@they-juanreina/compost-cli 0.1.0-rc.1

package/dist/lib/provisionNative.js

@@ -0,0 +1,105 @@
+import { spawnSync } from 'node:child_process';
+import { existsSync } from 'node:fs';
+import { join } from 'node:path';
+import { CompostError } from '../errors.js';
+import { findRepoTranscriberDir, managedVenvDir } from './nativeRuntime.js';
+/**
+ * Provision the managed native transcription venv (#183) so `compost transcribe`
+ * runs native (Parakeet + pyannote on Metal) with zero flags on a fresh
+ * Apple-Silicon machine. Side-effectful, so it's an explicit opt-in
+ * (`compost setup --provision-native`), separate from the read-only doctor.
+ *
+ * The dep list mirrors the transcriber's `[native]` extra in pyproject.toml; a
+ * unit test asserts they stay in sync. We install the deps (not the package
+ * editable) to avoid depending on a build backend — the `app` package is found
+ * at run time via the entrypoint's cwd.
+ */
+/** Native ASR deps — keep in sync with transcriber/pyproject.toml `[native]`. */
+export const NATIVE_DEPS = [
+    'parakeet-mlx',
+    'pyannote.audio>=3.3',
+    'silero-vad',
+    'torchaudio>=2.3',
+    'ffmpeg-python',
+];
+const PY_CANDIDATES = ['python3.11', 'python3.12', 'python3'];
+const VERIFY_IMPORTS = 'import parakeet_mlx, pyannote.audio, silero_vad, torchaudio';
+const MIN_MINOR = 11;
+const MAX_MINOR = 12; // transcriber requires-python = >=3.11,<3.13
+const defaultExec = (cmd, args, opts) => {
+    const res = spawnSync(cmd, args, {
+        encoding: 'utf8',
+        // stream → live stdout (download progress) but still capture stderr so a
+        // failure has a diagnostic (inheriting stderr would null it out).
+        stdio: opts?.stream ? ['ignore', 'inherit', 'pipe'] : 'pipe',
+        maxBuffer: 64 * 1024 * 1024,
+    });
+    return { ok: !res.error && res.status === 0, stdout: res.stdout ?? '', stderr: res.stderr ?? '' };
+};
+function pyVersion(exec, bin) {
+    const r = exec(bin, ['--version']);
+    if (!r.ok)
+        return null;
+    const m = /Python (\d+)\.(\d+)/.exec(`${r.stdout} ${r.stderr}`);
+    return m ? [Number(m[1]), Number(m[2])] : null;
+}
+function resolvePython(exec, explicit) {
+    for (const c of explicit ? [explicit] : PY_CANDIDATES) {
+        const v = pyVersion(exec, c);
+        if (v && v[0] === 3 && v[1] >= MIN_MINOR && v[1] <= MAX_MINOR)
+            return c;
+    }
+    throw new CompostError('INVALID_INPUT', `no Python 3.${MIN_MINOR}–3.${MAX_MINOR} found (tried ${(explicit ? [explicit] : PY_CANDIDATES).join(', ')}). ` +
+        'Install one — e.g. `brew install python@3.11` — or pass --python-bin.');
+}
+/** Create + populate the managed native venv. Idempotent: a venv that already
+ * imports the deps is a no-op unless `force`. */
+export function provisionNativeVenv(opts = {}) {
+    const env = opts.env ?? process.env;
+    const exists = opts.exists ?? existsSync;
+    const exec = opts.exec ?? defaultExec;
+    const steps = [];
+    const transcriberDir = opts.transcriberDir ?? env.COMPOST_TRANSCRIBER_DIR ?? findRepoTranscriberDir(undefined, exists);
+    if (transcriberDir === undefined || !exists(join(transcriberDir, 'pyproject.toml'))) {
+        throw new CompostError('INVALID_INPUT', 'could not locate the transcriber/ package. Pass --transcriber-dir or set COMPOST_TRANSCRIBER_DIR.');
+    }
+    const venvDir = managedVenvDir(env);
+    const venvPython = join(venvDir, 'bin', 'python');
+    if (!opts.force && exists(venvPython) && exec(venvPython, ['-c', VERIFY_IMPORTS]).ok) {
+        return {
+            status: 'already-ready',
+            venvDir,
+            venvPython,
+            pythonBin: null,
+            transcriberDir,
+            steps: ['venv present; native deps already import'],
+        };
+    }
+    // Only resolve a host interpreter when we actually need to CREATE the venv;
+    // reinstalling deps into an existing venv uses that venv's python, not the host.
+    let pythonBin = null;
+    if (opts.force || !exists(venvPython)) {
+        pythonBin = resolvePython(exec, opts.pythonBin);
+        steps.push(`interpreter ${pythonBin}`);
+        const mk = exec(pythonBin, ['-m', 'venv', ...(opts.force ? ['--clear'] : []), venvDir]);
+        if (!mk.ok)
+            throw new CompostError('PROVIDER_ERROR', `venv create failed: ${mk.stderr.slice(-300)}`);
+        steps.push(`created venv at ${venvDir}`);
+    }
+    else {
+        steps.push('venv present but deps missing — reinstalling');
+    }
+    exec(venvPython, ['-m', 'pip', 'install', '-q', '-U', 'pip']); // best-effort
+    const install = exec(venvPython, ['-m', 'pip', 'install', ...NATIVE_DEPS], { stream: true });
+    if (!install.ok) {
+        throw new CompostError('PROVIDER_ERROR', `pip install failed: ${install.stderr.slice(-400) || '(see pip output above)'}`);
+    }
+    steps.push(`installed ${NATIVE_DEPS.length} native deps`);
+    const verify = exec(venvPython, ['-c', VERIFY_IMPORTS]);
+    if (!verify.ok) {
+        throw new CompostError('PROVIDER_ERROR', `native deps did not import after install: ${verify.stderr.slice(-300)}`);
+    }
+    steps.push('verified parakeet-mlx + pyannote + silero + torchaudio import');
+    return { status: 'provisioned', venvDir, venvPython, pythonBin, transcriberDir, steps };
+}
+//# sourceMappingURL=provisionNative.js.map

