17:00.378Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/25/d5/16e41afbfa450cde85a3b7ec599bebefaef16b5c6ba4ec49a3532336ed72/cryptography-46.0.3-cp311-abi3-manylinux_2_28_ppc64le.whl", hash = "sha256:5ecfccd2329e37e9b7112a888e76d9feca2347f12f37918facbb893d7bb88ee8", size = 4984707, upload-time = "2025-10-15T23:17:01.98Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/c9/56/e7e69b427c3878352c2fb9b450bd0e19ed552753491d39d7d0a2f5226d41/cryptography-46.0.3-cp311-abi3-manylinux_2_28_x86_64.whl", hash = "sha256:a2c0cd47381a3229c403062f764160d57d4d175e022c1df84e168c6251a22eec", size = 4482541, upload-time = "2025-10-15T23:17:04.078Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/78/f6/50736d40d97e8483172f1bb6e698895b92a223dba513b0ca6f06b2365339/cryptography-46.0.3-cp311-abi3-manylinux_2_34_aarch64.whl", hash = "sha256:549e234ff32571b1f4076ac269fcce7a808d3bf98b76c8dd560e42dbc66d7d91", size = 4299464, upload-time = "2025-10-15T23:17:05.483Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/00/de/d8e26b1a855f19d9994a19c702fa2e93b0456beccbcfe437eda00e0701f2/cryptography-46.0.3-cp311-abi3-manylinux_2_34_ppc64le.whl", hash = "sha256:c0a7bb1a68a5d3471880e264621346c48665b3bf1c3759d682fc0864c540bd9e", size = 4950838, upload-time = "2025-10-15T23:17:07.425Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/8f/29/798fc4ec461a1c9e9f735f2fc58741b0daae30688f41b2497dcbc9ed1355/cryptography-46.0.3-cp311-abi3-manylinux_2_34_x86_64.whl", hash = "sha256:10b01676fc208c3e6feeb25a8b83d81767e8059e1fe86e1dc62d10a3018fa926", size = 4481596, upload-time = "2025-10-15T23:17:09.343Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/15/8d/03cd48b20a573adfff7652b76271078e3045b9f49387920e7f1f631d125e/cryptography-46.0.3-cp311-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:0abf1ffd6e57c67e92af68330d05760b7b7efb243aab8377e583284dbab72c71", size = 4426782, upload-time = "2025-10-15T23:17:11.22Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/fa/b1/ebacbfe53317d55cf33165bda24c86523497a6881f339f9aae5c2e13e57b/cryptography-46.0.3-cp311-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:a04bee9ab6a4da801eb9b51f1b708a1b5b5c9eb48c03f74198464c66f0d344ac", size = 4698381, upload-time = "2025-10-15T23:17:12.829Z" },
|
||||
@@ -323,10 +315,8 @@ wheels = [
|
||||
{ url = "https://files.pythonhosted.org/packages/c5/fd/bc1daf8230eaa075184cbbf5f8cd00ba9db4fd32d63fb83da4671b72ed8a/cryptography-46.0.3-cp314-cp314t-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:39b6755623145ad5eff1dab323f4eae2a32a77a7abef2c5089a04a3d04366715", size = 4435078, upload-time = "2025-10-15T23:17:23.042Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/82/98/d3bd5407ce4c60017f8ff9e63ffee4200ab3e23fe05b765cab805a7db008/cryptography-46.0.3-cp314-cp314t-manylinux_2_28_aarch64.whl", hash = "sha256:db391fa7c66df6762ee3f00c95a89e6d428f4d60e7abc8328f4fe155b5ac6e54", size = 4293460, upload-time = "2025-10-15T23:17:24.885Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/26/e9/e23e7900983c2b8af7a08098db406cf989d7f09caea7897e347598d4cd5b/cryptography-46.0.3-cp314-cp314t-manylinux_2_28_armv7l.manylinux_2_31_armv7l.whl", hash = "sha256:78a97cf6a8839a48c49271cdcbd5cf37ca2c1d6b7fdd86cc864f302b5e9bf459", size = 3995237, upload-time = "2025-10-15T23:17:26.449Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/91/15/af68c509d4a138cfe299d0d7ddb14afba15233223ebd933b4bbdbc7155d3/cryptography-46.0.3-cp314-cp314t-manylinux_2_28_ppc64le.whl", hash = "sha256:dfb781ff7eaa91a6f7fd41776ec37c5853c795d3b358d4896fdbb5df168af422", size = 4967344, upload-time = "2025-10-15T23:17:28.06Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/ca/e3/8643d077c53868b681af077edf6b3cb58288b5423610f21c62aadcbe99f4/cryptography-46.0.3-cp314-cp314t-manylinux_2_28_x86_64.whl", hash = "sha256:6f61efb26e76c45c4a227835ddeae96d83624fb0d29eb5df5b96e14ed1a0afb7", size = 4466564, upload-time = "2025-10-15T23:17:29.665Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/0e/43/c1e8726fa59c236ff477ff2b5dc071e54b21e5a1e51aa2cee1676f1c986f/cryptography-46.0.3-cp314-cp314t-manylinux_2_34_aarch64.whl", hash = "sha256:23b1a8f26e43f47ceb6d6a43115f33a5a37d57df4ea0ca295b780ae8546e8044", size = 4292415, upload-time = "2025-10-15T23:17:31.686Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/42/f9/2f8fefdb1aee8a8e3256a0568cffc4e6d517b256a2fe97a029b3f1b9fe7e/cryptography-46.0.3-cp314-cp314t-manylinux_2_34_ppc64le.whl", hash = "sha256:b419ae593c86b87014b9be7396b385491ad7f320bde96826d0dd174459e54665", size = 4931457, upload-time = "2025-10-15T23:17:33.478Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/79/30/9b54127a9a778ccd6d27c3da7563e9f2d341826075ceab89ae3b41bf5be2/cryptography-46.0.3-cp314-cp314t-manylinux_2_34_x86_64.whl", hash = "sha256:50fc3343ac490c6b08c0cf0d704e881d0d660be923fd3076db3e932007e726e3", size = 4466074, upload-time = "2025-10-15T23:17:35.158Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/ac/68/b4f4a10928e26c941b1b6a179143af9f4d27d88fe84a6a3c53592d2e76bf/cryptography-46.0.3-cp314-cp314t-musllinux_1_2_aarch64.whl", hash = "sha256:22d7e97932f511d6b0b04f2bfd818d73dcd5928db509460aaf48384778eb6d20", size = 4420569, upload-time = "2025-10-15T23:17:37.188Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/a3/49/3746dab4c0d1979888f125226357d3262a6dd40e114ac29e3d2abdf1ec55/cryptography-46.0.3-cp314-cp314t-musllinux_1_2_x86_64.whl", hash = "sha256:d55f3dffadd674514ad19451161118fd010988540cee43d8bc20675e775925de", size = 4681941, upload-time = "2025-10-15T23:17:39.236Z" },
|
||||
@@ -334,10 +324,8 @@ wheels = [
|
||||
{ url = "https://files.pythonhosted.org/packages/26/42/fa8389d4478368743e24e61eea78846a0006caffaf72ea24a15159215a14/cryptography-46.0.3-cp38-abi3-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:15ab9b093e8f09daab0f2159bb7e47532596075139dd74365da52ecc9cb46c5d", size = 4440029, upload-time = "2025-10-15T23:17:49.837Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/5f/eb/f483db0ec5ac040824f269e93dd2bd8a21ecd1027e77ad7bdf6914f2fd80/cryptography-46.0.3-cp38-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:46acf53b40ea38f9c6c229599a4a13f0d46a6c3fa9ef19fc1a124d62e338dfa0", size = 4297222, upload-time = "2025-10-15T23:17:51.357Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/fd/cf/da9502c4e1912cb1da3807ea3618a6829bee8207456fbbeebc361ec38ba3/cryptography-46.0.3-cp38-abi3-manylinux_2_28_armv7l.manylinux_2_31_armv7l.whl", hash = "sha256:10ca84c4668d066a9878890047f03546f3ae0a6b8b39b697457b7757aaf18dbc", size = 4012280, upload-time = "2025-10-15T23:17:52.964Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/6b/8f/9adb86b93330e0df8b3dcf03eae67c33ba89958fc2e03862ef1ac2b42465/cryptography-46.0.3-cp38-abi3-manylinux_2_28_ppc64le.whl", hash = "sha256:36e627112085bb3b81b19fed209c05ce2a52ee8b15d161b7c643a7d5a88491f3", size = 4978958, upload-time = "2025-10-15T23:17:54.965Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/d1/a0/5fa77988289c34bdb9f913f5606ecc9ada1adb5ae870bd0d1054a7021cc4/cryptography-46.0.3-cp38-abi3-manylinux_2_28_x86_64.whl", hash = "sha256:1000713389b75c449a6e979ffc7dcc8ac90b437048766cef052d4d30b8220971", size = 4473714, upload-time = "2025-10-15T23:17:56.754Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/14/e5/fc82d72a58d41c393697aa18c9abe5ae1214ff6f2a5c18ac470f92777895/cryptography-46.0.3-cp38-abi3-manylinux_2_34_aarch64.whl", hash = "sha256:b02cf04496f6576afffef5ddd04a0cb7d49cf6be16a9059d793a30b035f6b6ac", size = 4296970, upload-time = "2025-10-15T23:17:58.588Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/78/06/5663ed35438d0b09056973994f1aec467492b33bd31da36e468b01ec1097/cryptography-46.0.3-cp38-abi3-manylinux_2_34_ppc64le.whl", hash = "sha256:71e842ec9bc7abf543b47cf86b9a743baa95f4677d22baa4c7d5c69e49e9bc04", size = 4940236, upload-time = "2025-10-15T23:18:00.897Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/fc/59/873633f3f2dcd8a053b8dd1d38f783043b5fce589c0f6988bf55ef57e43e/cryptography-46.0.3-cp38-abi3-manylinux_2_34_x86_64.whl", hash = "sha256:402b58fc32614f00980b66d6e56a5b4118e6cb362ae8f3fda141ba4689bd4506", size = 4472642, upload-time = "2025-10-15T23:18:02.749Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/3d/39/8e71f3930e40f6877737d6f69248cf74d4e34b886a3967d32f919cc50d3b/cryptography-46.0.3-cp38-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:ef639cb3372f69ec44915fafcd6698b6cc78fbe0c2ea41be867f6ed612811963", size = 4423126, upload-time = "2025-10-15T23:18:04.85Z" },
|
||||
{ url = "https://files.pythonhosted.org/packages/cd/c7/f65027c2810e14c3e7268353b1681932b87e5a48e65505d8cc17c99e36ae/cryptography-46.0.3-cp38-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:3b51b8ca4f1c6453d8829e1eb7299499ca7f313900dd4d89a24b8b87c0a780d4", size = 4686573, upload-time = "2025-10-15T23:18:06.908Z" },
|
||||
|
||||
Reference in New Issue
Block a user