qmd/src/llm.ts

1599 lines
52 KiB
TypeScript
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

/**
* llm.ts - LLM abstraction layer for QMD using node-llama-cpp
*
* Provides embeddings, text generation, and reranking using local GGUF models.
*/
import {
getLlama,
resolveModelFile,
LlamaChatSession,
LlamaLogLevel,
type Llama,
type LlamaModel,
type LlamaEmbeddingContext,
type Token as LlamaToken,
} from "node-llama-cpp";
import { homedir } from "os";
import { join } from "path";
import { existsSync, mkdirSync, statSync, unlinkSync, readdirSync, readFileSync, writeFileSync } from "fs";
// =============================================================================
// Embedding Formatting Functions
// =============================================================================
/**
* Detect if a model URI uses the Qwen3-Embedding format.
* Qwen3-Embedding uses a different prompting style than nomic/embeddinggemma.
*/
export function isQwen3EmbeddingModel(modelUri: string): boolean {
return /qwen.*embed/i.test(modelUri) || /embed.*qwen/i.test(modelUri);
}
/**
* Format a query for embedding.
* Uses nomic-style task prefix format for embeddinggemma (default).
* Uses Qwen3-Embedding instruct format when a Qwen embedding model is active.
*/
export function formatQueryForEmbedding(query: string, modelUri?: string): string {
const uri = modelUri ?? process.env.QMD_EMBED_MODEL ?? DEFAULT_EMBED_MODEL;
if (isQwen3EmbeddingModel(uri)) {
return `Instruct: Retrieve relevant documents for the given query\nQuery: ${query}`;
}
return `task: search result | query: ${query}`;
}
/**
* Format a document for embedding.
* Uses nomic-style format with title and text fields (default).
* Qwen3-Embedding encodes documents as raw text without special prefixes.
*/
export function formatDocForEmbedding(text: string, title?: string, modelUri?: string): string {
const uri = modelUri ?? process.env.QMD_EMBED_MODEL ?? DEFAULT_EMBED_MODEL;
if (isQwen3EmbeddingModel(uri)) {
// Qwen3-Embedding: documents are raw text, no task prefix
return title ? `${title}\n${text}` : text;
}
return `title: ${title || "none"} | text: ${text}`;
}
// =============================================================================
// Types
// =============================================================================
/**
* Token with log probability
*/
export type TokenLogProb = {
token: string;
logprob: number;
};
/**
* Embedding result
*/
export type EmbeddingResult = {
embedding: number[];
model: string;
};
/**
* Generation result with optional logprobs
*/
export type GenerateResult = {
text: string;
model: string;
logprobs?: TokenLogProb[];
done: boolean;
};
/**
* Rerank result for a single document
*/
export type RerankDocumentResult = {
file: string;
score: number;
index: number;
};
/**
* Batch rerank result
*/
export type RerankResult = {
results: RerankDocumentResult[];
model: string;
};
/**
* Model info
*/
export type ModelInfo = {
name: string;
exists: boolean;
path?: string;
};
/**
* Options for embedding
*/
export type EmbedOptions = {
model?: string;
isQuery?: boolean;
title?: string;
};
/**
* Options for text generation
*/
export type GenerateOptions = {
model?: string;
maxTokens?: number;
temperature?: number;
};
/**
* Options for reranking
*/
export type RerankOptions = {
model?: string;
};
/**
* Options for LLM sessions
*/
export type LLMSessionOptions = {
/** Max session duration in ms (default: 10 minutes) */
maxDuration?: number;
/** External abort signal */
signal?: AbortSignal;
/** Debug name for logging */
name?: string;
};
/**
* Session interface for scoped LLM access with lifecycle guarantees
*/
export interface ILLMSession {
embed(text: string, options?: EmbedOptions): Promise<EmbeddingResult | null>;
embedBatch(texts: string[], options?: EmbedOptions): Promise<(EmbeddingResult | null)[]>;
expandQuery(query: string, options?: { context?: string; includeLexical?: boolean }): Promise<Queryable[]>;
rerank(query: string, documents: RerankDocument[], options?: RerankOptions): Promise<RerankResult>;
/** Whether this session is still valid (not released or aborted) */
readonly isValid: boolean;
/** Abort signal for this session (aborts on release or maxDuration) */
readonly signal: AbortSignal;
}
/**
* Supported query types for different search backends
*/
export type QueryType = 'lex' | 'vec' | 'hyde';
/**
* A single query and its target backend type
*/
export type Queryable = {
type: QueryType;
text: string;
};
/**
* Document to rerank
*/
export type RerankDocument = {
file: string;
text: string;
title?: string;
};
// =============================================================================
// Model Configuration
// =============================================================================
// HuggingFace model URIs for node-llama-cpp
// Format: hf:<user>/<repo>/<file>
// Override via QMD_EMBED_MODEL env var (e.g. hf:Qwen/Qwen3-Embedding-0.6B-GGUF/Qwen3-Embedding-0.6B-Q8_0.gguf)
const DEFAULT_EMBED_MODEL = "hf:ggml-org/embeddinggemma-300M-GGUF/embeddinggemma-300M-Q8_0.gguf";
const DEFAULT_RERANK_MODEL = "hf:ggml-org/Qwen3-Reranker-0.6B-Q8_0-GGUF/qwen3-reranker-0.6b-q8_0.gguf";
// const DEFAULT_GENERATE_MODEL = "hf:ggml-org/Qwen3-0.6B-GGUF/Qwen3-0.6B-Q8_0.gguf";
const DEFAULT_GENERATE_MODEL = "hf:tobil/qmd-query-expansion-1.7B-gguf/qmd-query-expansion-1.7B-q4_k_m.gguf";
