Merge commit 'd803bfe2b1fe7f5e219e50ac20d6801a0a58ac75' as 'vendor/ruvector'

vendor/ruvector/npm/packages/ruvllm-wasm/README.md

@@ -0,0 +1,276 @@
+# @ruvector/ruvllm-wasm
+
+[![npm version](https://img.shields.io/npm/v/@ruvector/ruvllm-wasm.svg)](https://www.npmjs.com/package/@ruvector/ruvllm-wasm)
+[![npm downloads](https://img.shields.io/npm/dt/@ruvector/ruvllm-wasm.svg)](https://www.npmjs.com/package/@ruvector/ruvllm-wasm)
+[![npm downloads/month](https://img.shields.io/npm/dm/@ruvector/ruvllm-wasm.svg)](https://www.npmjs.com/package/@ruvector/ruvllm-wasm)
+[![License](https://img.shields.io/npm/l/@ruvector/ruvllm-wasm.svg)](https://github.com/ruvnet/ruvector/blob/main/LICENSE)
+[![TypeScript](https://img.shields.io/badge/TypeScript-5.0-blue.svg)](https://www.typescriptlang.org/)
+
+**Run large language models directly in the browser** using WebAssembly with optional WebGPU acceleration for faster inference.
+
+## Features
+
+- **Browser-Native** - No server required, runs entirely client-side
+- **WebGPU Acceleration** - 10-50x faster inference with GPU support
+- **GGUF Models** - Load quantized models for efficient browser inference
+- **Streaming** - Real-time token streaming for responsive UX
+- **IndexedDB Caching** - Cache models locally for instant reload
+- **Privacy-First** - All processing happens on-device
+- **SIMD Support** - Optimized WASM with SIMD instructions
+- **Multi-Threading** - Parallel inference with SharedArrayBuffer
+
+## Installation
+
+```bash
+npm install @ruvector/ruvllm-wasm
+```
+
+## Quick Start
+
+```typescript
+import { RuvLLMWasm, checkWebGPU } from '@ruvector/ruvllm-wasm';
+
+// Check browser capabilities
+const webgpu = await checkWebGPU();
+console.log('WebGPU:', webgpu); // 'available' | 'unavailable' | 'not_supported'
+
+// Create instance with WebGPU (if available)
+const llm = await RuvLLMWasm.create({
+  useWebGPU: true,
+  memoryLimit: 4096, // 4GB max
+});
+
+// Load a model (with progress tracking)
+await llm.loadModel('https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF/resolve/main/tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf', {
+  onProgress: (loaded, total) => {
+    console.log(`Loading: ${Math.round(loaded / total * 100)}%`);
+  }
+});
+
+// Generate text
+const result = await llm.generate('What is the capital of France?', {
+  maxTokens: 100,
+  temperature: 0.7,
+});
+
+console.log(result.text);
+console.log(`${result.stats.tokensPerSecond.toFixed(1)} tokens/sec`);
+```
+
+## Streaming Tokens
+
+```typescript
+// Stream tokens as they're generated
+await llm.generate('Tell me a story about a robot', {
+  maxTokens: 200,
+  stream: true,
+}, (token, done) => {
+  process.stdout.write(token);
+  if (done) console.log('\n--- Done ---');
+});
+```
+
+## Chat Interface
+
+```typescript
+import { ChatMessage } from '@ruvector/ruvllm-wasm';
+
+const messages: ChatMessage[] = [
+  { role: 'system', content: 'You are a helpful assistant.' },
+  { role: 'user', content: 'What is 2 + 2?' },
+];
+
+const response = await llm.chat(messages, {
+  maxTokens: 100,
+  temperature: 0.5,
+});
+
+console.log(response.text); // "2 + 2 equals 4."
+```
+
+## React Hook Example
+
+```tsx
+import { useState, useEffect } from 'react';
+import { RuvLLMWasm, LoadingStatus } from '@ruvector/ruvllm-wasm';
+
+function useLLM(modelUrl: string) {
+  const [llm, setLLM] = useState<RuvLLMWasm | null>(null);
+  const [status, setStatus] = useState<LoadingStatus>('idle');
+  const [progress, setProgress] = useState(0);
+
+  useEffect(() => {
+    let instance: RuvLLMWasm;
+
+    async function init() {
+      instance = await RuvLLMWasm.create({ useWebGPU: true });
+      setStatus('downloading');
+
+      await instance.loadModel(modelUrl, {
+        onProgress: (loaded, total) => setProgress(loaded / total),
+      });
+
+      setStatus('ready');
+      setLLM(instance);
+    }
+
+    init();
+    return () => instance?.unload();
+  }, [modelUrl]);
+
+  return { llm, status, progress };
+}
+
+// Usage
+function ChatApp() {
+  const { llm, status, progress } = useLLM('https://example.com/model.gguf');
+  const [response, setResponse] = useState('');
+
+  if (status !== 'ready') {
+    return <div>Loading: {Math.round(progress * 100)}%</div>;
+  }
+
+  const generate = async () => {
+    const result = await llm!.generate('Hello!', { maxTokens: 50 });
+    setResponse(result.text);
+  };
+
+  return (
+    <div>
+      <button onClick={generate}>Generate</button>
+      <p>{response}</p>
+    </div>
+  );
+}
+```
+
+## Browser Requirements
+
+| Feature | Required | Benefit |
+|---------|----------|---------|
+| WebAssembly | Yes | Core execution |
+| WebGPU | No (recommended) | 10-50x faster |
+| SharedArrayBuffer | No | Multi-threading |
+| SIMD | No | 2-4x faster math |
+
+### Check Capabilities
+
+```typescript
+import { getCapabilities } from '@ruvector/ruvllm-wasm';
+
+const caps = await getCapabilities();
+console.log(caps);
+// {
+//   webgpu: 'available',
+//   sharedArrayBuffer: true,
+//   simd: true,
+//   crossOriginIsolated: true
+// }
+```
+
+### Enable SharedArrayBuffer
+
+Add these headers to your server:
+
+```
+Cross-Origin-Opener-Policy: same-origin
+Cross-Origin-Embedder-Policy: require-corp
+```
+
+## API Reference
+
+### `RuvLLMWasm.create(options?)`
+
+Create a new instance.
+
+```typescript
+const llm = await RuvLLMWasm.create({
+  useWebGPU: true,      // Enable WebGPU acceleration
+  threads: 4,           // CPU threads (requires SharedArrayBuffer)
+  memoryLimit: 4096,    // Max memory in MB
+});
+```
+
+### `loadModel(source, options?)`
+
+Load a GGUF model.
+
+```typescript
+await llm.loadModel(url, {
+  onProgress: (loaded, total) => { /* ... */ }
+});
+```
+
+### `generate(prompt, config?, onToken?)`
+
+Generate text completion.
+
+```typescript
+const result = await llm.generate('Hello', {
+  maxTokens: 100,
+  temperature: 0.7,
+  topP: 0.9,
+  topK: 40,
+  repetitionPenalty: 1.1,
+  stopSequences: ['\n\n'],
+  stream: true,
+}, (token, done) => { /* ... */ });
+```
+
+### `chat(messages, config?, onToken?)`
+
+Chat completion with message history.
+
+```typescript
+const result = await llm.chat([
+  { role: 'system', content: 'You are helpful.' },
+  { role: 'user', content: 'Hi!' },
+], { maxTokens: 100 });
+```
+
+### `unload()`
+
+Free memory and unload model.
+
+```typescript
+llm.unload();
+```
+
+## Recommended Models
+
+Small models suitable for browser inference:
+
+| Model | Size | Use Case |
+|-------|------|----------|
+| TinyLlama-1.1B-Q4 | ~700 MB | General chat |
+| Phi-2-Q4 | ~1.6 GB | Code, reasoning |
+| Qwen2-0.5B-Q4 | ~400 MB | Fast responses |
+| StableLM-Zephyr-3B-Q4 | ~2 GB | Quality chat |
+
+## Performance Tips
+
+1. **Use WebGPU** - Check support and enable for 10-50x speedup
+2. **Smaller models** - Q4_K_M quantization balances quality/size
+3. **Cache models** - IndexedDB caching avoids re-downloads
+4. **Limit context** - Smaller context = faster inference
+5. **Stream tokens** - Better UX with progressive output
+
+## Related Packages
+
+- [@ruvector/ruvllm](https://www.npmjs.com/package/@ruvector/ruvllm) - Node.js LLM library
+- [@ruvector/ruvllm-cli](https://www.npmjs.com/package/@ruvector/ruvllm-cli) - CLI tool
+- [ruvector](https://www.npmjs.com/package/ruvector) - Vector database
+
+## Documentation
+
+- [WASM Crate](https://github.com/ruvnet/ruvector/tree/main/crates/ruvllm-wasm)
+- [API Reference](https://docs.rs/ruvllm-wasm)
+- [Examples](https://github.com/ruvnet/ruvector/tree/main/examples/ruvLLM)
+
+## License
+
+MIT OR Apache-2.0
+
+---
+
+**Part of the [RuVector](https://github.com/ruvnet/ruvector) ecosystem** - High-performance vector database with self-learning capabilities.

