@saluzi/codegraph 0.2.0 → 0.2.1

package/dist/db/wal-valve.js

@@ -0,0 +1,208 @@
+"use strict";
+/**
+ * WAL checkpoint valve — bounds WAL growth while auto-checkpointing is
+ * deferred during a bulk index (#1231).
+ *
+ * Why deferral: SQLite's default `wal_autocheckpoint` (1000 pages) re-writes
+ * hot B-tree/FTS pages into the main DB file over and over during a bulk
+ * index — measured at ~95% of ALL disk I/O, and the difference between 45s
+ * and 19+ minutes on HDD-class storage (150 random IOPS). Deferring
+ * checkpoints turns the store into pure sequential WAL appends; each backfill
+ * pass writes distinct pages once, in page order (≈ sequential).
+ *
+ * Why a valve: unbounded deferral is its own failure mode, both measured in
+ * the #1231 repro. The WAL duplicates hot pages per COMMIT, so it grows far
+ * faster than the DB (5.9GB WAL for a ~340MB DB on a 3.3k-file index) —
+ * filling the disk, and poisoning every subsequent read that must page
+ * through it (the first resolution-phase read blocked the main thread >60s
+ * and the #850 liveness watchdog killed the healthy index). The valve
+ * watches WAL growth on a timer and, past a soft threshold, backfills with
+ * `PRAGMA wal_checkpoint(PASSIVE)` on a worker-thread connection — PASSIVE
+ * never blocks the writer, and off-thread means the main thread (and the
+ * watchdog heartbeat) keep turning regardless of how long a backfill takes.
+ *
+ * The load-bearing subtlety: a WAL file's SIZE never shrinks. After a full
+ * backfill, the writer's next commit RESTARTS the WAL from the top and the
+ * frames recycle inside the same file — so raw size says nothing about the
+ * un-backfilled backlog, and a size-triggered valve degenerates into firing
+ * (and pausing the writer) forever once the file passes its threshold
+ * (measured: guava crawled at ~9min per 160 files). Instead the valve
+ * tracks `sizeAtLastFullBackfill` — refreshed whenever a checkpoint reports
+ * `log === checkpointed` (everything backfilled) — and triggers on GROWTH
+ * beyond that baseline, which only happens when genuinely un-backfilled
+ * frames push past the file's high-water mark.
+ *
+ * Backpressure: if the writer outruns the checkpointer past a hard cap of
+ * growth (2× soft), {@link backpressure} pauses the writer (at a safe,
+ * between-transactions boundary) until a FULL backfill lands. One in-flight
+ * pass is not enough: on a disk saturated by the writer, every concurrent
+ * PASSIVE pass is already stale by the time it finishes (the writer appended
+ * past its snapshot), so neither SQLite's WAL wrap nor the baseline ever
+ * trigger and the WAL grows without bound (measured: 5.9GB on guava at 150
+ * IOPS, then a >60s read stall and a watchdog kill). With the writer parked,
+ * the next pass covers everything, the WAL wraps on the following commit,
+ * and the pause is the disk's honest catch-up cost — the correct terminal
+ * mode when hardware genuinely can't keep up with the append rate.
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.WalCheckpointValve = void 0;
+exports.resolveWalValveMb = resolveWalValveMb;
+/** Soft WAL-growth threshold (MB) that triggers an off-thread passive checkpoint. */
+const DEFAULT_WAL_VALVE_MB = 256;
+/** Hard cap = this × soft threshold; past it the writer pauses for a full backfill. */
+const HARD_CAP_MULTIPLIER = 2;
+/** Passes attempted per writer pause before giving up (a pinned reader could stall forever). */
+const MAX_PAUSED_BACKFILL_PASSES = 20;
+/** How often the timer looks at the WAL file size. */
+const CHECK_INTERVAL_MS = 2000;
+/**
+ * Resolve the valve's soft threshold from the `CODEGRAPH_WAL_VALVE_MB`
+ * override; non-numeric / non-positive values fall back to the default.
