Disabling model discovery for now.

package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "localm",
-  "version": "1.1.15",
+  "version": "1.1.35",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "localm",
-      "version": "1.1.15",
+      "version": "1.1.35",
       "license": "ISC",
       "dependencies": {
         "@huggingface/transformers": "^3.7.2",

package.json

@@ -1,7 +1,7 @@
 {
   "name": "localm",
-  "version": "1.1.35",
-  "description": "Chat application https://tty.wtf/#/Models//%60%60%60/const+xen+=+'Xenova%2Fllama2.c-stories15M';/const+lots+=+(await+Promise.all(%5B1,+2,+3,+4,+5,+6,+7,+8,+9,+10,+11,+12,+13,+14,+15,+16,+17,+18,+19,+20%5D.map(i+=%3E+fetch(/++'https:%2F%2Fhuggingface.co%2Fapi%2Fmodels%3Ffull=true%26limit=1000%26offset='+%2B+i+*+1000).then(x+=%3E+x.json())))).reduce((all,+arr)+=%3E+all.concat(arr));/%60%60%60////%60%60%60/const+tagFiltered+=+lots.filter(mod+=%3E+%2Ftransformers.js%2Fi.test(mod.tags.join('+')));/%60%60%60////%60%60%60/const+libFilter+=+await+fetch('https:%2F%2Fhuggingface.co%2Fapi%2Fmodels%3Ffull=true%26limit=1000%26gated=false%26library=transformers.js').then(x=%3Ex.json())/%60%60%60////%60%60%60/libFilter.filter(mod+=%3E+%2Ftransformers.js%2Fi.test(mod.tags.join('+')));/%60%60%60//////%60%60%60/++const+dataStr+=+'%5B%5Cn'+%2B+lots.map(entry+=%3E+JSON.stringify(entry)).join(',%5Cn')+%2B+'%5Cn%5D';/++const+blob+=+new+Blob(%5BdataStr%5D,+%7Btype:+'application%2Fjson'%7D);/++const+url+=+URL.createObjectURL(blob);/const+a+=+document.createElement('a');/a.href+=+url;/a.textContent+=+'Lots.json';/a.download+=+'lots.json';/a;/%60%60%60////%60%60%60/const+probeHuggingFaceModel+=+async+function+probeHuggingFaceModel(mod)+%7B/++const+baseUrl+=+%60https:%2F%2Fhuggingface.co%2F$%7Bmod.id%7D%2Fresolve%2Fmain%2F%60;/++const+essentialFiles+=+%5B/++++'config.json',/%2F%2F++++'model.onnx',/%2F%2F++++'model_quantized.onnx',/++++'tokenizer.json',/++++'tokenizer_config.json'/++%5D;//++let+withFiles+=+%7B/++++count:+0,/++++files:+%7B%7D,/++++...mod/++%7D;//++for+(const+file+of+essentialFiles)+%7B/++++const+fileUrl+=+%60$%7BbaseUrl%7D$%7Bfile%7D%60;/++++try+%7B/++++++const+response+=+await+fetch(/++++++++fileUrl,/++++++++%2F%2F%7B+method:+'HEAD'+%7D/++++++);/++++++if(response.ok)+%7B/++++++++withFiles.count%2B%2B;/++++++++withFiles.files%5Bfile%5D+=+true;/++++++%7D+else+%7B/++++++++withFiles.files%5Bfile%5D=+false;/++++++%7D/++++%7D+catch(error)+%7B/++++++withFiles.files%5Bfile%5D+=+error.message;/++++%7D/++%7D//++return+withFiles;/%7D//const+probeChecked+=+await+Promise.all(/++tagFiltered.map(mod+=%3E+probeHuggingFaceModel(mod)));//%60%60%60//",
+  "version": "1.1.37",
+  "description": "Chat application",
   "scripts": {
     "build": "esbuild src/index.js --target=es6 --bundle --sourcemap --outfile=./index.js --format=iife --external:fs --external:path --external:child_process --external:ws --external:katex/dist/katex.min.css",
     "start": "npm run build -- --watch --serve=0.0.0.0:8812 --servedir=. --serve-fallback=index.html",

src/app/worker-connection.js

@@ -7,7 +7,6 @@ export function workerConnection() {
 
   const connection = {
     loaded: workerLoaded.then(worker => ({ env: worker.env })),
-    listModels,
     loadModel,
     runPrompt,
     listChatModels
@@ -75,12 +74,6 @@ export function workerConnection() {
     });
   }
 