package/dist/lib/provisionNative.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"provisionNative.js","sourceRoot":"","sources":["../../src/lib/provisionNative.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,MAAM,oBAAoB,CAAA;AAC9C,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAA;AACpC,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AAEhC,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAA;AAC3C,OAAO,EAAE,sBAAsB,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAA;AAE3E;;;;;;;;;;GAUG;AAEH,iFAAiF;AACjF,MAAM,CAAC,MAAM,WAAW,GAAG;IACzB,cAAc;IACd,qBAAqB;IACrB,YAAY;IACZ,iBAAiB;IACjB,eAAe;CAChB,CAAA;AAED,MAAM,aAAa,GAAG,CAAC,YAAY,EAAE,YAAY,EAAE,SAAS,CAAC,CAAA;AAC7D,MAAM,cAAc,GAAG,6DAA6D,CAAA;AACpF,MAAM,SAAS,GAAG,EAAE,CAAA;AACpB,MAAM,SAAS,GAAG,EAAE,CAAA,CAAC,6CAA6C;AAoClE,MAAM,WAAW,GAAkB,CAAC,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE;IACrD,MAAM,GAAG,GAAG,SAAS,CAAC,GAAG,EAAE,IAAI,EAAE;QAC/B,QAAQ,EAAE,MAAM;QAChB,yEAAyE;QACzE,kEAAkE;QAClE,KAAK,EAAE,IAAI,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE,SAAS,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM;QAC5D,SAAS,EAAE,EAAE,GAAG,IAAI,GAAG,IAAI;KAC5B,CAAC,CAAA;IACF,OAAO,EAAE,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,IAAI,EAAE,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,IAAI,EAAE,EAAE,CAAA;AACnG,CAAC,CAAA;AAED,SAAS,SAAS,CAAC,IAAmB,EAAE,GAAW;IACjD,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC,WAAW,CAAC,CAAC,CAAA;IAClC,IAAI,CAAC,CAAC,CAAC,EAAE;QAAE,OAAO,IAAI,CAAA;IACtB,MAAM,CAAC,GAAG,qBAAqB,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC,CAAA;IAC/D,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;AAChD,CAAC;AAED,SAAS,aAAa,CAAC,IAAmB,EAAE,QAAiB;IAC3D,KAAK,MAAM,CAAC,IAAI,QAAQ,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,aAAa,EAAE,CAAC;QACtD,MAAM,CAAC,GAAG,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC,CAAA;QAC5B,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,SAAS,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,SAAS;YAAE,OAAO,CAAC,CAAA;IACzE,CAAC;IACD,MAAM,IAAI,YAAY,CACpB,eAAe,EACf,eAAe,SAAS,MAAM,SAAS,iBAAiB,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK;QAC7G,uEAAuE,CAC1E,CAAA;AACH,CAAC;AAED;iDACiD;AACjD,MAAM,UAAU,mBAAmB,CAAC,OAA+B,EAAE;IACnE,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,IAAI,OAAO,CAAC,GAAG,CAAA;IACnC,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,IAAI,UAAU,CAAA;IACxC,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,IAAI,WAAW,CAAA;IACrC,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,MAAM,cAAc,GAClB,IAAI,CAAC,cAAc,IAAI,GAAG,CAAC,uBAAuB,IAAI,sBAAsB,CAAC,SAAS,EAAE,MAAM,CAAC,CAAA;IACjG,IAAI,cAAc,KAAK,SAAS,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,cAAc,EAAE,gBAAgB,CAAC,CAAC,EAAE,CAAC;QACpF,MAAM,IAAI,YAAY,CACpB,eAAe,EACf,mGAAmG,CACpG,CAAA;IACH,CAAC;IAED,MAAM,OAAO,GAAG,cAAc,CAAC,GAAG,CAAC,CAAA;IACnC,MAAM,UAAU,GAAG,IAAI,CAAC,OAAO,EAAE,KAAK,EAAE,QAAQ,CAAC,CAAA;IAEjD,IAAI,CAAC,IAAI,CAAC,KAAK,IAAI,MAAM,CAAC,UAAU,CAAC,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC,IAAI,EAAE,cAAc,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC;QACrF,OAAO;YACL,MAAM,EAAE,eAAe;YACvB,OAAO;YACP,UAAU;YACV,SAAS,EAAE,IAAI;YACf,cAAc;YACd,KAAK,EAAE,CAAC,0CAA0C,CAAC;SACpD,CAAA;IACH,CAAC;IAED,4EAA4E;IAC5E,iFAAiF;IACjF,IAAI,SAAS,GAAkB,IAAI,CAAA;IACnC,IAAI,IAAI,CAAC,KAAK,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC;QACtC,SAAS,GAAG,aAAa,CAAC,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,CAAA;QAC/C,KAAK,CAAC,IAAI,CAAC,eAAe,SAAS,EAAE,CAAC,CAAA;QACtC,MAAM,EAAE,GAAG,IAAI,CAAC,SAAS,EAAE,CAAC,IAAI,EAAE,MAAM,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,OAAO,CAAC,CAAC,CAAA;QACvF,IAAI,CAAC,EAAE,CAAC,EAAE;YACR,MAAM,IAAI,YAAY,CAAC,gBAAgB,EAAE,uBAAuB,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAA;QAC1F,KAAK,CAAC,IAAI,CAAC,mBAAmB,OAAO,EAAE,CAAC,CAAA;IAC1C,CAAC;SAAM,CAAC;QACN,KAAK,CAAC,IAAI,CAAC,8CAA8C,CAAC,CAAA;IAC5D,CAAC;IAED,IAAI,CAAC,UAAU,EAAE,CAAC,IAAI,EAAE,KAAK,EAAE,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC,CAAA,CAAC,cAAc;IAC5E,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC,IAAI,EAAE,KAAK,EAAE,SAAS,EAAE,GAAG,WAAW,CAAC,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAA;IAC5F,IAAI,CAAC,OAAO,CAAC,EAAE,EAAE,CAAC;QAChB,MAAM,IAAI,YAAY,CACpB,gBAAgB,EAChB,uBAAuB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,IAAI,wBAAwB,EAAE,CAChF,CAAA;IACH,CAAC;IACD,KAAK,CAAC,IAAI,CAAC,aAAa,WAAW,CAAC,MAAM,cAAc,CAAC,CAAA;IAEzD,MAAM,MAAM,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC,IAAI,EAAE,cAAc,CAAC,CAAC,CAAA;IACvD,IAAI,CAAC,MAAM,CAAC,EAAE,EAAE,CAAC;QACf,MAAM,IAAI,YAAY,CACpB,gBAAgB,EAChB,6CAA6C,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,EAAE,CACzE,CAAA;IACH,CAAC;IACD,KAAK,CAAC,IAAI,CAAC,+DAA+D,CAAC,CAAA;IAE3E,OAAO,EAAE,MAAM,EAAE,aAAa,EAAE,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,cAAc,EAAE,KAAK,EAAE,CAAA;AACzF,CAAC"}

package/dist/lib/queue.d.ts

@@ -0,0 +1,36 @@
+export type JobKind = 'transcribe' | 'legacy-ingest';
+export type JobStatus = 'queued' | 'running' | 'done' | 'failed';
+export interface Job {
+    id: number;
+    kind: JobKind;
+    source_path: string;
+    payload: Record<string, unknown>;
+    status: JobStatus;
+    attempts: number;
+    created_at: string;
+    updated_at: string;
+    error: string | null;
+}
+export interface QueueOptions {
+    now?: () => Date;
+}
+export declare class JobQueue {
+    private readonly db;
+    private readonly now;
+    constructor(dbPath: string, opts?: QueueOptions);
+    /** Enqueue a job. Idempotent on (kind, source_path): returns the existing
+     * job id when already present (resumable folder ingest relies on this). */
+    enqueue(kind: JobKind, sourcePath: string, payload?: Record<string, unknown>): {
+        id: number;
+        inserted: boolean;
+    };
+    /** Atomically claim the oldest queued job (status → running, attempts++).
+     * Pass a kind to claim only jobs of that kind (per-worker draining). */
+    claim(kind?: JobKind): Job | null;
+    complete(id: number): void;
+    fail(id: number, error: string, maxAttempts?: number): void;
+    list(status?: JobStatus): Job[];
+    close(): void;
+}
+export declare function stateDbPath(seedPath: string): string;
+//# sourceMappingURL=queue.d.ts.map

