youtube-knowledge-mcp 2.0.2 → 2.2.0

package/dist/utils/brain-ingest.d.ts

@@ -0,0 +1,57 @@
+import type { BrainChunk, BrainVideoState } from '../brain-schemas.js';
+import { type VideoListItem } from './youtube.js';
+/**
+ * Reading one video into passages.
+ *
+ * Nothing here throws for anything to do with the video itself. A channel with
+ * a members-only upload, a live stream still processing, or an episode nobody
+ * captioned is still a channel worth having a brain for; which videos those
+ * were belongs in the manifest, not in an exception that loses the other two
+ * hundred.
+ *
+ * Metadata is read before the transcript, in that order deliberately: it is the
+ * cheaper request, it carries the date and length the filters need, and a video
+ * ruled out by them then costs one request instead of two.
+ */
+/**
+ * A transcript that chunks past this is not a transcript. Recording the video
+ * as failed keeps the brain buildable; letting it through does not.
+ *
+ * The brain-wide ceiling is what bounds memory, and it was chosen by measuring
+ * rather than by feel: a corpus of 100,000 passages is 81MB on disk, and the
+ * heaviest thing done to it — the phrase pass — peaked at 767MB of heap. Raising
+ * it means measuring again.
+ */
+export declare const MAX_CHUNKS_PER_VIDEO = 2000;
+export declare const MAX_CHUNKS_PER_BRAIN = 100000;
+/** Recorded against a video the size guards refused, alongside yt-dlp's own codes. */
+export declare const TOO_LARGE = "TOO_LARGE";
+export declare const BRAIN_FULL = "BRAIN_FULL";
+export interface VideoFilters {
+    /** Ignore anything published before this date (YYYY-MM-DD). */
+    since?: string;
+    /** Ignore anything shorter, to leave out shorts and clips. */
+    minDurationSeconds: number;
+}
+export interface IngestOptions extends VideoFilters {
+    language: string;
+}
+export interface IngestResult {
+    state: BrainVideoState;
+    chunks: BrainChunk[];
+}
+export declare function ingestVideo(video: VideoListItem, chunksSoFar: number, options: IngestOptions): Promise<IngestResult>;
+/**
+ * Whether the filters rule this video out, or `undefined` when its recorded
+ * state cannot say.
+ *
+ * A build asks this of the manifest before deciding what to read, which is what
+ * lets a changed filter take effect without re-fetching anything: once a video
+ * has been looked at, its real date and length are on disk and the question is
+ * answerable from there. `undefined` is the honest answer for a video only ever
+ * seen in a flat listing — not a licence to assume it qualifies.
+ */
+export declare function isExcluded(state: BrainVideoState, filters: VideoFilters): boolean | undefined;
+export declare function pendingState(video: VideoListItem): BrainVideoState;
+export declare function excluded(state: BrainVideoState): BrainVideoState;
+//# sourceMappingURL=brain-ingest.d.ts.map

package/dist/utils/brain-ingest.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-ingest.d.ts","sourceRoot":"","sources":["../../src/utils/brain-ingest.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AAKvE,OAAO,EAAkC,KAAK,aAAa,EAAE,MAAM,cAAc,CAAC;AAElF;;;;;;;;;;;;GAYG;AAEH;;;;;;;;GAQG;AACH,eAAO,MAAM,oBAAoB,OAAO,CAAC;AACzC,eAAO,MAAM,oBAAoB,SAAU,CAAC;AAE5C,sFAAsF;AACtF,eAAO,MAAM,SAAS,cAAc,CAAC;AACrC,eAAO,MAAM,UAAU,eAAe,CAAC;AAEvC,MAAM,WAAW,YAAY;IAC3B,+DAA+D;IAC/D,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,8DAA8D;IAC9D,kBAAkB,EAAE,MAAM,CAAC;CAC5B;AAED,MAAM,WAAW,aAAc,SAAQ,YAAY;IACjD,QAAQ,EAAE,MAAM,CAAC;CAClB;AAED,MAAM,WAAW,YAAY;IAC3B,KAAK,EAAE,eAAe,CAAC;IACvB,MAAM,EAAE,UAAU,EAAE,CAAC;CACtB;AAED,wBAAsB,WAAW,CAC/B,KAAK,EAAE,aAAa,EACpB,WAAW,EAAE,MAAM,EACnB,OAAO,EAAE,aAAa,GACrB,OAAO,CAAC,YAAY,CAAC,CAoDvB;AAED;;;;;;;;;GASG;AACH,wBAAgB,UAAU,CAAC,KAAK,EAAE,eAAe,EAAE,OAAO,EAAE,YAAY,GAAG,OAAO,GAAG,SAAS,CAe7F;AAED,wBAAgB,YAAY,CAAC,KAAK,EAAE,aAAa,GAAG,eAAe,CAWlE;AAED,wBAAgB,QAAQ,CAAC,KAAK,EAAE,eAAe,GAAG,eAAe,CAEhE"}