+ */
+function resolveWalValveMb(envVal) {
+    if (envVal !== undefined && envVal !== '') {
+        const n = Number(envVal);
+        if (Number.isFinite(n) && n > 0)
+            return Math.floor(n);
+    }
+    return DEFAULT_WAL_VALVE_MB;
+}
+class WalCheckpointValve {
+    db;
+    intervalMs;
+    log;
+    timer = null;
+    inflight = null;
+    /** Writer pause in progress (hard cap breached): passes loop until a full backfill. */
+    pause = null;
+    /**
+     * WAL file size observed when a checkpoint last reported the ENTIRE WAL
+     * backfilled. Growth is measured against this baseline — see the header
+     * comment for why absolute size cannot be used.
+     */
+    sizeAtLastFullBackfill = 0;
+    softBytes;
+    hardBytes;
+    constructor(db, softMb = resolveWalValveMb(process.env.CODEGRAPH_WAL_VALVE_MB), intervalMs = CHECK_INTERVAL_MS, log = () => { }) {
+        this.db = db;
+        this.intervalMs = intervalMs;
+        this.log = log;
+        this.softBytes = softMb * 1024 * 1024;
+        this.hardBytes = this.softBytes * HARD_CAP_MULTIPLIER;
+    }
+    mb(n) {
+        return `${Math.round(n / 1024 / 1024)}MB`;
+    }
+    /** Un-backfilled growth estimate: bytes the WAL has grown past the last full backfill. */
+    growthBytes() {
+        return this.db.getWalSizeBytes() - this.sizeAtLastFullBackfill;
+    }
+    /** Begin watching the WAL. Idempotent; the timer never holds the loop open. */
+    start() {
+        if (this.timer)
+            return;
+        this.timer = setInterval(() => this.check(), this.intervalMs);
+        this.timer.unref?.();
+    }
+    /** Stop watching. Any in-flight checkpoint keeps running — await drain(). */
+    stop() {
+        if (this.timer) {
+            clearInterval(this.timer);
+            this.timer = null;
+        }
+    }
+    /** One poll: fire an off-thread passive checkpoint when growth passes the soft threshold. */
+    check() {
+        if (!this.pause && !this.inflight && this.growthBytes() > this.softBytes)
+            this.fire();
+    }
+    /**
+     * Writer-side backstop, called at a between-transactions boundary. Returns
+     * null (no wait) while growth is under the hard cap; past it, returns a
+     * promise that resolves only once a FULL backfill has landed — see the
+     * header comment for why a single pass is not enough on a saturated disk.
+     */
+    backpressure() {
+        if (this.pause)
+            return this.pause;
+        if (this.growthBytes() <= this.hardBytes)
+            return null;
+        this.log(`backpressure: wal=${this.mb(this.db.getWalSizeBytes())} baseline=${this.mb(this.sizeAtLastFullBackfill)} — pausing writer for full backfill`);
+        const t0 = Date.now();
+        this.pause = this.backfillFully().finally(() => {
+            this.pause = null;
+            this.log(`backpressure released after ${Date.now() - t0}ms: wal=${this.mb(this.db.getWalSizeBytes())} baseline=${this.mb(this.sizeAtLastFullBackfill)}`);
+        });
+        return this.pause;
+    }
+    /** Await any in-flight checkpoint and writer pause. */
+    async drain() {
+        while (this.pause || this.inflight) {
+            if (this.pause)
+                await this.pause;
+            if (this.inflight)
+                await this.inflight;
+        }
+    }
+    /**
+     * Phase-boundary fold: backfill the ENTIRE WAL now (off-thread, awaited).
+     * Called between bulk phases — e.g. after parsing, before resolution's
+     * first reads — so the next phase never pages a bulk-write-sized WAL on
+     * the main thread (the post-parse read against a multi-GB WAL is what
+     * blew the #850 watchdog's 60s window in the #1231 repro). The await
+     * keeps the event loop (and the watchdog heartbeat) turning.