vendor/ruvector/npm/packages/ruvllm-wasm/package.json

@@ -0,0 +1,68 @@
+{
+  "name": "@ruvector/ruvllm-wasm",
+  "version": "0.1.0",
+  "description": "WASM bindings for browser-based LLM inference - run AI models directly in the browser with WebGPU acceleration",
+  "main": "dist/index.js",
+  "types": "dist/index.d.ts",
+  "exports": {
+    ".": {
+      "import": {
+        "types": "./dist/index.d.ts",
+        "default": "./dist/index.js"
+      },
+      "require": {
+        "types": "./dist/index.d.ts",
+        "default": "./dist/index.js"
+      }
+    }
+  },
+  "scripts": {
+    "build": "tsc",
+    "prepublishOnly": "npm run build",
+    "test": "node --test test/*.test.js",
+    "typecheck": "tsc --noEmit",
+    "clean": "rm -rf dist"
+  },
+  "devDependencies": {
+    "@types/node": "^20.19.30",
+    "@webgpu/types": "^0.1.69",
+    "typescript": "^5.9.3"
+  },
+  "keywords": [
+    "llm",
+    "wasm",
+    "webassembly",
+    "browser",
+    "inference",
+    "webgpu",
+    "ai",
+    "machine-learning",
+    "edge",
+    "offline",
+    "ruvector",
+    "ruvllm",
+    "transformers"
+  ],
+  "author": "rUv Team <team@ruv.io>",
+  "license": "MIT OR Apache-2.0",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/ruvnet/ruvector.git",
+    "directory": "npm/packages/ruvllm-wasm"
+  },
+  "homepage": "https://github.com/ruvnet/ruvector/tree/main/crates/ruvllm-wasm",
+  "bugs": {
+    "url": "https://github.com/ruvnet/ruvector/issues"
+  },
+  "engines": {
+    "node": ">= 18"
+  },
+  "publishConfig": {
+    "registry": "https://registry.npmjs.org/",
+    "access": "public"
+  },
+  "files": [
+    "dist",
+    "README.md"
+  ]
+}