package/dist/lib/queue.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"queue.d.ts","sourceRoot":"","sources":["../../src/lib/queue.ts"],"names":[],"mappings":"AAKA,MAAM,MAAM,OAAO,GAAG,YAAY,GAAG,eAAe,CAAA;AACpD,MAAM,MAAM,SAAS,GAAG,QAAQ,GAAG,SAAS,GAAG,MAAM,GAAG,QAAQ,CAAA;AAEhE,MAAM,WAAW,GAAG;IAClB,EAAE,EAAE,MAAM,CAAA;IACV,IAAI,EAAE,OAAO,CAAA;IACb,WAAW,EAAE,MAAM,CAAA;IACnB,OAAO,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;IAChC,MAAM,EAAE,SAAS,CAAA;IACjB,QAAQ,EAAE,MAAM,CAAA;IAChB,UAAU,EAAE,MAAM,CAAA;IAClB,UAAU,EAAE,MAAM,CAAA;IAClB,KAAK,EAAE,MAAM,GAAG,IAAI,CAAA;CACrB;AA4CD,MAAM,WAAW,YAAY;IAC3B,GAAG,CAAC,EAAE,MAAM,IAAI,CAAA;CACjB;AAED,qBAAa,QAAQ;IACnB,OAAO,CAAC,QAAQ,CAAC,EAAE,CAAmB;IACtC,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAY;gBAEpB,MAAM,EAAE,MAAM,EAAE,IAAI,GAAE,YAAiB;IAQnD;+EAC2E;IAC3E,OAAO,CACL,IAAI,EAAE,OAAO,EACb,UAAU,EAAE,MAAM,EAClB,OAAO,GAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAM,GACpC;QACD,EAAE,EAAE,MAAM,CAAA;QACV,QAAQ,EAAE,OAAO,CAAA;KAClB;IAeD;4EACwE;IACxE,KAAK,CAAC,IAAI,CAAC,EAAE,OAAO,GAAG,GAAG,GAAG,IAAI;IAuBjC,QAAQ,CAAC,EAAE,EAAE,MAAM,GAAG,IAAI;IAM1B,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,WAAW,SAAI,GAAG,IAAI;IAUtD,IAAI,CAAC,MAAM,CAAC,EAAE,SAAS,GAAG,GAAG,EAAE;IAS/B,KAAK,IAAI,IAAI;CAGd;AAED,wBAAgB,WAAW,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,CAEpD"}

package/dist/lib/queue.js

@@ -0,0 +1,102 @@
+import { mkdirSync } from 'node:fs';
+import { dirname, join } from 'node:path';
+import Database from 'better-sqlite3';
+const SCHEMA = `
+CREATE TABLE IF NOT EXISTS jobs (
+  id          INTEGER PRIMARY KEY AUTOINCREMENT,
+  kind        TEXT NOT NULL,
+  source_path TEXT NOT NULL,
+  payload     TEXT NOT NULL DEFAULT '{}',
+  status      TEXT NOT NULL DEFAULT 'queued',
+  attempts    INTEGER NOT NULL DEFAULT 0,
+  created_at  TEXT NOT NULL,
+  updated_at  TEXT NOT NULL,
+  error       TEXT,
+  UNIQUE (kind, source_path)
+);
+CREATE INDEX IF NOT EXISTS idx_jobs_status ON jobs(status);
+`;
+function rowToJob(r) {
+    return {
+        id: r.id,
+        kind: r.kind,
+        source_path: r.source_path,
+        payload: JSON.parse(r.payload),
+        status: r.status,
+        attempts: r.attempts,
+        created_at: r.created_at,
+        updated_at: r.updated_at,
+        error: r.error,
+    };
+}
+export class JobQueue {
+    db;
+    now;
+    constructor(dbPath, opts = {}) {
+        if (dbPath !== ':memory:')
+            mkdirSync(dirname(dbPath), { recursive: true });
+        this.db = new Database(dbPath);
+        this.db.pragma('journal_mode = WAL');
+        this.db.exec(SCHEMA);
+        this.now = opts.now ?? (() => new Date());
+    }
+    /** Enqueue a job. Idempotent on (kind, source_path): returns the existing
+     * job id when already present (resumable folder ingest relies on this). */
+    enqueue(kind, sourcePath, payload = {}) {
+        const ts = this.now().toISOString();
+        const existing = this.db
+            .prepare('SELECT id FROM jobs WHERE kind = ? AND source_path = ?')
+            .get(kind, sourcePath);
+        if (existing !== undefined)
+            return { id: existing.id, inserted: false };
+        const info = this.db
+            .prepare(`INSERT INTO jobs (kind, source_path, payload, status, created_at, updated_at)
+         VALUES (?, ?, ?, 'queued', ?, ?)`)
+            .run(kind, sourcePath, JSON.stringify(payload), ts, ts);
+        return { id: Number(info.lastInsertRowid), inserted: true };
+    }
+    /** Atomically claim the oldest queued job (status → running, attempts++).
+     * Pass a kind to claim only jobs of that kind (per-worker draining). */
+    claim(kind) {
+        const tx = this.db.transaction(() => {
+            const row = (kind === undefined
+                ? this.db.prepare("SELECT * FROM jobs WHERE status = 'queued' ORDER BY id LIMIT 1").get()
+                : this.db
+                    .prepare("SELECT * FROM jobs WHERE status = 'queued' AND kind = ? ORDER BY id LIMIT 1")
+                    .get(kind));
+            if (row === undefined)
+                return null;
+            const ts = this.now().toISOString();
+            this.db
+                .prepare("UPDATE jobs SET status = 'running', attempts = attempts + 1, updated_at = ? WHERE id = ?")
+                .run(ts, row.id);
+            return rowToJob({ ...row, status: 'running', attempts: row.attempts + 1 });
+        });
+        return tx();
+    }
+    complete(id) {
+        this.db
+            .prepare("UPDATE jobs SET status = 'done', updated_at = ?, error = NULL WHERE id = ?")
+            .run(this.now().toISOString(), id);
+    }
+    fail(id, error, maxAttempts = 3) {
+        const row = this.db.prepare('SELECT attempts FROM jobs WHERE id = ?').get(id);
+        const status = row !== undefined && row.attempts >= maxAttempts ? 'failed' : 'queued';
+        this.db
+            .prepare('UPDATE jobs SET status = ?, updated_at = ?, error = ? WHERE id = ?')
+            .run(status, this.now().toISOString(), error, id);
+    }
+    list(status) {
+        const rows = (status === undefined
+            ? this.db.prepare('SELECT * FROM jobs ORDER BY id').all()
+            : this.db.prepare('SELECT * FROM jobs WHERE status = ? ORDER BY id').all(status));
+        return rows.map(rowToJob);
+    }
+    close() {
+        this.db.close();
+    }
+}
+export function stateDbPath(seedPath) {
+    return join(seedPath, '.compost', 'state.sqlite');
+}
+//# sourceMappingURL=queue.js.map