// Alternative generation models for query expansion:
// LiquidAI LFM2 - hybrid architecture optimized for edge/on-device inference
// Use these as base for fine-tuning with configs/sft_lfm2.yaml
export const LFM2_GENERATE_MODEL = "hf:LiquidAI/LFM2-1.2B-GGUF/LFM2-1.2B-Q4_K_M.gguf";
export const LFM2_INSTRUCT_MODEL = "hf:LiquidAI/LFM2.5-1.2B-Instruct-GGUF/LFM2.5-1.2B-Instruct-Q4_K_M.gguf";
export const DEFAULT_EMBED_MODEL_URI = DEFAULT_EMBED_MODEL;
export const DEFAULT_RERANK_MODEL_URI = DEFAULT_RERANK_MODEL;
export const DEFAULT_GENERATE_MODEL_URI = DEFAULT_GENERATE_MODEL;
// Local model cache directory
const MODEL_CACHE_DIR = process.env.XDG_CACHE_HOME
? join(process.env.XDG_CACHE_HOME, "qmd", "models")
: join(homedir(), ".cache", "qmd", "models");
export const DEFAULT_MODEL_CACHE_DIR = MODEL_CACHE_DIR;
export type PullResult = {
model: string;
path: string;
sizeBytes: number;
refreshed: boolean;
};
type HfRef = {
repo: string;
file: string;
};
function parseHfUri(model: string): HfRef | null {
if (!model.startsWith("hf:")) return null;
const without = model.slice(3);
const parts = without.split("/");
if (parts.length < 3) return null;
const repo = parts.slice(0, 2).join("/");
const file = parts.slice(2).join("/");
return { repo, file };
}
async function getRemoteEtag(ref: HfRef): Promise<string | null> {
const url = `https://huggingface.co/${ref.repo}/resolve/main/${ref.file}`;
try {
const resp = await fetch(url, { method: "HEAD" });
if (!resp.ok) return null;
const etag = resp.headers.get("etag");
return etag || null;
} catch {
return null;
}
}
export async function pullModels(
models: string[],
options: { refresh?: boolean; cacheDir?: string } = {}
): Promise<PullResult[]> {
const cacheDir = options.cacheDir || MODEL_CACHE_DIR;
if (!existsSync(cacheDir)) {
mkdirSync(cacheDir, { recursive: true });
}
const results: PullResult[] = [];
for (const model of models) {
let refreshed = false;
const hfRef = parseHfUri(model);
const filename = model.split("/").pop();
const entries = readdirSync(cacheDir, { withFileTypes: true });
const cached = filename
? entries
.filter((entry) => entry.isFile() && entry.name.includes(filename))
.map((entry) => join(cacheDir, entry.name))
: [];
if (hfRef && filename) {
const etagPath = join(cacheDir, `${filename}.etag`);
const remoteEtag = await getRemoteEtag(hfRef);
const localEtag = existsSync(etagPath)
? readFileSync(etagPath, "utf-8").trim()
: null;
const shouldRefresh =
options.refresh || !remoteEtag || remoteEtag !== localEtag || cached.length === 0;
if (shouldRefresh) {
for (const candidate of cached) {
if (existsSync(candidate)) unlinkSync(candidate);
}
if (existsSync(etagPath)) unlinkSync(etagPath);
refreshed = cached.length > 0;
}
} else if (options.refresh && filename) {
for (const candidate of cached) {
if (existsSync(candidate)) unlinkSync(candidate);
refreshed = true;
}
}
const path = await resolveModelFile(model, cacheDir);
const sizeBytes = existsSync(path) ? statSync(path).size : 0;
if (hfRef && filename) {
const remoteEtag = await getRemoteEtag(hfRef);
if (remoteEtag) {
const etagPath = join(cacheDir, `${filename}.etag`);
writeFileSync(etagPath, remoteEtag + "\n", "utf-8");
}
}
results.push({ model, path, sizeBytes, refreshed });
}
return results;
}
// =============================================================================
// LLM Interface
// =============================================================================
/**
* Abstract LLM interface - implement this for different backends
*/
export interface LLM {
/**
* Get embeddings for text
*/
embed(text: string, options?: EmbedOptions): Promise<EmbeddingResult | null>;
/**
* Generate text completion
*/
generate(prompt: string, options?: GenerateOptions): Promise<GenerateResult | null>;
/**
* Check if a model exists/is available
*/
modelExists(model: string): Promise<ModelInfo>;
/**
* Expand a search query into multiple variations for different backends.
* Returns a list of Queryable objects.
*/
expandQuery(query: string, options?: { context?: string, includeLexical?: boolean }): Promise<Queryable[]>;
/**
* Rerank documents by relevance to a query
* Returns list of documents with relevance scores (higher = more relevant)
*/
rerank(query: string, documents: RerankDocument[], options?: RerankOptions): Promise<RerankResult>;
/**
* Dispose of resources
*/
dispose(): Promise<void>;
}
// =============================================================================
// node-llama-cpp Implementation
// =============================================================================
export type LlamaCppConfig = {
embedModel?: string;
generateModel?: string;
rerankModel?: string;
modelCacheDir?: string;
/**
* Context size used for query expansion generation contexts.
* Default: 2048. Can also be set via QMD_EXPAND_CONTEXT_SIZE.
*/
expandContextSize?: number;
/**
* Inactivity timeout in ms before unloading contexts (default: 2 minutes, 0 to disable).
*
* Per node-llama-cpp lifecycle guidance, we prefer keeping models loaded and only disposing
* contexts when idle, since contexts (and their sequences) are the heavy per-session objects.
* @see https://node-llama-cpp.withcat.ai/guide/objects-lifecycle
*/
inactivityTimeoutMs?: number;
/**
* Whether to dispose models on inactivity (default: false).
*
* Keeping models loaded avoids repeated VRAM thrash; set to true only if you need aggressive
* memory reclaim.