vendor/ruvector/npm/packages/ruvllm-wasm/src/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;AAEH,OAAO,EACL,YAAY,EACZ,aAAa,EACb,iBAAiB,EACjB,aAAa,EACb,UAAU,EACV,gBAAgB,EAChB,aAAa,EACb,gBAAgB,EAChB,cAAc,EACd,WAAW,EACX,gBAAgB,EAChB,gBAAgB,GACjB,MAAM,YAAY,CAAC;AAEpB,sBAAsB;AACtB,eAAO,MAAM,OAAO,UAAU,CAAC;AAE/B;;GAEG;AACH,wBAAsB,WAAW,IAAI,OAAO,CAAC,OAAO,YAAY,EAAE,YAAY,CAAC,CAkB9E;AAED;;GAEG;AACH,wBAAgB,sBAAsB,IAAI,OAAO,CAEhD;AAED;;GAEG;AACH,wBAAsB,SAAS,IAAI,OAAO,CAAC,OAAO,CAAC,CAclD;AAED;;GAEG;AACH,wBAAsB,eAAe,IAAI,OAAO,CAAC;IAC/C,MAAM,EAAE,OAAO,YAAY,EAAE,YAAY,CAAC;IAC1C,iBAAiB,EAAE,OAAO,CAAC;IAC3B,IAAI,EAAE,OAAO,CAAC;IACd,mBAAmB,EAAE,OAAO,CAAC;CAC9B,CAAC,CAYD;AAED;;GAEG;AACH,wBAAgB,cAAc,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAWpD;AAED;;GAEG;AACH,wBAAgB,cAAc,CAAC,aAAa,EAAE,MAAM,GAAG;IACrD,OAAO,EAAE,MAAM,CAAC;IAChB,WAAW,EAAE,MAAM,CAAC;CACrB,CAQA;AAED;;;GAGG;AACH,qBAAa,UAAU;IACrB,OAAO,CAAC,MAAM,CAAkC;IAChD,OAAO,CAAC,MAAM,CAAoF;IAElG,OAAO;IAIP;;OAEG;WACU,MAAM,CAAC,OAAO,CAAC,EAAE;QAC5B,SAAS,CAAC,EAAE,OAAO,CAAC;QACpB,OAAO,CAAC,EAAE,MAAM,CAAC;QACjB,WAAW,CAAC,EAAE,MAAM,CAAC;KACtB,GAAG,OAAO,CAAC,UAAU,CAAC;IAqBvB;;OAEG;IACH,SAAS,IAAI,OAAO,YAAY,EAAE,aAAa;IAI/C;;OAEG;IACG,SAAS,CACb,MAAM,EAAE,MAAM,GAAG,WAAW,EAC5B,OAAO,CAAC,EAAE;QACR,UAAU,CAAC,EAAE,OAAO,YAAY,EAAE,gBAAgB,CAAC;KACpD,GACA,OAAO,CAAC,OAAO,YAAY,EAAE,aAAa,CAAC;IAuB9C;;OAEG;IACG,QAAQ,CACZ,MAAM,EAAE,MAAM,EACd,MAAM,CAAC,EAAE,OAAO,YAAY,EAAE,gBAAgB,EAC9C,OAAO,CAAC,EAAE,OAAO,YAAY,EAAE,aAAa,GAC3C,OAAO,CAAC,OAAO,YAAY,EAAE,gBAAgB,CAAC;IAkBjD;;OAEG;IACG,IAAI,CACR,QAAQ,EAAE,OAAO,YAAY,EAAE,WAAW,EAAE,EAC5C,MAAM,CAAC,EAAE,OAAO,YAAY,EAAE,gBAAgB,EAC9C,OAAO,CAAC,EAAE,OAAO,YAAY,EAAE,aAAa,GAC3C,OAAO,CAAC,OAAO,YAAY,EAAE,gBAAgB,CAAC;IAQjD;;OAEG;IACH,MAAM,IAAI,IAAI;CAGf"}