package/dist/lib/queue.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"queue.js","sourceRoot":"","sources":["../../src/lib/queue.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,MAAM,SAAS,CAAA;AACnC,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AAEzC,OAAO,QAAQ,MAAM,gBAAgB,CAAA;AAiBrC,MAAM,MAAM,GAAG;;;;;;;;;;;;;;CAcd,CAAA;AAcD,SAAS,QAAQ,CAAC,CAAS;IACzB,OAAO;QACL,EAAE,EAAE,CAAC,CAAC,EAAE;QACR,IAAI,EAAE,CAAC,CAAC,IAAe;QACvB,WAAW,EAAE,CAAC,CAAC,WAAW;QAC1B,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,CAA4B;QACzD,MAAM,EAAE,CAAC,CAAC,MAAmB;QAC7B,QAAQ,EAAE,CAAC,CAAC,QAAQ;QACpB,UAAU,EAAE,CAAC,CAAC,UAAU;QACxB,UAAU,EAAE,CAAC,CAAC,UAAU;QACxB,KAAK,EAAE,CAAC,CAAC,KAAK;KACf,CAAA;AACH,CAAC;AAMD,MAAM,OAAO,QAAQ;IACF,EAAE,CAAmB;IACrB,GAAG,CAAY;IAEhC,YAAY,MAAc,EAAE,OAAqB,EAAE;QACjD,IAAI,MAAM,KAAK,UAAU;YAAE,SAAS,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAA;QAC1E,IAAI,CAAC,EAAE,GAAG,IAAI,QAAQ,CAAC,MAAM,CAAC,CAAA;QAC9B,IAAI,CAAC,EAAE,CAAC,MAAM,CAAC,oBAAoB,CAAC,CAAA;QACpC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QACpB,IAAI,CAAC,GAAG,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC,IAAI,IAAI,EAAE,CAAC,CAAA;IAC3C,CAAC;IAED;+EAC2E;IAC3E,OAAO,CACL,IAAa,EACb,UAAkB,EAClB,UAAmC,EAAE;QAKrC,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC,WAAW,EAAE,CAAA;QACnC,MAAM,QAAQ,GAAG,IAAI,CAAC,EAAE;aACrB,OAAO,CAAC,wDAAwD,CAAC;aACjE,GAAG,CAAC,IAAI,EAAE,UAAU,CAA+B,CAAA;QACtD,IAAI,QAAQ,KAAK,SAAS;YAAE,OAAO,EAAE,EAAE,EAAE,QAAQ,CAAC,EAAE,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAA;QACvE,MAAM,IAAI,GAAG,IAAI,CAAC,EAAE;aACjB,OAAO,CACN;0CACkC,CACnC;aACA,GAAG,CAAC,IAAI,EAAE,UAAU,EAAE,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC,CAAA;QACzD,OAAO,EAAE,EAAE,EAAE,MAAM,CAAC,IAAI,CAAC,eAAe,CAAC,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAA;IAC7D,CAAC;IAED;4EACwE;IACxE,KAAK,CAAC,IAAc;QAClB,MAAM,EAAE,GAAG,IAAI,CAAC,EAAE,CAAC,WAAW,CAAC,GAAG,EAAE;YAClC,MAAM,GAAG,GAAG,CACV,IAAI,KAAK,SAAS;gBAChB,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,OAAO,CAAC,gEAAgE,CAAC,CAAC,GAAG,EAAE;gBACzF,CAAC,CAAC,IAAI,CAAC,EAAE;qBACJ,OAAO,CACN,6EAA6E,CAC9E;qBACA,GAAG,CAAC,IAAI,CAAC,CACK,CAAA;YACvB,IAAI,GAAG,KAAK,SAAS;gBAAE,OAAO,IAAI,CAAA;YAClC,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC,WAAW,EAAE,CAAA;YACnC,IAAI,CAAC,EAAE;iBACJ,OAAO,CACN,0FAA0F,CAC3F;iBACA,GAAG,CAAC,EAAE,EAAE,GAAG,CAAC,EAAE,CAAC,CAAA;YAClB,OAAO,QAAQ,CAAC,EAAE,GAAG,GAAG,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,GAAG,CAAC,QAAQ,GAAG,CAAC,EAAE,CAAC,CAAA;QAC5E,CAAC,CAAC,CAAA;QACF,OAAO,EAAE,EAAE,CAAA;IACb,CAAC;IAED,QAAQ,CAAC,EAAU;QACjB,IAAI,CAAC,EAAE;aACJ,OAAO,CAAC,4EAA4E,CAAC;aACrF,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,CAAC,WAAW,EAAE,EAAE,EAAE,CAAC,CAAA;IACtC,CAAC;IAED,IAAI,CAAC,EAAU,EAAE,KAAa,EAAE,WAAW,GAAG,CAAC;QAC7C,MAAM,GAAG,GAAG,IAAI,CAAC,EAAE,CAAC,OAAO,CAAC,wCAAwC,CAAC,CAAC,GAAG,CAAC,EAAE,CAE/D,CAAA;QACb,MAAM,MAAM,GAAG,GAAG,KAAK,SAAS,IAAI,GAAG,CAAC,QAAQ,IAAI,WAAW,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAA;QACrF,IAAI,CAAC,EAAE;aACJ,OAAO,CAAC,oEAAoE,CAAC;aAC7E,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,GAAG,EAAE,CAAC,WAAW,EAAE,EAAE,KAAK,EAAE,EAAE,CAAC,CAAA;IACrD,CAAC;IAED,IAAI,CAAC,MAAkB;QACrB,MAAM,IAAI,GAAG,CACX,MAAM,KAAK,SAAS;YAClB,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,OAAO,CAAC,gCAAgC,CAAC,CAAC,GAAG,EAAE;YACzD,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,OAAO,CAAC,iDAAiD,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,CACvE,CAAA;QACb,OAAO,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;IAC3B,CAAC;IAED,KAAK;QACH,IAAI,CAAC,EAAE,CAAC,KAAK,EAAE,CAAA;IACjB,CAAC;CACF;AAED,MAAM,UAAU,WAAW,CAAC,QAAgB;IAC1C,OAAO,IAAI,CAAC,QAAQ,EAAE,UAAU,EAAE,cAAc,CAAC,CAAA;AACnD,CAAC"}

