Compare commits

..

18 Commits

Author SHA1 Message Date
Alex Cheema
19a21e9065 feat(dashboard): add light/dark mode toggle with warm parchment palette
Adds a theme system to the EXO dashboard with a "Mission Control, Dawn
Shift" light mode — warm parchment backgrounds (oklch(0.97 0.015 80))
and deep amber/brass accents (oklch(0.50 0.14 65)) that feel premium
rather than cold.

Changes:
- dashboard/src/lib/stores/theme.svelte.ts: new Svelte 5 rune store,
  persists choice to localStorage under 'exo-theme'
- dashboard/src/app.html: FOUC prevention — html starts as class="dark",
  inline script reads localStorage and switches to class="light" before
  first paint
- dashboard/src/routes/+layout.svelte: calls theme.init() on mount to
  sync rune state with the DOM class
- dashboard/src/lib/components/HeaderNav.svelte: sun/moon toggle button
  in the right nav area
- dashboard/src/app.css: full html.light palette + utility overrides
  (scrollbar, logo filter, graph links, scanlines, etc.)

No new npm dependencies — avoids mode-watcher entirely.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-02-18 12:36:51 -08:00
Alex Cheema
6c322ebb72 feat: only show thinking toggle for models that support it (#1497)
## Summary
- Adds `thinking_toggle` capability to 26 model cards that support
toggling thinking mode on/off
- GPT-OSS models (20b, 120b) excluded — they always think and don't
support toggling
- Dashboard UI updated to check for `thinking_toggle` capability before
showing the toggle button

## Test plan
- [x] `uv run basedpyright` — 0 errors
- [x] `uv run ruff check` — all checks passed
- [x] `nix fmt` — 0 files changed
- [x] `uv run pytest` — 188 passed, 0 failed
- [x] Security review passed (no secrets, eval/exec, innerHTML, or dep
changes)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-18 17:05:00 +00:00
vskiwi
2ebe6216b4 feat: add explicit --offline mode for air-gapped clusters (#1525)
## Motivation

Closes #1510

There is currently no reliable way to run exo on an air-gapped or offline cluster where models are pre-staged on local disks. The two existing mechanisms — `--no-downloads` and `HF_HUB_OFFLINE=1` — each cover only a subset of the problem:

1. **`--no-downloads` blocks model loading**: When passed, `DownloadCoordinator` is not created. No `NodeDownloadProgress` events are ever emitted, so `_model_needs_download()` in `plan.py` perpetually returns `DownloadModel`, short-circuiting `_load_model()` and preventing the model from ever being loaded.

2. **`HF_HUB_OFFLINE=1` doesn't cover exo's aiohttp code**: exo's download pipeline primarily uses raw `aiohttp` for HTTP operations (file list fetching, file downloads, HEAD verification), not the `huggingface_hub` library. These calls will attempt connections and time out on air-gapped networks.

3. **`skip_internet` is not propagated to `download_file_with_retry()`**: Even when `internet_connection = False`, the `_download_file()` function still makes HTTP HEAD calls via `file_meta()` to verify local files and unconditionally attempts downloads for missing files.

## Changes

### `src/exo/main.py`
- Add `--offline` flag to `Args` with env var detection (`EXO_OFFLINE=1`, `HF_HUB_OFFLINE=1`)
- Pass `offline` to `DownloadCoordinator` at creation and re-creation (election loop)

### `src/exo/download/coordinator.py`
- Add `offline: bool = False` field
- In offline mode: set `internet_connection = False` immediately in `__post_init__`, skip `_test_internet_connection()` ping (avoids 3s timeout), skip `_check_internet_connection` periodic loop
- In `_start_download()`: if model is not fully available locally, emit `DownloadFailed` with clear message instead of starting a download task

### `src/exo/download/download_utils.py`
- Add `skip_internet: bool` parameter to `download_file_with_retry()` and `_download_file()`
- When `skip_internet=True` in `_download_file()`: return local file immediately without HTTP HEAD verification; raise `FileNotFoundError` for missing files
- Propagate `skip_internet` from `download_shard()` to `download_file_with_retry()`

### `src/exo/download/tests/test_offline_mode.py` (new)
- 8 tests covering `_download_file`, `download_file_with_retry`, and `fetch_file_list_with_cache` in offline mode

## Why It Works

Unlike `--no-downloads` which disables `DownloadCoordinator` entirely, `--offline` keeps the coordinator running in a restricted mode. The existing `_emit_existing_download_progress()` disk scanner still runs every 60 seconds, emitting `DownloadCompleted` events for pre-staged models. These events flow through the event-sourcing pipeline and populate `state.downloads`, which unblocks `_model_needs_download()` in `plan.py` — no changes to the planning logic required.

```
--offline flag
  → DownloadCoordinator (offline mode)
    → Skip 1.1.1.1 ping, internet_connection = False
    → _emit_existing_download_progress scans disk
      → Emits DownloadCompleted for pre-staged models
        → _model_needs_download sees DownloadCompleted
          → _load_model proceeds normally
```

## Test Plan

### Automated Testing
- `ruff check` — passes
- 8 new tests in `test_offline_mode.py` — all pass
- 11 existing download tests in `test_download_verification.py` — all pass (no regressions)

### Manual Testing
1. Pre-stage a model on disk (e.g., `~/.exo/models/mlx-community--Qwen3-0.6B-4bit/`)
2. Start exo with `--offline` (or `EXO_OFFLINE=1`)
3. Place an instance via API or dashboard
4. Verify: model loads into memory and inference works without any network calls

### Environment
- macOS (Apple Silicon), multi-node cluster with Thunderbolt interconnect
- Models pre-staged via rsync / NFS mount
2026-02-18 16:18:09 +00:00
ciaranbor
f54c80b121 Ciaran/image edit api (#1500)
## Motivation

- Image editing previously ignored input image dimensions, always
defaulting to 1024x1024
- Size dropdown was hidden in edit mode, giving users no control over
output dimensions
- Portrait/landscape presets used non-standard aspect ratios (1024x1365
/ 1365x1024)

## Changes

- Added "auto" size option that uses input image dimensions for edits,
defaults to 1024x1024 for generation
- Introduced ImageSize Literal type and normalize_image_size() validator
(replaces raw str size fields)
  - Updated portrait/landscape presets to standard 1024x1536 / 1536x1024
  - Made size selector visible in edit mode (previously hidden)
  - Default size changed from "1024x1024" to "auto"

## Why It Works

- "auto" reads actual input image dimensions via PIL at generation time,
so edits preserve the original aspect ratio
- Pydantic field_validator on both ImageGenerationTaskParams and
ImageEditsTaskParams normalizes None → "auto", keeping the API
backward-compatible

## Test Plan

### Manual Testing

- Verify image edits output at the input image's native resolution when
size is "auto"
- Verify size dropdown appears and works in both generate and edit modes
2026-02-18 16:05:39 +00:00
rltakashige
48b8f86395 Add support for GLM 5 (#1526)
## Motivation

Add GLM 5 support in favor of #1513 

## Changes

<!-- Describe what you changed in detail -->

## Why It Works

<!-- Explain why your approach solves the problem -->

## Test Plan

### Manual Testing
<!-- Hardware: (e.g., MacBook Pro M1 Max 32GB, Mac Mini M2 16GB,
connected via Thunderbolt 4) -->
<!-- What you did: -->
<!-- - -->

### Automated Testing
<!-- Describe changes to automated tests, or how existing tests cover
this change -->
<!-- - -->
2026-02-18 14:04:06 +00:00
Evan
5cbd6377a2 prioritize official model cards over custom model cards
our old model card search path would override official model cards with
custom model cards - our packaged model cards should always be the
default here
2026-02-18 13:20:05 +00:00
Evan Quiney
8f01523ddb remove dead code (#1496) 2026-02-18 11:43:27 +00:00
Alex Cheema
3addeadea8 Update mlx-lm to 0.30.7 (#1520)
## Summary
- Bumps `mlx-lm` from 0.30.6 to 0.30.7 in `pyproject.toml` and `uv.lock`

## Test plan
- [x] `uv lock` resolves successfully
- [x] `basedpyright` — no new errors (63 pre-existing in unrelated
`test_tool_call_tracker.py`)
- [x] `ruff check` — all checks passed
- [x] `nix fmt` — no formatting changes
- [x] `pytest` — 188 passed, 1 skipped

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-18 11:14:23 +00:00
rltakashige
f2be929211 Leo/address rdma gpu locks 2 (#1515)
Same as #1489 . Had to revert and redo thanks to Claude.

---------

Co-authored-by: Jake Hillion <jake@hillion.co.uk>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-17 14:00:52 -08:00
rltakashige
83af8c63fa Revert "Use custom fork that resolves GPU locks" (#1502)
Reverts exo-explore/exo#1489

Goddammit Claude...
2026-02-17 18:18:54 +00:00
Evan Quiney
eccc6298d1 Revert "Add MetaInstance declarative layer (#1447)"
This reverts commit a962a28afc.
2026-02-17 18:11:47 +00:00
Evan Quiney
c8997217cf Revert "feat: better onboarding UX for new users (#1479)"
This reverts commit 490d2e46ba.
2026-02-17 18:02:32 +00:00
Alex Cheema
490d2e46ba feat: better onboarding UX for new users (#1479)
## Summary

- **Auto-open dashboard** in browser on first launch (uses
`~/.exo/.dashboard_opened` marker)
- **Welcome overlay** with "Choose a Model" CTA button when no model
instance is running
- **Tutorial progress messages** during model download → loading → ready
lifecycle stages
- **Fix conversation sidebar** text contrast — bumped to white text,
added active state background
- **Simplify technical jargon** — sharding/instance type/min nodes
hidden behind collapsible "Advanced Options" toggle; strategy display
hidden behind debug mode
- **Polished DMG installer** with drag-to-Applications layout, custom
branded background, and AppleScript-configured window positioning

## Test plan

- [ ] Launch exo for the first time (delete `~/.exo/.dashboard_opened`
to simulate) — browser should auto-open
- [ ] Verify welcome overlay appears on topology when no model is loaded
- [ ] Launch a model and verify download/loading/ready messages appear
in instance cards
- [ ] Check conversation sidebar text is readable (white on dark, yellow
when active)
- [ ] Verify "Advanced Options" toggle hides/shows sharding controls
- [ ] Build DMG with `packaging/dmg/create-dmg.sh` and verify
drag-to-Applications layout

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-17 17:52:49 +00:00
rltakashige
facf2d4d03 Use custom fork that resolves GPU locks (#1489)
## Motivation

There is an issue on Macs that means that an explicit synchronization is
necessary for memory to be updated from L1 cache. This means that GPU
locks can occur when a spin wait does not see the updated timestamp.

## Changes

Updated in my own personal fork.

## Why It Works

https://github.com/ARM-software/acle/releases

## Test Plan

### Manual Testing
Tested manually that no GPU locks occur (even with multiple simultaneous
instances running) and that the performance differential is negligible
(267 vs 269 tps on Llama 3.2 1B at an approx 10k context.)


------------------------------------------------------
I have seen a GPU lock, specifically when sending a particularly large
chat completion while the model was loading. However, I have since been
unable to reproduce and this may be something I did wrong. Please do
create an issue and tag me if any GPU locks do occur.

---------

Co-authored-by: Jake Hillion <jake@hillion.co.uk>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-17 17:48:43 +00:00
Alex Cheema
a962a28afc Add MetaInstance declarative layer (#1447)
## Motivation

Users currently manage instances directly, which means if a node
disconnects or connections break, the instance dies and nothing
recreates it. MetaInstance is a declarative primitive: "ensure an
instance matching these parameters always exists." The reconciler
watches for unhealthy or missing backing instances and re-places them
automatically.

## Changes

- **MetaInstance type** (`meta_instance.py`): declarative constraint
with `model_id`, `min_nodes`, optional `node_ids`, and `sharding`
- **Reconciler** (`reconcile.py`): `find_unsatisfied_meta_instances`
checks which MetaInstances lack a healthy backing instance,
`try_place_for_meta_instance` creates one
- **Master loop** (`main.py`): periodically reconciles unsatisfied
MetaInstances; immediate placement on `CreateMetaInstance` command
- **API** (`api.py`): `create_meta_instance` / `delete_meta_instance` /
`GET /meta_instances` endpoints; delete cascades to backing instances
with task cancellation
- **Binding via `meta_instance_id` on Instance** (`instances.py`): no
separate binding event or backing map — the instance carries its parent
MetaInstance ID directly, eliminating race conditions in the reconciler
- **Dashboard**: sidebar shows MetaInstances with their backing instance
status; orphan instances (created directly) still shown separately
- **Tests**: constraint matching, connection health, unsatisfied
detection, exclusive binding, cascade delete with task cancellation

### Recent improvements

- **fix: cancel active tasks on cascade delete** — `DeleteMetaInstance`
now emits `TaskStatusUpdated(Cancelled)` for any Pending/Running tasks
on backing instances before emitting `InstanceDeleted`. Previously,
cascade-deleting backing instances left orphaned task references in
state.
- **Lifecycle logging** — added `logger.info`/`logger.warning` for:
`CreateMetaInstance` (model, min_nodes, sharding), `DeleteMetaInstance`
(with cascade count), reconciler placement success/failure, and retry
decisions with attempt counts in `InstanceHealthReconciler`.
- **GET `/meta_instances` endpoint** — lists all meta-instances without
needing to fetch full state.
- **2 regression tests** — `test_cascade_delete_cancels_active_tasks`
and `test_cascade_delete_skips_completed_tasks` verify the
cascade-delete event sequence.

## Why It Works

Putting `meta_instance_id` on `BaseInstance` makes binding inherent to
instance creation. When the reconciler creates an instance for a
MetaInstance, it tags it via `model_copy`. When the instance is deleted,
the binding disappears with it. This avoids the two bugs that a separate
binding mechanism would introduce:
1. Stale exclusion sets — the reconciler loop can't accidentally bind
two MetaInstances to the same instance
2. Delete ordering race — no window between deleting an instance and its
binding where the reconciler could re-place

## Test Plan

### Manual Testing
<!-- Hardware: (e.g., MacBook Pro M1 Max 32GB, Mac Mini M2 16GB,
connected via Thunderbolt 4) -->
- Created MetaInstance via dashboard, verified instance placed
- Verified delete cascades (deleting MetaInstance removes backing
instance)
- Verified orphan instances still work independently

### Automated Testing
- 30 tests in `test_meta_instance_edge_cases.py`: lifecycle, retry
logic, error handling, concurrent operations, cascade delete with task
cancellation
- 24 tests in `test_reconcile.py`: constraint matching, connection
health (single/multi-node, edge removal, IP changes), unsatisfied
detection, exclusive binding, idempotency
- All 261 tests pass
- basedpyright 0 errors, ruff clean, dashboard builds

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-17 09:48:19 -08:00
Alex Cheema
db79c350c1 Fix graceful process shutdown in macOS app (#1372)
## Motivation

Fixes #1370

When the macOS app stops exo, GPU/system memory isn't released. This
happens because:

1. The macOS app calls `process.terminate()` (SIGTERM) but the Python
process only registers a graceful shutdown handler for SIGINT, not
SIGTERM. SIGTERM's default Python behavior raises `SystemExit` which
bypasses the cleanup cascade (runner subprocess MLX cleanup via
`mx.clear_cache()`, channel closing, etc.).
2. The app doesn't wait for the process to actually finish cleanup — it
immediately nils out the process reference.