vendor/ruvector/npm/packages/ruvllm-wasm/src/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["index.ts"],"names":[],"mappings":";AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4BG;;;AAuBH,kCAkBC;AAKD,wDAEC;AAKD,8BAcC;AAKD,0CAiBC;AAKD,wCAWC;AAKD,wCAWC;AAvHD,uCAaoB;AAZlB,wGAAA,YAAY,OAAA;AACZ,yGAAA,aAAa,OAAA;AACb,6GAAA,iBAAiB,OAAA;AAYnB,sBAAsB;AACT,QAAA,OAAO,GAAG,OAAO,CAAC;AAE/B;;GAEG;AACI,KAAK,UAAU,WAAW;IAC/B,IAAI,OAAO,SAAS,KAAK,WAAW,EAAE,CAAC;QACrC,OAAO,eAAoD,CAAC;IAC9D,CAAC;IAED,IAAI,CAAC,CAAC,KAAK,IAAI,SAAS,CAAC,EAAE,CAAC;QAC1B,OAAO,eAAoD,CAAC;IAC9D,CAAC;IAED,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,MAAO,SAAiB,CAAC,GAAG,CAAC,cAAc,EAAE,CAAC;QAC9D,IAAI,OAAO,EAAE,CAAC;YACZ,OAAO,WAAgD,CAAC;QAC1D,CAAC;QACD,OAAO,aAAkD,CAAC;IAC5D,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,aAAkD,CAAC;IAC5D,CAAC;AACH,CAAC;AAED;;GAEG;AACH,SAAgB,sBAAsB;IACpC,OAAO,OAAO,iBAAiB,KAAK,WAAW,CAAC;AAClD,CAAC;AAED;;GAEG;AACI,KAAK,UAAU,SAAS;IAC7B,IAAI,CAAC;QACH,8BAA8B;QAC9B,MAAM,QAAQ,GAAG,IAAI,UAAU,CAAC;YAC9B,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI;YAC9C,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI;YAC9C,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI;YAC9C,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI;SACnC,CAAC,CAAC;QACH,MAAM,WAAW,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;QACpC,OAAO,IAAI,CAAC;IACd,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,KAAK,CAAC;IACf,CAAC;AACH,CAAC;AAED;;GAEG;AACI,KAAK,UAAU,eAAe;IAMnC,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QACvC,WAAW,EAAE;QACb,SAAS,EAAE;KACZ,CAAC,CAAC;IAEH,OAAO;QACL,MAAM;QACN,iBAAiB,EAAE,sBAAsB,EAAE;QAC3C,IAAI;QACJ,mBAAmB,EAAE,OAAO,mBAAmB,KAAK,WAAW,IAAI,mBAAmB;KACvF,CAAC;AACJ,CAAC;AAED;;GAEG;AACH,SAAgB,cAAc,CAAC,KAAa;IAC1C,MAAM,KAAK,GAAG,CAAC,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;IACtC,IAAI,IAAI,GAAG,KAAK,CAAC;IACjB,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,OAAO,IAAI,IAAI,IAAI,IAAI,SAAS,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpD,IAAI,IAAI,IAAI,CAAC;QACb,SAAS,EAAE,CAAC;IACd,CAAC;IAED,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC;AAClD,CAAC;AAED;;GAEG;AACH,SAAgB,cAAc,CAAC,aAAqB;IAIlD,sCAAsC;IACtC,MAAM,UAAU,GAAG,aAAa,GAAG,CAAC,IAAI,GAAG,IAAI,CAAC,CAAC;IAEjD,OAAO;QACL,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,GAAG,CAAC,EAAE,eAAe;QACrD,WAAW,EAAE,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,GAAG,CAAC,EAAE,4BAA4B;KACvE,CAAC;AACJ,CAAC;AAED;;;GAGG;AACH,MAAa,UAAU;IAIrB,YAAoB,MAAuC;QAFnD,WAAM,GAAuC,MAA4C,CAAC;QAGhG,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACvB,CAAC;IAED;;OAEG;IACH,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,OAInB;QACC,MAAM,MAAM,GAAoC;YAC9C,OAAO,EAAE,OAAO,EAAE,OAAO;YACzB,WAAW,EAAE,OAAO,EAAE,WAAW;YACjC,IAAI,EAAE,MAAM,SAAS,EAAE;YACvB,WAAW,EAAE,IAAI;SAClB,CAAC;QAEF,IAAI,OAAO,EAAE,SAAS,EAAE,CAAC;YACvB,MAAM,YAAY,GAAG,MAAM,WAAW,EAAE,CAAC;YACzC,IAAI,YAAY,KAAK,WAAW,EAAE,CAAC;gBACjC,MAAM,OAAO,GAAG,MAAO,SAAiB,CAAC,GAAG,CAAC,cAAc,EAAE,CAAC;gBAC9D,IAAI,OAAO,EAAE,CAAC;oBACZ,MAAM,CAAC,MAAM,GAAG,MAAM,OAAO,CAAC,aAAa,EAAE,CAAC;gBAChD,CAAC;YACH,CAAC;QACH,CAAC;QAED,OAAO,IAAI,UAAU,CAAC,MAAM,CAAC,CAAC;IAChC,CAAC;IAED;;OAEG;IACH,SAAS;QACP,OAAO,IAAI,CAAC,MAAM,CAAC;IACrB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,SAAS,CACb,MAA4B,EAC5B,OAEC;QAED,IAAI,CAAC,MAAM,GAAG,SAA+C,CAAC;QAE9D,2DAA2D;QAC3D,OAAO,CAAC,GAAG,CAAC,qBAAqB,EAAE,OAAO,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC;QACxF,OAAO,CAAC,GAAG,CAAC,2DAA2D,CAAC,CAAC;QACzE,OAAO,CAAC,GAAG,CAAC,gCAAgC,CAAC,CAAC;QAE9C,IAAI,CAAC,MAAM,GAAG,OAA6C,CAAC;QAE5D,OAAO;YACL,IAAI,EAAE,aAAa;YACnB,YAAY,EAAE,OAAiD;YAC/D,UAAU,EAAE,IAAI;YAChB,aAAa,EAAE,IAAI;YACnB,SAAS,EAAE,KAAK;YAChB,YAAY,EAAE,IAAI;YAClB,SAAS,EAAE,EAAE;YACb,YAAY,EAAE,QAAQ;YACtB,QAAQ,EAAE,CAAC;SACZ,CAAC;IACJ,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,QAAQ,CACZ,MAAc,EACd,MAA8C,EAC9C,OAA4C;QAE5C,OAAO,CAAC,GAAG,CAAC,yBAAyB,EAAE,MAAM,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,KAAK,CAAC,CAAC;QACxE,OAAO,CAAC,GAAG,CAAC,wDAAwD,CAAC,CAAC;QAEtE,OAAO;YACL,IAAI,EAAE,8DAA8D;YACpE,KAAK,EAAE;gBACL,eAAe,EAAE,CAAC;gBAClB,gBAAgB,EAAE,CAAC;gBACnB,SAAS,EAAE,CAAC;gBACZ,eAAe,EAAE,CAAC;gBAClB,YAAY,EAAE,CAAC;gBACf,UAAU,EAAE,CAAC;aACd;YACD,YAAY,EAAE,MAAM;SACrB,CAAC;IACJ,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,IAAI,CACR,QAA4C,EAC5C,MAA8C,EAC9C,OAA4C;QAE5C,MAAM,MAAM,GAAG,QAAQ;aACpB,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,OAAO,EAAE,CAAC;aACnC,IAAI,CAAC,IAAI,CAAC,CAAC;QAEd,OAAO,IAAI,CAAC,QAAQ,CAAC,MAAM,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC;IAChD,CAAC;IAED;;OAEG;IACH,MAAM;QACJ,IAAI,CAAC,MAAM,GAAG,MAA4C,CAAC;IAC7D,CAAC;CACF;AAxHD,gCAwHC"}