package/dist/lib/retrieve.d.ts

@@ -0,0 +1,51 @@
+import { type Chunk, type DenseRetriever, type EvidenceSet, type ScoredChunk } from 'compost-retrieval';
+import { type CompostConfig } from './config.js';
+/**
+ * Shared retrieval primitives for `compost search` (retrieval-only) and
+ * `compost chat` (retrieval + LLM synthesis). Extracted so the two commands
+ * never diverge on how the corpus is loaded, chunked, and ranked.
+ *
+ * Retrieval is BM25-only today: HybridRetriever's dense slot is null because
+ * nothing reads the LanceDB index the embed-worker writes (#137). Wiring
+ * LanceDBRetriever into this shared path is a follow-up that upgrades both
+ * `search` and `chat` ranking at once — see the tracking issue.
+ */
+export interface LoadedCorpus {
+    chunks: Chunk[];
+    evidence: EvidenceSet;
+    seedName: string;
+}
+/**
+ * Load every session transcript under a seed into:
+ *   - `chunks`: chunked text for retrieval (utterance + neighbor windows + …)
+ *   - `evidence`: utterance_id → {session_id, text}, for citation validation
+ */
+export declare function loadSeedCorpus(seedPath: string): LoadedCorpus;
+export type RetrievalMode = 'hybrid' | 'bm25';
+export interface RetrieveOptions {
+    topK?: number;
+    /** Dense retriever (LanceDB). When provided, retrieval is BM25 ∪ dense fused
+     * via RRF. When null/omitted, BM25-only. Inject via buildDenseRetriever. */
+    dense?: DenseRetriever | null;
+}
+export interface RetrieveResult {
+    retrieved: ScoredChunk[];
+    corpus: LoadedCorpus;
+    mode: RetrievalMode;
+}
+/**
+ * Rank corpus chunks against a query. Returns ScoredChunks (highest first).
+ * Empty array when the seed has no indexed sessions or nothing matches.
+ *
+ * Hybrid (BM25 + dense via RRF) when a dense retriever is supplied; BM25-only
+ * otherwise. The `mode` field reports which ran.
+ */
+export declare function retrieveChunks(seedPath: string, query: string, opts?: RetrieveOptions): Promise<RetrieveResult>;
+/**
+ * Build a dense retriever for a seed, or return null when dense isn't available
+ * — no LanceDB index yet, the native binary absent, or the embeddings provider
+ * unreachable. Never throws: a null result means the caller does BM25-only, so
+ * `compost search` always works even with Ollama down.
+ */
+export declare function buildDenseRetriever(seedPath: string, config?: CompostConfig): Promise<DenseRetriever | null>;
+//# sourceMappingURL=retrieve.d.ts.map

package/dist/lib/retrieve.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"retrieve.d.ts","sourceRoot":"","sources":["../../src/lib/retrieve.ts"],"names":[],"mappings":"AAGA,OAAO,EAEL,KAAK,KAAK,EAGV,KAAK,cAAc,EACnB,KAAK,WAAW,EAGhB,KAAK,WAAW,EACjB,MAAM,mBAAmB,CAAA;AAG1B,OAAO,EAAE,KAAK,aAAa,EAAc,MAAM,aAAa,CAAA;AAE5D;;;;;;;;;GASG;AAEH,MAAM,WAAW,YAAY;IAC3B,MAAM,EAAE,KAAK,EAAE,CAAA;IACf,QAAQ,EAAE,WAAW,CAAA;IACrB,QAAQ,EAAE,MAAM,CAAA;CACjB;AAED;;;;GAIG;AACH,wBAAgB,cAAc,CAAC,QAAQ,EAAE,MAAM,GAAG,YAAY,CAoB7D;AAED,MAAM,MAAM,aAAa,GAAG,QAAQ,GAAG,MAAM,CAAA;AAE7C,MAAM,WAAW,eAAe;IAC9B,IAAI,CAAC,EAAE,MAAM,CAAA;IACb;gFAC4E;IAC5E,KAAK,CAAC,EAAE,cAAc,GAAG,IAAI,CAAA;CAC9B;AAED,MAAM,WAAW,cAAc;IAC7B,SAAS,EAAE,WAAW,EAAE,CAAA;IACxB,MAAM,EAAE,YAAY,CAAA;IACpB,IAAI,EAAE,aAAa,CAAA;CACpB;AAED;;;;;;GAMG;AACH,wBAAsB,cAAc,CAClC,QAAQ,EAAE,MAAM,EAChB,KAAK,EAAE,MAAM,EACb,IAAI,GAAE,eAAoB,GACzB,OAAO,CAAC,cAAc,CAAC,CAWzB;AAWD;;;;;GAKG;AACH,wBAAsB,mBAAmB,CACvC,QAAQ,EAAE,MAAM,EAChB,MAAM,CAAC,EAAE,aAAa,GACrB,OAAO,CAAC,cAAc,GAAG,IAAI,CAAC,CAyBhC"}

package/dist/lib/retrieve.js

@@ -0,0 +1,92 @@
+import { existsSync, readdirSync, readFileSync, statSync } from 'node:fs';
+import { join } from 'node:path';
+import { BM25Index, chunkTranscript, HybridRetriever, openLanceDBForRead, } from 'compost-retrieval';
+import { LLMAdapter } from '../llm/adapter.js';
+import { loadConfig } from './config.js';
+/**
+ * Load every session transcript under a seed into:
+ *   - `chunks`: chunked text for retrieval (utterance + neighbor windows + …)
+ *   - `evidence`: utterance_id → {session_id, text}, for citation validation
+ */
+export function loadSeedCorpus(seedPath) {
+    const seedName = seedPath.split('/').pop() ?? 'seed';
+    const sessionsDir = join(seedPath, 'sessions');
+    const evidence = new Map();
+    const allChunks = [];
+    if (!existsSync(sessionsDir))
+        return { chunks: [], evidence, seedName };
+    for (const entry of readdirSync(sessionsDir)) {
+        if (entry.startsWith('.') || entry === '_inbox')
+            continue;
+        const tPath = join(sessionsDir, entry, 'transcript.json');
+        if (!existsSync(tPath) || !statSync(tPath).isFile())
+            continue;
+        const transcript = JSON.parse(readFileSync(tPath, 'utf8'));
+        for (const u of transcript.utterances) {
+            evidence.set(u.id, { session_id: transcript.session_id, text: u.text });
+        }
+        allChunks.push(...chunkTranscript(transcript, { seed: seedName }));
+    }
+    return { chunks: allChunks, evidence, seedName };
+}
+/**
+ * Rank corpus chunks against a query. Returns ScoredChunks (highest first).
+ * Empty array when the seed has no indexed sessions or nothing matches.
+ *
+ * Hybrid (BM25 + dense via RRF) when a dense retriever is supplied; BM25-only
+ * otherwise. The `mode` field reports which ran.
+ */
+export async function retrieveChunks(seedPath, query, opts = {}) {
+    const corpus = loadSeedCorpus(seedPath);
+    const dense = opts.dense ?? null;
+    const mode = dense ? 'hybrid' : 'bm25';
+    if (corpus.chunks.length === 0)
+        return { retrieved: [], corpus, mode };
+    const bm25 = new BM25Index();
+    bm25.addAll(corpus.chunks);
+    const retriever = new HybridRetriever(bm25, dense);
+    const retrieved = await retriever.retrieve(query, { topK: opts.topK ?? 8 });
+    return { retrieved, corpus, mode };
+}
+/** Embedding-vector dimension by model tag (mirrors the embed-worker's table). */
+const DIM_BY_MODEL = {
+    'bge-m3': 1024,
+    'bge-m3:q4_k_m': 1024,
+    'mxbai-embed-large': 1024,
+    'nomic-embed-text': 768,
+    'all-minilm': 384,
+};
+/**
+ * Build a dense retriever for a seed, or return null when dense isn't available
+ * — no LanceDB index yet, the native binary absent, or the embeddings provider
+ * unreachable. Never throws: a null result means the caller does BM25-only, so
+ * `compost search` always works even with Ollama down.
+ */
+export async function buildDenseRetriever(seedPath, config) {
+    const uri = join(seedPath, '.compost', 'vectors.lancedb');
+    if (!existsSync(uri))
+        return null;
+    let adapter;
+    try {
+        adapter = new LLMAdapter(config ?? loadConfig(seedPath));
+    }
+    catch {
+        return null;
+    }
+    const embedQuery = async (q) => {
+        const resp = await adapter.embed('embeddings', [q]);
+        const vec = resp.vectors[0];
+        if (vec === undefined)
+            throw new Error('embeddings provider returned no vector');
+        return vec;
+    };
+    // openLanceDBForRead is null-safe (missing index / absent binary); we still
+    // wrap in try/catch so a provider/config error degrades to BM25, not a throw.
+    try {
+        return await openLanceDBForRead(uri, embedQuery);
+    }
+    catch {
+        return null;
+    }
+}
+//# sourceMappingURL=retrieve.js.map