*/
disposeModelsOnInactivity?: boolean;
};
/**
* LLM implementation using node-llama-cpp
*/
// Default inactivity timeout: 5 minutes (keep models warm during typical search sessions)
const DEFAULT_INACTIVITY_TIMEOUT_MS = 5 * 60 * 1000;
const DEFAULT_EXPAND_CONTEXT_SIZE = 2048;
type LlamaGpuMode = "auto" | "metal" | "vulkan" | "cuda" | false;
export function resolveLlamaGpuMode(envValue = process.env.QMD_LLAMA_GPU): LlamaGpuMode {
const normalized = envValue?.trim().toLowerCase() ?? "";
if (!normalized) return "auto";
if (["false", "off", "none", "disable", "disabled", "0"].includes(normalized)) return false;
if (normalized === "metal" || normalized === "vulkan" || normalized === "cuda") return normalized;
process.stderr.write(`QMD Warning: invalid QMD_LLAMA_GPU="${envValue}", using auto GPU selection.\n`);
return "auto";
}
function resolveExpandContextSize(configValue?: number): number {
if (configValue !== undefined) {
if (!Number.isInteger(configValue) || configValue <= 0) {
throw new Error(`Invalid expandContextSize: ${configValue}. Must be a positive integer.`);
}
return configValue;
}
const envValue = process.env.QMD_EXPAND_CONTEXT_SIZE?.trim();
if (!envValue) return DEFAULT_EXPAND_CONTEXT_SIZE;
const parsed = Number.parseInt(envValue, 10);
if (!Number.isInteger(parsed) || parsed <= 0) {
process.stderr.write(
`QMD Warning: invalid QMD_EXPAND_CONTEXT_SIZE="${envValue}", using default ${DEFAULT_EXPAND_CONTEXT_SIZE}.\n`
);
return DEFAULT_EXPAND_CONTEXT_SIZE;
}
return parsed;
}
export class LlamaCpp implements LLM {
private readonly _ciMode = !!process.env.CI;
private llama: Llama | null = null;
private embedModel: LlamaModel | null = null;
private embedContexts: LlamaEmbeddingContext[] = [];
private generateModel: LlamaModel | null = null;
private rerankModel: LlamaModel | null = null;
private rerankContexts: Awaited<ReturnType<LlamaModel["createRankingContext"]>>[] = [];
private embedModelUri: string;
private generateModelUri: string;
private rerankModelUri: string;
private modelCacheDir: string;
private expandContextSize: number;
// Ensure we don't load the same model/context concurrently (which can allocate duplicate VRAM).
private embedModelLoadPromise: Promise<LlamaModel> | null = null;
private generateModelLoadPromise: Promise<LlamaModel> | null = null;
private rerankModelLoadPromise: Promise<LlamaModel> | null = null;
// Inactivity timer for auto-unloading models
private inactivityTimer: ReturnType<typeof setTimeout> | null = null;
private inactivityTimeoutMs: number;
private disposeModelsOnInactivity: boolean;
// Track disposal state to prevent double-dispose
private disposed = false;
constructor(config: LlamaCppConfig = {}) {
this.embedModelUri = config.embedModel || process.env.QMD_EMBED_MODEL || DEFAULT_EMBED_MODEL;
this.generateModelUri = config.generateModel || process.env.QMD_GENERATE_MODEL || DEFAULT_GENERATE_MODEL;
this.rerankModelUri = config.rerankModel || process.env.QMD_RERANK_MODEL || DEFAULT_RERANK_MODEL;
this.modelCacheDir = config.modelCacheDir || MODEL_CACHE_DIR;
this.expandContextSize = resolveExpandContextSize(config.expandContextSize);
this.inactivityTimeoutMs = config.inactivityTimeoutMs ?? DEFAULT_INACTIVITY_TIMEOUT_MS;
this.disposeModelsOnInactivity = config.disposeModelsOnInactivity ?? false;
}
get embedModelName(): string {
return this.embedModelUri;
}
/**
* Reset the inactivity timer. Called after each model operation.
* When timer fires, models are unloaded to free memory (if no active sessions).
*/
private touchActivity(): void {
// Clear existing timer
if (this.inactivityTimer) {
clearTimeout(this.inactivityTimer);
this.inactivityTimer = null;
}
// Only set timer if we have disposable contexts and timeout is enabled
if (this.inactivityTimeoutMs > 0 && this.hasLoadedContexts()) {
this.inactivityTimer = setTimeout(() => {
// Check if session manager allows unloading
// canUnloadLLM is defined later in this file - it checks the session manager
// We use dynamic import pattern to avoid circular dependency issues
if (typeof canUnloadLLM === 'function' && !canUnloadLLM()) {
// Active sessions/operations - reschedule timer
this.touchActivity();
return;
}
this.unloadIdleResources().catch(err => {
console.error("Error unloading idle resources:", err);
});
}, this.inactivityTimeoutMs);
// Don't keep process alive just for this timer
this.inactivityTimer.unref();
}
}
/**
* Check if any contexts are currently loaded (and therefore worth unloading on inactivity).
*/
private hasLoadedContexts(): boolean {
return !!(this.embedContexts.length > 0 || this.rerankContexts.length > 0);
}
/**
* Unload idle resources but keep the instance alive for future use.
*
* By default, this disposes contexts (and their dependent sequences), while keeping models loaded.
* This matches the intended lifecycle: model → context → sequence, where contexts are per-session.
*/
async unloadIdleResources(): Promise<void> {
// Don't unload if already disposed
if (this.disposed) {
return;
}
// Clear timer
if (this.inactivityTimer) {
clearTimeout(this.inactivityTimer);
this.inactivityTimer = null;
}
// Dispose contexts first
for (const ctx of this.embedContexts) {
await ctx.dispose();
}
this.embedContexts = [];
for (const ctx of this.rerankContexts) {
await ctx.dispose();
}
this.rerankContexts = [];
// Optionally dispose models too (opt-in)
if (this.disposeModelsOnInactivity) {
if (this.embedModel) {
await this.embedModel.dispose();
this.embedModel = null;
}
if (this.generateModel) {
await this.generateModel.dispose();
this.generateModel = null;
}
if (this.rerankModel) {
await this.rerankModel.dispose();
this.rerankModel = null;
}
// Reset load promises so models can be reloaded later
this.embedModelLoadPromise = null;
this.generateModelLoadPromise = null;
this.rerankModelLoadPromise = null;
}
// Note: We keep llama instance alive - it's lightweight
}
/**
* Ensure model cache directory exists
*/
private ensureModelCacheDir(): void {
if (!existsSync(this.modelCacheDir)) {
mkdirSync(this.modelCacheDir, { recursive: true });
}
}
/**
* Initialize the llama instance (lazy)
*/
private async ensureLlama(allowBuild = true): Promise<Llama> {
if (!this.llama) {
const gpuMode = resolveLlamaGpuMode();
const loadLlama = async (gpu: LlamaGpuMode) =>
await getLlama({
build: allowBuild ? "autoAttempt" : "never",
logLevel: LlamaLogLevel.error,
gpu,
skipDownload: !allowBuild,
});
let llama: Llama;
if (gpuMode === false) {
llama = await loadLlama(false);
} else {
try {
llama = await loadLlama(gpuMode);
} catch (err) {
// GPU backend (e.g. Vulkan on headless/driverless machines) can throw at init.