vendor/ruvector/npm/packages/ruvllm-wasm/src/index.ts

@@ -0,0 +1,276 @@
+/**
+ * @ruvector/ruvllm-wasm - Browser LLM Inference with WebAssembly
+ *
+ * Run large language models directly in the browser using WebAssembly
+ * with optional WebGPU acceleration for faster inference.
+ *
+ * @example
+ * ```typescript
+ * import { RuvLLMWasm } from '@ruvector/ruvllm-wasm';
+ *
+ * // Initialize with WebGPU (if available)
+ * const llm = await RuvLLMWasm.create({ useWebGPU: true });
+ *
+ * // Load a model
+ * await llm.loadModel('https://example.com/model.gguf', {
+ *   onProgress: (loaded, total) => console.log(`${loaded}/${total}`)
+ * });
+ *
+ * // Generate text
+ * const result = await llm.generate('Hello, world!', {
+ *   maxTokens: 100,
+ *   temperature: 0.7,
+ * });
+ *
+ * console.log(result.text);
+ * ```
+ *
+ * @packageDocumentation
+ */
+
+export {
+  WebGPUStatus,
+  LoadingStatus,
+  ModelArchitecture,
+  ModelMetadata,
+  WASMConfig,
+  GenerationConfig,
+  TokenCallback,
+  ProgressCallback,
+  InferenceStats,
+  ChatMessage,
+  CompletionResult,
+  DownloadProgress,
+} from './types.js';
+
+/** Package version */
+export const VERSION = '0.1.0';
+
+/**
+ * Check WebGPU availability
+ */
+export async function checkWebGPU(): Promise<import('./types.js').WebGPUStatus> {
+  if (typeof navigator === 'undefined') {
+    return 'not_supported' as import('./types.js').WebGPUStatus;
+  }
+
+  if (!('gpu' in navigator)) {
+    return 'not_supported' as import('./types.js').WebGPUStatus;
+  }
+
+  try {
+    const adapter = await (navigator as any).gpu.requestAdapter();
+    if (adapter) {
+      return 'available' as import('./types.js').WebGPUStatus;
+    }
+    return 'unavailable' as import('./types.js').WebGPUStatus;
+  } catch {
+    return 'unavailable' as import('./types.js').WebGPUStatus;
+  }
+}
+
+/**
+ * Check SharedArrayBuffer support (required for threading)
+ */
+export function checkSharedArrayBuffer(): boolean {
+  return typeof SharedArrayBuffer !== 'undefined';
+}
+
+/**
+ * Check SIMD support
+ */
+export async function checkSIMD(): Promise<boolean> {
+  try {
+    // Check for WASM SIMD support
+    const simdTest = new Uint8Array([
+      0x00, 0x61, 0x73, 0x6d, 0x01, 0x00, 0x00, 0x00,
+      0x01, 0x05, 0x01, 0x60, 0x00, 0x01, 0x7b, 0x03,
+      0x02, 0x01, 0x00, 0x0a, 0x0a, 0x01, 0x08, 0x00,
+      0x41, 0x00, 0xfd, 0x0f, 0x00, 0x0b,
+    ]);
+    await WebAssembly.compile(simdTest);
+    return true;
+  } catch {
+    return false;
+  }
+}
+
+/**
+ * Get browser capabilities for LLM inference
+ */
+export async function getCapabilities(): Promise<{
+  webgpu: import('./types.js').WebGPUStatus;
+  sharedArrayBuffer: boolean;
+  simd: boolean;
+  crossOriginIsolated: boolean;
+}> {
+  const [webgpu, simd] = await Promise.all([
+    checkWebGPU(),
+    checkSIMD(),
+  ]);
+
+  return {
+    webgpu,
+    sharedArrayBuffer: checkSharedArrayBuffer(),
+    simd,
+    crossOriginIsolated: typeof crossOriginIsolated !== 'undefined' && crossOriginIsolated,
+  };
+}
+
+/**
+ * Format file size for display
+ */
+export function formatFileSize(bytes: number): string {
+  const units = ['B', 'KB', 'MB', 'GB'];
+  let size = bytes;
+  let unitIndex = 0;
+
+  while (size >= 1024 && unitIndex < units.length - 1) {
+    size /= 1024;
+    unitIndex++;
+  }
+
+  return `${size.toFixed(1)} ${units[unitIndex]}`;
+}
+
+/**
+ * Estimate memory requirements for a model
+ */
+export function estimateMemory(fileSizeBytes: number): {
+  minimum: number;
+  recommended: number;
+} {
+  // Rough estimates based on model size
+  const fileSizeMB = fileSizeBytes / (1024 * 1024);
+
+  return {
+    minimum: Math.ceil(fileSizeMB * 1.2), // 20% overhead
+    recommended: Math.ceil(fileSizeMB * 1.5), // 50% overhead for KV cache
+  };
+}
+
+/**
+ * RuvLLM WASM class placeholder
+ * Full implementation requires WASM binary from ruvllm-wasm crate
+ */
+export class RuvLLMWasm {
+  private config: import('./types.js').WASMConfig;
+  private status: import('./types.js').LoadingStatus = 'idle' as import('./types.js').LoadingStatus;
+
+  private constructor(config: import('./types.js').WASMConfig) {
+    this.config = config;
+  }
+
+  /**
+   * Create a new RuvLLMWasm instance
+   */
+  static async create(options?: {
+    useWebGPU?: boolean;
+    threads?: number;
+    memoryLimit?: number;
+  }): Promise<RuvLLMWasm> {
+    const config: import('./types.js').WASMConfig = {
+      threads: options?.threads,
+      memoryLimit: options?.memoryLimit,
+      simd: await checkSIMD(),
+      cacheModels: true,
+    };
+
+    if (options?.useWebGPU) {
+      const webgpuStatus = await checkWebGPU();
+      if (webgpuStatus === 'available') {
+        const adapter = await (navigator as any).gpu.requestAdapter();
+        if (adapter) {
+          config.device = await adapter.requestDevice();
+        }
+      }
+    }
+
+    return new RuvLLMWasm(config);
+  }
+
+  /**
+   * Get current loading status
+   */
+  getStatus(): import('./types.js').LoadingStatus {
+    return this.status;
+  }
+
+  /**
+   * Load a model from URL or ArrayBuffer
+   */
+  async loadModel(
+    source: string | ArrayBuffer,
+    options?: {
+      onProgress?: import('./types.js').ProgressCallback;
+    }
+  ): Promise<import('./types.js').ModelMetadata> {
+    this.status = 'loading' as import('./types.js').LoadingStatus;
+
+    // Placeholder - actual implementation requires WASM binary
+    console.log('Loading model from:', typeof source === 'string' ? source : 'ArrayBuffer');
+    console.log('Note: Full model loading requires the ruvllm-wasm binary.');
+    console.log('Build from: crates/ruvllm-wasm');
+
+    this.status = 'ready' as import('./types.js').LoadingStatus;
+
+    return {
+      name: 'placeholder',
+      architecture: 'llama' as import('./types.js').ModelArchitecture,
+      parameters: '0B',
+      contextLength: 2048,
+      vocabSize: 32000,
+      embeddingDim: 2048,
+      numLayers: 22,
+      quantization: 'q4_k_m',
+      fileSize: 0,
+    };
+  }
+
+  /**
+   * Generate text completion
+   */
+  async generate(
+    prompt: string,
+    config?: import('./types.js').GenerationConfig,
+    onToken?: import('./types.js').TokenCallback
+  ): Promise<import('./types.js').CompletionResult> {
+    console.log('Generating with prompt:', prompt.substring(0, 50) + '...');
+    console.log('Note: Full generation requires the ruvllm-wasm binary.');
+
+    return {
+      text: '[Placeholder - build ruvllm-wasm crate for actual inference]',
+      stats: {
+        tokensGenerated: 0,
+        timeToFirstToken: 0,
+        totalTime: 0,
+        tokensPerSecond: 0,
+        promptTokens: 0,
+        memoryUsed: 0,
+      },
+      finishReason: 'stop',
+    };
+  }
+
+  /**
+   * Chat completion with message history
+   */
+  async chat(
+    messages: import('./types.js').ChatMessage[],
+    config?: import('./types.js').GenerationConfig,
+    onToken?: import('./types.js').TokenCallback
+  ): Promise<import('./types.js').CompletionResult> {
+    const prompt = messages
+      .map(m => `${m.role}: ${m.content}`)
+      .join('\n');
+
+    return this.generate(prompt, config, onToken);
+  }
+
+  /**
+   * Unload model and free memory
+   */
+  unload(): void {
+    this.status = 'idle' as import('./types.js').LoadingStatus;
+  }
+}