package/dist/lib/retrieve.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"retrieve.js","sourceRoot":"","sources":["../../src/lib/retrieve.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,WAAW,EAAE,YAAY,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAA;AACzE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AAEhC,OAAO,EACL,SAAS,EAGT,eAAe,EAGf,eAAe,EACf,kBAAkB,GAEnB,MAAM,mBAAmB,CAAA;AAE1B,OAAO,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAA;AAC9C,OAAO,EAAsB,UAAU,EAAE,MAAM,aAAa,CAAA;AAmB5D;;;;GAIG;AACH,MAAM,UAAU,cAAc,CAAC,QAAgB;IAC7C,MAAM,QAAQ,GAAG,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,MAAM,CAAA;IACpD,MAAM,WAAW,GAAG,IAAI,CAAC,QAAQ,EAAE,UAAU,CAAC,CAAA;IAC9C,MAAM,QAAQ,GAAgB,IAAI,GAAG,EAAE,CAAA;IACvC,MAAM,SAAS,GAAY,EAAE,CAAA;IAC7B,IAAI,CAAC,UAAU,CAAC,WAAW,CAAC;QAAE,OAAO,EAAE,MAAM,EAAE,EAAE,EAAE,QAAQ,EAAE,QAAQ,EAAE,CAAA;IAEvE,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,WAAW,CAAC,EAAE,CAAC;QAC7C,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,KAAK,KAAK,QAAQ;YAAE,SAAQ;QACzD,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,KAAK,EAAE,iBAAiB,CAAC,CAAA;QACzD,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,MAAM,EAAE;YAAE,SAAQ;QAC7D,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,KAAK,EAAE,MAAM,CAAC,CAExD,CAAA;QACD,KAAK,MAAM,CAAC,IAAI,UAAU,CAAC,UAAU,EAAE,CAAC;YACtC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,UAAU,EAAE,UAAU,CAAC,UAAU,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;QACzE,CAAC;QACD,SAAS,CAAC,IAAI,CAAC,GAAG,eAAe,CAAC,UAAU,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC,CAAA;IACpE,CAAC;IACD,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,QAAQ,EAAE,CAAA;AAClD,CAAC;AAiBD;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,cAAc,CAClC,QAAgB,EAChB,KAAa,EACb,OAAwB,EAAE;IAE1B,MAAM,MAAM,GAAG,cAAc,CAAC,QAAQ,CAAC,CAAA;IACvC,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,IAAI,IAAI,CAAA;IAChC,MAAM,IAAI,GAAkB,KAAK,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAA;IACrD,IAAI,MAAM,CAAC,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,SAAS,EAAE,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,CAAA;IAEtE,MAAM,IAAI,GAAG,IAAI,SAAS,EAAE,CAAA;IAC5B,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAA;IAC1B,MAAM,SAAS,GAAG,IAAI,eAAe,CAAC,IAAI,EAAE,KAAK,CAAC,CAAA;IAClD,MAAM,SAAS,GAAG,MAAM,SAAS,CAAC,QAAQ,CAAC,KAAK,EAAE,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC,EAAE,CAAC,CAAA;IAC3E,OAAO,EAAE,SAAS,EAAE,MAAM,EAAE,IAAI,EAAE,CAAA;AACpC,CAAC;AAED,kFAAkF;AAClF,MAAM,YAAY,GAA2B;IAC3C,QAAQ,EAAE,IAAI;IACd,eAAe,EAAE,IAAI;IACrB,mBAAmB,EAAE,IAAI;IACzB,kBAAkB,EAAE,GAAG;IACvB,YAAY,EAAE,GAAG;CAClB,CAAA;AAED;;;;;GAKG;AACH,MAAM,CAAC,KAAK,UAAU,mBAAmB,CACvC,QAAgB,EAChB,MAAsB;IAEtB,MAAM,GAAG,GAAG,IAAI,CAAC,QAAQ,EAAE,UAAU,EAAE,iBAAiB,CAAC,CAAA;IACzD,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAA;IAEjC,IAAI,OAAmB,CAAA;IACvB,IAAI,CAAC;QACH,OAAO,GAAG,IAAI,UAAU,CAAC,MAAM,IAAI,UAAU,CAAC,QAAQ,CAAC,CAAC,CAAA;IAC1D,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAA;IACb,CAAC;IAED,MAAM,UAAU,GAAG,KAAK,EAAE,CAAS,EAAqB,EAAE;QACxD,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,KAAK,CAAC,YAAY,EAAE,CAAC,CAAC,CAAC,CAAC,CAAA;QACnD,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAA;QAC3B,IAAI,GAAG,KAAK,SAAS;YAAE,MAAM,IAAI,KAAK,CAAC,wCAAwC,CAAC,CAAA;QAChF,OAAO,GAAG,CAAA;IACZ,CAAC,CAAA;IAED,4EAA4E;IAC5E,8EAA8E;IAC9E,IAAI,CAAC;QACH,OAAO,MAAM,kBAAkB,CAAC,GAAG,EAAE,UAAU,CAAC,CAAA;IAClD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAA;IACb,CAAC;AACH,CAAC"}

package/dist/lib/saturate.d.ts

@@ -0,0 +1,21 @@
+export interface SessionWithThemes {
+    id: string;
+    themes: string[];
+}
+export interface GatherOptions {
+    cwd?: string;
+    seed?: string;
+}
+/**
+ * Walk a seed's coded corpus and return, per session in chronological order,
+ * the themes it contributed to. The join is theme.codes → code.evidence
+ * (highlight ids) → highlight.session_id. A theme is "from" a session if any
+ * of its codes is evidenced by a highlight in that session.
+ *
+ * Sessions are ordered by directory name (S001, S002 …) — compost's session
+ * ids are assigned sequentially at ingest, so lexicographic order is also
+ * chronological. The returned shape matches the input contract of
+ * `saturationPulse()` in compost-retrieval.
+ */
+export declare function gatherSessionsWithThemes(opts?: GatherOptions): SessionWithThemes[];
+//# sourceMappingURL=saturate.d.ts.map

package/dist/lib/saturate.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"saturate.d.ts","sourceRoot":"","sources":["../../src/lib/saturate.ts"],"names":[],"mappings":"AAOA,MAAM,WAAW,iBAAiB;IAChC,EAAE,EAAE,MAAM,CAAA;IACV,MAAM,EAAE,MAAM,EAAE,CAAA;CACjB;AAED,MAAM,WAAW,aAAa;IAC5B,GAAG,CAAC,EAAE,MAAM,CAAA;IACZ,IAAI,CAAC,EAAE,MAAM,CAAA;CACd;AAED;;;;;;;;;;GAUG;AACH,wBAAgB,wBAAwB,CAAC,IAAI,GAAE,aAAkB,GAAG,iBAAiB,EAAE,CAyCtF"}