+     */
+    async foldNow() {
+        await this.drain();
+        if (this.growthBytes() <= 0)
+            return;
+        this.log(`foldNow: wal=${this.mb(this.db.getWalSizeBytes())} baseline=${this.mb(this.sizeAtLastFullBackfill)}`);
+        this.pause = this.backfillFully().finally(() => { this.pause = null; });
+        await this.pause;
+    }
+    /**
+     * With the writer parked on the returned promise, loop passive passes until
+     * one reports the entire WAL backfilled (typically the second: the first
+     * drains the pass that was already running against a stale snapshot). Gives
+     * up after a bounded number of passes — e.g. a reader pinning the WAL —
+     * because unbounded WAL growth degrades; a wedged writer never recovers.
+     */
+    async backfillFully() {
+        for (let i = 0; i < MAX_PAUSED_BACKFILL_PASSES; i++) {
+            if (this.inflight)
+                await this.inflight; // fold in the stale in-flight pass first
+            const res = await this.db.checkpointWalPassive();
+            if (!res)
+                return; // checkpoint machinery unavailable — don't spin
+            this.log(`backfill pass ${i + 1}: busy=${res.busy} log=${res.log} checkpointed=${res.checkpointed} wal=${this.mb(this.db.getWalSizeBytes())}`);
+            if (res.busy === 0 && res.log === res.checkpointed) {
+                this.sizeAtLastFullBackfill = this.db.getWalSizeBytes();
+                return;
+            }
+        }
+        this.log(`backfill gave up after ${MAX_PAUSED_BACKFILL_PASSES} passes — WAL stays unbounded this cycle`);
+    }
+    fire() {
+        const p = this.db
+            .checkpointWalPassive()
+            .then((res) => {
+            // Full backfill (busy 0, every log frame checkpointed) ⇒ the writer's
+            // next commit wraps the WAL; the file's current size becomes the new
+            // growth baseline. A partial pass (writer appended during it, or a
+            // read transaction pinned frames) leaves the baseline alone, so the
+            // next tick fires again and copies the remainder. In non-WAL mode
+            // SQLite reports log = checkpointed = -1, which is harmless here.
+            if (res && res.busy === 0 && res.log === res.checkpointed) {
+                this.sizeAtLastFullBackfill = this.db.getWalSizeBytes();
+            }
+        })
+            .catch(() => { })
+            .finally(() => {
+            if (this.inflight === p)
+                this.inflight = null;
+        });
+        this.inflight = p;
+    }
+}
+exports.WalCheckpointValve = WalCheckpointValve;
+//# sourceMappingURL=wal-valve.js.map