vendor/ruvector/npm/packages/ruvllm-wasm/src/types.d.ts

@@ -0,0 +1,123 @@
+/**
+ * RuvLLM WASM Types
+ * Types for browser-based LLM inference
+ */
+/** WebGPU availability status */
+export declare enum WebGPUStatus {
+    Available = "available",
+    Unavailable = "unavailable",
+    NotSupported = "not_supported"
+}
+/** Model loading status */
+export declare enum LoadingStatus {
+    Idle = "idle",
+    Downloading = "downloading",
+    Loading = "loading",
+    Ready = "ready",
+    Error = "error"
+}
+/** Supported model architectures */
+export declare enum ModelArchitecture {
+    Llama = "llama",
+    Mistral = "mistral",
+    Phi = "phi",
+    Qwen = "qwen",
+    Gemma = "gemma",
+    StableLM = "stablelm"
+}
+/** Model metadata */
+export interface ModelMetadata {
+    /** Model name */
+    name: string;
+    /** Model architecture */
+    architecture: ModelArchitecture;
+    /** Number of parameters */
+    parameters: string;
+    /** Context length */
+    contextLength: number;
+    /** Vocabulary size */
+    vocabSize: number;
+    /** Embedding dimension */
+    embeddingDim: number;
+    /** Number of layers */
+    numLayers: number;
+    /** Quantization type */
+    quantization: string;
+    /** File size in bytes */
+    fileSize: number;
+}
+/** WASM module configuration */
+export interface WASMConfig {
+    /** WebGPU device (optional) */
+    device?: GPUDevice;
+    /** Number of threads (SharedArrayBuffer required) */
+    threads?: number;
+    /** SIMD enabled */
+    simd?: boolean;
+    /** Memory limit in MB */
+    memoryLimit?: number;
+    /** Cache models in IndexedDB */
+    cacheModels?: boolean;
+}
+/** Generation configuration */
+export interface GenerationConfig {
+    /** Maximum tokens to generate */
+    maxTokens?: number;
+    /** Temperature (0-2) */
+    temperature?: number;
+    /** Top-p sampling */
+    topP?: number;
+    /** Top-k sampling */
+    topK?: number;
+    /** Repetition penalty */
+    repetitionPenalty?: number;
+    /** Stop sequences */
+    stopSequences?: string[];
+    /** Stream tokens as generated */
+    stream?: boolean;
+}
+/** Token callback for streaming */
+export type TokenCallback = (token: string, done: boolean) => void;
+/** Progress callback for model loading */
+export type ProgressCallback = (loaded: number, total: number) => void;
+/** Inference statistics */
+export interface InferenceStats {
+    /** Tokens generated */
+    tokensGenerated: number;
+    /** Time to first token (ms) */
+    timeToFirstToken: number;
+    /** Total time (ms) */
+    totalTime: number;
+    /** Tokens per second */
+    tokensPerSecond: number;
+    /** Prompt tokens */
+    promptTokens: number;
+    /** Memory used (MB) */
+    memoryUsed: number;
+}
+/** Chat message */
+export interface ChatMessage {
+    role: 'system' | 'user' | 'assistant';
+    content: string;
+}
+/** Completion result */
+export interface CompletionResult {
+    /** Generated text */
+    text: string;
+    /** Inference statistics */
+    stats: InferenceStats;
+    /** Finish reason */
+    finishReason: 'stop' | 'length' | 'error';
+}
+/** Model download progress */
+export interface DownloadProgress {
+    /** Bytes downloaded */
+    loaded: number;
+    /** Total bytes */
+    total: number;
+    /** Download speed (bytes/sec) */
+    speed: number;
+    /** Estimated time remaining (seconds) */
+    eta: number;
+}
+//# sourceMappingURL=types.d.ts.map