## Changes

**`src/exo/main.py`**: Register SIGTERM handler alongside SIGINT so the
graceful shutdown cascade (`Node.shutdown()` → cancel task group →
worker/runner cleanup → `mx.clear_cache()` + `gc.collect()`) runs
regardless of which signal is received.

**`app/EXO/EXO/ExoProcessController.swift`**: Replace immediate
`process.terminate()` with escalating shutdown per @Evanev7's
suggestion:
1. Send SIGINT via `process.interrupt()` — triggers the registered
Python handler for graceful cleanup
2. Wait up to 5 seconds for the process to exit
3. If still running, escalate to SIGTERM via `process.terminate()`
4. Wait up to 3 seconds
5. If still running, force kill via SIGKILL

The escalation runs in a detached `Task` so the UI updates immediately
(status → stopped) without blocking.

## Why It Works

The root cause is that SIGTERM wasn't triggering the graceful shutdown
path. By registering a SIGTERM handler in Python and sending SIGINT
first from the macOS app, the process gets a chance to run the full
cleanup cascade: cancelling the task group, shutting down runners (which
call `del model; mx.clear_cache(); gc.collect()`), closing channels, and
flushing logs. The escalation to SIGTERM and SIGKILL ensures the process
always terminates even if graceful shutdown hangs.

## Test Plan

### Manual Testing
<!-- Hardware: Mac Studio M4 Max 128GB -->
- Start exo via macOS app, load a model, run inference
- Stop via the toggle switch, verify memory is released without
requiring a system restart
- Test rapid stop/start (restart) to ensure no race conditions

### Automated Testing
- `uv run basedpyright` — 0 errors
- `uv run ruff check` — passes
- `nix fmt` — no changes

---------

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Co-authored-by: Evan Quiney <evanev7@gmail.com>
2026-02-17 09:03:54 -08:00
Alex Cheema
d6301ed593 dashboard: redesign downloads page as model×node table (#1465)
## Motivation

The current downloads page uses a node-centric card grid layout that is
messy and hard to read — the same model across different nodes appears
in separate cards, and deep nesting wastes space. This makes it
difficult to quickly see which models are on which nodes.

## Changes

Rewrote the downloads page
(`dashboard/src/routes/downloads/+page.svelte`) from a card grid to a
clean table layout:

- **Rows** = models (unique across all nodes)
- **Columns** = nodes (with disk free shown in header)
- **Cells** show status at a glance:
  -  Green checkmark + size for completed downloads
  - 🟡 Yellow percentage + mini progress bar + speed for active downloads
  - `...` for pending downloads
  -  Red X for failed downloads
  - `--` for models not present on a node
- Delete/download action buttons appear on row hover
- Model name column is sticky on horizontal scroll (for many-node
clusters)
- Models sorted by number of nodes with completed downloads
- Imported shared utilities from `$lib/utils/downloads` instead of
inline re-implementations

### Backend: model directory in download events

- Added `model_directory` field to `BaseDownloadProgress` so all
download status events include the on-disk path
- Added `_model_dir()` helper to `DownloadCoordinator` to compute the
path from `EXO_MODELS_DIR`
- Dashboard uses this to show file location and enable "open in Finder"
for completed downloads

### Info modal

- Clicking a model name opens an info modal showing card details
(family, quantization, capabilities, storage size, layer count, tensor
parallelism support)

### Other fixes

- Fixed model name truncation in the table
- Excluded `tests/start_distributed_test.py` from pytest collection (CLI
script that calls `sys.exit()` at import time)

## Test Plan

- [x] `uv run basedpyright` — 0 errors
- [x] `uv run ruff check` — all passed
- [x] `nix fmt` — clean
- [x] `uv run pytest` — 188 passed, 1 skipped

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-17 14:31:47 +00:00
Evan Quiney
6d1ca6689b don't time out node identities (#1493)
currently nodes leaving and rejoining the cluster can lose their identity. We have no need to delete this data on node timing out, so let's just persist it.
2026-02-17 11:48:28 +00:00
93 changed files with 3886 additions and 1342 deletions

View File

@@ -0,0 +1,46 @@
"""Type stubs for mlx_lm.models.glm_moe_dsa"""
from dataclasses import dataclass
from typing import Any, Dict, Optional
from .base import BaseModelArgs
from .deepseek_v32 import Model as DSV32Model
@dataclass
class ModelArgs(BaseModelArgs):
model_type: str
vocab_size: int
hidden_size: int
index_head_dim: int
index_n_heads: int
index_topk: int
intermediate_size: int
moe_intermediate_size: int
num_hidden_layers: int
num_attention_heads: int
num_key_value_heads: int
n_shared_experts: Optional[int]
n_routed_experts: Optional[int]
routed_scaling_factor: float
kv_lora_rank: int
q_lora_rank: int
qk_rope_head_dim: int
v_head_dim: int
qk_nope_head_dim: int
topk_method: str
scoring_func: str
norm_topk_prob: bool
n_group: int
topk_group: int
num_experts_per_tok: int
moe_layer_freq: int
first_k_dense_replace: int
max_position_embeddings: int
rms_norm_eps: float
rope_parameters: Dict[str, Any]
attention_bias: bool
rope_scaling: Dict[str, Any] | None
rope_theta: float | None
class Model(DSV32Model):
def __init__(self, config: ModelArgs) -> None: ...

48
Cargo.lock generated
View File

@@ -673,6 +673,17 @@ 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"
@@ -876,23 +887,32 @@ dependencies = [
name = "exo_pyo3_bindings"
version = "0.0.1"
dependencies = [
"delegate",
"env_logger",
"futures-lite",
"extend",
"futures",
"libp2p",
"log",
"networking",
"pin-project",
"pyo3",
"pyo3-async-runtimes",
"pyo3-log",
"pyo3-stub-gen",
"tokio",
"util",
]
[[package]]
name = "fastrand"
version = "2.3.0"
name = "extend"
version = "1.2.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "37909eebbb50d72f9059c3b6d82c0463f2ff062c9e95845c43a6c9c0355411be"
checksum = "311a6d2f1f9d60bff73d2c78a0af97ed27f79672f15c238192a5bbb64db56d00"
dependencies = [
"proc-macro2",
"quote",
"syn 2.0.111",
]
[[package]]
name = "ff"
@@ -1002,10 +1022,7 @@ 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",
]
@@ -1721,6 +1738,12 @@ 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"
@@ -2727,10 +2750,17 @@ dependencies = [
name = "networking"
version = "0.0.1"
dependencies = [
"delegate",
"either",
"extend",
"futures",
"futures-timer",
"keccak-const",
"libp2p",
"log",
"tokio",
"tracing-subscriber",
"util",
]
[[package]]
@@ -4560,6 +4590,10 @@ 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"

View File

@@ -3,6 +3,7 @@ resolver = "3"
members = [
"rust/networking",
"rust/exo_pyo3_bindings",
"rust/util",
]
[workspace.package]
@@ -23,18 +24,34 @@ opt-level = 3
[workspace.dependencies]
## Crate members as common dependencies
networking = { path = "rust/networking" }
util = { path = "rust/util" }
# Macro dependecies
extend = "1.2"
delegate = "0.13"
pin-project = "1"
# Utility dependencies
keccak-const = "0.2"
# Async dependencies
tokio = "1.46"
futures = "0.3"
futures-timer = "3.0"
# Data structures
either = "1.15"
# Tracing/logging
log = "0.4"
# networking
libp2p = "0.56"
libp2p-tcp = "0.44"
[workspace.lints.rust]
static_mut_refs = "warn"
static_mut_refs = "warn" # Or use "warn" instead of deny
incomplete_features = "allow"
# Clippy's lint category level configurations;
# every member crate needs to inherit these by adding
@@ -55,3 +72,64 @@ 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"

View File

@@ -1,5 +1,5 @@
# Missed things
[X] Log namespace on start in exo/main.py
[X] Log EXO_LIBP2P_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

View File

@@ -72,16 +72,23 @@ There are two ways to run exo:
### Run from Source (macOS)
If you have [Nix](https://nixos.org/) installed, you can skip most of the steps below and run exo directly (after accepting the Cachix cache):
```bash
nix run .#exo
```
**Prerequisites:**
- [Xcode](https://developer.apple.com/xcode/) (provides the Metal ToolChain required for MLX compilation)
- [brew](https://github.com/Homebrew/brew) (for simple package management on macOS)
```bash
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
```
- [uv](https://github.com/astral-sh/uv) (for Python dependency management)
- [macmon](https://github.com/vladkens/macmon) (for hardware monitoring on Apple Silicon)
- [node](https://github.com/nodejs/node) (for building the dashboard)
```bash
brew install uv macmon node
```
@@ -199,14 +206,14 @@ 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 `--namespace` cli arg:
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:
- **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 `--namespace` argument when running from source)
- **Configuration**: Access this setting in the app's Advanced settings (or set the `EXO_LIBP2P_NAMESPACE` environment variable when running from source)
The namespace is logged on startup for debugging purposes.
@@ -418,4 +425,4 @@ 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.

View File

@@ -82,7 +82,6 @@ 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
@@ -127,11 +126,37 @@ final class ExoProcessController: ObservableObject {
return
}
process.terminationHandler = nil
if process.isRunning {
process.terminate()
}
self.process = nil
status = .stopped
guard process.isRunning else {
self.process = nil
return
}
let proc = process
self.process = nil
Task.detached {
proc.interrupt()
for _ in 0..<50 {
if !proc.isRunning { return }
try? await Task.sleep(nanoseconds: 100_000_000)
}
if proc.isRunning {
proc.terminate()
}
for _ in 0..<30 {
if !proc.isRunning { return }
try? await Task.sleep(nanoseconds: 100_000_000)
}
if proc.isRunning {
kill(proc.processIdentifier, SIGKILL)
}
}
}
func restart() {
@@ -217,6 +242,7 @@ 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
}

