fix(p2p): print the p2p-worker command, not the pre-v4 worker one

Root cause: #9124 moved the llama.cpp P2P worker under `p2p-worker` and
made `worker` the generic distributed worker, but the hint printed after
`run --p2p`, the Swarm page's command builder and the docs kept the old
`local-ai worker p2p-llama-cpp-rpc`, which the parser now rejects with
"unexpected argument p2p-llama-cpp-rpc". The MLX shard command had the
same stale prefix.

The hint becomes a constant that a test parses against the real CLI
tree, so the printed command cannot drift from the command set again.

Fixes #13032

Assisted-by: Claude:claude-fable-5-1
Signed-off-by: Vivek1898 <viveksingh27795@gmail.com>

core/cli/p2p_worker_hint_test.go

@@ -0,0 +1,36 @@
+package cli
+
+import (
+	"strings"
+
+	"github.com/alecthomas/kong"
+	. "github.com/onsi/ginkgo/v2"
+	. "github.com/onsi/gomega"
+
+	"github.com/mudler/LocalAI/core/config"
+)
+
+var _ = Describe("p2p worker hint", func() {
+	It("names a command the CLI accepts", func() {
+		// Parse a copy so the package-level CLI is left untouched. The vars
+		// are the ones cmd/local-ai/main.go supplies; the tags cannot be
+		// resolved without them.
+		app := CLI
+		k, err := kong.New(&app, kong.Vars{
+			"basepath":             kong.ExpandPath("."),
+			"generatedcontentpath": DefaultGeneratedContentPath(),
+			"uploadpath":           DefaultUploadPath(),
+			"galleries":            config.DefaultGalleriesJSON,
+			"backends":             config.DefaultBackendGalleriesJSON,
+			"version":              "test",
+		})
+		Expect(err).ToNot(HaveOccurred())
+
+		args := strings.Fields(p2pWorkerCommand)
+		Expect(args[0]).To(Equal("local-ai"))
+
+		ctx, err := k.Parse(args[1:])
+		Expect(err).ToNot(HaveOccurred())
+		Expect(ctx.Command()).To(Equal("p2p-worker p2p-llama-cpp-rpc"))
+	})
+})

core/cli/run.go

@@ -32,6 +32,13 @@ import (
 // When renaming flags, add the old name as an alias for backward compatibility
 // and document the deprecation in the help text.
 
+// p2pWorkerCommand is the command a user runs on another node to join the
+// P2P network with a llama.cpp worker. It is printed after `run --p2p`
+// generates a token, and a test parses it against the CLI so it cannot drift
+// from the command tree again (it named the pre-v4 `worker` command for a
+// release cycle).
+const p2pWorkerCommand = "local-ai p2p-worker p2p-llama-cpp-rpc"
+
 type RunCMD struct {
 	AddonHostOps              string `name:"addon-host-ops" env:"LOCALAI_ADDON_HOST_OPS" default:"fine-tuning=finetune.,quantization=quantize.,sandbox=workspace.,sandbox=egress.,office=workspace.,web=workspace.,telegram=connector." help:"Add-ons that may call host operations, as a comma list of addon-id=operation-prefix. An add-on that is not listed gets no host operation channel. Empty gives none" group:"addons"`
 	AddonsEnabled             bool   `name:"addons-enabled" env:"LOCALAI_ADDONS_ENABLED" default:"true" help:"Run the add-on owner, which installs and supervises add-ons. Turn it off with --addons-enabled=false" group:"addons"`
@@ -627,7 +634,7 @@ func (r *RunCMD) Run(ctx *cliContext.Context) (result error) {
 			fmt.Println(token)
 
 			xlog.Info("To use the token, you can run the following command in another node or terminal:")
-			fmt.Printf("export TOKEN=\"%s\"\nlocal-ai worker p2p-llama-cpp-rpc\n", token)
+			fmt.Printf("export TOKEN=\"%s\"\n%s\n", token, p2pWorkerCommand)
 		}
 		opts = append(opts, config.WithP2PToken(token))
 	}