// Fall back to CPU so qmd still works.
process.stderr.write(
`QMD Warning: GPU init failed${gpuMode === "auto" ? "" : ` for QMD_LLAMA_GPU=${gpuMode}`} (${err instanceof Error ? err.message : String(err)}), falling back to CPU.\n`
);
llama = await loadLlama(false);
}
}
if (llama.gpu === false) {
process.stderr.write(
"QMD Warning: no GPU acceleration, running on CPU (slow). Run 'qmd status' for details.\n"
);
}
this.llama = llama;
}
return this.llama;
}
/**
* Resolve a model URI to a local path, downloading if needed
*/
private async resolveModel(modelUri: string): Promise<string> {
this.ensureModelCacheDir();
// resolveModelFile handles HF URIs and downloads to the cache dir
return await resolveModelFile(modelUri, this.modelCacheDir);
}
/**
* Load embedding model (lazy)
*/
private async ensureEmbedModel(): Promise<LlamaModel> {
if (this.embedModel) {
return this.embedModel;
}
if (this.embedModelLoadPromise) {
return await this.embedModelLoadPromise;
}
this.embedModelLoadPromise = (async () => {
const llama = await this.ensureLlama();
const modelPath = await this.resolveModel(this.embedModelUri);
const model = await llama.loadModel({ modelPath });
this.embedModel = model;
// Model loading counts as activity - ping to keep alive
this.touchActivity();
return model;
})();
try {
return await this.embedModelLoadPromise;
} finally {
// Keep the resolved model cached; clear only the in-flight promise.
this.embedModelLoadPromise = null;
}
}
/**
* Compute how many parallel contexts to create.
*
* GPU: constrained by VRAM (25% of free, capped at 8).
* CPU: constrained by cores. Splitting threads across contexts enables
* true parallelism (each context runs on its own cores). Use at most
* half the math cores, with at least 4 threads per context.
*/
private async computeParallelism(perContextMB: number): Promise<number> {
const llama = await this.ensureLlama();
if (llama.gpu) {
try {
const vram = await llama.getVramState();
const freeMB = vram.free / (1024 * 1024);
const maxByVram = Math.floor((freeMB * 0.25) / perContextMB);
return Math.max(1, Math.min(8, maxByVram));
} catch {
return 2;
}
}
// CPU: split cores across contexts. At least 4 threads per context.
const cores = llama.cpuMathCores || 4;
const maxContexts = Math.floor(cores / 4);
return Math.max(1, Math.min(4, maxContexts));
}
/**
* Get the number of threads each context should use, given N parallel contexts.
* Splits available math cores evenly across contexts.
*/
private async threadsPerContext(parallelism: number): Promise<number> {
const llama = await this.ensureLlama();
if (llama.gpu) return 0; // GPU: let the library decide
const cores = llama.cpuMathCores || 4;
return Math.max(1, Math.floor(cores / parallelism));
}
/**
* Load embedding contexts (lazy). Creates multiple for parallel embedding.
* Uses promise guard to prevent concurrent context creation race condition.
*/
private embedContextsCreatePromise: Promise<LlamaEmbeddingContext[]> | null = null;
private async ensureEmbedContexts(): Promise<LlamaEmbeddingContext[]> {
if (this.embedContexts.length > 0) {
this.touchActivity();
return this.embedContexts;
}
if (this.embedContextsCreatePromise) {
return await this.embedContextsCreatePromise;
}
this.embedContextsCreatePromise = (async () => {
const model = await this.ensureEmbedModel();
// Embed contexts are ~143 MB each (nomic-embed 2048 ctx)
const n = await this.computeParallelism(150);
const threads = await this.threadsPerContext(n);
for (let i = 0; i < n; i++) {
try {
this.embedContexts.push(await model.createEmbeddingContext({
contextSize: LlamaCpp.EMBED_CONTEXT_SIZE,
...(threads > 0 ? { threads } : {}),
}));
} catch {
if (this.embedContexts.length === 0) throw new Error("Failed to create any embedding context");
break;
}
}
this.touchActivity();
return this.embedContexts;
})();
try {
return await this.embedContextsCreatePromise;
} finally {
this.embedContextsCreatePromise = null;
}
}
/**
* Get a single embed context (for single-embed calls). Uses first from pool.
*/
private async ensureEmbedContext(): Promise<LlamaEmbeddingContext> {
const contexts = await this.ensureEmbedContexts();
return contexts[0]!;
}
/**
* Load generation model (lazy) - context is created fresh per call
*/
private async ensureGenerateModel(): Promise<LlamaModel> {
if (!this.generateModel) {
if (this.generateModelLoadPromise) {
return await this.generateModelLoadPromise;
}
this.generateModelLoadPromise = (async () => {
const llama = await this.ensureLlama();
const modelPath = await this.resolveModel(this.generateModelUri);
const model = await llama.loadModel({ modelPath });
this.generateModel = model;
return model;
})();
try {
await this.generateModelLoadPromise;
} finally {
this.generateModelLoadPromise = null;
}
}
this.touchActivity();
if (!this.generateModel) {
throw new Error("Generate model not loaded");
}
return this.generateModel;
}
/**
* Load rerank model (lazy)
*/
private async ensureRerankModel(): Promise<LlamaModel> {
if (this.rerankModel) {
return this.rerankModel;
}
if (this.rerankModelLoadPromise) {
return await this.rerankModelLoadPromise;
}
this.rerankModelLoadPromise = (async () => {
const llama = await this.ensureLlama();
const modelPath = await this.resolveModel(this.rerankModelUri);
const model = await llama.loadModel({ modelPath });
this.rerankModel = model;
// Model loading counts as activity - ping to keep alive
this.touchActivity();
return model;
})();
try {
return await this.rerankModelLoadPromise;
} finally {
this.rerankModelLoadPromise = null;
}
}
/**
* Load rerank contexts (lazy). Creates multiple contexts for parallel ranking.