vendor/ruvector/npm/packages/ruvllm-wasm/src/types.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["types.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,iCAAiC;AACjC,oBAAY,YAAY;IACtB,SAAS,cAAc;IACvB,WAAW,gBAAgB;IAC3B,YAAY,kBAAkB;CAC/B;AAED,2BAA2B;AAC3B,oBAAY,aAAa;IACvB,IAAI,SAAS;IACb,WAAW,gBAAgB;IAC3B,OAAO,YAAY;IACnB,KAAK,UAAU;IACf,KAAK,UAAU;CAChB;AAED,oCAAoC;AACpC,oBAAY,iBAAiB;IAC3B,KAAK,UAAU;IACf,OAAO,YAAY;IACnB,GAAG,QAAQ;IACX,IAAI,SAAS;IACb,KAAK,UAAU;IACf,QAAQ,aAAa;CACtB;AAED,qBAAqB;AACrB,MAAM,WAAW,aAAa;IAC5B,iBAAiB;IACjB,IAAI,EAAE,MAAM,CAAC;IACb,yBAAyB;IACzB,YAAY,EAAE,iBAAiB,CAAC;IAChC,2BAA2B;IAC3B,UAAU,EAAE,MAAM,CAAC;IACnB,qBAAqB;IACrB,aAAa,EAAE,MAAM,CAAC;IACtB,sBAAsB;IACtB,SAAS,EAAE,MAAM,CAAC;IAClB,0BAA0B;IAC1B,YAAY,EAAE,MAAM,CAAC;IACrB,uBAAuB;IACvB,SAAS,EAAE,MAAM,CAAC;IAClB,wBAAwB;IACxB,YAAY,EAAE,MAAM,CAAC;IACrB,yBAAyB;IACzB,QAAQ,EAAE,MAAM,CAAC;CAClB;AAED,gCAAgC;AAChC,MAAM,WAAW,UAAU;IACzB,+BAA+B;IAC/B,MAAM,CAAC,EAAE,SAAS,CAAC;IACnB,qDAAqD;IACrD,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,mBAAmB;IACnB,IAAI,CAAC,EAAE,OAAO,CAAC;IACf,yBAAyB;IACzB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,gCAAgC;IAChC,WAAW,CAAC,EAAE,OAAO,CAAC;CACvB;AAED,+BAA+B;AAC/B,MAAM,WAAW,gBAAgB;IAC/B,iCAAiC;IACjC,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,wBAAwB;IACxB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,qBAAqB;IACrB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,qBAAqB;IACrB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,yBAAyB;IACzB,iBAAiB,CAAC,EAAE,MAAM,CAAC;IAC3B,qBAAqB;IACrB,aAAa,CAAC,EAAE,MAAM,EAAE,CAAC;IACzB,iCAAiC;IACjC,MAAM,CAAC,EAAE,OAAO,CAAC;CAClB;AAED,mCAAmC;AACnC,MAAM,MAAM,aAAa,GAAG,CAAC,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,KAAK,IAAI,CAAC;AAEnE,0CAA0C;AAC1C,MAAM,MAAM,gBAAgB,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,KAAK,IAAI,CAAC;AAEvE,2BAA2B;AAC3B,MAAM,WAAW,cAAc;IAC7B,uBAAuB;IACvB,eAAe,EAAE,MAAM,CAAC;IACxB,+BAA+B;IAC/B,gBAAgB,EAAE,MAAM,CAAC;IACzB,sBAAsB;IACtB,SAAS,EAAE,MAAM,CAAC;IAClB,wBAAwB;IACxB,eAAe,EAAE,MAAM,CAAC;IACxB,oBAAoB;IACpB,YAAY,EAAE,MAAM,CAAC;IACrB,uBAAuB;IACvB,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,mBAAmB;AACnB,MAAM,WAAW,WAAW;IAC1B,IAAI,EAAE,QAAQ,GAAG,MAAM,GAAG,WAAW,CAAC;IACtC,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,wBAAwB;AACxB,MAAM,WAAW,gBAAgB;IAC/B,qBAAqB;IACrB,IAAI,EAAE,MAAM,CAAC;IACb,2BAA2B;IAC3B,KAAK,EAAE,cAAc,CAAC;IACtB,oBAAoB;IACpB,YAAY,EAAE,MAAM,GAAG,QAAQ,GAAG,OAAO,CAAC;CAC3C;AAED,8BAA8B;AAC9B,MAAM,WAAW,gBAAgB;IAC/B,uBAAuB;IACvB,MAAM,EAAE,MAAM,CAAC;IACf,kBAAkB;IAClB,KAAK,EAAE,MAAM,CAAC;IACd,iCAAiC;IACjC,KAAK,EAAE,MAAM,CAAC;IACd,yCAAyC;IACzC,GAAG,EAAE,MAAM,CAAC;CACb"}

vendor/ruvector/npm/packages/ruvllm-wasm/src/types.js

@@ -0,0 +1,34 @@
+"use strict";
+/**
+ * RuvLLM WASM Types
+ * Types for browser-based LLM inference
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.ModelArchitecture = exports.LoadingStatus = exports.WebGPUStatus = void 0;
+/** WebGPU availability status */
+var WebGPUStatus;
+(function (WebGPUStatus) {
+    WebGPUStatus["Available"] = "available";
+    WebGPUStatus["Unavailable"] = "unavailable";
+    WebGPUStatus["NotSupported"] = "not_supported";
+})(WebGPUStatus || (exports.WebGPUStatus = WebGPUStatus = {}));
+/** Model loading status */
+var LoadingStatus;
+(function (LoadingStatus) {
+    LoadingStatus["Idle"] = "idle";
+    LoadingStatus["Downloading"] = "downloading";
+    LoadingStatus["Loading"] = "loading";
+    LoadingStatus["Ready"] = "ready";
+    LoadingStatus["Error"] = "error";
+})(LoadingStatus || (exports.LoadingStatus = LoadingStatus = {}));
+/** Supported model architectures */
+var ModelArchitecture;
+(function (ModelArchitecture) {
+    ModelArchitecture["Llama"] = "llama";
+    ModelArchitecture["Mistral"] = "mistral";
+    ModelArchitecture["Phi"] = "phi";
+    ModelArchitecture["Qwen"] = "qwen";
+    ModelArchitecture["Gemma"] = "gemma";
+    ModelArchitecture["StableLM"] = "stablelm";
+})(ModelArchitecture || (exports.ModelArchitecture = ModelArchitecture = {}));
+//# sourceMappingURL=types.js.map