core/http/react-ui/e2e/add-node.spec.js

@@ -111,9 +111,9 @@ test.describe('Add a node', () => {
     test('a memory shard offers llama.cpp RPC and MLX', async ({ page }) => {
       await mockSwarm(page, { p2p })
       await page.goto('/app/nodes/add?join=shard')
-      await expect(command(page)).toContainText('worker p2p-llama-cpp-rpc', { timeout: 15_000 })
+      await expect(command(page)).toContainText('p2p-worker p2p-llama-cpp-rpc', { timeout: 15_000 })
       await page.getByRole('radio', { name: 'MLX (Apple Silicon)' }).click()
-      await expect(command(page)).toContainText('worker p2p-mlx')
+      await expect(command(page)).toContainText('p2p-worker p2p-mlx')
       await expect(command(page)).toContainText('latest-metal-darwin-arm64')
       // MLX runs on a Mac: GPU build targets do not apply.
       await expect(page.getByRole('radio', { name: 'CUDA 12', exact: true })).toHaveCount(0)

core/http/react-ui/src/utils/swarm.js

@@ -249,8 +249,8 @@ export function workerCommand({ flavor, nodeType, option, dev, frontend, natsUrl
 export function p2pCommand({ method, option, dev, token, flavor }) {
   const suffix = method === 'peer' ? 'run --federated --p2p'
     : method === 'server' ? 'federated'
-      : method === 'mlx' ? 'worker p2p-mlx'
-        : 'worker p2p-llama-cpp-rpc'
+      : method === 'mlx' ? 'p2p-worker p2p-mlx'
+        : 'p2p-worker p2p-llama-cpp-rpc'
   const name = method === 'peer' ? 'local-ai' : method === 'server' ? 'local-ai-federated' : method === 'mlx' ? 'local-ai-mlx-worker' : 'local-ai-worker'
   const tokenValue = token || 'your-token-here'
   if (flavor === 'cli') return `TOKEN="${tokenValue}" local-ai ${suffix}`

core/http/react-ui/src/utils/swarm.test.js

@@ -140,5 +140,6 @@ test('builds join commands that carry no secret', () => {
   assert.match(p2pCommand({ method: 'peer', option, dev: false, token: 'abc', flavor: 'docker' }), /run --federated --p2p$/)
   assert.match(p2pCommand({ method: 'mlx', option, dev: true, token: '', flavor: 'docker' }), /your-token-here/)
   assert.match(p2pCommand({ method: 'server', option, dev: false, token: 't', flavor: 'docker' }), /--name local-ai-federated[\s\S]*federated$/)
-  assert.match(p2pCommand({ method: 'shard', option, dev: false, token: 't', flavor: 'cli' }), /^TOKEN="t" local-ai worker p2p-llama-cpp-rpc$/)
+  assert.match(p2pCommand({ method: 'shard', option, dev: false, token: 't', flavor: 'cli' }), /^TOKEN="t" local-ai p2p-worker p2p-llama-cpp-rpc$/)
+  assert.match(p2pCommand({ method: 'mlx', option, dev: false, token: 't', flavor: 'cli' }), /^TOKEN="t" local-ai p2p-worker p2p-mlx$/)
 })

docs/content/features/distributed_inferencing.md

@@ -119,7 +119,7 @@ To reuse the same token later, restart the server with `--p2ptoken` or `P2P_TOKE
 2. Start the workers. Copy the `local-ai` binary to other hosts and run as many workers as needed using the token:
 
 ```bash
-TOKEN=XXX ./local-ai worker p2p-llama-cpp-rpc --llama-cpp-args="-m <memory>" 
+TOKEN=XXX ./local-ai p2p-worker p2p-llama-cpp-rpc --llama-cpp-args="-m <memory>" 
 ```
 
 (Note: You can also supply the token via command-line arguments)

docs/content/features/mlx-distributed.md

@@ -43,7 +43,7 @@ On each additional Mac:
 docker run -ti --net host \
   -e TOKEN="<your-token>" \
   --name local-ai-mlx-worker \
-  localai/localai:latest-metal-darwin-arm64 worker p2p-mlx
+  localai/localai:latest-metal-darwin-arm64 p2p-worker p2p-mlx
 ```
 
 Workers auto-register on the P2P network. The LocalAI server discovers them and generates a hostfile for MLX distributed.
@@ -192,7 +192,7 @@ Starts a worker or manual rank 0 process.
 | `--addr` | `MLX_DISTRIBUTED_ADDR` | `localhost:50051` | gRPC API listen address (rank 0 only, for LocalAI or external access) |
 | `--coordinator` | `MLX_JACCL_COORDINATOR` | | JACCL coordinator `ip:port` - rank 0's address for RDMA setup (all ranks must use the same value) |
 
-### `worker p2p-mlx`
+### `p2p-worker p2p-mlx`
 
 P2P mode - auto-discovers peers and generates hostfile.