package/dist/lib/saturate.js

@@ -0,0 +1,132 @@
+import { existsSync, readdirSync, readFileSync, statSync } from 'node:fs';
+import { basename, join, resolve } from 'node:path';
+import { CompostError } from '../errors.js';
+import { listCanonicalSessionIds } from './canonicalSessions.js';
+import { resolveSeedPath } from './seedResolve.js';
+/**
+ * Walk a seed's coded corpus and return, per session in chronological order,
+ * the themes it contributed to. The join is theme.codes → code.evidence
+ * (highlight ids) → highlight.session_id. A theme is "from" a session if any
+ * of its codes is evidenced by a highlight in that session.
+ *
+ * Sessions are ordered by directory name (S001, S002 …) — compost's session
+ * ids are assigned sequentially at ingest, so lexicographic order is also
+ * chronological. The returned shape matches the input contract of
+ * `saturationPulse()` in compost-retrieval.
+ */
+export function gatherSessionsWithThemes(opts = {}) {
+    const cwd = opts.cwd ?? process.cwd();
+    const seedPath = resolveSeedPath(cwd, opts.seed);
+    const highlightToSession = readHighlightSessions(join(seedPath, 'highlights'));
+    const codeToHighlights = readCodeEvidence(join(seedPath, 'codebook'));
+    const themeToCodes = readThemeCodes(join(seedPath, 'synthesis', 'themes'));
+    const sessionToThemes = new Map();
+    // Single source of truth (#166): the same canonical-session predicate
+    // `compost status` uses — `S\d+` names OR a transcript.json/source.* file.
+    // Pre-fix: every subdir of sessions/ (incl. legacy Attachments/, Transcripts/)
+    // was counted, inflating the session set vs status's view.
+    for (const sessionId of listCanonicalSessionIds(join(seedPath, 'sessions'))) {
+        sessionToThemes.set(sessionId, new Set());
+    }
+    for (const [themeId, codeIds] of themeToCodes) {
+        for (const codeId of codeIds) {
+            const highlightIds = codeToHighlights.get(codeId) ?? [];
+            for (const hid of highlightIds) {
+                const sessionId = highlightToSession.get(hid);
+                if (sessionId === undefined)
+                    continue;
+                const bucket = sessionToThemes.get(sessionId);
+                if (bucket === undefined) {
+                    // Highlight references a session that isn't a directory under
+                    // sessions/ — surface it so a broken corpus fails loudly rather
+                    // than silently distorting the saturation curve.
+                    throw new CompostError('SCHEMA_VIOLATION', `Highlight ${hid} references session ${sessionId}, but Seeds/${basename(seedPath)}/sessions/${sessionId}/ does not exist.`);
+                }
+                bucket.add(themeId);
+            }
+        }
+    }
+    return [...sessionToThemes.entries()]
+        .sort(([a], [b]) => a.localeCompare(b))
+        .map(([id, themes]) => ({ id, themes: [...themes].sort() }));
+}
+function readHighlightSessions(dir) {
+    const out = new Map();
+    for (const fm of readFrontmatters(dir)) {
+        const id = fm.scalars.get('id');
+        const sessionId = fm.scalars.get('session_id');
+        if (id !== undefined && sessionId !== undefined)
+            out.set(id, sessionId);
+    }
+    return out;
+}
+function readCodeEvidence(dir) {
+    const out = new Map();
+    for (const fm of readFrontmatters(dir)) {
+        const id = fm.scalars.get('id');
+        const evidence = fm.arrays.get('evidence');
+        if (id !== undefined && evidence !== undefined)
+            out.set(id, evidence);
+    }
+    return out;
+}
+function readThemeCodes(dir) {
+    const out = new Map();
+    for (const fm of readFrontmatters(dir)) {
+        const id = fm.scalars.get('id');
+        const codes = fm.arrays.get('codes');
+        if (id !== undefined && codes !== undefined)
+            out.set(id, codes);
+    }
+    return out;
+}
+function readFrontmatters(dir) {
+    if (!existsSync(dir))
+        return [];
+    const out = [];
+    for (const entry of readdirSync(dir)) {
+        if (!entry.endsWith('.md'))
+            continue;
+        const path = resolve(dir, entry);
+        if (!statSync(path).isFile())
+            continue;
+        const fm = parseFrontmatter(readFileSync(path, 'utf8'));
+        if (fm !== null)
+            out.push(fm);
+    }
+    return out;
+}
+function parseFrontmatter(content) {
+    const match = content.match(/^---\n([\s\S]*?)\n---/);
+    if (match === null)
+        return null;
+    const scalars = new Map();
+    const arrays = new Map();
+    for (const line of (match[1] ?? '').split('\n')) {
+        const m = line.match(/^([A-Za-z_][A-Za-z0-9_-]*):\s*(.*)$/);
+        if (m === null)
+            continue;
+        const key = m[1];
+        const raw = (m[2] ?? '').trim();
+        if (key === undefined || raw.length === 0)
+            continue;
+        const inline = raw.match(/^\[(.*)\]$/);
+        if (inline !== null) {
+            arrays.set(key, (inline[1] ?? '')
+                .split(',')
+                .map((s) => stripQuotes(s.trim()))
+                .filter((s) => s.length > 0));
+        }
+        else {
+            scalars.set(key, stripQuotes(raw));
+        }
+    }
+    return { scalars, arrays };
+}
+function stripQuotes(s) {
+    if ((s.startsWith('"') && s.endsWith('"')) || (s.startsWith("'") && s.endsWith("'"))) {
+        return s.slice(1, -1);
+    }
+    return s;
+}
+//# sourceMappingURL=saturate.js.map