package/dist/db/wal-valve.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"wal-valve.js","sourceRoot":"","sources":["../../src/db/wal-valve.ts"],"names":[],"mappings":";AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA4CG;;;AAiBH,8CAMC;AAnBD,qFAAqF;AACrF,MAAM,oBAAoB,GAAG,GAAG,CAAC;AACjC,uFAAuF;AACvF,MAAM,mBAAmB,GAAG,CAAC,CAAC;AAC9B,gGAAgG;AAChG,MAAM,0BAA0B,GAAG,EAAE,CAAC;AACtC,sDAAsD;AACtD,MAAM,iBAAiB,GAAG,IAAI,CAAC;AAE/B;;;GAGG;AACH,SAAgB,iBAAiB,CAAC,MAA0B;IAC1D,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,KAAK,EAAE,EAAE,CAAC;QAC1C,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;QACzB,IAAI,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC;YAAE,OAAO,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACxD,CAAC;IACD,OAAO,oBAAoB,CAAC;AAC9B,CAAC;AAED,MAAa,kBAAkB;IAeV;IAEA;IACA;IAjBX,KAAK,GAA0C,IAAI,CAAC;IACpD,QAAQ,GAAyB,IAAI,CAAC;IAC9C,uFAAuF;IAC/E,KAAK,GAAyB,IAAI,CAAC;IAC3C;;;;OAIG;IACK,sBAAsB,GAAG,CAAC,CAAC;IAClB,SAAS,CAAS;IAClB,SAAS,CAAS;IAEnC,YACmB,EAAsB,EACvC,SAAiB,iBAAiB,CAAC,OAAO,CAAC,GAAG,CAAC,sBAAsB,CAAC,EACrD,aAAqB,iBAAiB,EACtC,MAA6B,GAAG,EAAE,GAAE,CAAC;QAHrC,OAAE,GAAF,EAAE,CAAoB;QAEtB,eAAU,GAAV,UAAU,CAA4B;QACtC,QAAG,GAAH,GAAG,CAAkC;QAEtD,IAAI,CAAC,SAAS,GAAG,MAAM,GAAG,IAAI,GAAG,IAAI,CAAC;QACtC,IAAI,CAAC,SAAS,GAAG,IAAI,CAAC,SAAS,GAAG,mBAAmB,CAAC;IACxD,CAAC;IAEO,EAAE,CAAC,CAAS;QAClB,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;IAC5C,CAAC;IAED,0FAA0F;IAClF,WAAW;QACjB,OAAO,IAAI,CAAC,EAAE,CAAC,eAAe,EAAE,GAAG,IAAI,CAAC,sBAAsB,CAAC;IACjE,CAAC;IAED,+EAA+E;IAC/E,KAAK;QACH,IAAI,IAAI,CAAC,KAAK;YAAE,OAAO;QACvB,IAAI,CAAC,KAAK,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,KAAK,EAAE,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC;QAC9D,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,EAAE,CAAC;IACvB,CAAC;IAED,6EAA6E;IAC7E,IAAI;QACF,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YACf,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAC1B,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC;QACpB,CAAC;IACH,CAAC;IAED,6FAA6F;IAC7F,KAAK;QACH,IAAI,CAAC,IAAI,CAAC,KAAK,IAAI,CAAC,IAAI,CAAC,QAAQ,IAAI,IAAI,CAAC,WAAW,EAAE,GAAG,IAAI,CAAC,SAAS;YAAE,IAAI,CAAC,IAAI,EAAE,CAAC;IACxF,CAAC;IAED;;;;;OAKG;IACH,YAAY;QACV,IAAI,IAAI,CAAC,KAAK;YAAE,OAAO,IAAI,CAAC,KAAK,CAAC;QAClC,IAAI,IAAI,CAAC,WAAW,EAAE,IAAI,IAAI,CAAC,SAAS;YAAE,OAAO,IAAI,CAAC;QACtD,IAAI,CAAC,GAAG,CAAC,qBAAqB,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,eAAe,EAAE,CAAC,aAAa,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,sBAAsB,CAAC,qCAAqC,CAAC,CAAC;QACxJ,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QACtB,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,aAAa,EAAE,CAAC,OAAO,CAAC,GAAG,EAAE;YAC7C,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC;YAClB,IAAI,CAAC,GAAG,CAAC,+BAA+B,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE,WAAW,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,eAAe,EAAE,CAAC,aAAa,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,sBAAsB,CAAC,EAAE,CAAC,CAAC;QAC3J,CAAC,CAAC,CAAC;QACH,OAAO,IAAI,CAAC,KAAK,CAAC;IACpB,CAAC;IAED,uDAAuD;IACvD,KAAK,CAAC,KAAK;QACT,OAAO,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YACnC,IAAI,IAAI,CAAC,KAAK;gBAAE,MAAM,IAAI,CAAC,KAAK,CAAC;YACjC,IAAI,IAAI,CAAC,QAAQ;gBAAE,MAAM,IAAI,CAAC,QAAQ,CAAC;QACzC,CAAC;IACH,CAAC;IAED;;;;;;;OAOG;IACH,KAAK,CAAC,OAAO;QACX,MAAM,IAAI,CAAC,KAAK,EAAE,CAAC;QACnB,IAAI,IAAI,CAAC,WAAW,EAAE,IAAI,CAAC;YAAE,OAAO;QACpC,IAAI,CAAC,GAAG,CAAC,gBAAgB,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,eAAe,EAAE,CAAC,aAAa,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,sBAAsB,CAAC,EAAE,CAAC,CAAC;QAChH,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,aAAa,EAAE,CAAC,OAAO,CAAC,GAAG,EAAE,GAAG,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;QACxE,MAAM,IAAI,CAAC,KAAK,CAAC;IACnB,CAAC;IAED;;;;;;OAMG;IACK,KAAK,CAAC,aAAa;QACzB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,0BAA0B,EAAE,CAAC,EAAE,EAAE,CAAC;YACpD,IAAI,IAAI,CAAC,QAAQ;gBAAE,MAAM,IAAI,CAAC,QAAQ,CAAC,CAAC,yCAAyC;YACjF,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,EAAE,CAAC,oBAAoB,EAAE,CAAC;YACjD,IAAI,CAAC,GAAG;gBAAE,OAAO,CAAC,gDAAgD;YAClE,IAAI,CAAC,GAAG,CAAC,iBAAiB,CAAC,GAAG,CAAC,UAAU,GAAG,CAAC,IAAI,QAAQ,GAAG,CAAC,GAAG,iBAAiB,GAAG,CAAC,YAAY,QAAQ,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,eAAe,EAAE,CAAC,EAAE,CAAC,CAAC;YAC/I,IAAI,GAAG,CAAC,IAAI,KAAK,CAAC,IAAI,GAAG,CAAC,GAAG,KAAK,GAAG,CAAC,YAAY,EAAE,CAAC;gBACnD,IAAI,CAAC,sBAAsB,GAAG,IAAI,CAAC,EAAE,CAAC,eAAe,EAAE,CAAC;gBACxD,OAAO;YACT,CAAC;QACH,CAAC;QACD,IAAI,CAAC,GAAG,CAAC,0BAA0B,0BAA0B,0CAA0C,CAAC,CAAC;IAC3G,CAAC;IAEO,IAAI;QACV,MAAM,CAAC,GAAG,IAAI,CAAC,EAAE;aACd,oBAAoB,EAAE;aACtB,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE;YACZ,sEAAsE;YACtE,qEAAqE;YACrE,mEAAmE;YACnE,oEAAoE;YACpE,kEAAkE;YAClE,kEAAkE;YAClE,IAAI,GAAG,IAAI,GAAG,CAAC,IAAI,KAAK,CAAC,IAAI,GAAG,CAAC,GAAG,KAAK,GAAG,CAAC,YAAY,EAAE,CAAC;gBAC1D,IAAI,CAAC,sBAAsB,GAAG,IAAI,CAAC,EAAE,CAAC,eAAe,EAAE,CAAC;YAC1D,CAAC;QACH,CAAC,CAAC;aACD,KAAK,CAAC,GAAG,EAAE,GAAqB,CAAC,CAAC;aAClC,OAAO,CAAC,GAAG,EAAE;YACZ,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC;gBAAE,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC;QAChD,CAAC,CAAC,CAAC;QACL,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAC;IACpB,CAAC;CACF;AAxID,gDAwIC"}

package/dist/directory.d.ts

@@ -0,0 +1,183 @@
+/**
+ * Directory Management
+ *
+ * Manages the .codegraph/ directory structure for CodeGraph data.
+ */
+/**
+ * Resolve the per-project data directory name, honoring the `CODEGRAPH_DIR`
+ * environment override (default `.codegraph`). The override is a single path
+ * segment that lives in the project root.
+ *
+ * Why this exists: two environments that share one working tree must NOT share
+ * one `.codegraph/` — most concretely Windows-native and WSL (issue #636). The
+ * daemon lockfile (`.codegraph/daemon.pid`) records a platform-specific pid and
+ * socket path (a Windows named pipe vs a WSL Unix socket), and SQLite file
+ * locking across the WSL2 ↔ Windows filesystem boundary is unreliable, so two
+ * daemons sharing one index risks corruption. Setting `CODEGRAPH_DIR=.codegraph-win`
+ * on one side gives each environment its own index in the same tree.