View File

@@ -16,9 +16,10 @@
/* Gotham-inspired accent colors */
--exo-grid: oklch(0.25 0 0);
--exo-scanline: oklch(0.15 0 0);
--exo-glow-yellow: 0 0 20px oklch(0.85 0.18 85 / 0.3);
--exo-glow-yellow-strong: 0 0 40px oklch(0.85 0.18 85 / 0.5);
--exo-glow-yellow: oklch(0.85 0.18 85 / 0.3);
--exo-glow-yellow-strong: oklch(0.85 0.18 85 / 0.5);
--exo-bg-hover: oklch(0.18 0 0);
/* Theme Variables */
--radius: 0.375rem;
--background: var(--exo-black);
@@ -41,6 +42,237 @@
--ring: var(--exo-yellow);
}
/* ============================================================
LIGHT THEME — "Mission Control, Dawn Shift"
Warm parchment + deep amber. Applied when <html> has .light class.
============================================================ */
html.light {
/* EXO brand palette — warm amber shift */
--exo-black: oklch(0.97 0.015 80);
--exo-dark-gray: oklch(0.92 0.012 80);
--exo-medium-gray: oklch(0.83 0.009 78);
--exo-light-gray: oklch(0.50 0.018 75);
--exo-yellow: oklch(0.50 0.14 65);
--exo-yellow-darker: oklch(0.40 0.13 65);
--exo-yellow-glow: oklch(0.60 0.14 65);
--exo-grid: oklch(0.88 0.009 80);
--exo-scanline: oklch(0.93 0.010 80);
--exo-glow-yellow: oklch(0.50 0.14 65 / 0.12);
--exo-glow-yellow-strong: oklch(0.50 0.14 65 / 0.22);
--exo-bg-hover: oklch(0.89 0.010 80);
/* Semantic tokens */
--background: oklch(0.97 0.015 80);
--foreground: oklch(0.13 0.015 75);
--card: oklch(0.92 0.012 80);
--card-foreground: oklch(0.13 0.015 75);
--popover: oklch(0.95 0.012 80);
--popover-foreground: oklch(0.13 0.015 75);
--primary: oklch(0.50 0.14 65);
--primary-foreground: oklch(0.97 0.015 80);
--secondary: oklch(0.88 0.008 80);
--secondary-foreground: oklch(0.15 0.012 75);
--muted: oklch(0.90 0.009 80);
--muted-foreground: oklch(0.50 0.018 75);
--accent: oklch(0.88 0.008 80);
--accent-foreground: oklch(0.15 0.012 75);
--destructive: oklch(0.52 0.22 25);
--border: oklch(0.84 0.007 78);
--input: oklch(0.87 0.008 80);
--ring: oklch(0.50 0.14 65);
}
/* ============================================================
LIGHT MODE UTILITY OVERRIDES
============================================================ */
html.light {
& .text-white,
& .text-white\/90,
& .text-white\/80,
& .text-white\/70 {
color: var(--foreground) !important;
}
& .text-white\/60,
& .text-white\/50 {
color: color-mix(in oklch, var(--foreground) 60%, transparent) !important;
}
& .text-white\/40,
& .text-white\/30 {
color: color-mix(in oklch, var(--foreground) 38%, transparent) !important;
}
& .bg-black\/80,
& .bg-black\/60,
& .bg-black\/50,
& .bg-black\/40 {
background-color: oklch(0.90 0.010 80 / 0.7) !important;
}
& [class*="bg-exo-black/"] {
background-color: oklch(0.90 0.010 80 / 0.6) !important;
}
& [class*="shadow-black"] {
--tw-shadow-color: oklch(0.30 0.010 75 / 0.10) !important;
}
& ::-webkit-scrollbar-track {
background: oklch(0.93 0.010 80) !important;
}
& ::-webkit-scrollbar-thumb {
background: oklch(0.76 0.010 78) !important;
}
& ::-webkit-scrollbar-thumb:hover {
background: oklch(0.50 0.14 65 / 0.6) !important;
}
& .command-panel {
background: linear-gradient(
180deg,
oklch(0.94 0.012 80 / 0.96) 0%,
oklch(0.91 0.010 80 / 0.98) 100%
) !important;
border-color: oklch(0.82 0.008 78) !important;
box-shadow:
inset 0 1px 0 oklch(1 0 0 / 0.6),
0 4px 20px oklch(0.30 0.010 75 / 0.08) !important;
}
& .glow-text {
text-shadow:
0 0 12px oklch(0.50 0.14 65 / 0.20),
0 1px 3px oklch(0.30 0.010 75 / 0.12) !important;
}
& .grid-bg {
background-image:
linear-gradient(oklch(0.75 0.008 78 / 0.25) 1px, transparent 1px),
linear-gradient(90deg, oklch(0.75 0.008 78 / 0.25) 1px, transparent 1px) !important;
}
& .scanlines::before {
background: repeating-linear-gradient(
0deg,
transparent,
transparent 2px,
oklch(0.50 0.010 78 / 0.018) 2px,
oklch(0.50 0.010 78 / 0.018) 4px
) !important;
}
& .crt-screen {
background: radial-gradient(
ellipse at center,
oklch(0.95 0.012 80) 0%,
oklch(0.92 0.010 80) 50%,
oklch(0.89 0.009 80) 100%
) !important;
box-shadow:
inset 0 0 60px oklch(0.30 0.010 75 / 0.04),
0 0 30px oklch(0.50 0.14 65 / 0.04) !important;
}
& .graph-link {
stroke: oklch(0.50 0.018 75 / 0.45) !important;
filter: none !important;
}
& .graph-link-active {
stroke: oklch(0.50 0.14 65 / 0.75) !important;
filter: none !important;
}
& .shooting-stars {
display: none !important;
}
& img[alt="EXO"] {
filter: brightness(0) drop-shadow(0 0 6px oklch(0.30 0.010 75 / 0.10)) !important;
}
& .text-red-400 { color: oklch(0.52 0.22 25) !important; }
& .text-green-400 { color: oklch(0.48 0.17 155) !important; }
& .text-blue-200,
& .text-blue-300,
& .text-blue-400 { color: oklch(0.48 0.17 250) !important; }
& .bg-red-500\/10 { background-color: oklch(0.52 0.22 25 / 0.07) !important; }
& .bg-red-500\/20 { background-color: oklch(0.52 0.22 25 / 0.11) !important; }
& .bg-red-500\/30 { background-color: oklch(0.52 0.22 25 / 0.14) !important; }
& textarea,
& input[type="text"] { color: var(--foreground) !important; }
& textarea::placeholder,
& input::placeholder { color: oklch(0.50 0.012 78 / 0.55) !important; }
& .code-block-wrapper,
& .math-display-wrapper {
background: oklch(0.95 0.010 80) !important;
border-color: oklch(0.83 0.007 78) !important;
}
& .code-block-header,
& .math-display-header {
background: oklch(0.91 0.009 80) !important;
border-color: oklch(0.85 0.007 78) !important;
}
& .inline-code {
background: oklch(0.89 0.009 80) !important;
color: oklch(0.20 0.012 75) !important;
}
& blockquote { background: oklch(0.93 0.010 80) !important; }
& th {
background: oklch(0.90 0.009 80) !important;
border-color: oklch(0.80 0.007 78) !important;
}
& td { border-color: oklch(0.84 0.007 78) !important; }
& hr { border-color: oklch(0.84 0.007 78) !important; }
& .hljs { color: oklch(0.22 0.012 75) !important; }
& .hljs-keyword, & .hljs-selector-tag, & .hljs-literal, & .hljs-section, & .hljs-link {
color: oklch(0.45 0.18 300) !important;
}
& .hljs-string, & .hljs-title, & .hljs-name, & .hljs-type,
& .hljs-attribute, & .hljs-symbol, & .hljs-bullet, & .hljs-addition,
& .hljs-variable, & .hljs-template-tag, & .hljs-template-variable {
color: oklch(0.45 0.14 65) !important;
}
& .hljs-comment, & .hljs-quote, & .hljs-deletion, & .hljs-meta {
color: oklch(0.55 0.010 78) !important;
}
& .hljs-number, & .hljs-regexp, & .hljs-built_in {
color: oklch(0.45 0.15 160) !important;
}
& .hljs-function, & .hljs-class .hljs-title {
color: oklch(0.42 0.17 240) !important;
}
& .katex, & .katex .mord, & .katex .minner, & .katex .mop,
& .katex .mbin, & .katex .mrel, & .katex .mpunct {
color: oklch(0.15 0.012 75) !important;
}
& .katex .frac-line, & .katex .overline-line, & .katex .underline-line,
& .katex .hline, & .katex .rule {
border-color: oklch(0.25 0.012 75) !important;
background: oklch(0.25 0.012 75) !important;
}
& .katex svg { fill: oklch(0.25 0.012 75) !important; stroke: oklch(0.25 0.012 75) !important; }
& .katex svg path { stroke: oklch(0.25 0.012 75) !important; }
& .katex .mopen, & .katex .mclose,
& .katex .delimsizing, & [class^="katex .delim-size"] {
color: oklch(0.35 0.012 75) !important;
}
& .latex-proof { background: oklch(0.96 0.010 80) !important; border-left-color: oklch(0.72 0.010 78) !important; }
& .latex-proof-header { color: oklch(0.22 0.012 75) !important; }
& .latex-proof-content { color: oklch(0.15 0.012 75) !important; }
& .latex-proof-content::after { color: oklch(0.48 0.012 75) !important; }
& .latex-theorem { background: oklch(0.94 0.010 80) !important; border-color: oklch(0.80 0.008 78) !important; }
& .latex-diagram-placeholder {
background: oklch(0.96 0.010 80) !important;
border-color: oklch(0.80 0.008 78) !important;
color: oklch(0.38 0.012 75) !important;
}
}
@theme inline {
--radius-sm: calc(var(--radius) - 2px);
--radius-md: var(--radius);

View File

@@ -1,7 +1,15 @@
<!doctype html>
<html lang="en">
<html lang="en" class="dark">
<head>
<meta charset="utf-8" />
<script>
try {
if (localStorage.getItem('exo-theme') === 'light') {
document.documentElement.classList.remove('dark');
document.documentElement.classList.add('light');
}
} catch (_) {}
</script>
<link rel="icon" href="%sveltekit.assets%/favicon.ico" />
<meta name="viewport" content="width=device-width, initial-scale=1" />
<title>EXO</title>

View File

@@ -103,7 +103,7 @@
const modelSupportsThinking = $derived(() => {
if (!currentModel) return false;
const caps = modelCapabilities[currentModel] || [];
return caps.includes("thinking") && caps.includes("text");
return caps.includes("thinking_toggle") && caps.includes("text");
});
const isEditOnlyWithoutImage = $derived(

View File

@@ -1,5 +1,6 @@
<script lang="ts">
import { browser } from "$app/environment";
import { theme } from "$lib/stores/theme.svelte";
export let showHome = true;
export let onHome: (() => void) | null = null;
@@ -79,10 +80,48 @@
/>
</button>
<!-- Right: Home + Downloads -->
<!-- Right: Theme toggle + Home + Downloads -->
<div
class="absolute right-6 top-1/2 -translate-y-1/2 flex items-center gap-4"
>
<button
onclick={() => theme.toggle()}
class="p-2 rounded border border-exo-medium-gray/40 hover:border-exo-yellow/50 transition-colors cursor-pointer"
title={theme.isLight ? "Switch to dark mode" : "Switch to light mode"}
aria-label={theme.isLight
? "Switch to dark mode"
: "Switch to light mode"}
>
{#if theme.isLight}
<svg
class="w-4 h-4 text-exo-light-gray"
fill="none"
viewBox="0 0 24 24"
stroke="currentColor"
stroke-width="2"
>
<path
stroke-linecap="round"
stroke-linejoin="round"
d="M21 12.79A9 9 0 1111.21 3a7 7 0 009.79 9.79z"
/>
</svg>
{:else}
<svg
class="w-4 h-4 text-exo-light-gray"
fill="none"
viewBox="0 0 24 24"
stroke="currentColor"
stroke-width="2"
>
<circle cx="12" cy="12" r="5" />
<path
stroke-linecap="round"
d="M12 1v2m0 18v2M4.22 4.22l1.42 1.42m12.72 12.72l1.42 1.42M1 12h2m18 0h2M4.22 19.78l1.42-1.42M18.36 5.64l1.42-1.42"
/>
</svg>
{/if}
</button>
{#if showHome}
<button
onclick={handleHome}

View File

@@ -59,13 +59,14 @@
}
const sizeOptions: ImageGenerationParams["size"][] = [
"auto",
"512x512",
"768x768",
"1024x1024",
"1024x768",
"768x1024",
"1024x1365",
"1365x1024",
"1024x1536",
"1536x1024",
];
const qualityOptions: ImageGenerationParams["quality"][] = [
@@ -176,92 +177,90 @@
<div class="border-b border-exo-medium-gray/30 px-3 py-2">
<!-- Basic params row -->
<div class="flex items-center gap-3 flex-wrap">
<!-- Size (hidden in edit mode - output size comes from input image) -->
{#if !isEditMode}
<div class="flex items-center gap-1.5">
<span class="text-xs text-exo-light-gray uppercase tracking-wider"
>SIZE:</span
<!-- Size -->
<div class="flex items-center gap-1.5">
<span class="text-xs text-exo-light-gray uppercase tracking-wider"
>SIZE:</span
>
<div class="relative">
<button
bind:this={sizeButtonRef}
type="button"
onclick={() => (isSizeDropdownOpen = !isSizeDropdownOpen)}
class="bg-exo-medium-gray/50 border border-exo-yellow/30 rounded pl-2 pr-6 py-1 text-xs font-mono text-exo-yellow cursor-pointer transition-all duration-200 hover:border-exo-yellow/50 focus:outline-none focus:border-exo-yellow/70 {isSizeDropdownOpen
? 'border-exo-yellow/70'
: ''}"
>
<div class="relative">
<button
bind:this={sizeButtonRef}
type="button"
onclick={() => (isSizeDropdownOpen = !isSizeDropdownOpen)}
class="bg-exo-medium-gray/50 border border-exo-yellow/30 rounded pl-2 pr-6 py-1 text-xs font-mono text-exo-yellow cursor-pointer transition-all duration-200 hover:border-exo-yellow/50 focus:outline-none focus:border-exo-yellow/70 {isSizeDropdownOpen
? 'border-exo-yellow/70'
: ''}"
{params.size.toUpperCase()}
</button>
<div
class="absolute right-1.5 top-1/2 -translate-y-1/2 pointer-events-none transition-transform duration-200 {isSizeDropdownOpen
? 'rotate-180'
: ''}"
>
<svg
class="w-3 h-3 text-exo-yellow/60"
fill="none"
viewBox="0 0 24 24"
stroke="currentColor"
>
{params.size}
</button>
<div
class="absolute right-1.5 top-1/2 -translate-y-1/2 pointer-events-none transition-transform duration-200 {isSizeDropdownOpen
? 'rotate-180'
: ''}"
>
<svg
class="w-3 h-3 text-exo-yellow/60"
fill="none"
viewBox="0 0 24 24"
stroke="currentColor"
>
<path
stroke-linecap="round"
stroke-linejoin="round"
stroke-width="2"
d="M19 9l-7 7-7-7"
/>
</svg>
<path
stroke-linecap="round"
stroke-linejoin="round"
stroke-width="2"
d="M19 9l-7 7-7-7"
/>
</svg>
</div>
</div>
{#if isSizeDropdownOpen}
<!-- Backdrop to close dropdown -->
<button
type="button"
class="fixed inset-0 z-[9998] cursor-default"
onclick={() => (isSizeDropdownOpen = false)}
aria-label="Close dropdown"
></button>
<!-- Dropdown Panel - fixed positioning to escape overflow:hidden -->
<div
class="fixed bg-exo-dark-gray border border-exo-yellow/30 rounded shadow-lg shadow-black/50 z-[9999] max-h-48 overflow-y-auto overflow-x-hidden min-w-max"
style="bottom: calc(100vh - {sizeDropdownPosition()
.top}px + 4px); left: {sizeDropdownPosition().left}px;"
>
<div class="py-1">
{#each sizeOptions as size}
<button
type="button"
onclick={() => selectSize(size)}
class="w-full px-3 py-1.5 text-left text-xs font-mono tracking-wide transition-colors duration-100 flex items-center gap-2 {params.size ===
size
? 'bg-transparent text-exo-yellow'
: 'text-exo-light-gray hover:text-exo-yellow'}"
>
{#if params.size === size}
<svg
class="w-3 h-3 flex-shrink-0"
fill="currentColor"
viewBox="0 0 20 20"
>
<path
fill-rule="evenodd"
d="M16.707 5.293a1 1 0 010 1.414l-8 8a1 1 0 01-1.414 0l-4-4a1 1 0 011.414-1.414L8 12.586l7.293-7.293a1 1 0 011.414 0z"
clip-rule="evenodd"
/>
</svg>
{:else}
<span class="w-3"></span>
{/if}
<span>{size.toUpperCase()}</span>
</button>
{/each}
</div>
</div>
{#if isSizeDropdownOpen}
<!-- Backdrop to close dropdown -->
<button
type="button"
class="fixed inset-0 z-[9998] cursor-default"
onclick={() => (isSizeDropdownOpen = false)}
aria-label="Close dropdown"
></button>
<!-- Dropdown Panel - fixed positioning to escape overflow:hidden -->
<div
class="fixed bg-exo-dark-gray border border-exo-yellow/30 rounded shadow-lg shadow-black/50 z-[9999] max-h-48 overflow-y-auto min-w-max"
style="bottom: calc(100vh - {sizeDropdownPosition()
.top}px + 4px); left: {sizeDropdownPosition().left}px;"
>
<div class="py-1">
{#each sizeOptions as size}
<button
type="button"
onclick={() => selectSize(size)}
class="w-full px-3 py-1.5 text-left text-xs font-mono tracking-wide transition-colors duration-100 flex items-center gap-2 {params.size ===
size
? 'bg-transparent text-exo-yellow'
: 'text-exo-light-gray hover:text-exo-yellow'}"
>
{#if params.size === size}
<svg
class="w-3 h-3 flex-shrink-0"
fill="currentColor"
viewBox="0 0 20 20"
>
<path
fill-rule="evenodd"
d="M16.707 5.293a1 1 0 010 1.414l-8 8a1 1 0 01-1.414 0l-4-4a1 1 0 011.414-1.414L8 12.586l7.293-7.293a1 1 0 011.414 0z"
clip-rule="evenodd"
/>
</svg>
{:else}
<span class="w-3"></span>
{/if}
<span>{size}</span>
</button>
{/each}
</div>
</div>
{/if}
</div>
{/if}
{/if}
</div>
<!-- Quality -->
<div class="flex items-center gap-1.5">
@@ -311,7 +310,7 @@
<!-- Dropdown Panel - fixed positioning to escape overflow:hidden -->
<div
class="fixed bg-exo-dark-gray border border-exo-yellow/30 rounded shadow-lg shadow-black/50 z-[9999] max-h-48 overflow-y-auto min-w-max"
class="fixed bg-exo-dark-gray border border-exo-yellow/30 rounded shadow-lg shadow-black/50 z-[9999] max-h-48 overflow-y-auto overflow-x-hidden min-w-max"
style="bottom: calc(100vh - {qualityDropdownPosition()
.top}px + 4px); left: {qualityDropdownPosition().left}px;"
>