package/dist/lib/saturate.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"saturate.js","sourceRoot":"","sources":["../../src/lib/saturate.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,WAAW,EAAE,YAAY,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAA;AACzE,OAAO,EAAE,QAAQ,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAA;AAEnD,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAA;AAC3C,OAAO,EAAE,uBAAuB,EAAE,MAAM,wBAAwB,CAAA;AAChE,OAAO,EAAE,eAAe,EAAE,MAAM,kBAAkB,CAAA;AAYlD;;;;;;;;;;GAUG;AACH,MAAM,UAAU,wBAAwB,CAAC,OAAsB,EAAE;IAC/D,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,IAAI,OAAO,CAAC,GAAG,EAAE,CAAA;IACrC,MAAM,QAAQ,GAAG,eAAe,CAAC,GAAG,EAAE,IAAI,CAAC,IAAI,CAAC,CAAA;IAEhD,MAAM,kBAAkB,GAAG,qBAAqB,CAAC,IAAI,CAAC,QAAQ,EAAE,YAAY,CAAC,CAAC,CAAA;IAC9E,MAAM,gBAAgB,GAAG,gBAAgB,CAAC,IAAI,CAAC,QAAQ,EAAE,UAAU,CAAC,CAAC,CAAA;IACrE,MAAM,YAAY,GAAG,cAAc,CAAC,IAAI,CAAC,QAAQ,EAAE,WAAW,EAAE,QAAQ,CAAC,CAAC,CAAA;IAE1E,MAAM,eAAe,GAAG,IAAI,GAAG,EAAuB,CAAA;IACtD,sEAAsE;IACtE,2EAA2E;IAC3E,+EAA+E;IAC/E,2DAA2D;IAC3D,KAAK,MAAM,SAAS,IAAI,uBAAuB,CAAC,IAAI,CAAC,QAAQ,EAAE,UAAU,CAAC,CAAC,EAAE,CAAC;QAC5E,eAAe,CAAC,GAAG,CAAC,SAAS,EAAE,IAAI,GAAG,EAAE,CAAC,CAAA;IAC3C,CAAC;IAED,KAAK,MAAM,CAAC,OAAO,EAAE,OAAO,CAAC,IAAI,YAAY,EAAE,CAAC;QAC9C,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,MAAM,YAAY,GAAG,gBAAgB,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,CAAA;YACvD,KAAK,MAAM,GAAG,IAAI,YAAY,EAAE,CAAC;gBAC/B,MAAM,SAAS,GAAG,kBAAkB,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;gBAC7C,IAAI,SAAS,KAAK,SAAS;oBAAE,SAAQ;gBACrC,MAAM,MAAM,GAAG,eAAe,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;gBAC7C,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;oBACzB,8DAA8D;oBAC9D,gEAAgE;oBAChE,iDAAiD;oBACjD,MAAM,IAAI,YAAY,CACpB,kBAAkB,EAClB,aAAa,GAAG,uBAAuB,SAAS,eAAe,QAAQ,CAAC,QAAQ,CAAC,aAAa,SAAS,mBAAmB,CAC3H,CAAA;gBACH,CAAC;gBACD,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;YACrB,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,CAAC,GAAG,eAAe,CAAC,OAAO,EAAE,CAAC;SAClC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC;SACtC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,MAAM,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,GAAG,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAA;AAChE,CAAC;AAED,SAAS,qBAAqB,CAAC,GAAW;IACxC,MAAM,GAAG,GAAG,IAAI,GAAG,EAAkB,CAAA;IACrC,KAAK,MAAM,EAAE,IAAI,gBAAgB,CAAC,GAAG,CAAC,EAAE,CAAC;QACvC,MAAM,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QAC/B,MAAM,SAAS,GAAG,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,CAAA;QAC9C,IAAI,EAAE,KAAK,SAAS,IAAI,SAAS,KAAK,SAAS;YAAE,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,SAAS,CAAC,CAAA;IACzE,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,SAAS,gBAAgB,CAAC,GAAW;IACnC,MAAM,GAAG,GAAG,IAAI,GAAG,EAAoB,CAAA;IACvC,KAAK,MAAM,EAAE,IAAI,gBAAgB,CAAC,GAAG,CAAC,EAAE,CAAC;QACvC,MAAM,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QAC/B,MAAM,QAAQ,GAAG,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,UAAU,CAAC,CAAA;QAC1C,IAAI,EAAE,KAAK,SAAS,IAAI,QAAQ,KAAK,SAAS;YAAE,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,QAAQ,CAAC,CAAA;IACvE,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,SAAS,cAAc,CAAC,GAAW;IACjC,MAAM,GAAG,GAAG,IAAI,GAAG,EAAoB,CAAA;IACvC,KAAK,MAAM,EAAE,IAAI,gBAAgB,CAAC,GAAG,CAAC,EAAE,CAAC;QACvC,MAAM,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QAC/B,MAAM,KAAK,GAAG,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;QACpC,IAAI,EAAE,KAAK,SAAS,IAAI,KAAK,KAAK,SAAS;YAAE,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,KAAK,CAAC,CAAA;IACjE,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAOD,SAAS,gBAAgB,CAAC,GAAW;IACnC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,EAAE,CAAA;IAC/B,MAAM,GAAG,GAAkB,EAAE,CAAA;IAC7B,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;QACrC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC;YAAE,SAAQ;QACpC,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,EAAE,KAAK,CAAC,CAAA;QAChC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE;YAAE,SAAQ;QACtC,MAAM,EAAE,GAAG,gBAAgB,CAAC,YAAY,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,CAAA;QACvD,IAAI,EAAE,KAAK,IAAI;YAAE,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;IAC/B,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,SAAS,gBAAgB,CAAC,OAAe;IACvC,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,uBAAuB,CAAC,CAAA;IACpD,IAAI,KAAK,KAAK,IAAI;QAAE,OAAO,IAAI,CAAA;IAC/B,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAA;IACzC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAoB,CAAA;IAC1C,KAAK,MAAM,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;QAChD,MAAM,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,qCAAqC,CAAC,CAAA;QAC3D,IAAI,CAAC,KAAK,IAAI;YAAE,SAAQ;QACxB,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAA;QAChB,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;QAC/B,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC;YAAE,SAAQ;QACnD,MAAM,MAAM,GAAG,GAAG,CAAC,KAAK,CAAC,YAAY,CAAC,CAAA;QACtC,IAAI,MAAM,KAAK,IAAI,EAAE,CAAC;YACpB,MAAM,CAAC,GAAG,CACR,GAAG,EACH,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;iBACd,KAAK,CAAC,GAAG,CAAC;iBACV,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;iBACjC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAC/B,CAAA;QACH,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,WAAW,CAAC,GAAG,CAAC,CAAC,CAAA;QACpC,CAAC;IACH,CAAC;IACD,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,CAAA;AAC5B,CAAC;AAED,SAAS,WAAW,CAAC,CAAS;IAC5B,IAAI,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;QACrF,OAAO,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;IACvB,CAAC;IACD,OAAO,CAAC,CAAA;AACV,CAAC"}

package/dist/lib/schemas.generated.d.ts

@@ -0,0 +1,15 @@
+/**
+ * GENERATED — DO NOT EDIT. Run `pnpm --filter compost-cli run build` to regenerate.
+ *
+ * Embeds every schema/*.json file as a typed constant so validate.ts can use
+ * them without filesystem IO. The CLI can be published standalone without
+ * the schema/ tree because the schemas are now part of the compiled bundle.
+ */
+export declare const ANSWER_SCHEMA: Record<string, unknown>;
+export declare const CUES_TAXONOMY: Record<string, unknown>;
+export declare const EVENTS_SCHEMA: Record<string, unknown>;
+export declare const FRAMES_TAXONOMY: Record<string, unknown>;
+export declare const STATUS_SCHEMA: Record<string, unknown>;
+export declare const TRANSCRIPT_SCHEMA: Record<string, unknown>;
+export declare const ALL_SCHEMAS: Record<string, Record<string, unknown>>;
+//# sourceMappingURL=schemas.generated.d.ts.map

package/dist/lib/schemas.generated.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"schemas.generated.d.ts","sourceRoot":"","sources":["../../src/lib/schemas.generated.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAGH,eAAO,MAAM,aAAa,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAyDjD,CAAA;AAED,eAAO,MAAM,aAAa,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAsDjD,CAAA;AAED,eAAO,MAAM,aAAa,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CA0KjD,CAAA;AAED,eAAO,MAAM,eAAe,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAmDnD,CAAA;AAED,eAAO,MAAM,aAAa,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CA+IjD,CAAA;AAED,eAAO,MAAM,iBAAiB,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CA6drD,CAAA;AAED,eAAO,MAAM,WAAW,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAO/D,CAAA"}