package/dist/utils/brain-ingest.js

@@ -0,0 +1,127 @@
+import { chunkTranscript } from './brain-chunks.js';
+import { throwIfAborted } from './context.js';
+import { asYouTubeError } from './errors.js';
+import { countWords } from './transcript.js';
+import { getTranscript, getVideoDetails } from './youtube.js';
+/**
+ * Reading one video into passages.
+ *
+ * Nothing here throws for anything to do with the video itself. A channel with
+ * a members-only upload, a live stream still processing, or an episode nobody
+ * captioned is still a channel worth having a brain for; which videos those
+ * were belongs in the manifest, not in an exception that loses the other two
+ * hundred.
+ *
+ * Metadata is read before the transcript, in that order deliberately: it is the
+ * cheaper request, it carries the date and length the filters need, and a video
+ * ruled out by them then costs one request instead of two.
+ */
+/**
+ * A transcript that chunks past this is not a transcript. Recording the video
+ * as failed keeps the brain buildable; letting it through does not.
+ *
+ * The brain-wide ceiling is what bounds memory, and it was chosen by measuring
+ * rather than by feel: a corpus of 100,000 passages is 81MB on disk, and the
+ * heaviest thing done to it — the phrase pass — peaked at 767MB of heap. Raising
+ * it means measuring again.
+ */
+export const MAX_CHUNKS_PER_VIDEO = 2000;
+export const MAX_CHUNKS_PER_BRAIN = 100_000;
+/** Recorded against a video the size guards refused, alongside yt-dlp's own codes. */
+export const TOO_LARGE = 'TOO_LARGE';
+export const BRAIN_FULL = 'BRAIN_FULL';
+export async function ingestVideo(video, chunksSoFar, options) {
+    const listed = pendingState(video);
+    try {
+        // The listing is flat, which is what makes enumerating a thousand videos one
+        // request rather than a thousand — and flat entries carry no date. These are
+        // the real values, and the chapters come back in the same request.
+        const details = await getVideoDetails(video.id);
+        const known = {
+            ...listed,
+            uploadDate: details.uploadDate,
+            durationSeconds: details.durationSeconds,
+        };
+        if (isExcluded(known, options) === true)
+            return { state: excluded(known), chunks: [] };
+        const transcript = await getTranscript(video.id, { language: options.language });
+        const chunks = chunkTranscript({
+            videoId: video.id,
+            title: video.title,
+            segments: transcript.segments,
+            chapters: details.chapters,
+        });
+        if (chunks.length > MAX_CHUNKS_PER_VIDEO)
+            return refused(known, TOO_LARGE);
+        if (chunksSoFar + chunks.length > MAX_CHUNKS_PER_BRAIN)
+            return refused(known, BRAIN_FULL);
+        return {
+            state: {
+                ...known,
+                state: 'indexed',
+                chunkCount: chunks.length,
+                wordCount: chunks.reduce((total, chunk) => total + countWords(chunk.text), 0),
+            },
+            chunks,
+        };
+    }
+    catch (error) {
+        // Cancellation is the client's decision, not this video's problem. Asking
+        // the signal rather than classifying the error catches it however it was
+        // raised — an aborted spawn, a rejected read, or the check between videos.
+        throwIfAborted();
+        const failure = asYouTubeError(error);
+        return {
+            state: failure.code === 'NO_CAPTIONS'
+                ? { ...listed, state: 'no-captions' }
+                : { ...listed, state: 'failed', error: failure.code },
+            chunks: [],
+        };
+    }
+}
+/**
+ * Whether the filters rule this video out, or `undefined` when its recorded
+ * state cannot say.
+ *
+ * A build asks this of the manifest before deciding what to read, which is what
+ * lets a changed filter take effect without re-fetching anything: once a video
+ * has been looked at, its real date and length are on disk and the question is
+ * answerable from there. `undefined` is the honest answer for a video only ever
+ * seen in a flat listing — not a licence to assume it qualifies.
+ */
+export function isExcluded(state, filters) {
+    const { since, minDurationSeconds } = filters;
+    let undecided = false;
+    if (minDurationSeconds > 0) {
+        if (state.durationSeconds <= 0)
+            undecided = true;
+        else if (state.durationSeconds < minDurationSeconds)
+            return true;
+    }
+    if (since !== undefined) {
+        if (state.uploadDate === '')
+            undecided = true;
+        else if (state.uploadDate < since)
+            return true;
+    }
+    return undecided ? undefined : false;
+}
+export function pendingState(video) {
+    return {
+        videoId: video.id,
+        title: video.title,
+        url: video.url,
+        uploadDate: video.uploadDate,
+        durationSeconds: video.duration,
+        state: 'pending',
+        chunkCount: 0,
+        wordCount: 0,
+    };
+}
+export function excluded(state) {
+    return { ...state, state: 'excluded', chunkCount: 0, wordCount: 0 };
+}
+function refused(state, reason) {
+    return { state: { ...state, state: 'failed', error: reason }, chunks: [] };
+}
+//# sourceMappingURL=brain-ingest.js.map