-  async function listModels() {
-    await workerLoaded;
-    const { send } = await workerLoaded;
-    return send({ type: 'listModels' });
-  }
-
   /**
    * List and classify chat-capable models via worker. Returns a promise and accepts an onProgress callback.
    * @param {object} params

src/worker/boot-worker.js

@@ -3,6 +3,8 @@
 import { ModelCache } from './model-cache';
 import { listChatModelsIterator } from './list-chat-models.js';
 
+import curatedList from './curated-model-list.json' assert { type: 'json' };
+
 export function bootWorker() {
   const modelCache = new ModelCache();
   let selectedModel = modelCache.knownModels[0];
@@ -27,9 +29,7 @@ export function bootWorker() {
   async function handleMessage({ data }) {
     const { id } = data;
     try {
-      if (data.type === 'listModels') {
-        self.postMessage({ id, type: 'response', result: modelCache.knownModels });
-      } else if (data.type === 'listChatModels') {
+      if (data.type === 'listChatModels') {
         // kick off the long-running listing/classification task
         handleListChatModels(data).catch(err => {
           self.postMessage({ id, type: 'error', error: String(err) });
@@ -74,6 +74,10 @@ export function bootWorker() {
 
   // Implementation of the listChatModels worker action using the async-iterator action.
   async function handleListChatModels({ id, params = {} }) {
+
+    self.postMessage({ id, type: 'response', result: { models: curatedList } });
+    return;
+
     const iterator = listChatModelsIterator(params);
     let sawDone = false;
     // batching buffer
@@ -99,26 +103,28 @@ export function bootWorker() {
       }
     }
 
-    activeTasks.set(id, { abort: () => { try { iterator.return(); } catch (e) {} } });
+    activeTasks.set(id, { abort: () => iterator.return() });
+    let lastBatchDelta;
     try {
       for await (const delta of iterator) {
-        try { enqueueProgress(delta); } catch (e) {}
+        try { enqueueProgress(delta); } catch (e) { }
+        if (delta.models) lastBatchDelta = delta;
         if (delta && delta.status === 'done') {
           sawDone = true;
-          // flush any remaining progress messages synchronously
-          try { flushBatch(); } catch (e) {}
-          try { self.postMessage({ id, type: 'response', result: { models: delta.models, meta: delta.meta } }); } catch (e) {}
         }
       }
 
+      // flush any remaining progress messages synchronously
+      flushBatch();
       if (!sawDone) {
         // iterator exited early (likely cancelled)
-        try { flushBatch(); } catch (e) {}
-        try { self.postMessage({ id, type: 'response', result: { cancelled: true } }); } catch (e) {}
+        self.postMessage({ id, type: 'response', result: { cancelled: true } });
+      } else {
+        self.postMessage({ id, type: 'response', result: lastBatchDelta });
       }
     } catch (err) {
-      try { flushBatch(); } catch (e) {}
-      try { self.postMessage({ id, type: 'error', error: String(err), code: err.code || null }); } catch (e) {}
+      flushBatch();
+      self.postMessage({ id, type: 'error', error: String(err), code: err.code || null });
     } finally {
       activeTasks.delete(id);
     }