View File

@@ -306,13 +306,14 @@ const IMAGE_PARAMS_STORAGE_KEY = "exo-image-generation-params";
export interface ImageGenerationParams {
// Basic params
size:
| "auto"
| "512x512"
| "768x768"
| "1024x1024"
| "1024x768"
| "768x1024"
| "1024x1365"
| "1365x1024";
| "1024x1536"
| "1536x1024";
quality: "low" | "medium" | "high";
outputFormat: "png" | "jpeg";
numImages: number;
@@ -336,7 +337,7 @@ export interface EditingImage {
}
const DEFAULT_IMAGE_PARAMS: ImageGenerationParams = {
size: "1024x1024",
size: "auto",
quality: "medium",
outputFormat: "png",
numImages: 1,

View File

@@ -0,0 +1,28 @@
import { browser } from "$app/environment";
let _isLight = $state(false);
export const theme = {
get isLight() {
return _isLight;
},
init() {
if (!browser) return;
_isLight = document.documentElement.classList.contains("light");
},
toggle() {
if (!browser) return;
_isLight = !_isLight;
if (_isLight) {
document.documentElement.classList.remove("dark");
document.documentElement.classList.add("light");
localStorage.setItem("exo-theme", "light");
} else {
document.documentElement.classList.remove("light");
document.documentElement.classList.add("dark");
localStorage.setItem("exo-theme", "dark");
}
},
};

View File

@@ -1,7 +1,13 @@
<script lang="ts">
import "../app.css";
import { onMount } from "svelte";
import { theme } from "$lib/stores/theme.svelte";
let { children } = $props();
onMount(() => {
theme.init();
});
</script>
<svelte:head>

View File

File diff suppressed because it is too large Load Diff

View File

@@ -115,7 +115,7 @@
packages = lib.optionalAttrs pkgs.stdenv.hostPlatform.isDarwin (
let
uvLock = builtins.fromTOML (builtins.readFile ./uv.lock);
mlxPackage = builtins.head (builtins.filter (p: p.name == "mlx") uvLock.package);
mlxPackage = builtins.head (builtins.filter (p: p.name == "mlx" && p.source ? git) uvLock.package);
uvLockMlxVersion = mlxPackage.version;
in
{

View File

@@ -41,16 +41,16 @@ let
mlx = stdenv.mkDerivation rec {
pname = "mlx";
version = let v = "0.30.6"; in
version = let v = "0.30.7.dev20260218+14841977"; in
assert v == uvLockMlxVersion || throw "MLX version mismatch: nix/mlx.nix has ${v} but uv.lock has ${uvLockMlxVersion}. Update both the version and hash in nix/mlx.nix.";
v;
pyproject = true;
src = fetchFromGitHub {
owner = "ml-explore";
repo = "mlx";
tag = "v${version}";
hash = "sha256-avD5EGhwgmPdXLAyQSqTO6AXk/W3ziH+f6AetjK3Sdo=";
owner = "rltakashige";
repo = "mlx-jaccl-fix-small-recv";
rev = "1484197707f35186ad3bd614357c7c47fdf86ebc";
hash = "sha256-FupCMoK/SF/ldfKuvMSAKECcOP8c+ANgkQlPZttDsLk=";
};
patches = [

View File

@@ -17,9 +17,9 @@ dependencies = [
"loguru>=0.7.3",
"exo_pyo3_bindings", # rust bindings
"anyio==4.11.0",
"mlx==0.30.6; sys_platform == 'darwin'",
"mlx; sys_platform == 'darwin'",
"mlx[cpu]==0.30.6; sys_platform == 'linux'",
"mlx-lm==0.30.6",
"mlx-lm==0.30.7",
"tiktoken>=0.12.0", # required for kimi k2 tokenizer
"hypercorn>=0.18.0",
"openai-harmony>=0.0.8",
@@ -64,6 +64,7 @@ members = [
[tool.uv.sources]
exo_pyo3_bindings = { workspace = true }
mlx = { git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git", branch = "address-rdma-gpu-locks", marker = "sys_platform == 'darwin'" }
#mlx-lm = { git = "https://github.com/davidmcc73/mlx-lm", branch = "stable" }
# Uncomment to use local mlx/mlx-lm development versions:
# mlx = { path = "/Users/Shared/mlx", editable=true }
@@ -132,7 +133,7 @@ markers = [
env = [
"EXO_TESTS=1"
]
addopts = "-m 'not slow'"
addopts = "-m 'not slow' --ignore=tests/start_distributed_test.py"
filterwarnings = [
"ignore:builtin type Swig:DeprecationWarning",
]

View File

@@ -58,6 +58,21 @@
lib.optionalAttrs pkgs.stdenv.hostPlatform.isLinux (
(lib.mapAttrs (_: ignoreMissing) nvidiaPackages) // {
mlx = ignoreMissing prev.mlx;
mlx-cuda-13 = prev.mlx-cuda-13.overrideAttrs (old: {
buildInputs = (old.buildInputs or [ ]) ++ [
final.nvidia-cublas
final.nvidia-cuda-nvrtc
final.nvidia-cudnn-cu13
final.nvidia-nccl-cu13
];
preFixup = ''
addAutoPatchelfSearchPath ${final.nvidia-cublas}
addAutoPatchelfSearchPath ${final.nvidia-cuda-nvrtc}
addAutoPatchelfSearchPath ${final.nvidia-cudnn-cu13}
addAutoPatchelfSearchPath ${final.nvidia-nccl-cu13}
'';
autoPatchelfIgnoreMissingDeps = [ "libcuda.so.1" ];
});
torch = ignoreMissing prev.torch;
triton = ignoreMissing prev.triton;
}
@@ -74,14 +89,25 @@
linuxOverlay
]
);
exoVenv = pythonSet.mkVirtualEnv "exo-env" workspace.deps.default;
# mlx-cpu and mlx-cuda-13 both ship mlx/ site-packages files; keep first.
# mlx-cpu/mlx-cuda-13 and nvidia-cudnn-cu12/cu13 ship overlapping files.
venvCollisionPaths = lib.optionals pkgs.stdenv.hostPlatform.isLinux [
"lib/python3.13/site-packages/mlx*"
"lib/python3.13/site-packages/nvidia*"
];
exoVenv = (pythonSet.mkVirtualEnv "exo-env" workspace.deps.default).overrideAttrs {
venvIgnoreCollisions = venvCollisionPaths;
};
# Virtual environment with dev dependencies for testing
testVenv = pythonSet.mkVirtualEnv "exo-test-env" (
testVenv = (pythonSet.mkVirtualEnv "exo-test-env" (
workspace.deps.default // {
exo = [ "dev" ]; # Include pytest, pytest-asyncio, pytest-env
}
);
)).overrideAttrs {
venvIgnoreCollisions = venvCollisionPaths;
};
mkPythonScript = name: path: pkgs.writeShellApplication {
inherit name;

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "deepseek"
quantization = "4bit"
base_model = "DeepSeek V3.1"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 405874409472

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "deepseek"
quantization = "8bit"
base_model = "DeepSeek V3.1"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 765577920512

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "glm"
quantization = "8bit"
base_model = "GLM 4.5 Air"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 122406567936

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "glm"
quantization = "bf16"
base_model = "GLM 4.5 Air"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 229780750336

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "glm"
quantization = "4bit"
base_model = "GLM 4.7"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 198556925568

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "glm"
quantization = "6bit"
base_model = "GLM 4.7"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 286737579648

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "glm"
quantization = "8bit"
base_model = "GLM 4.7"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 396963397248

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "glm"
quantization = "4bit"
base_model = "GLM 4.7 Flash"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 19327352832

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "glm"
quantization = "5bit"
base_model = "GLM 4.7 Flash"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 22548578304

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "glm"
quantization = "6bit"
base_model = "GLM 4.7 Flash"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 26843545600

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "glm"
quantization = "8bit"
base_model = "GLM 4.7 Flash"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 34359738368

View File

@@ -0,0 +1,12 @@
model_id = "mlx-community/GLM-5-8bit-MXFP8"
n_layers = 78
hidden_size = 6144
supports_tensor = true
tasks = ["TextGeneration"]
family = "glm"
quantization = "8bit"
base_model = "GLM-5"
capabilities = ["text", "thinking"]
[storage_size]
in_bytes = 790517400864

View File

@@ -0,0 +1,12 @@
model_id = "mlx-community/GLM-5-MXFP4-Q8"
n_layers = 78
hidden_size = 6144
supports_tensor = true
tasks = ["TextGeneration"]
family = "glm"
quantization = "MXFP4-Q8"
base_model = "GLM-5"
capabilities = ["text", "thinking"]
[storage_size]
in_bytes = 405478939008

View File

@@ -0,0 +1,12 @@
model_id = "mlx-community/GLM-5"
n_layers = 78
hidden_size = 6144
supports_tensor = true
tasks = ["TextGeneration"]
family = "glm"
quantization = "bf16"
base_model = "GLM-5"
capabilities = ["text", "thinking"]
[storage_size]
in_bytes = 1487822475264

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "kimi"
quantization = ""
base_model = "Kimi K2"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 706522120192

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "kimi"
quantization = ""
base_model = "Kimi K2.5"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 662498705408

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "minimax"
quantization = "3bit"
base_model = "MiniMax M2.1"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 100086644736

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "minimax"
quantization = "8bit"
base_model = "MiniMax M2.1"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 242986745856

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "qwen"
quantization = "4bit"
base_model = "Qwen3 0.6B"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 342884352

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "qwen"
quantization = "8bit"
base_model = "Qwen3 0.6B"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 698351616

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "qwen"
quantization = "4bit"
base_model = "Qwen3 235B"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 141733920768

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "qwen"
quantization = "8bit"
base_model = "Qwen3 235B"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 268435456000

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "qwen"
quantization = "4bit"
base_model = "Qwen3 30B"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 17612931072

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "qwen"
quantization = "8bit"
base_model = "Qwen3 30B"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 33279705088

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "qwen"
quantization = "4bit"
base_model = "Qwen3 Next 80B"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 47080074240

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "qwen"
quantization = "8bit"
base_model = "Qwen3 Next 80B"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 88814387200

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "step"
quantization = "4bit"
base_model = "Step 3.5 Flash"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 114572190076

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "step"
quantization = "6bit"
base_model = "Step 3.5 Flash"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 159039627774

View File

@@ -6,7 +6,7 @@ tasks = ["TextGeneration"]
family = "step"
quantization = "8bit"
base_model = "Step 3.5 Flash"
capabilities = ["text", "thinking"]
capabilities = ["text", "thinking", "thinking_toggle"]
[storage_size]
in_bytes = 209082699847

2
rust/clippy.toml Normal file
View File

@@ -0,0 +1,2 @@
# we can manually exclude false-positive lint errors for dual packages (if in dependencies)
#allowed-duplicate-crates = ["hashbrown"]

View File

@@ -26,11 +26,11 @@ networking = { workspace = true }
# interop
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)
# "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
@@ -42,18 +42,21 @@ 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 }
pin-project = { workspace = true }
# async runtime
tokio = { workspace = true, features = ["full", "tracing"] }
futures-lite = "2.6.1"
futures = { workspace = true }
# utility dependencies
util = { workspace = true }
# Tracing
#tracing = "0.1"
#tracing-subscriber = "0.3"
#console-subscriber = "0.1.5"
#tracing-log = "0.2.0"
log = { workspace = true }
env_logger = "0.11"
# Networking
libp2p = { workspace = true, features = ["full"] }

View File

@@ -2,39 +2,220 @@
# ruff: noqa: E501, F401
import builtins
import enum
import typing
@typing.final
class Keypair:
@staticmethod
def generate() -> Keypair:
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"""
Generate a new ed25519 keypair
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:
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.
"""
@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 to a protobuf structure.
Encode a private key as protobuf structure.
"""
def to_peer_id(self) -> PeerId:
r"""
Convert the `Keypair` into the corresponding `PeerId`.
"""
def to_string(self) -> builtins.str: ...
@typing.final
class PyPeer:
class Multiaddr:
r"""
Representation of a Multiaddr.
"""
@staticmethod
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: ...
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.
"""
@typing.final
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]]: ...
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 = ...