+ *
+ * Read live (not captured at load) so it is both process-accurate and testable.
+ * An override that isn't a plain directory name — empty, containing a path
+ * separator, `.`, `..`/traversal, or absolute — is ignored (we keep the
+ * default) rather than risk writing the index outside the project or into the
+ * project root itself; we warn once to stderr so the misconfiguration is seen.
+ */
+export declare function codeGraphDirName(): string;
+/**
+ * CodeGraph directory name — a load-time snapshot of {@link codeGraphDirName}.
+ * A running process's environment is fixed, so this equals the live value;
+ * it's kept as a stable string export for backward compatibility. Internal code
+ * resolves the name through {@link codeGraphDirName} / {@link getCodeGraphDir}
+ * so the `CODEGRAPH_DIR` override always applies.
+ */
+export declare const CODEGRAPH_DIR: string;
+/**
+ * Is `name` (a single path segment) a CodeGraph data directory? Matches the
+ * default `.codegraph`, the active `CODEGRAPH_DIR` override, and any
+ * `.codegraph-*` sibling. File-watching and the indexer skip ALL of these, so
+ * when two environments share one working tree (Windows + WSL, issue #636)
+ * neither indexes or watches the other's index directory.
+ */
+export declare function isCodeGraphDataDir(name: string): boolean;
+/**
+ * Get the .codegraph directory path for a project
+ */
+export declare function getCodeGraphDir(projectRoot: string): string;
+/**
+ * Check if a project has been initialized with CodeGraph
+ * Requires both .codegraph/ directory AND codegraph.db to exist
+ */
+export declare function isInitialized(projectRoot: string): boolean;
+/**
+ * Find the nearest parent directory containing .codegraph/
+ *
+ * Walks up from the given path to find a CodeGraph-initialized project,
+ * similar to how git finds .git/ directories.
+ *
+ * @param startPath - Directory to start searching from
+ * @returns The project root containing .codegraph/, or null if not found
+ */
+/**
+ * Reason a directory is unsafe to use as an index ROOT, or null when it's fine.
+ *
+ * Indexing your home directory or a filesystem root drags in caches, `Library`,
+ * every other project, etc. — a multi-GB index, constant file-watcher churn, and
+ * (pre-1.0 on macOS) a file-descriptor blowup that exhausted `kern.maxfiles` and
+ * took unrelated apps / the whole machine down (#845). The classic trigger:
+ * running the installer or `codegraph init` from `$HOME`, which auto-indexes the
+ * current directory. These are never intended project roots, so the installer
+ * and `init`/`index` refuse them (overridable with `--force`).
+ *
+ * Pure-ish (reads only `os.homedir()` + realpath) so it's easy to unit-test.
+ * The returned string is a human phrase that slots into "… looks like {reason}".
+ */
+export declare function unsafeIndexRootReason(projectRoot: string): string | null;
+export declare function findNearestCodeGraphRoot(startPath: string): string | null;
+/**
+ * Indexed sub-project roots beneath `root` (bounded breadth-first scan). For
+ * the monorepo case behind #964: the index lives in a CHILD
+ * (`packages/x/.codegraph/`), not at the workspace root the agent's cwd points
+ * at. Descent stops at the first indexed directory on a branch (a project's
+ * own sub-dirs aren't separate projects) and is bounded by depth + count so it
+ * never turns into a full-tree crawl on a large repo.
+ */
+export declare function findIndexedSubprojectRoots(root: string, opts?: {
+    maxDepth?: number;
+    max?: number;
+}): string[];
+/**
+ * Does `prompt` contain an explicit structural keyword? A keyword is a strong,
+ * self-contained signal, so the front-load hook fires on it directly — no graph
+ * check needed. (A *code-token* match, by contrast, is only a candidate the
+ * hook verifies against the graph first; see {@link extractCodeTokens}.)
+ * Coverage is multilingual (#994, #1126): the ~29 languages with the largest
+ * developer populations, across Latin, Cyrillic, Greek, CJK, Hangul, Arabic,
+ * Hebrew, Thai, and Devanagari scripts. Languages beyond the keyword table
+ * still fire through the language-agnostic code-token path.