* Each context has its own sequence, so they can evaluate independently.
*
* Tuning choices:
* - contextSize 1024: reranking chunks are ~800 tokens max, 1024 is plenty
* - flashAttention: ~20% less VRAM per context (568 vs 711 MB)
* - Combined: drops from 11.6 GB (auto, no flash) to 568 MB per context (20×)
*/
// Qwen3 reranker template adds ~200 tokens overhead (system prompt, tags, etc.)
// Default 2048 was too small for longer documents (e.g. session transcripts,
// CJK text, or large markdown files) — callers hit "input lengths exceed
// context size" errors even after truncation because the overhead estimate
// was insufficient. 4096 comfortably fits the largest real-world chunks
// while staying well below the 40 960-token auto size.
// Override with QMD_RERANK_CONTEXT_SIZE env var if you need more headroom.
private static readonly RERANK_CONTEXT_SIZE: number = (() => {
const v = parseInt(process.env.QMD_RERANK_CONTEXT_SIZE ?? "", 10);
return Number.isFinite(v) && v > 0 ? v : 4096;
})();
private static readonly EMBED_CONTEXT_SIZE: number = (() => {
const v = parseInt(process.env.QMD_EMBED_CONTEXT_SIZE ?? "", 10);
return Number.isFinite(v) && v > 0 ? v : 2048;
})();
private async ensureRerankContexts(): Promise<Awaited<ReturnType<LlamaModel["createRankingContext"]>>[]> {
if (this.rerankContexts.length === 0) {
const model = await this.ensureRerankModel();
// ~960 MB per context with flash attention at contextSize 2048
const n = Math.min(await this.computeParallelism(1000), 4);
const threads = await this.threadsPerContext(n);
for (let i = 0; i < n; i++) {
try {
this.rerankContexts.push(await model.createRankingContext({
contextSize: LlamaCpp.RERANK_CONTEXT_SIZE,
flashAttention: true,
...(threads > 0 ? { threads } : {}),
} as any));
} catch {
if (this.rerankContexts.length === 0) {
// Flash attention might not be supported — retry without it
try {
this.rerankContexts.push(await model.createRankingContext({
contextSize: LlamaCpp.RERANK_CONTEXT_SIZE,
...(threads > 0 ? { threads } : {}),
}));
} catch {
throw new Error("Failed to create any rerank context");
}
}
break;
}
}
}
this.touchActivity();
return this.rerankContexts;
}
// ==========================================================================
// Tokenization
// ==========================================================================
/**
* Tokenize text using the embedding model's tokenizer
* Returns tokenizer tokens (opaque type from node-llama-cpp)
*/
async tokenize(text: string): Promise<readonly LlamaToken[]> {
await this.ensureEmbedContext(); // Ensure model is loaded
if (!this.embedModel) {
throw new Error("Embed model not loaded");
}
return this.embedModel.tokenize(text);
}
/**
* Count tokens in text using the embedding model's tokenizer
*/
async countTokens(text: string): Promise<number> {
const tokens = await this.tokenize(text);
return tokens.length;
}
/**
* Detokenize token IDs back to text
*/
async detokenize(tokens: readonly LlamaToken[]): Promise<string> {
await this.ensureEmbedContext();
if (!this.embedModel) {
throw new Error("Embed model not loaded");
}
return this.embedModel.detokenize(tokens);
}
// ==========================================================================
// Core API methods
// ==========================================================================
/**
* Truncate text to fit within the embedding model's context window.
* Uses the model's own tokenizer for accurate token counting, then
* detokenizes back to text if truncation is needed.
* Returns the (possibly truncated) text and whether truncation occurred.
*/
private async truncateToContextSize(text: string): Promise<{ text: string; truncated: boolean }> {
if (!this.embedModel) return { text, truncated: false };
const maxTokens = this.embedModel.trainContextSize;
if (maxTokens <= 0) return { text, truncated: false };
const tokens = this.embedModel.tokenize(text);
if (tokens.length <= maxTokens) return { text, truncated: false };
// Leave a small margin (4 tokens) for BOS/EOS overhead
const safeLimit = Math.max(1, maxTokens - 4);
const truncatedTokens = tokens.slice(0, safeLimit);
const truncatedText = this.embedModel.detokenize(truncatedTokens);
return { text: truncatedText, truncated: true };
}
async embed(text: string, options: EmbedOptions = {}): Promise<EmbeddingResult | null> {
// Ping activity at start to keep models alive during this operation
this.touchActivity();
try {
const context = await this.ensureEmbedContext();
// Guard: truncate text that exceeds model context window to prevent GGML crash
const { text: safeText, truncated } = await this.truncateToContextSize(text);
if (truncated) {
console.warn(`⚠ Text truncated to fit embedding context (${this.embedModel?.trainContextSize} tokens)`);
}
const embedding = await context.getEmbeddingFor(safeText);
return {
embedding: Array.from(embedding.vector),
model: options.model ?? this.embedModelUri,
};
} catch (error) {
console.error("Embedding error:", error);
return null;
}
}
/**
* Batch embed multiple texts efficiently
* Uses Promise.all for parallel embedding - node-llama-cpp handles batching internally
*/
async embedBatch(texts: string[], options: EmbedOptions = {}): Promise<(EmbeddingResult | null)[]> {
if (this._ciMode) throw new Error("LLM operations are disabled in CI (set CI=true)");
// Ping activity at start to keep models alive during this operation
this.touchActivity();
if (texts.length === 0) return [];
try {
const contexts = await this.ensureEmbedContexts();
const n = contexts.length;
if (n === 1) {
// Single context: sequential (no point splitting)
const context = contexts[0]!;
const embeddings: ({ embedding: number[]; model: string } | null)[] = [];
for (const text of texts) {
try {
const { text: safeText, truncated } = await this.truncateToContextSize(text);
if (truncated) {
console.warn(`⚠ Batch text truncated to fit embedding context (${this.embedModel?.trainContextSize} tokens)`);
}
const embedding = await context.getEmbeddingFor(safeText);