View File

@@ -1,22 +1,38 @@
//! See: <https://pyo3.rs/v0.27.2/async-await.html#detaching-from-the-interpreter-across-await>
//! 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;
use pyo3::prelude::*;
use std::{
future::Future,
pin::{Pin, pin},
pin::Pin,
task::{Context, Poll},
};
pub struct AllowThreads<F>(pub(crate) F);
/// 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)
}
}
impl<F> Future for AllowThreads<F>
where
F: Future + Unpin + Send,
F::Output: Send,
F: Future + Ungil,
F::Output: Ungil,
{
type Output = F::Output;
fn poll(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Self::Output> {
fn poll(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Self::Output> {
let waker = cx.waker();
Python::attach(|py| py.detach(|| pin!(&mut self.0).poll(&mut Context::from_waker(waker))))
Python::attach(|py| py.detach(|| self.project().0.poll(&mut Context::from_waker(waker))))
}
}

View File

@@ -1,42 +1,191 @@
//! TODO: crate documentation
pub(crate) mod allow_threading;
//!
//! this is here as a placeholder documentation
//!
//!
// 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;
pub(crate) mod networking;
pub(crate) mod take_once {
use std::sync::Mutex;
pub(crate) mod pylibp2p;
pub struct TakeOnce<T>(Mutex<Option<T>>);
impl<T> TakeOnce<T> {
pub fn new(t: T) -> Self {
Self(Mutex::new(Some(t)))
use crate::networking::networking_submodule;
use crate::pylibp2p::ident::ident_submodule;
use crate::pylibp2p::multiaddr::multiaddr_submodule;
use pyo3::prelude::PyModule;
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::marker::Tuple;
pub trait SendFn<Args: Tuple + Send + 'static, Output> =
Fn<Args, Output = Output> + Send + 'static;
}
/// 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::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::attach(|py| PyBytes::new(py, self).unbind())
}
pub fn take(&self) -> Option<T> {
match self.0.try_lock() {
Ok(mut o) => o.take(),
Err(_) => None,
}
#[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::attach(|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()),
}
}
}
}
use pyo3::prelude::*;
use pyo3_stub_gen::define_stub_info_gatherer;
/// 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")]
pub fn networking_module(m: &Bound<'_, PyModule>) -> PyResult<()> {
fn main_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);
m.add_class::<networking::PyPeer>()?;
m.add_class::<networking::PyKeypair>()?;
// 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: ...
Ok(())
}

View File

@@ -1,214 +1,571 @@
use crate::allow_threading::AllowThreads;
use crate::take_once::TakeOnce;
#![allow(
clippy::multiple_inherent_impl,
clippy::unnecessary_wraps,
clippy::unused_self,
clippy::needless_pass_by_value
)]
use std::pin::pin;
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;
use libp2p::gossipsub::{IdentTopic, Message, MessageId, PublishError};
use libp2p::swarm::SwarmEvent;
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 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};
mod exception {
use pyo3::types::PyTuple;
use pyo3::{exceptions::PyException, prelude::*};
use pyo3_stub_gen::derive::*;
#[gen_stub_pyclass]
#[pyclass(name = "Keypair", frozen)]
#[derive(Clone)]
pub struct PyKeypair(Keypair);
#[gen_stub_pyclass]
#[pyclass(frozen, extends=PyException, name="NoPeersSubscribedToTopicError")]
pub struct PyNoPeersSubscribedToTopicError {}
#[gen_stub_pymethods]
#[pymethods]
impl PyKeypair {
/// Generate a new ed25519 keypair
#[staticmethod]
fn generate() -> Self {
Self(Keypair::generate_ed25519())
}
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.";
/// 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"))
}
/// 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())),
/// 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???
}
}
fn to_string(&self) -> String {
self.0.public().to_peer_id().to_base58()
#[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()
}
}
}
struct PeerBuilder(
String,
Keypair,
mpsc::Sender<FromSwarm>,
mpsc::Receiver<ToSwarm>,
);
/// 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,
}
#[gen_stub_pyclass]
#[pyclass]
pub struct PyPeer {
peer: TakeOnce<PeerBuilder>,
to_swarm: mpsc::Sender<ToSwarm>,
from_swarm: Mutex<mpsc::Receiver<FromSwarm>>,
#[pyclass(frozen, name = "ConnectionUpdate")]
#[derive(Debug, Clone)]
struct PyConnectionUpdate {
/// Whether this is a connection or disconnection event
#[pyo3(get)]
update_type: PyConnectionUpdateType,
/// Identity of the peer that we have connected to or disconnected from.
#[pyo3(get)]
peer_id: PyPeerId,
/// Remote connection's IPv4 address.
#[pyo3(get)]
remote_ipv4: String,
/// 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 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:?}");
}
}
}
}
}
log::info!("RUST: networking task stopped");
}
#[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.")
}
}
#[gen_stub_pymethods]
#[pymethods]
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),
})
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,
))
}
#[gen_stub(skip)]
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"))
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.to_task_tx = None; // Using Option<T> as a trick to force channel to be dropped
}
// ---- 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,
})
.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,
}
.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())?
}
async fn subscribe(&self, topic: String) -> PyResult<()> {
self.to_swarm
.send(ToSwarm::Subscribe(topic))
/// 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(|_| 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"))
.map_err(|_| PyErr::receiver_channel_closed())
}
#[gen_stub(skip)]
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
})
))
/// 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::attach(|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
{
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)),
};
}
.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()
// }
}
// 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: ...
"#
}
}
pub fn networking_submodule(m: &Bound<'_, PyModule>) -> PyResult<()> {
m.add_class::<exception::PyNoPeersSubscribedToTopicError>()?;
m.add_class::<exception::PyAllQueuesFullError>()?;
#[gen_stub_pyclass]
#[pyclass]
pub struct PySwarmEvent(FromSwarm);
m.add_class::<PyConnectionUpdateType>()?;
m.add_class::<PyConnectionUpdate>()?;
m.add_class::<PyConnectionUpdateType>()?;
m.add_class::<PyNetworkingHandle>()?;
#[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
}
}
Ok(())
}

View File

@@ -0,0 +1,159 @@
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(())
}

View File

@@ -0,0 +1,8 @@
//! 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;

View File

@@ -0,0 +1,81 @@
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(())
}

View File

@@ -0,0 +1,54 @@
#[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;
}
}

View File

@@ -13,11 +13,22 @@ path = "src/lib.rs"
workspace = true
[dependencies]
# datastructures
either = { workspace = true }
# macro dependencies
extend = { workspace = true }
delegate = { workspace = true }
# async
tokio = { workspace = true, features = ["full"] }
futures = { workspace = true }
futures-timer = { workspace = true }
# utility dependencies
util = { workspace = true }
tracing-subscriber = { version = "0.3.19", features = ["default", "env-filter"] }
keccak-const = { workspace = true }
# tracing/logging
log = { workspace = true }

View File

@@ -1,6 +1,6 @@
use libp2p::identity;
use networking::{self, FromSwarm, ToSwarm};
use tokio::sync::mpsc;
use futures::stream::StreamExt as _;
use libp2p::{gossipsub, identity, swarm::SwarmEvent};
use networking::{discovery, swarm};
use tokio::{io, io::AsyncBufReadExt as _, select};
use tracing_subscriber::EnvFilter;
use tracing_subscriber::filter::LevelFilter;
@@ -12,51 +12,63 @@ async fn main() {
.try_init();
// Configure swarm
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");
let mut swarm =
swarm::create_swarm(identity::Keypair::generate_ed25519()).expect("Swarm creation failed");
// 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() => {
_ = to_swarm.send(ToSwarm::Message("chatting".to_string(), line.into_bytes())).await;
if let Err(e) = swarm
.behaviour_mut().gossipsub
.publish(topic.clone(), line.as_bytes()) {
println!("Publish error: {e:?}");
}
}
event = from_swarm.recv() => match event {
event = swarm.select_next_some() => match event {
// on gossipsub incoming
Some(FromSwarm::Message(peer_id,_, data)) => println!(
"\n\nGot message: '{}' from peer: {peer_id}\n\n",
String::from_utf8_lossy(&data),
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),
),
// on discovery
Some(FromSwarm::Discovered(peer_id)) => {
println!("\n\nConnected to: {peer_id}\n\n");
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::Expired(peer_id)) => {
println!("\n\nDisconnected from: {peer_id}\n\n");
}
Some(FromSwarm::PublishError(e)) => eprintln!("\n\nError {e:?}\n\n"),
None => break,
// 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:?}"); }
}
}
}
_ = jh.await;
}

View File

@@ -0,0 +1,127 @@
// 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::error::Error;
use std::time::Duration;
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);
}
}
}
}
}

View File

@@ -0,0 +1,44 @@
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

View File

@@ -0,0 +1,382 @@
use crate::ext::MultiaddrExt;
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
}
}

View File

@@ -1,299 +1,44 @@
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;
//! TODO: crate documentation
//!
//! this is here as a placeholder documentation
//!
//!
pub mod discovery;
pub mod swarm;
#[derive(Debug)]
pub struct ListenError;
/// Namespace for all the type/trait aliases used by this crate.
pub(crate) mod alias {
use std::error::Error;
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),
pub type AnyError = Box<dyn Error + Send + Sync + 'static>;
pub type AnyResult<T> = Result<T, AnyError>;
}
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();
/// 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;
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(|_| ())?;
#[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,
}
}
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,
} else {
return None;
};
let Some(Protocol::Tcp(port)) = ps.next() else {
return None;
};
Some((ip, port))
}
}
}
// TODO: more tests
#[cfg(test)]
mod tests {
use super::*;
use tokio::time::{Duration, timeout};
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;
}
}

View File

@@ -0,0 +1,143 @@
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")
}
}

View File

@@ -0,0 +1,7 @@
// maybe this will hold test in the future...??
#[cfg(test)]
mod tests {
#[test]
fn does_nothing() {}
}

2
rust/rust-toolchain.toml Normal file
View File

@@ -0,0 +1,2 @@
[toolchain]
channel = "nightly"

15
rust/util/Cargo.toml Normal file
View File

@@ -0,0 +1,15 @@
[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
rust/util/src/lib.rs Normal file
View File

@@ -0,0 +1 @@
pub mod wakerdeque;

View File

@@ -0,0 +1,55 @@
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);
}
}

View File

@@ -14,6 +14,7 @@ from exo.download.download_utils import (
map_repo_download_progress_to_download_progress_data,
)
from exo.download.shard_downloader import ShardDownloader
from exo.shared.constants import EXO_MODELS_DIR
from exo.shared.models.model_cards import ModelId
from exo.shared.types.commands import (
CancelDownload,
@@ -46,6 +47,7 @@ class DownloadCoordinator:
download_command_receiver: Receiver[ForwarderDownloadCommand]
local_event_sender: Sender[ForwarderEvent]
event_index_counter: Iterator[int]
offline: bool = False
# Local state
download_status: dict[ModelId, DownloadProgress] = field(default_factory=dict)
@@ -61,8 +63,13 @@ class DownloadCoordinator:
def __post_init__(self) -> None:
self.event_sender, self.event_receiver = channel[Event]()
if self.offline:
self.shard_downloader.set_internet_connection(False)
self.shard_downloader.on_progress(self._download_progress_callback)
def _model_dir(self, model_id: ModelId) -> str:
return str(EXO_MODELS_DIR / model_id.normalize())
async def _download_progress_callback(
self, callback_shard: ShardMetadata, progress: RepoDownloadProgress
) -> None:
@@ -74,6 +81,7 @@ class DownloadCoordinator:
shard_metadata=callback_shard,
node_id=self.node_id,
total_bytes=progress.total_bytes,
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = completed
await self.event_sender.send(
@@ -93,6 +101,7 @@ class DownloadCoordinator:
download_progress=map_repo_download_progress_to_download_progress_data(
progress
),
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = ongoing
await self.event_sender.send(
@@ -101,13 +110,17 @@ class DownloadCoordinator:
self._last_progress_time[model_id] = current_time()
async def run(self) -> None:
logger.info("Starting DownloadCoordinator")
self._test_internet_connection()
logger.info(
f"Starting DownloadCoordinator{' (offline mode)' if self.offline else ''}"
)
if not self.offline:
self._test_internet_connection()
async with self._tg as tg:
tg.start_soon(self._command_processor)
tg.start_soon(self._forward_events)
tg.start_soon(self._emit_existing_download_progress)
tg.start_soon(self._check_internet_connection)
if not self.offline:
tg.start_soon(self._check_internet_connection)
def _test_internet_connection(self) -> None:
try:
@@ -170,7 +183,11 @@ class DownloadCoordinator:
return
# Emit pending status
progress = DownloadPending(shard_metadata=shard, node_id=self.node_id)
progress = DownloadPending(
shard_metadata=shard,
node_id=self.node_id,
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = progress
await self.event_sender.send(NodeDownloadProgress(download_progress=progress))
@@ -184,6 +201,7 @@ class DownloadCoordinator:
shard_metadata=shard,
node_id=self.node_id,
total_bytes=initial_progress.total_bytes,
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = completed
await self.event_sender.send(
@@ -191,6 +209,20 @@ class DownloadCoordinator:
)
return
if self.offline:
logger.warning(
f"Offline mode: model {model_id} is not fully available locally, cannot download"
)
failed = DownloadFailed(
shard_metadata=shard,
node_id=self.node_id,
error_message=f"Model files not found locally in offline mode: {model_id}",
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = failed
await self.event_sender.send(NodeDownloadProgress(download_progress=failed))
return
# Start actual download
self._start_download_task(shard, initial_progress)
@@ -206,6 +238,7 @@ class DownloadCoordinator:
download_progress=map_repo_download_progress_to_download_progress_data(
initial_progress
),
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = status
self.event_sender.send_nowait(NodeDownloadProgress(download_progress=status))
@@ -219,6 +252,7 @@ class DownloadCoordinator:
shard_metadata=shard,
node_id=self.node_id,
error_message=str(e),
model_directory=self._model_dir(model_id),
)
self.download_status[model_id] = failed
await self.event_sender.send(
@@ -253,6 +287,7 @@ class DownloadCoordinator:
pending = DownloadPending(
shard_metadata=current_status.shard_metadata,
node_id=self.node_id,
model_directory=self._model_dir(model_id),
)
await self.event_sender.send(
NodeDownloadProgress(download_progress=pending)
@@ -295,11 +330,18 @@ class DownloadCoordinator:
node_id=self.node_id,
shard_metadata=progress.shard,
total_bytes=progress.total_bytes,
model_directory=self._model_dir(
progress.shard.model_card.model_id
),
)
elif progress.status in ["in_progress", "not_started"]:
if progress.downloaded_bytes_this_session.in_bytes == 0:
status = DownloadPending(
node_id=self.node_id, shard_metadata=progress.shard
node_id=self.node_id,
shard_metadata=progress.shard,
model_directory=self._model_dir(
progress.shard.model_card.model_id
),
)
else:
status = DownloadOngoing(
@@ -308,6 +350,9 @@ class DownloadCoordinator:
download_progress=map_repo_download_progress_to_download_progress_data(
progress
),
model_directory=self._model_dir(
progress.shard.model_card.model_id
),
)
else:
continue

View File

@@ -448,12 +448,13 @@ async def download_file_with_retry(
target_dir: Path,
on_progress: Callable[[int, int, bool], None] = lambda _, __, ___: None,
on_connection_lost: Callable[[], None] = lambda: None,
skip_internet: bool = False,
) -> Path:
n_attempts = 3
for attempt in range(n_attempts):
try:
return await _download_file(
model_id, revision, path, target_dir, on_progress
model_id, revision, path, target_dir, on_progress, skip_internet
)
except HuggingFaceAuthenticationError:
raise
@@ -487,10 +488,14 @@ async def _download_file(
path: str,
target_dir: Path,
on_progress: Callable[[int, int, bool], None] = lambda _, __, ___: None,
skip_internet: bool = False,
) -> Path:
target_path = target_dir / path
if await aios.path.exists(target_path):
if skip_internet:
return target_path
local_size = (await aios.stat(target_path)).st_size
# Try to verify against remote, but allow offline operation
@@ -510,6 +515,11 @@ async def _download_file(
)
return target_path
if skip_internet:
raise FileNotFoundError(
f"File {path} not found locally and cannot download in offline mode"
)
await aios.makedirs((target_dir / path).parent, exist_ok=True)
length, etag = await file_meta(model_id, revision, path)
remote_hash = etag[:-5] if etag.endswith("-gzip") else etag
@@ -814,6 +824,7 @@ async def download_shard(
file, curr_bytes, total_bytes, is_renamed
),
on_connection_lost=on_connection_lost,
skip_internet=skip_internet,
)
if not skip_download:

View File

@@ -0,0 +1,230 @@
"""Tests for offline/air-gapped mode."""
from collections.abc import AsyncIterator
from pathlib import Path
from unittest.mock import AsyncMock, patch
import aiofiles
import aiofiles.os as aios
import pytest
from exo.download.download_utils import (
_download_file, # pyright: ignore[reportPrivateUsage]
download_file_with_retry,
fetch_file_list_with_cache,
)
from exo.shared.types.common import ModelId
from exo.shared.types.worker.downloads import FileListEntry
@pytest.fixture
def model_id() -> ModelId:
return ModelId("test-org/test-model")
@pytest.fixture
async def temp_models_dir(tmp_path: Path) -> AsyncIterator[Path]:
models_dir = tmp_path / "models"
await aios.makedirs(models_dir, exist_ok=True)
with patch("exo.download.download_utils.EXO_MODELS_DIR", models_dir):
yield models_dir
class TestDownloadFileOffline:
"""Tests for _download_file with skip_internet=True."""
async def test_returns_local_file_without_http_verification(
self, model_id: ModelId, tmp_path: Path
) -> None:
"""When skip_internet=True and file exists locally, return it immediately
without making any HTTP calls (no file_meta verification)."""
target_dir = tmp_path / "downloads"
await aios.makedirs(target_dir, exist_ok=True)
local_file = target_dir / "model.safetensors"
async with aiofiles.open(local_file, "wb") as f:
await f.write(b"model weights data")
with patch(
"exo.download.download_utils.file_meta",
new_callable=AsyncMock,
) as mock_file_meta:
result = await _download_file(
model_id,
"main",
"model.safetensors",
target_dir,
skip_internet=True,
)
assert result == local_file
mock_file_meta.assert_not_called()
async def test_raises_file_not_found_for_missing_file(
self, model_id: ModelId, tmp_path: Path
) -> None:
"""When skip_internet=True and file does NOT exist locally,
raise FileNotFoundError instead of attempting download."""
target_dir = tmp_path / "downloads"
await aios.makedirs(target_dir, exist_ok=True)
with pytest.raises(FileNotFoundError, match="offline mode"):
await _download_file(
model_id,
"main",
"missing_model.safetensors",
target_dir,
skip_internet=True,
)
async def test_returns_local_file_in_subdirectory(
self, model_id: ModelId, tmp_path: Path
) -> None:
"""When skip_internet=True and file exists in a subdirectory,
return it without HTTP calls."""
target_dir = tmp_path / "downloads"
subdir = target_dir / "transformer"
await aios.makedirs(subdir, exist_ok=True)
local_file = subdir / "diffusion_pytorch_model.safetensors"
async with aiofiles.open(local_file, "wb") as f:
await f.write(b"weights")
with patch(
"exo.download.download_utils.file_meta",
new_callable=AsyncMock,
) as mock_file_meta:
result = await _download_file(
model_id,
"main",
"transformer/diffusion_pytorch_model.safetensors",
target_dir,
skip_internet=True,
)
assert result == local_file
mock_file_meta.assert_not_called()
class TestDownloadFileWithRetryOffline:
"""Tests for download_file_with_retry with skip_internet=True."""
async def test_propagates_skip_internet_to_download_file(
self, model_id: ModelId, tmp_path: Path
) -> None:
"""Verify skip_internet is passed through to _download_file."""
target_dir = tmp_path / "downloads"
await aios.makedirs(target_dir, exist_ok=True)
local_file = target_dir / "config.json"
async with aiofiles.open(local_file, "wb") as f:
await f.write(b'{"model_type": "qwen2"}')
with patch(
"exo.download.download_utils.file_meta",
new_callable=AsyncMock,
) as mock_file_meta:
result = await download_file_with_retry(
model_id,
"main",
"config.json",
target_dir,
skip_internet=True,
)
assert result == local_file
mock_file_meta.assert_not_called()
async def test_file_not_found_does_not_retry(
self, model_id: ModelId, tmp_path: Path
) -> None:
"""FileNotFoundError from offline mode should not trigger retries."""
target_dir = tmp_path / "downloads"
await aios.makedirs(target_dir, exist_ok=True)
with pytest.raises(FileNotFoundError):
await download_file_with_retry(
model_id,
"main",
"nonexistent.safetensors",
target_dir,
skip_internet=True,
)
class TestFetchFileListOffline:
"""Tests for fetch_file_list_with_cache with skip_internet=True."""
async def test_uses_cached_file_list(
self, model_id: ModelId, temp_models_dir: Path
) -> None:
"""When skip_internet=True and cache file exists, use it without network."""
from pydantic import TypeAdapter
cache_dir = temp_models_dir / "caches" / model_id.normalize()
await aios.makedirs(cache_dir, exist_ok=True)
cached_list = [
FileListEntry(type="file", path="model.safetensors", size=1000),
FileListEntry(type="file", path="config.json", size=200),
]
cache_file = cache_dir / f"{model_id.normalize()}--main--file_list.json"
async with aiofiles.open(cache_file, "w") as f:
await f.write(
TypeAdapter(list[FileListEntry]).dump_json(cached_list).decode()
)
with patch(
"exo.download.download_utils.fetch_file_list_with_retry",
new_callable=AsyncMock,
) as mock_fetch:
result = await fetch_file_list_with_cache(
model_id, "main", skip_internet=True
)
assert result == cached_list
mock_fetch.assert_not_called()
async def test_falls_back_to_local_directory_scan(
self, model_id: ModelId, temp_models_dir: Path
) -> None:
"""When skip_internet=True and no cache but local files exist,
build file list from local directory."""
import json
model_dir = temp_models_dir / model_id.normalize()
await aios.makedirs(model_dir, exist_ok=True)
async with aiofiles.open(model_dir / "config.json", "w") as f:
await f.write('{"model_type": "qwen2"}')
index_data = {
"metadata": {},
"weight_map": {"model.layers.0.weight": "model.safetensors"},
}
async with aiofiles.open(model_dir / "model.safetensors.index.json", "w") as f:
await f.write(json.dumps(index_data))
async with aiofiles.open(model_dir / "model.safetensors", "wb") as f:
await f.write(b"x" * 500)
with patch(
"exo.download.download_utils.fetch_file_list_with_retry",
new_callable=AsyncMock,
) as mock_fetch:
result = await fetch_file_list_with_cache(
model_id, "main", skip_internet=True
)
mock_fetch.assert_not_called()
paths = {entry.path for entry in result}
assert "config.json" in paths
assert "model.safetensors" in paths
async def test_raises_when_no_cache_and_no_local_files(
self, model_id: ModelId, temp_models_dir: Path
) -> None:
"""When skip_internet=True and neither cache nor local files exist,
raise FileNotFoundError."""
with pytest.raises(FileNotFoundError, match="No internet"):
await fetch_file_list_with_cache(model_id, "main", skip_internet=True)

View File

@@ -1,5 +1,4 @@
import argparse
import importlib.metadata
import itertools
import multiprocessing as mp
import os
@@ -40,14 +39,15 @@ class Node:
node_id: NodeId
event_index_counter: Iterator[int]
offline: bool
_tg: TaskGroup = field(init=False, default_factory=anyio.create_task_group)
@classmethod
async def create(cls, args: "Args") -> "Self":
keypair = get_node_id_keypair()
node_id = NodeId(keypair.to_string())
node_id = NodeId(keypair.to_peer_id().to_base58())
session_id = SessionId(master_node_id=node_id, election_clock=0)
router = Router.create(keypair, namespace=args.namespace)
router = Router.create(keypair)
await router.register_topic(topics.GLOBAL_EVENTS)
await router.register_topic(topics.LOCAL_EVENTS)
await router.register_topic(topics.COMMANDS)
@@ -69,11 +69,12 @@ class Node:
download_command_receiver=router.receiver(topics.DOWNLOAD_COMMANDS),
local_event_sender=router.sender(topics.LOCAL_EVENTS),
event_index_counter=event_index_counter,
offline=args.offline,
)
else:
download_coordinator = None
if not args.no_api:
if args.spawn_api:
api = API(
node_id,
session_id,
@@ -133,10 +134,13 @@ class Node:
api,
node_id,
event_index_counter,
args.offline,
)
async def run(self):
async with self._tg as tg:
signal.signal(signal.SIGINT, lambda _, __: self.shutdown())
signal.signal(signal.SIGTERM, lambda _, __: self.shutdown())
tg.start_soon(self.router.run)
tg.start_soon(self.election.run)
if self.download_coordinator:
@@ -148,8 +152,6 @@ class Node:
if self.api:
tg.start_soon(self.api.run)
tg.start_soon(self._elect_loop)
signal.signal(signal.SIGINT, lambda _, __: self.shutdown())
signal.signal(signal.SIGTERM, lambda _, __: self.shutdown())
def shutdown(self):
# if this is our second call to shutdown, just sys.exit
@@ -223,6 +225,7 @@ class Node:
),
local_event_sender=self.router.sender(topics.LOCAL_EVENTS),
event_index_counter=self.event_index_counter,
offline=self.offline,
)
self._tg.start_soon(self.download_coordinator.run)
if self.worker:
@@ -259,7 +262,10 @@ def main():
# TODO: Refactor the current verbosity system
logger_setup(EXO_LOG, args.verbosity)
logger.info("Starting EXO")
logger.info(f"Namespace: {args.namespace}")
logger.info(f"EXO_LIBP2P_NAMESPACE: {os.getenv('EXO_LIBP2P_NAMESPACE')}")
if args.offline:
logger.info("Running in OFFLINE mode — no internet checks, local models only")
# Set FAST_SYNCH override env var for runner subprocesses
if args.fast_synch is True:
@@ -276,13 +282,14 @@ def main():
class Args(CamelCaseModel):
verbosity: int
force_master: bool
no_api: bool
api_port: PositiveInt
verbosity: int = 0
force_master: bool = False
spawn_api: bool = False
api_port: PositiveInt = 52415
tb_only: bool = False
no_worker: bool = False
no_downloads: bool = False
namespace: str
offline: bool = False
fast_synch: bool | None = None # None = auto, True = force on, False = force off
@classmethod
@@ -312,15 +319,14 @@ class Args(CamelCaseModel):
)
parser.add_argument(
"--no-api",
action="store_true",
help="Disable the API server for this node",
action="store_false",
dest="spawn_api",
)
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",
@@ -332,9 +338,9 @@ class Args(CamelCaseModel):
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",
"--offline",
action="store_true",
help="Run in offline/air-gapped mode: skip internet checks, use only pre-staged local models",
)
fast_synch_group = parser.add_mutually_exclusive_group()
fast_synch_group.add_argument(

View File

@@ -85,6 +85,7 @@ from exo.shared.types.api import (
ImageGenerationTaskParams,
ImageListItem,
ImageListResponse,
ImageSize,
ModelList,
ModelListModel,
PlaceInstanceParams,
@@ -100,6 +101,7 @@ from exo.shared.types.api import (
TraceRankStats,
TraceResponse,
TraceStatsResponse,
normalize_image_size,
)
from exo.shared.types.chunks import (
ErrorChunk,
@@ -751,9 +753,11 @@ class API:
When stream=True and partial_images > 0, returns a StreamingResponse
with SSE-formatted events for partial and final images.
"""
payload.model = await self._validate_image_model(ModelId(payload.model))
payload = payload.model_copy(
update={"advanced_params": _ensure_seed(payload.advanced_params)}
update={
"model": await self._validate_image_model(ModelId(payload.model)),
"advanced_params": _ensure_seed(payload.advanced_params),
}
)
command = ImageGeneration(
@@ -1009,12 +1013,13 @@ class API:
async def bench_image_generations(
self, request: Request, payload: BenchImageGenerationTaskParams
) -> BenchImageGenerationResponse:
payload.model = await self._validate_image_model(ModelId(payload.model))
payload.stream = False
payload.partial_images = 0
payload = payload.model_copy(
update={"advanced_params": _ensure_seed(payload.advanced_params)}
update={
"model": await self._validate_image_model(ModelId(payload.model)),
"stream": False,
"partial_images": 0,
"advanced_params": _ensure_seed(payload.advanced_params),
}
)
command = ImageGeneration(
@@ -1035,7 +1040,7 @@ class API:
prompt: str,
model: ModelId,
n: int,
size: str,
size: ImageSize,
response_format: Literal["url", "b64_json"],
input_fidelity: Literal["low", "high"],
stream: bool,
@@ -1105,7 +1110,7 @@ class API:
prompt: str = Form(...),
model: str = Form(...),
n: int = Form(1),
size: str = Form("1024x1024"),
size: str | None = Form(None),
response_format: Literal["url", "b64_json"] = Form("b64_json"),
input_fidelity: Literal["low", "high"] = Form("low"),
stream: str = Form("false"),
@@ -1131,7 +1136,7 @@ class API:
prompt=prompt,
model=ModelId(model),
n=n,
size=size,
size=normalize_image_size(size),
response_format=response_format,
input_fidelity=input_fidelity,
stream=stream_bool,
@@ -1167,7 +1172,7 @@ class API:
prompt: str = Form(...),
model: str = Form(...),
n: int = Form(1),
size: str = Form("1024x1024"),
size: str | None = Form(None),
response_format: Literal["url", "b64_json"] = Form("b64_json"),
input_fidelity: Literal["low", "high"] = Form("low"),
quality: Literal["high", "medium", "low"] = Form("medium"),
@@ -1187,7 +1192,7 @@ class API:
prompt=prompt,
model=ModelId(model),
n=n,
size=size,
size=normalize_image_size(size),
response_format=response_format,
input_fidelity=input_fidelity,
stream=False,

View File

@@ -396,7 +396,7 @@ class Master:
await self._handle_traces_collected(event)
continue
logger.trace(f"Master indexing event: {str(event)[:100]}")
logger.debug(f"Master indexing event: {str(event)[:100]}")
indexed = IndexedEvent(event=event, idx=len(self._event_log))
self.state = apply(self.state, indexed)