+ */
+export declare function hasStructuralKeyword(prompt: string): boolean;
+/**
+ * Identifier-shaped tokens in `prompt` — camelCase / PascalCase-with-inner-cap,
+ * snake_case, a `name(` call, or the two sides of an `a.b` member access. Naming
+ * a symbol is a code question whatever the surrounding human language, and these
+ * shapes almost never occur in ordinary prose, so they catch the common
+ * "<symbol> 的调用链?" / "where is <symbol> 定義" prompts no keyword list would.
+ *
+ * These are *candidates*, not a verdict: a tech brand like `JavaScript` or
+ * `GitHub` is identifier-shaped too, so the front-load hook checks each token
+ * against the actual index ({@link getNodesByName}) and only fires when one is a
+ * real symbol here — otherwise a brand-name prompt would inject ~16KB of
+ * low-relevance context (issue #994 follow-up). A doc/data filename ("README.md")
+ * is excluded from the member-access form since it's a file reference, not a symbol.
+ */
+export declare function extractCodeTokens(prompt: string): string[];
+/**
+ * Cheap, graph-free candidate gate for the front-load hook: could `prompt` be a
+ * structural / flow / impact / "where-how" question worth front-loading context
+ * for? True on an explicit keyword in any covered language (#994, #1126) OR an
+ * identifier-shaped token. A keyword is sufficient to fire on its own; a
+ * token-only match is only a candidate the hook then verifies against the graph
+ * (a brand name like `JavaScript` is token-shaped but isn't a symbol). Every
+ * non-candidate prompt ("fix this typo", in any language) stays a zero-cost no-op.
+ */
+export declare function isStructuralPrompt(prompt: string): boolean;
+/**
+ * What the front-load hook should do for a prompt issued from a directory.
+ */
+export interface FrontloadPlan {
+    /** Open + explore this project and inject its source as context. `null` when
+     *  there's no single project to front-load (none indexed, or several indexed
+     *  sub-projects with no clear match — see {@link nudgeProjects}). */
+    exploreRoot: string | null;
+    /** Indexed sub-projects to surface in a "pass `projectPath`" nudge: the rest
+     *  of a monorepo's indexed projects alongside `exploreRoot`, or — when no one
+     *  project clearly matches — the full list (with `exploreRoot` null). */
+    nudgeProjects: string[];
+    /** True when the plan came from scanning DOWN into sub-projects (cwd itself
+     *  is not under any index) — the monorepo case, where a follow-up
+     *  `codegraph_explore` needs an explicit `projectPath`. */
+    viaSubScan: boolean;
+}
+/**
+ * Decide what the front-load hook injects for a `prompt` issued from `cwd`,
+ * shaped by where the `.codegraph/` index(es) actually are:
+ *   1. **cwd (or an ancestor) is indexed** → front-load that project. The
+ *      normal single-project / nested-file case.
+ *   2. **cwd isn't indexed but looks like a workspace root** → the indexes live
+ *      in sub-projects (the monorepo case behind #964). One indexed
+ *      sub-project → front-load it; several → front-load the one the prompt
+ *      names (by relative path like `packages/api`, or package directory name)
+ *      and nudge about the rest; several with no match → nudge the full list so
+ *      the agent passes `projectPath`, rather than guessing wrong.
+ *   3. **nothing indexed reachable** → do nothing (the agent's own tools apply).
+ */
+export declare function planFrontload(cwd: string, prompt: string): FrontloadPlan;
+/**
+ * Create the .codegraph directory structure
+ * Note: Only throws if codegraph.db already exists, not just if .codegraph/ exists.