this.touchActivity();
embeddings.push({ embedding: Array.from(embedding.vector), model: options.model ?? this.embedModelUri });
} catch (err) {
console.error("Embedding error for text:", err);
embeddings.push(null);
}
}
return embeddings;
}
// Multiple contexts: split texts across contexts for parallel evaluation
const chunkSize = Math.ceil(texts.length / n);
const chunks = Array.from({ length: n }, (_, i) =>
texts.slice(i * chunkSize, (i + 1) * chunkSize)
);
const chunkResults = await Promise.all(
chunks.map(async (chunk, i) => {
const ctx = contexts[i]!;
const results: (EmbeddingResult | null)[] = [];
for (const text of chunk) {
try {
const { text: safeText, truncated } = await this.truncateToContextSize(text);
if (truncated) {
console.warn(`⚠ Batch text truncated to fit embedding context (${this.embedModel?.trainContextSize} tokens)`);
}
const embedding = await ctx.getEmbeddingFor(safeText);
this.touchActivity();
results.push({ embedding: Array.from(embedding.vector), model: options.model ?? this.embedModelUri });
} catch (err) {
console.error("Embedding error for text:", err);
results.push(null);
}
}
return results;
})
);
return chunkResults.flat();
} catch (error) {
console.error("Batch embedding error:", error);
return texts.map(() => null);
}
}
async generate(prompt: string, options: GenerateOptions = {}): Promise<GenerateResult | null> {
if (this._ciMode) throw new Error("LLM operations are disabled in CI (set CI=true)");
// Ping activity at start to keep models alive during this operation
this.touchActivity();
// Ensure model is loaded
await this.ensureGenerateModel();
// Create fresh context -> sequence -> session for each call
const context = await this.generateModel!.createContext();
const sequence = context.getSequence();
const session = new LlamaChatSession({ contextSequence: sequence });
const maxTokens = options.maxTokens ?? 150;
// Qwen3 recommends temp=0.7, topP=0.8, topK=20 for non-thinking mode
// DO NOT use greedy decoding (temp=0) - causes repetition loops
const temperature = options.temperature ?? 0.7;
let result = "";
try {
await session.prompt(prompt, {
maxTokens,
temperature,
topK: 20,
topP: 0.8,
onTextChunk: (text) => {
result += text;
},
});
return {
text: result,
model: this.generateModelUri,
done: true,
};
} finally {
// Dispose context (which disposes dependent sequences/sessions per lifecycle rules)
await context.dispose();
}
}
async modelExists(modelUri: string): Promise<ModelInfo> {
// For HuggingFace URIs, we assume they exist
// For local paths, check if file exists
if (modelUri.startsWith("hf:")) {
return { name: modelUri, exists: true };
}
const exists = existsSync(modelUri);
return {
name: modelUri,
exists,
path: exists ? modelUri : undefined,
};
}
// ==========================================================================
// High-level abstractions
// ==========================================================================
async expandQuery(query: string, options: { context?: string, includeLexical?: boolean, intent?: string } = {}): Promise<Queryable[]> {
if (this._ciMode) throw new Error("LLM operations are disabled in CI (set CI=true)");
// Ping activity at start to keep models alive during this operation
this.touchActivity();
const llama = await this.ensureLlama();
await this.ensureGenerateModel();
const includeLexical = options.includeLexical ?? true;
const context = options.context;
const grammar = await llama.createGrammar({
grammar: `
root ::= line+
line ::= type ": " content "\\n"
type ::= "lex" | "vec" | "hyde"
content ::= [^\\n]+
`
});
const intent = options.intent;
const prompt = intent
? `/no_think Expand this search query: ${query}\nQuery intent: ${intent}`
: `/no_think Expand this search query: ${query}`;
// Create a bounded context for expansion to prevent large default VRAM allocations.
const genContext = await this.generateModel!.createContext({
contextSize: this.expandContextSize,
});
const sequence = genContext.getSequence();
const session = new LlamaChatSession({ contextSequence: sequence });
try {
// Qwen3 recommended settings for non-thinking mode:
// temp=0.7, topP=0.8, topK=20, presence_penalty for repetition
// DO NOT use greedy decoding (temp=0) - causes infinite loops
const result = await session.prompt(prompt, {
grammar,
maxTokens: 600,
temperature: 0.7,
topK: 20,
topP: 0.8,
repeatPenalty: {
lastTokens: 64,
presencePenalty: 0.5,
},
});
const lines = result.trim().split("\n");
const queryLower = query.toLowerCase();
const queryTerms = queryLower.replace(/[^a-z0-9\s]/g, " ").split(/\s+/).filter(Boolean);
const hasQueryTerm = (text: string): boolean => {
const lower = text.toLowerCase();
if (queryTerms.length === 0) return true;
return queryTerms.some(term => lower.includes(term));
};
const queryables: Queryable[] = lines.map(line => {
const colonIdx = line.indexOf(":");
if (colonIdx === -1) return null;
const type = line.slice(0, colonIdx).trim();
if (type !== 'lex' && type !== 'vec' && type !== 'hyde') return null;
const text = line.slice(colonIdx + 1).trim();
if (!hasQueryTerm(text)) return null;
return { type: type as QueryType, text };
}).filter((q): q is Queryable => q !== null);
// Filter out lex entries if not requested
const filtered = includeLexical ? queryables : queryables.filter(q => q.type !== 'lex');
if (filtered.length > 0) return filtered;
const fallback: Queryable[] = [
{ type: 'hyde', text: `Information about ${query}` },
{ type: 'lex', text: query },
{ type: 'vec', text: query },
];
return includeLexical ? fallback : fallback.filter(q => q.type !== 'lex');
} catch (error) {
console.error("Structured query expansion failed:", error);
// Fallback to original query
const fallback: Queryable[] = [{ type: 'vec', text: query }];
if (includeLexical) fallback.unshift({ type: 'lex', text: query });
return fallback;
} finally {
await genContext.dispose();
}
}
// Qwen3 reranker chat template overhead (system prompt, tags, separators).