View File

@@ -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_string())
node_id = NodeId(keypair.to_peer_id().to_base58())
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_string()}_sender")
sender_node_id = NodeId(f"{keypair.to_peer_id().to_base58()}_sender")
# inject a NodeGatheredInfo event
logger.info("inject a NodeGatheredInfo event")
await local_event_sender.send(

View File

@@ -1,9 +1,37 @@
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
expired: bool
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,
)

View File

@@ -1,5 +1,5 @@
from copy import copy
from dataclasses import dataclass, field
from itertools import count
from math import inf
from os import PathLike
from pathlib import Path
@@ -14,14 +14,15 @@ from anyio import (
)
from anyio.abc import TaskGroup
from exo_pyo3_bindings import (
AllQueuesFullError,
Keypair,
PyPeer,
NetworkingHandle,
NoPeersSubscribedToTopicError,
)
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
@@ -98,32 +99,28 @@ class TopicRouter[T: CamelCaseModel]:
)
@dataclass
class Router:
_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 __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))
def create(cls, identity: Keypair) -> "Router":
return cls(handle=NetworkingHandle(identity))
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
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))
@@ -151,18 +148,14 @@ 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:
@@ -177,58 +170,47 @@ class Router:
self._tg.cancel_scope.cancel()
async def _networking_subscribe(self, topic: str):
await self._peer.subscribe(topic)
await self._net.gossipsub_subscribe(topic)
logger.info(f"Subscribed to {topic}")
async def _networking_unsubscribe(self, topic: str):
await self._peer.unsubscribe(topic)
await self._net.gossipsub_unsubscribe(topic)
logger.info(f"Unsubscribed from {topic}")
async def _networking_recv(self):
while True:
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}")
topic, data = await self._net.gossipsub_recv()
logger.trace(f"Received message on {topic} with payload {data}")
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(payload)
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)
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._peer.send(topic, data)
except RuntimeError:
break
await self._net.gossipsub_publish(topic, data)
except NoPeersSubscribedToTopicError:
pass
except AllQueuesFullError:
logger.warning(f"All peer queues full, dropping message on {topic}")
def get_node_id_keypair(
@@ -239,7 +221,7 @@ 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()
return Keypair.generate_ed25519()
def lock_path(path: str | bytes | PathLike[str] | PathLike[bytes]) -> Path:
return Path(str(path) + ".lock")

View File

@@ -218,11 +218,6 @@ def apply_node_timed_out(event: NodeTimedOut, state: State) -> State:
key: value for key, value in state.downloads.items() if key != event.node_id
}
# Clean up all granular node mappings
node_identities = {
key: value
for key, value in state.node_identities.items()
if key != event.node_id
}
node_memory = {
key: value for key, value in state.node_memory.items() if key != event.node_id
}
@@ -263,7 +258,6 @@ def apply_node_timed_out(event: NodeTimedOut, state: State) -> State:
"downloads": downloads,
"topology": topology,
"last_seen": last_seen,
"node_identities": node_identities,
"node_memory": node_memory,
"node_disk": node_disk,
"node_system": node_system,

View File

@@ -44,7 +44,8 @@ async def _refresh_card_cache():
async for toml_file in path.rglob("*.toml"):
try:
card = await ModelCard.load_from_path(toml_file)
_card_cache[card.model_id] = card
if card.model_id not in _card_cache:
_card_cache[card.model_id] = card
except (ValidationError, TOMLKitError):
pass
@@ -182,6 +183,7 @@ class ConfigData(BaseModel):
def supports_tensor(self) -> bool:
return self.architectures in [
["Glm4MoeLiteForCausalLM"],
["GlmMoeDsaForCausalLM"],
["DeepseekV32ForCausalLM"],
["DeepseekV3ForCausalLM"],
["Qwen3NextForCausalLM"],

View File

@@ -1,7 +1,7 @@
import pytest
from anyio import create_task_group, fail_after, move_on_after
from exo.routing.connection_message import ConnectionMessage
from exo.routing.connection_message import ConnectionMessage, ConnectionMessageType
from exo.shared.election import Election, ElectionMessage, ElectionResult
from exo.shared.types.commands import ForwarderCommand, TestCommand
from exo.shared.types.common import NodeId, SessionId
@@ -330,7 +330,9 @@ async def test_connection_message_triggers_new_round_broadcast() -> None:
await cm_tx.send(
ConnectionMessage(
node_id=NodeId(),
expired=False,
connection_type=ConnectionMessageType.Connected,
remote_ipv4="",
remote_tcp_port=0,
)
)

View File

@@ -1,9 +1,9 @@
import time
from collections.abc import Generator
from typing import Annotated, Any, Literal
from typing import Annotated, Any, Literal, get_args
from uuid import uuid4
from pydantic import BaseModel, Field
from pydantic import BaseModel, Field, field_validator
from exo.shared.models.model_cards import ModelCard, ModelId
from exo.shared.types.common import CommandId, NodeId
@@ -262,6 +262,27 @@ class DeleteInstanceResponse(BaseModel):
instance_id: InstanceId
ImageSize = Literal[
"auto",
"512x512",
"768x768",
"1024x768",
"768x1024",
"1024x1024",
"1024x1536",
"1536x1024",
]
def normalize_image_size(v: object) -> ImageSize:
"""Shared validator for ImageSize fields: maps None → "auto" and rejects invalid values."""
if v is None:
return "auto"
if v not in get_args(ImageSize):
raise ValueError(f"Invalid size: {v!r}. Must be one of {get_args(ImageSize)}")
return v # pyright: ignore[reportReturnType]
class AdvancedImageParams(BaseModel):
seed: Annotated[int, Field(ge=0)] | None = None
num_inference_steps: Annotated[int, Field(ge=1, le=100)] | None = None
@@ -281,7 +302,7 @@ class ImageGenerationTaskParams(BaseModel):
partial_images: int | None = 0
quality: Literal["high", "medium", "low"] | None = "medium"
response_format: Literal["url", "b64_json"] | None = "b64_json"
size: str | None = "1024x1024"
size: ImageSize = "auto"
stream: bool | None = False
style: str | None = "vivid"
user: str | None = None
@@ -289,6 +310,11 @@ class ImageGenerationTaskParams(BaseModel):
# Internal flag for benchmark mode - set by API, preserved through serialization
bench: bool = False
@field_validator("size", mode="before")
@classmethod
def normalize_size(cls, v: object) -> ImageSize:
return normalize_image_size(v)
class BenchImageGenerationTaskParams(ImageGenerationTaskParams):
bench: bool = True
@@ -305,13 +331,18 @@ class ImageEditsTaskParams(BaseModel):
quality: Literal["high", "medium", "low"] | None = "medium"
output_format: Literal["png", "jpeg", "webp"] = "png"
response_format: Literal["url", "b64_json"] | None = "b64_json"
size: str | None = "1024x1024"
size: ImageSize = "auto"
image_strength: float | None = 0.7
stream: bool = False
partial_images: int | None = 0
advanced_params: AdvancedImageParams | None = None
bench: bool = False
@field_validator("size", mode="before")
@classmethod
def normalize_size(cls, v: object) -> ImageSize:
return normalize_image_size(v)
def __repr_args__(self) -> Generator[tuple[str, Any], None, None]:
for name, value in super().__repr_args__(): # pyright: ignore[reportAny]
if name == "image_data":

View File

@@ -26,6 +26,7 @@ class DownloadProgressData(CamelCaseModel):
class BaseDownloadProgress(TaggedModel):
node_id: NodeId
shard_metadata: ShardMetadata
model_directory: str = ""
class DownloadPending(BaseDownloadProgress):

View File

@@ -14,6 +14,7 @@ from exo.shared.types.api import (
ImageEditsTaskParams,
ImageGenerationStats,
ImageGenerationTaskParams,
ImageSize,
)
from exo.shared.types.memory import Memory
from exo.shared.types.worker.runner_response import (
@@ -23,9 +24,9 @@ from exo.shared.types.worker.runner_response import (
from exo.worker.engines.image.distributed_model import DistributedImageModel
def parse_size(size_str: str | None) -> tuple[int, int]:
def parse_size(size_str: ImageSize) -> tuple[int, int]:
"""Parse size parameter like '1024x1024' to (width, height) tuple."""
if not size_str:
if size_str == "auto":
return (1024, 1024)
try:
@@ -109,6 +110,9 @@ def generate_image(
# Decode base64 image data and save to temp file
image_path = Path(tmpdir) / "input.png"
image_path.write_bytes(base64.b64decode(task.image_data))
if task.size == "auto":
with Image.open(image_path) as img:
width, height = img.size
for image_num in range(num_images):
# Increment seed for each image to ensure unique results

View File

@@ -163,11 +163,14 @@ class PipelineLastLayer(CustomMlxLayer):
output, (self.r + 1) % self.s, group=self.group
)
if cache is not None:
cache.keys = mx.depends(cache.keys, output) # type: ignore[reportUnknownMemberType]
# CacheList (used by MLA models like DeepSeekV32, GLM MoE DSA)
# doesn't have .keys directly; access via first sub-cache.
_cache = cache[0] if hasattr(cache, "caches") else cache # type: ignore
_cache.keys = mx.depends(_cache.keys, output) # type: ignore
if self.is_prefill:
mx.eval(output)
if cache is not None:
mx.eval(cache.keys) # type: ignore
mx.eval(_cache.keys) # type: ignore
if not self.is_prefill:
output = mx.distributed.all_gather(output, group=self.group)[
@@ -307,7 +310,9 @@ def patch_pipeline_model[T](model: T, group: mx.distributed.Group) -> T:
# Add dependency to last cache entry to ensure distributed ops are evaluated
if cache is not None:
cache[-1].state = mx.depends(cache[-1].state, logits) # type: ignore
last = cache[-1] # type: ignore
dep_cache = last[0] if hasattr(last, "caches") else last # type: ignore
dep_cache.keys = mx.depends(dep_cache.keys, logits) # type: ignore
return logits
@@ -333,7 +338,9 @@ def patch_tensor_model[T](model: T) -> T:
# Add dependency to last cache entry to ensure distributed ops are evaluated
if cache is not None and len(cache) > 0: # pyright: ignore[reportAny]
cache[-1].state = mx.depends(cache[-1].state, logits) # pyright: ignore[reportAny,reportUnknownMemberType]
last = cache[-1] # pyright: ignore[reportAny]
dep_cache = last[0] if hasattr(last, "caches") else last # pyright: ignore[reportAny]
dep_cache.keys = mx.depends(dep_cache.keys, logits) # pyright: ignore[reportAny,reportUnknownMemberType]
return logits
@@ -547,10 +554,12 @@ class DeepSeekShardingStrategy(TensorParallelShardingStrategy):
on_timeout: TimeoutCallback | None,
) -> nn.Module:
model = cast(DeepseekV3Model, model)
for layer in model.layers:
eval_with_timeout(
layer.parameters(), timeout_seconds / len(model.layers), on_timeout
)
# Shard the self attention
if layer.self_attn.q_lora_rank is None:
layer.self_attn.q_proj = self.all_to_sharded_linear(
@@ -581,12 +590,18 @@ class DeepSeekShardingStrategy(TensorParallelShardingStrategy):
layer.mlp.down_proj = self.sharded_to_all_linear(layer.mlp.down_proj)
layer.mlp.up_proj = self.all_to_sharded_linear(layer.mlp.up_proj)
# Shard the MoE. Shard in place since the MoE should be responsible
# for aggregating the results.
# Shard the MoE.
else:
self.all_to_sharded_linear_in_place(layer.mlp.shared_experts.gate_proj)
self.sharded_to_all_linear_in_place(layer.mlp.shared_experts.down_proj)
self.all_to_sharded_linear_in_place(layer.mlp.shared_experts.up_proj)
if getattr(layer.mlp, "shared_experts", None) is not None:
self.all_to_sharded_linear_in_place(
layer.mlp.shared_experts.gate_proj
)
self.sharded_to_all_linear_in_place(
layer.mlp.shared_experts.down_proj
)
self.all_to_sharded_linear_in_place(
layer.mlp.shared_experts.up_proj
)
self.all_to_sharded_linear_in_place(layer.mlp.switch_mlp.gate_proj)
self.sharded_to_all_linear_in_place(layer.mlp.switch_mlp.down_proj)
self.all_to_sharded_linear_in_place(layer.mlp.switch_mlp.up_proj)
@@ -779,8 +794,7 @@ class MiniMaxShardingStrategy(TensorParallelShardingStrategy):
layer.self_attn = WrappedMiniMaxAttention(layer.self_attn, self.group) # pyright: ignore[reportAttributeAccessIssue,reportArgumentType]
# Shard the MoE. Shard in place since the MoE should be responsible
# for aggregating the results.
# Shard the MoE.
self.all_to_sharded_linear_in_place(
layer.block_sparse_moe.switch_mlp.gate_proj
)
@@ -893,8 +907,7 @@ class QwenShardingStrategy(TensorParallelShardingStrategy):
layer.self_attn.num_attention_heads //= self.N
layer.self_attn.num_key_value_heads //= self.N
# Shard the MoE. Shard in place since the MoE should be responsible
# for aggregating the results.
# Shard the MoE.
if isinstance(layer.mlp, (Qwen3MoeSparseMoeBlock, Qwen3NextSparseMoeBlock)):
self.all_to_sharded_linear_in_place(layer.mlp.switch_mlp.gate_proj)
self.sharded_to_all_linear_in_place(layer.mlp.switch_mlp.down_proj)

View File

@@ -57,6 +57,7 @@ def prefill(
sampler: Callable[[mx.array], mx.array],
prompt_tokens: mx.array,
cache: KVCacheType,
group: mx.distributed.Group | None = None,
) -> tuple[float, int, list[CacheSnapshot]]:
"""Prefill the KV cache with prompt tokens.
@@ -86,6 +87,9 @@ def prefill(
set_pipeline_prefill(model, is_prefill=True)
mx_barrier(group)
logger.info("Starting prefill")
# Use max_tokens=1 because max_tokens=0 does not work.
# We just throw away the generated token - we only care about filling the cache
for _ in stream_generate(
@@ -305,16 +309,9 @@ def mlx_generate(
)
max_stop_len = max((len(s) for s in stop_sequences), default=0)
mx_barrier(group)
logger.info("Starting prefill")
# Prefill cache with all tokens except the last one
prefill_tps, prefill_tokens, ssm_snapshots_list = prefill(
model,
tokenizer,
sampler,
prompt_tokens[:-1],
caches,
model, tokenizer, sampler, prompt_tokens[:-1], caches, group
)
cache_snapshots: list[CacheSnapshot] | None = ssm_snapshots_list or None
@@ -331,6 +328,7 @@ def mlx_generate(
think_start = tokenizer.think_start
think_end = tokenizer.think_end
logger.info("Starting decode")
mx_barrier(group)
for completion_tokens, out in enumerate(