+ */
+export declare function createDirectory(projectRoot: string): void;
+/**
+ * Remove the .codegraph directory
+ */
+export declare function removeDirectory(projectRoot: string): void;
+/**
+ * Get all files in the .codegraph directory
+ */
+export declare function listDirectoryContents(projectRoot: string): string[];
+/**
+ * Get the total size of the .codegraph directory in bytes
+ */
+export declare function getDirectorySize(projectRoot: string): number;
+/**
+ * Ensure a subdirectory exists within .codegraph
+ */
+export declare function ensureSubdirectory(projectRoot: string, subdirName: string): string;
+/**
+ * Check if the .codegraph directory has valid structure
+ */
+export declare function validateDirectory(projectRoot: string): {
+    valid: boolean;
+    errors: string[];
+};
+//# sourceMappingURL=directory.d.ts.map

package/dist/directory.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"directory.d.ts","sourceRoot":"","sources":["../src/directory.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAWH;;;;;;;;;;;;;;;;;;GAkBG;AACH,wBAAgB,gBAAgB,IAAI,MAAM,CAqBzC;AAED;;;;;;GAMG;AACH,eAAO,MAAM,aAAa,QAAqB,CAAC;AAEhD;;;;;;GAMG;AACH,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAMxD;AAED;;GAEG;AACH,wBAAgB,eAAe,CAAC,WAAW,EAAE,MAAM,GAAG,MAAM,CAE3D;AAED;;;GAGG;AACH,wBAAgB,aAAa,CAAC,WAAW,EAAE,MAAM,GAAG,OAAO,CAQ1D;AAED;;;;;;;;GAQG;AACH;;;;;;;;;;;;;GAaG;AACH,wBAAgB,qBAAqB,CAAC,WAAW,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CA8BxE;AAED,wBAAgB,wBAAwB,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAmBzE;AA2BD;;;;;;;GAOG;AACH,wBAAgB,0BAA0B,CACxC,IAAI,EAAE,MAAM,EACZ,IAAI,GAAE;IAAE,QAAQ,CAAC,EAAE,MAAM,CAAC;IAAC,GAAG,CAAC,EAAE,MAAM,CAAA;CAAO,GAC7C,MAAM,EAAE,CAmBV;AAsND;;;;;;;;;GASG;AACH,wBAAgB,oBAAoB,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAK5D;AAED;;;;;;;;;;;;;GAaG;AACH,wBAAgB,iBAAiB,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,EAAE,CAkB1D;AAED;;;;;;;;GAQG;AACH,wBAAgB,kBAAkB,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAE1D;AAED;;GAEG;AACH,MAAM,WAAW,aAAa;IAC5B;;yEAEqE;IACrE,WAAW,EAAE,MAAM,GAAG,IAAI,CAAC;IAC3B;;6EAEyE;IACzE,aAAa,EAAE,MAAM,EAAE,CAAC;IACxB;;+DAE2D;IAC3D,UAAU,EAAE,OAAO,CAAC;CACrB;AAED;;;;;;;;;;;;GAYG;AACH,wBAAgB,aAAa,CAAC,GAAG,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,aAAa,CAwCxE;AAyDD;;;GAGG;AACH,wBAAgB,eAAe,CAAC,WAAW,EAAE,MAAM,GAAG,IAAI,CAgBzD;AAED;;GAEG;AACH,wBAAgB,eAAe,CAAC,WAAW,EAAE,MAAM,GAAG,IAAI,CAuBzD;AAED;;GAEG;AACH,wBAAgB,qBAAqB,CAAC,WAAW,EAAE,MAAM,GAAG,MAAM,EAAE,CA8BnE;AAED;;GAEG;AACH,wBAAgB,gBAAgB,CAAC,WAAW,EAAE,MAAM,GAAG,MAAM,CA+B5D;AAED;;GAEG;AACH,wBAAgB,kBAAkB,CAAC,WAAW,EAAE,MAAM,EAAE,UAAU,EAAE,MAAM,GAAG,MAAM,CAYlF;AAED;;GAEG;AACH,wBAAgB,iBAAiB,CAAC,WAAW,EAAE,MAAM,GAAG;IACtD,KAAK,EAAE,OAAO,CAAC;IACf,MAAM,EAAE,MAAM,EAAE,CAAC;CAClB,CA6BA"}