src/worker/curated-model-list.json

@@ -0,0 +1,151 @@
+[
+  {
+    "id": "Xenova/llama2.c-stories15M",
+    "name": "Llama2.c Stories 15M",
+    "model_type": "llama",
+    "architectures": ["llama"],
+    "classification": "gen",
+    "confidence": "high",
+    "fetchStatus": "ok",
+    "hasTokenizer": true,
+    "hasOnnxModel": true,
+    "isTransformersJsReady": true,
+    "info": {
+      "display_name": "Llama2.c Stories 15M (Xenova)",
+      "params": "15M",
+      "params_count": 15000000,
+      "architecture": "LLaMA-family (tiny conversion for stories use via Xenova toolchain)",
+      "context_window": 2048,
+      "quantization": "likely float16/uint8/packed formats supported by Xenova runtime",
+      "hf_url": "https://huggingface.co/Xenova/llama2.c-stories15M",
+      "recommended_runtime": "transformers.js / Xenova runtime (wasm/webgpu)",
+      "is_mobile_capable": true,
+      "verified": false,
+      "assumed": true,
+      "notes": "Original curated metadata preserved; tokenizer/ONNX flags assumed for Xenova-converted tiny model."
+    }
+  },
+  {
+    "id": "Xenova/phi-3-mini-4k-instruct",
+    "name": "phi-3-mini-4k-instruct",
+    "model_type": "phi3",
+    "architectures": ["phi3"],
+    "classification": "gen",
+    "confidence": "high",
+    "fetchStatus": "ok",
+    "hasTokenizer": true,
+    "hasOnnxModel": true,
+    "isTransformersJsReady": true,
+    "info": {
+      "display_name": "Phi-3 Mini 4k Instruct (Xenova)",
+      "params": "mini (estimate)",
+      "params_count": null,
+      "architecture": "Phi family (Phi-3 mini conversion)",
+      "context_window": 4096,
+      "hf_url": "https://huggingface.co/Xenova/phi-3-mini-4k-instruct",
+      "recommended_runtime": "transformers.js / Xenova runtime (wasm/webgpu)",
+      "is_mobile_capable": true,
+      "verified": false,
+      "assumed": true,
+      "notes": "Parameter count and asset layout should be verified against HF API."
+    }
+  },
+  {
+    "id": "Xenova/all-MiniLM-L6-v2",
+    "name": "all-MiniLM-L6-v2",
+    "model_type": "sentence-transformers",
+    "architectures": ["MiniLM"],
+    "classification": "encoder",
+    "confidence": "high",
+    "fetchStatus": "ok",
+    "hasTokenizer": true,
+    "hasOnnxModel": true,
+    "isTransformersJsReady": true,
+    "info": {
+      "display_name": "all-MiniLM-L6-v2 (Xenova mirror)",
+      "params": "≈22M",
+      "params_count": 22000000,
+      "architecture": "MiniLM (distilled transformer)",
+      "context_window": 512,
+      "hf_url": "https://huggingface.co/Xenova/all-MiniLM-L6-v2",
+      "recommended_runtime": "transformers.js (wasm)",
+      "is_mobile_capable": true,
+      "verified": false,
+      "assumed": true,
+      "notes": "Encoder-style model (worker prefilter normally excludes sentence-transformers from generation lists)."
+    }
+  },
+  {
+    "id": "Xenova/distilgpt2",
+    "name": "distilgpt2",
+    "model_type": "gpt2",
+    "architectures": ["gpt2"],
+    "classification": "gen",
+    "confidence": "high",
+    "fetchStatus": "ok",
+    "hasTokenizer": true,
+    "hasOnnxModel": true,
+    "isTransformersJsReady": true,
+    "info": {
+      "display_name": "DistilGPT-2 (Xenova mirror)",
+      "params": "≈82M",
+      "params_count": 82000000,
+      "architecture": "GPT-2 distilled",
+      "context_window": 1024,
+      "hf_url": "https://huggingface.co/Xenova/distilgpt2",
+      "recommended_runtime": "transformers.js",
+      "is_mobile_capable": true,
+      "verified": false,
+      "assumed": true
+    }
+  },
+  {
+    "id": "Xenova/gpt2",
+    "name": "gpt2",
+    "model_type": "gpt2",
+    "architectures": ["gpt2"],
+    "classification": "gen",
+    "confidence": "high",
+    "fetchStatus": "ok",
+    "hasTokenizer": true,
+    "hasOnnxModel": true,
+    "isTransformersJsReady": true,
+    "info": {
+      "display_name": "GPT-2 (Xenova mirror)",
+      "params": "117M (base)",
+      "params_count": 117000000,
+      "architecture": "GPT-2",
+      "context_window": 1024,
+      "hf_url": "https://huggingface.co/Xenova/gpt2",
+      "recommended_runtime": "transformers.js",
+      "is_mobile_capable": true,
+      "verified": false,
+      "assumed": true
+    }
+  },
+  {
+    "id": "Xenova/qwen-2.5-small-instruct",
+    "name": "qwen-2.5-small-instruct",
+    "model_type": "qwen",
+    "architectures": ["qwen"],
+    "classification": "gen",
+    "confidence": "medium",
+    "fetchStatus": "ok",
+    "hasTokenizer": true,
+    "hasOnnxModel": false,
+    "isTransformersJsReady": false,
+    "info": {
+      "display_name": "Qwen-2.5 Small Instruct (Xenova mirror)",
+      "params": "≈2.5B (estimate)",
+      "params_count": null,
+      "architecture": "Qwen family",
+      "context_window": 8192,
+      "hf_url": "https://huggingface.co/Xenova/qwen-2.5-small-instruct",
+      "recommended_runtime": "Xenova runtime / transformers.js (quantized)",
+      "is_mobile_capable": false,
+      "verified": false,
+      "assumed": true,
+      "notes": "Marked conservatively as not transformers.js-ready / no ONNX in HF assets."
+    }
+  }
+]