View File

@@ -285,10 +285,12 @@ def get_eos_token_ids_for_model(model_id: ModelId) -> list[int] | None:
model_id_lower = model_id.lower()
if "kimi-k2" in model_id_lower:
return [163586]
elif "glm-4.7-flash" in model_id_lower:
elif "glm-5" in model_id_lower or "glm-4.7" in model_id_lower:
# For GLM-5 and GLM-4.7
# 154820: <|endoftext|>, 154827: <|user|>, 154829: <|observation|>
return [154820, 154827, 154829]
elif "glm" in model_id_lower:
# For GLM-4.5 and older
return [151336, 151329, 151338]
return None

View File

@@ -348,7 +348,7 @@ class Worker:
session=self.session_id,
event=event,
)
logger.trace(f"Worker published event {idx}: {str(event)[:100]}")
logger.debug(f"Worker published event {idx}: {str(event)[:100]}")
await self.local_event_sender.send(fe)
self.out_for_delivery[event.event_id] = fe

View File

@@ -191,7 +191,7 @@ class RunnerSupervisor:
logger.info("Checking runner's status")
if self.runner_process.is_alive():
logger.info("Runner was found to be alive, attempting to join process")
await to_thread.run_sync(self.runner_process.join, 1)
await to_thread.run_sync(self.runner_process.join, 5)
rc = self.runner_process.exitcode
logger.info(f"RunnerSupervisor exited with exit code {rc}")
if rc == 0:

View File

@@ -28,12 +28,12 @@ trap 'cleanup' EXIT INT TERM
for host; do
ssh -T -o BatchMode=yes -o ServerAliveInterval=30 "$host@$host" \
"/nix/var/nix/profiles/default/bin/nix build github:exo-explore/exo/$commit" &
"EXO_LIBP2P_NAMESPACE=$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@$host" \
"/nix/var/nix/profiles/default/bin/nix run github:exo-explore/exo/$commit -- --namespace $commit" &>/dev/null &
"EXO_LIBP2P_NAMESPACE=$commit /nix/var/nix/profiles/default/bin/nix run github:exo-explore/exo/$commit" &>/dev/null &
done
for host; do

View File

@@ -35,7 +35,7 @@ i=0
for host; do
colour=${colours[i++ % 4]}
ssh -T -o BatchMode=yes -o ServerAliveInterval=30 "$host@$host" \
"/nix/var/nix/profiles/default/bin/nix run github:exo-explore/exo/$commit -- --namespace $commit" |&
"EXO_LIBP2P_NAMESPACE=$commit /nix/var/nix/profiles/default/bin/nix run github:exo-explore/exo/$commit" |&
awk -v p="${colour}[${host}]${reset}" '{ print p $0; fflush() }' &
done

60
uv.lock generated
View File

@@ -193,14 +193,20 @@ 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" },
@@ -306,8 +312,10 @@ 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" },
@@ -315,8 +323,10 @@ 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" },
@@ -324,8 +334,10 @@ 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" },
@@ -365,8 +377,8 @@ dependencies = [
{ name = "hypercorn", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "loguru", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "mflux", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "mlx", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "mlx", extra = ["cpu"], marker = "sys_platform == 'linux'" },
{ name = "mlx", version = "0.30.6", source = { registry = "https://pypi.org/simple" }, extra = ["cpu"], marker = "sys_platform == 'linux'" },
{ name = "mlx", version = "0.30.7.dev20260218+14841977", source = { git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git?branch=address-rdma-gpu-locks#1484197707f35186ad3bd614357c7c47fdf86ebc" }, marker = "sys_platform == 'darwin'" },
{ name = "mlx-lm", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "msgspec", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "openai-harmony", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
@@ -404,9 +416,9 @@ requires-dist = [
{ name = "hypercorn", specifier = ">=0.18.0" },
{ name = "loguru", specifier = ">=0.7.3" },
{ name = "mflux", specifier = "==0.15.5" },
{ name = "mlx", marker = "sys_platform == 'darwin'", specifier = "==0.30.6" },
{ name = "mlx", marker = "sys_platform == 'darwin'", git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git?branch=address-rdma-gpu-locks" },
{ name = "mlx", extras = ["cpu"], marker = "sys_platform == 'linux'", specifier = "==0.30.6" },
{ name = "mlx-lm", specifier = "==0.30.6" },
{ name = "mlx-lm", specifier = "==0.30.7" },
{ name = "msgspec", specifier = ">=0.19.0" },
{ name = "openai-harmony", specifier = ">=0.0.8" },
{ name = "pillow", specifier = ">=11.0,<12.0" },
@@ -1008,8 +1020,8 @@ dependencies = [
{ name = "fonttools", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "huggingface-hub", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "matplotlib", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "mlx", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "mlx", extra = ["cuda13"], marker = "sys_platform == 'linux'" },
{ name = "mlx", version = "0.30.6", source = { registry = "https://pypi.org/simple" }, extra = ["cuda13"], marker = "sys_platform == 'linux'" },
{ name = "mlx", version = "0.30.7.dev20260218+14841977", source = { git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git?branch=address-rdma-gpu-locks#1484197707f35186ad3bd614357c7c47fdf86ebc" }, marker = "sys_platform == 'darwin'" },
{ name = "numpy", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "opencv-python", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "piexif", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
@@ -1036,18 +1048,12 @@ wheels = [
name = "mlx"
version = "0.30.6"
source = { registry = "https://pypi.org/simple" }
dependencies = [
{ name = "mlx-metal", marker = "sys_platform == 'darwin'" },
resolution-markers = [
"sys_platform == 'linux'",
]
wheels = [
{ url = "https://files.pythonhosted.org/packages/ae/5b/e460e144a34d5529e010056cccf50b538d56ed001473bc6b246018fd58cb/mlx-0.30.6-cp313-cp313-macosx_14_0_arm64.whl", hash = "sha256:ed86f8bffc174c2f259ca589ea25464c96cf69d1bb457074a2bf2ef53737e54f", size = 573515, upload-time = "2026-02-06T03:45:23.405Z" },
{ url = "https://files.pythonhosted.org/packages/60/25/69833fefb9a3fef30b56792b1bcd022496c4fea83e45411d289b77ef7546/mlx-0.30.6-cp313-cp313-macosx_15_0_arm64.whl", hash = "sha256:c52294958269e20f300639a17c1900ca8fc737d859ddda737f9811e94bd040e5", size = 573516, upload-time = "2026-02-06T03:45:24.618Z" },
{ url = "https://files.pythonhosted.org/packages/9c/6a/7e7fbeebc5cb51b6a5eba96b263a6298707bcbdc059f4b0b73e088bc3dea/mlx-0.30.6-cp313-cp313-macosx_26_0_arm64.whl", hash = "sha256:b5b6636f7c49a4d86d8ec82643b972f45a144a7a9f3a967b27b2e6e22cf71e6a", size = 573592, upload-time = "2026-02-06T03:45:25.928Z" },
{ url = "https://files.pythonhosted.org/packages/93/06/280f6f2ba80520a7109730425eda0d966658793aa0d02d8be8d351f75253/mlx-0.30.6-cp313-cp313-manylinux_2_35_aarch64.whl", hash = "sha256:67e6c9e30a9faeacc209917ef5523177cf9b086914b6b5d83ff886e4294b727d", size = 622011, upload-time = "2026-02-06T03:45:28.165Z" },
{ url = "https://files.pythonhosted.org/packages/fe/35/f872afbee9c079cc69924d9e9c46f5663adb7da58cba3511db082dd307c1/mlx-0.30.6-cp313-cp313-manylinux_2_35_x86_64.whl", hash = "sha256:47db8b16fcb6f6c5a47c0bdb24ed377b41237017ac93aa6cb6aa206c9bdf82e4", size = 663650, upload-time = "2026-02-06T03:45:30.315Z" },
{ url = "https://files.pythonhosted.org/packages/60/23/361dc7a5797634e4d7e9bdd6564c6b28f9b1246672632def2f91bf066b18/mlx-0.30.6-cp314-cp314-macosx_14_0_arm64.whl", hash = "sha256:78804a89dcff4a838f7c2da72392fe87a523e95122a3c840e53df019122aad45", size = 575028, upload-time = "2026-02-06T03:45:31.549Z" },
{ url = "https://files.pythonhosted.org/packages/a8/69/1854484d414171586814dfbe8def95f75c4ea2c7341ba13ba8ee675f7c62/mlx-0.30.6-cp314-cp314-macosx_15_0_arm64.whl", hash = "sha256:ec13584ab069665cc7ad34a05494d9291cd623aef6ae96be48875fc87cfc25d6", size = 575026, upload-time = "2026-02-06T03:45:33.072Z" },
{ url = "https://files.pythonhosted.org/packages/6b/b8/3adbc441924209a7e4c568308b2a0b54bd09aee6a68db5bae85304791e54/mlx-0.30.6-cp314-cp314-macosx_26_0_arm64.whl", hash = "sha256:b2c5e8a090a753ef99a1380a4d059c983083f36198864f6df9faaf1223d083df", size = 575041, upload-time = "2026-02-06T03:45:34.814Z" },
{ url = "https://files.pythonhosted.org/packages/3f/54/9d9e06804fb2088202a2cdf60458e00b221f71420bea285720b60f9e82b5/mlx-0.30.6-cp314-cp314-manylinux_2_35_aarch64.whl", hash = "sha256:9ceddede4af0de31d1f6b3099f70e5469d60cd7c546975dedbdbeab3519cab3f", size = 624002, upload-time = "2026-02-06T03:45:36Z" },
{ url = "https://files.pythonhosted.org/packages/42/92/3140a15a50cb1f9267a6552171e1dfa577861de53e093124bc43707f2a0e/mlx-0.30.6-cp314-cp314-manylinux_2_35_x86_64.whl", hash = "sha256:4a6ffd2d16728cf95f63a1b555d7c2eaeea686a0e6b73228bd265411cb5d77a4", size = 663569, upload-time = "2026-02-06T03:45:37.242Z" },
]
@@ -1060,6 +1066,14 @@ cuda13 = [
{ name = "mlx-cuda-13", marker = "sys_platform == 'linux'" },
]
[[package]]
name = "mlx"
version = "0.30.7.dev20260218+14841977"
source = { git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git?branch=address-rdma-gpu-locks#1484197707f35186ad3bd614357c7c47fdf86ebc" }
resolution-markers = [
"sys_platform == 'darwin'",
]
[[package]]
name = "mlx-cpu"
version = "0.30.6"
@@ -1086,30 +1100,20 @@ wheels = [
[[package]]
name = "mlx-lm"
version = "0.30.6"
version = "0.30.7"
source = { registry = "https://pypi.org/simple" }
dependencies = [
{ name = "jinja2", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "mlx", marker = "sys_platform == 'darwin'" },
{ name = "mlx", version = "0.30.7.dev20260218+14841977", source = { git = "https://github.com/rltakashige/mlx-jaccl-fix-small-recv.git?branch=address-rdma-gpu-locks#1484197707f35186ad3bd614357c7c47fdf86ebc" }, marker = "sys_platform == 'darwin'" },
{ name = "numpy", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "protobuf", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "pyyaml", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "sentencepiece", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
{ name = "transformers", marker = "sys_platform == 'darwin' or sys_platform == 'linux'" },
]
sdist = { url = "https://files.pythonhosted.org/packages/76/cb/815deddc8699b1f694d7e1f9cbed52934c03a8b49432c8add72932bb2f0b/mlx_lm-0.30.6.tar.gz", hash = "sha256:807e042d7040268f1b19190b7eaefd8b2efbff5590a65460974ad4225b91dda1", size = 271733, upload-time = "2026-02-04T21:27:45.741Z" }
sdist = { url = "https://files.pythonhosted.org/packages/66/0d/56542e2ae13ec6f542d3977d7cff89a205d4f6c5122e0ce23f33265f61c9/mlx_lm-0.30.7.tar.gz", hash = "sha256:e5f31ac58d9f2381f28e1ba639ff903e64f7cff1bdc245c0bc97f72264be329c", size = 275764, upload-time = "2026-02-12T18:41:11.86Z" }
wheels = [
{ url = "https://files.pythonhosted.org/packages/20/5f/01d281f1fa8a1521d5936659beb4f5ab1f32b463d059263cf9d4cef969d9/mlx_lm-0.30.6-py3-none-any.whl", hash = "sha256:a7405bd581eacc4bf8209d7a6b7f23629585a0d7c6740c2a97e51fee35b3b0e1", size = 379451, upload-time = "2026-02-04T21:27:43.222Z" },
]
[[package]]
name = "mlx-metal"
version = "0.30.6"
source = { registry = "https://pypi.org/simple" }
wheels = [
{ url = "https://files.pythonhosted.org/packages/f3/85/44406b521f920248fad621334d4dc15e77660a494edf890e7cbee33bf38d/mlx_metal-0.30.6-py3-none-macosx_14_0_arm64.whl", hash = "sha256:ea6d0c973def9a5b4f652cc77036237db3f88c9d0af63701d76b5fddde99b820", size = 38437818, upload-time = "2026-02-06T03:44:56.19Z" },
{ url = "https://files.pythonhosted.org/packages/d0/cb/10a516995f7d0c154b0d7e633c54b51e96977a86a355105b6474cfcbe0d0/mlx_metal-0.30.6-py3-none-macosx_15_0_arm64.whl", hash = "sha256:0f8cb94634d07e06a372d6ad9a090f38a18bab1ff19a140aede60eacf707bb94", size = 38433701, upload-time = "2026-02-06T03:44:59.678Z" },
{ url = "https://files.pythonhosted.org/packages/4c/7d/70cb272f7373c334709f210ed8420511fc9d64d05a7a646c0b3b94c29c04/mlx_metal-0.30.6-py3-none-macosx_26_0_arm64.whl", hash = "sha256:d761ae26304f2c4b454eeea7f612a56919d9e5e57dbb1dc0788f8e34aa6f41c2", size = 47718448, upload-time = "2026-02-06T03:45:03.133Z" },
{ url = "https://files.pythonhosted.org/packages/1e/17/a41c798a3d9cbdc47f39c6db5bba4c2cd199203ead26bf911cb03b644070/mlx_lm-0.30.7-py3-none-any.whl", hash = "sha256:17442a4bf01c4c2d3bca1e647712fe44f19890c3f1eadc8589d389e57b44b9bf", size = 386591, upload-time = "2026-02-12T18:41:10.236Z" },
]
[[package]]