// Measured at ~350 tokens on real queries; use 512 as a safe upper bound so
// the truncation budget never lets a document slip past the context limit.
private static readonly RERANK_TEMPLATE_OVERHEAD = 512;
private static readonly RERANK_TARGET_DOCS_PER_CONTEXT = 10;
async rerank(
query: string,
documents: RerankDocument[],
options: RerankOptions = {}
): Promise<RerankResult> {
if (this._ciMode) throw new Error("LLM operations are disabled in CI (set CI=true)");
// Ping activity at start to keep models alive during this operation
this.touchActivity();
const contexts = await this.ensureRerankContexts();
const model = await this.ensureRerankModel();
// Truncate documents that would exceed the rerank context size.
// Budget = contextSize - template overhead - query tokens
const queryTokens = model.tokenize(query).length;
const maxDocTokens = LlamaCpp.RERANK_CONTEXT_SIZE - LlamaCpp.RERANK_TEMPLATE_OVERHEAD - queryTokens;
const truncationCache = new Map<string, string>();
const truncatedDocs = documents.map((doc) => {
const cached = truncationCache.get(doc.text);
if (cached !== undefined) {
return cached === doc.text ? doc : { ...doc, text: cached };
}
const tokens = model.tokenize(doc.text);
const truncatedText = tokens.length <= maxDocTokens
? doc.text
: model.detokenize(tokens.slice(0, maxDocTokens));
truncationCache.set(doc.text, truncatedText);
if (truncatedText === doc.text) return doc;
return { ...doc, text: truncatedText };
});
// Deduplicate identical effective texts before scoring.
// This avoids redundant work for repeated chunks and fixes collisions where
// multiple docs map to the same chunk text.
const textToDocs = new Map<string, { file: string; index: number }[]>();
truncatedDocs.forEach((doc, index) => {
const existing = textToDocs.get(doc.text);
if (existing) {
existing.push({ file: doc.file, index });
} else {
textToDocs.set(doc.text, [{ file: doc.file, index }]);
}
});
// Extract just the text for ranking
const texts = Array.from(textToDocs.keys());
// Split documents across contexts for parallel evaluation.
// Each context has its own sequence with a lock, so parallelism comes
// from multiple contexts evaluating different chunks simultaneously.
const activeContextCount = Math.max(
1,
Math.min(
contexts.length,
Math.ceil(texts.length / LlamaCpp.RERANK_TARGET_DOCS_PER_CONTEXT)
)
);
const activeContexts = contexts.slice(0, activeContextCount);
const chunkSize = Math.ceil(texts.length / activeContexts.length);
const chunks = Array.from({ length: activeContexts.length }, (_, i) =>
texts.slice(i * chunkSize, (i + 1) * chunkSize)
).filter(chunk => chunk.length > 0);
const allScores = await Promise.all(
chunks.map((chunk, i) => activeContexts[i]!.rankAll(query, chunk))
);
// Reassemble scores in original order and sort
const flatScores = allScores.flat();
const ranked = texts
.map((text, i) => ({ document: text, score: flatScores[i]! }))
.sort((a, b) => b.score - a.score);
// Map back to our result format.
const results: RerankDocumentResult[] = [];
for (const item of ranked) {
const docInfos = textToDocs.get(item.document) ?? [];
for (const docInfo of docInfos) {
results.push({
file: docInfo.file,
score: item.score,
index: docInfo.index,
});
}
}
return {
results,
model: this.rerankModelUri,
};
}
/**
* Get device/GPU info for status display.
* Initializes llama if not already done.
*/
async getDeviceInfo(options: { allowBuild?: boolean } = {}): Promise<{
gpu: string | false;
gpuOffloading: boolean;
gpuDevices: string[];
vram?: { total: number; used: number; free: number };
cpuCores: number;
}> {
const llama = await this.ensureLlama(options.allowBuild ?? true);
const gpuDevices = await llama.getGpuDeviceNames();
let vram: { total: number; used: number; free: number } | undefined;
if (llama.gpu) {
try {
const state = await llama.getVramState();
vram = { total: state.total, used: state.used, free: state.free };
} catch { /* no vram info */ }
}
return {
gpu: llama.gpu,
gpuOffloading: llama.supportsGpuOffloading,
gpuDevices,
vram,
cpuCores: llama.cpuMathCores,
};
}
async dispose(): Promise<void> {
// Prevent double-dispose
if (this.disposed) {
return;
}
this.disposed = true;
// Clear inactivity timer
if (this.inactivityTimer) {
clearTimeout(this.inactivityTimer);
this.inactivityTimer = null;
}
// Disposing llama cascades to models and contexts automatically
// See: https://node-llama-cpp.withcat.ai/guide/objects-lifecycle
// Note: llama.dispose() can hang indefinitely, so we use a timeout
if (this.llama) {
const disposePromise = this.llama.dispose();
const timeoutPromise = new Promise<void>((resolve) => setTimeout(resolve, 1000));
await Promise.race([disposePromise, timeoutPromise]);
}
// Clear references
this.embedContexts = [];
this.rerankContexts = [];
this.embedModel = null;
this.generateModel = null;
this.rerankModel = null;
this.llama = null;
// Clear any in-flight load/create promises
this.embedModelLoadPromise = null;
this.embedContextsCreatePromise = null;
this.generateModelLoadPromise = null;
this.rerankModelLoadPromise = null;
}
}
// =============================================================================
// Session Management Layer
// =============================================================================
/**
* Manages LLM session lifecycle with reference counting.
* Coordinates with LlamaCpp idle timeout to prevent disposal during active sessions.
*/
class LLMSessionManager {
private llm: LlamaCpp;
private _activeSessionCount = 0;
private _inFlightOperations = 0;
constructor(llm: LlamaCpp) {
this.llm = llm;
}
get activeSessionCount(): number {
return this._activeSessionCount;
}
get inFlightOperations(): number {
return this._inFlightOperations;
}
/**
* Returns true only when both session count and in-flight operations are 0.
* Used by LlamaCpp to determine if idle unload is safe.
*/
canUnload(): boolean {
return this._activeSessionCount === 0 && this._inFlightOperations === 0;
}
acquire(): void {
this._activeSessionCount++;
}
release(): void {
this._activeSessionCount = Math.max(0, this._activeSessionCount - 1);
}
operationStart(): void {
this._inFlightOperations++;
}
operationEnd(): void {
this._inFlightOperations = Math.max(0, this._inFlightOperations - 1);
}
getLlamaCpp(): LlamaCpp {
return this.llm;
}
}
/**
* Error thrown when an operation is attempted on a released or aborted session.