vendor/ruvector/npm/packages/ruvllm-wasm/src/types.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.js","sourceRoot":"","sources":["types.ts"],"names":[],"mappings":";AAAA;;;GAGG;;;AAEH,iCAAiC;AACjC,IAAY,YAIX;AAJD,WAAY,YAAY;IACtB,uCAAuB,CAAA;IACvB,2CAA2B,CAAA;IAC3B,8CAA8B,CAAA;AAChC,CAAC,EAJW,YAAY,4BAAZ,YAAY,QAIvB;AAED,2BAA2B;AAC3B,IAAY,aAMX;AAND,WAAY,aAAa;IACvB,8BAAa,CAAA;IACb,4CAA2B,CAAA;IAC3B,oCAAmB,CAAA;IACnB,gCAAe,CAAA;IACf,gCAAe,CAAA;AACjB,CAAC,EANW,aAAa,6BAAb,aAAa,QAMxB;AAED,oCAAoC;AACpC,IAAY,iBAOX;AAPD,WAAY,iBAAiB;IAC3B,oCAAe,CAAA;IACf,wCAAmB,CAAA;IACnB,gCAAW,CAAA;IACX,kCAAa,CAAA;IACb,oCAAe,CAAA;IACf,0CAAqB,CAAA;AACvB,CAAC,EAPW,iBAAiB,iCAAjB,iBAAiB,QAO5B"}

vendor/ruvector/npm/packages/ruvllm-wasm/src/types.ts

@@ -0,0 +1,134 @@
+/**
+ * RuvLLM WASM Types
+ * Types for browser-based LLM inference
+ */
+
+/** WebGPU availability status */
+export enum WebGPUStatus {
+  Available = 'available',
+  Unavailable = 'unavailable',
+  NotSupported = 'not_supported',
+}
+
+/** Model loading status */
+export enum LoadingStatus {
+  Idle = 'idle',
+  Downloading = 'downloading',
+  Loading = 'loading',
+  Ready = 'ready',
+  Error = 'error',
+}
+
+/** Supported model architectures */
+export enum ModelArchitecture {
+  Llama = 'llama',
+  Mistral = 'mistral',
+  Phi = 'phi',
+  Qwen = 'qwen',
+  Gemma = 'gemma',
+  StableLM = 'stablelm',
+}
+
+/** Model metadata */
+export interface ModelMetadata {
+  /** Model name */
+  name: string;
+  /** Model architecture */
+  architecture: ModelArchitecture;
+  /** Number of parameters */
+  parameters: string;
+  /** Context length */
+  contextLength: number;
+  /** Vocabulary size */
+  vocabSize: number;
+  /** Embedding dimension */
+  embeddingDim: number;
+  /** Number of layers */
+  numLayers: number;
+  /** Quantization type */
+  quantization: string;
+  /** File size in bytes */
+  fileSize: number;
+}
+
+/** WASM module configuration */
+export interface WASMConfig {
+  /** WebGPU device (optional) */
+  device?: GPUDevice;
+  /** Number of threads (SharedArrayBuffer required) */
+  threads?: number;
+  /** SIMD enabled */
+  simd?: boolean;
+  /** Memory limit in MB */
+  memoryLimit?: number;
+  /** Cache models in IndexedDB */
+  cacheModels?: boolean;
+}
+
+/** Generation configuration */
+export interface GenerationConfig {
+  /** Maximum tokens to generate */
+  maxTokens?: number;
+  /** Temperature (0-2) */
+  temperature?: number;
+  /** Top-p sampling */
+  topP?: number;
+  /** Top-k sampling */
+  topK?: number;
+  /** Repetition penalty */
+  repetitionPenalty?: number;
+  /** Stop sequences */
+  stopSequences?: string[];
+  /** Stream tokens as generated */
+  stream?: boolean;
+}
+
+/** Token callback for streaming */
+export type TokenCallback = (token: string, done: boolean) => void;
+
+/** Progress callback for model loading */
+export type ProgressCallback = (loaded: number, total: number) => void;
+
+/** Inference statistics */
+export interface InferenceStats {
+  /** Tokens generated */
+  tokensGenerated: number;
+  /** Time to first token (ms) */
+  timeToFirstToken: number;
+  /** Total time (ms) */
+  totalTime: number;
+  /** Tokens per second */
+  tokensPerSecond: number;
+  /** Prompt tokens */
+  promptTokens: number;
+  /** Memory used (MB) */
+  memoryUsed: number;
+}
+
+/** Chat message */
+export interface ChatMessage {
+  role: 'system' | 'user' | 'assistant';
+  content: string;
+}
+
+/** Completion result */
+export interface CompletionResult {
+  /** Generated text */
+  text: string;
+  /** Inference statistics */
+  stats: InferenceStats;
+  /** Finish reason */
+  finishReason: 'stop' | 'length' | 'error';
+}
+
+/** Model download progress */
+export interface DownloadProgress {
+  /** Bytes downloaded */
+  loaded: number;
+  /** Total bytes */
+  total: number;
+  /** Download speed (bytes/sec) */
+  speed: number;
+  /** Estimated time remaining (seconds) */
+  eta: number;
+}

vendor/ruvector/npm/packages/ruvllm-wasm/tsconfig.json

@@ -0,0 +1,20 @@
+{
+  "compilerOptions": {
+    "target": "ES2022",
+    "module": "NodeNext",
+    "moduleResolution": "NodeNext",
+    "declaration": true,
+    "declarationMap": true,
+    "sourceMap": true,
+    "outDir": "./dist",
+    "rootDir": "./src",
+    "strict": true,
+    "esModuleInterop": true,
+    "skipLibCheck": true,
+    "forceConsistentCasingInFileNames": true,
+    "resolveJsonModule": true,
+    "types": ["node", "@webgpu/types"]
+  },
+  "include": ["src/**/*"],
+  "exclude": ["node_modules", "dist", "test"]
+}