*/
export class SessionReleasedError extends Error {
constructor(message = "LLM session has been released or aborted") {
super(message);
this.name = "SessionReleasedError";
}
}
/**
* Scoped LLM session with automatic lifecycle management.
* Wraps LlamaCpp methods with operation tracking and abort handling.
*/
class LLMSession implements ILLMSession {
private manager: LLMSessionManager;
private released = false;
private abortController: AbortController;
private maxDurationTimer: ReturnType<typeof setTimeout> | null = null;
private name: string;
constructor(manager: LLMSessionManager, options: LLMSessionOptions = {}) {
this.manager = manager;
this.name = options.name || "unnamed";
this.abortController = new AbortController();
// Link external abort signal if provided
if (options.signal) {
if (options.signal.aborted) {
this.abortController.abort(options.signal.reason);
} else {
options.signal.addEventListener("abort", () => {
this.abortController.abort(options.signal!.reason);
}, { once: true });
}
}
// Set up max duration timer
const maxDuration = options.maxDuration ?? 10 * 60 * 1000; // Default 10 minutes
if (maxDuration > 0) {
this.maxDurationTimer = setTimeout(() => {
this.abortController.abort(new Error(`Session "${this.name}" exceeded max duration of ${maxDuration}ms`));
}, maxDuration);
this.maxDurationTimer.unref(); // Don't keep process alive
}
// Acquire session lease
this.manager.acquire();
}
get isValid(): boolean {
return !this.released && !this.abortController.signal.aborted;
}
get signal(): AbortSignal {
return this.abortController.signal;
}
/**
* Release the session and decrement ref count.
* Called automatically by withLLMSession when the callback completes.
*/
release(): void {
if (this.released) return;
this.released = true;
if (this.maxDurationTimer) {
clearTimeout(this.maxDurationTimer);
this.maxDurationTimer = null;
}
this.abortController.abort(new Error("Session released"));
this.manager.release();
}
/**
* Wrap an operation with tracking and abort checking.
*/
private async withOperation<T>(fn: () => Promise<T>): Promise<T> {
if (!this.isValid) {
throw new SessionReleasedError();
}
this.manager.operationStart();
try {
// Check abort before starting
if (this.abortController.signal.aborted) {
throw new SessionReleasedError(
this.abortController.signal.reason?.message || "Session aborted"
);
}
return await fn();
} finally {
this.manager.operationEnd();
}
}
async embed(text: string, options?: EmbedOptions): Promise<EmbeddingResult | null> {
return this.withOperation(() => this.manager.getLlamaCpp().embed(text, options));
}
async embedBatch(texts: string[], options?: EmbedOptions): Promise<(EmbeddingResult | null)[]> {
return this.withOperation(() => this.manager.getLlamaCpp().embedBatch(texts, options));
}
async expandQuery(
query: string,
options?: { context?: string; includeLexical?: boolean }
): Promise<Queryable[]> {
return this.withOperation(() => this.manager.getLlamaCpp().expandQuery(query, options));
}
async rerank(
query: string,
documents: RerankDocument[],
options?: RerankOptions
): Promise<RerankResult> {
return this.withOperation(() => this.manager.getLlamaCpp().rerank(query, documents, options));
}
}
// Session manager for the default LlamaCpp instance
let defaultSessionManager: LLMSessionManager | null = null;
/**
* Get the session manager for the default LlamaCpp instance.
*/
function getSessionManager(): LLMSessionManager {
const llm = getDefaultLlamaCpp();
if (!defaultSessionManager || defaultSessionManager.getLlamaCpp() !== llm) {
defaultSessionManager = new LLMSessionManager(llm);
}
return defaultSessionManager;
}
/**
* Execute a function with a scoped LLM session.
* The session provides lifecycle guarantees - resources won't be disposed mid-operation.
*
* @example
* ```typescript
* await withLLMSession(async (session) => {
* const expanded = await session.expandQuery(query);
* const embeddings = await session.embedBatch(texts);
* const reranked = await session.rerank(query, docs);
* return reranked;
* }, { maxDuration: 10 * 60 * 1000, name: 'querySearch' });
* ```
*/
export async function withLLMSession<T>(
fn: (session: ILLMSession) => Promise<T>,
options?: LLMSessionOptions
): Promise<T> {
const manager = getSessionManager();
const session = new LLMSession(manager, options);
try {
return await fn(session);
} finally {
session.release();
}
}
/**
* Execute a function with a scoped LLM session using a specific LlamaCpp instance.
* Unlike withLLMSession, this does not use the global singleton.
*/
export async function withLLMSessionForLlm<T>(
llm: LlamaCpp,
fn: (session: ILLMSession) => Promise<T>,
options?: LLMSessionOptions
): Promise<T> {
const manager = new LLMSessionManager(llm);
const session = new LLMSession(manager, options);
try {
return await fn(session);
} finally {
session.release();
}
}
/**
* Check if idle unload is safe (no active sessions or operations).
* Used internally by LlamaCpp idle timer.
*/
export function canUnloadLLM(): boolean {
if (!defaultSessionManager) return true;
return defaultSessionManager.canUnload();
}
// =============================================================================
// Singleton for default LlamaCpp instance
// =============================================================================
let defaultLlamaCpp: LlamaCpp | null = null;
/**
* Get the default LlamaCpp instance (creates one if needed)
*/
export function getDefaultLlamaCpp(): LlamaCpp {
if (!defaultLlamaCpp) {
defaultLlamaCpp = new LlamaCpp();
}
return defaultLlamaCpp;
}
/**
* Set a custom default LlamaCpp instance (useful for testing)
*/
export function setDefaultLlamaCpp(llm: LlamaCpp | null): void {
defaultLlamaCpp = llm;
}
/**
* Dispose the default LlamaCpp instance if it exists.
* Call this before process exit to prevent NAPI crashes.
*/
export async function disposeDefaultLlamaCpp(): Promise<void> {
if (defaultLlamaCpp) {
await defaultLlamaCpp.dispose();
defaultLlamaCpp = null;
}
}