package/dist/utils/brain-ingest.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-ingest.js","sourceRoot":"","sources":["../../src/utils/brain-ingest.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,eAAe,EAAE,MAAM,mBAAmB,CAAC;AACpD,OAAO,EAAE,cAAc,EAAE,MAAM,cAAc,CAAC;AAC9C,OAAO,EAAE,cAAc,EAAE,MAAM,aAAa,CAAC;AAC7C,OAAO,EAAE,UAAU,EAAE,MAAM,iBAAiB,CAAC;AAC7C,OAAO,EAAE,aAAa,EAAE,eAAe,EAAsB,MAAM,cAAc,CAAC;AAElF;;;;;;;;;;;;GAYG;AAEH;;;;;;;;GAQG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAAG,IAAI,CAAC;AACzC,MAAM,CAAC,MAAM,oBAAoB,GAAG,OAAO,CAAC;AAE5C,sFAAsF;AACtF,MAAM,CAAC,MAAM,SAAS,GAAG,WAAW,CAAC;AACrC,MAAM,CAAC,MAAM,UAAU,GAAG,YAAY,CAAC;AAkBvC,MAAM,CAAC,KAAK,UAAU,WAAW,CAC/B,KAAoB,EACpB,WAAmB,EACnB,OAAsB;IAEtB,MAAM,MAAM,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;IAEnC,IAAI,CAAC;QACH,6EAA6E;QAC7E,6EAA6E;QAC7E,mEAAmE;QACnE,MAAM,OAAO,GAAG,MAAM,eAAe,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QAChD,MAAM,KAAK,GAAoB;YAC7B,GAAG,MAAM;YACT,UAAU,EAAE,OAAO,CAAC,UAAU;YAC9B,eAAe,EAAE,OAAO,CAAC,eAAe;SACzC,CAAC;QAEF,IAAI,UAAU,CAAC,KAAK,EAAE,OAAO,CAAC,KAAK,IAAI;YAAE,OAAO,EAAE,KAAK,EAAE,QAAQ,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC;QAEvF,MAAM,UAAU,GAAG,MAAM,aAAa,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,OAAO,CAAC,QAAQ,EAAE,CAAC,CAAC;QACjF,MAAM,MAAM,GAAG,eAAe,CAAC;YAC7B,OAAO,EAAE,KAAK,CAAC,EAAE;YACjB,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,QAAQ,EAAE,UAAU,CAAC,QAAQ;YAC7B,QAAQ,EAAE,OAAO,CAAC,QAAQ;SAC3B,CAAC,CAAC;QAEH,IAAI,MAAM,CAAC,MAAM,GAAG,oBAAoB;YAAE,OAAO,OAAO,CAAC,KAAK,EAAE,SAAS,CAAC,CAAC;QAC3E,IAAI,WAAW,GAAG,MAAM,CAAC,MAAM,GAAG,oBAAoB;YAAE,OAAO,OAAO,CAAC,KAAK,EAAE,UAAU,CAAC,CAAC;QAE1F,OAAO;YACL,KAAK,EAAE;gBACL,GAAG,KAAK;gBACR,KAAK,EAAE,SAAS;gBAChB,UAAU,EAAE,MAAM,CAAC,MAAM;gBACzB,SAAS,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;aAC9E;YACD,MAAM;SACP,CAAC;IACJ,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,0EAA0E;QAC1E,yEAAyE;QACzE,2EAA2E;QAC3E,cAAc,EAAE,CAAC;QAEjB,MAAM,OAAO,GAAG,cAAc,CAAC,KAAK,CAAC,CAAC;QAEtC,OAAO;YACL,KAAK,EACH,OAAO,CAAC,IAAI,KAAK,aAAa;gBAC5B,CAAC,CAAC,EAAE,GAAG,MAAM,EAAE,KAAK,EAAE,aAAa,EAAE;gBACrC,CAAC,CAAC,EAAE,GAAG,MAAM,EAAE,KAAK,EAAE,QAAQ,EAAE,KAAK,EAAE,OAAO,CAAC,IAAI,EAAE;YACzD,MAAM,EAAE,EAAE;SACX,CAAC;IACJ,CAAC;AACH,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,UAAU,UAAU,CAAC,KAAsB,EAAE,OAAqB;IACtE,MAAM,EAAE,KAAK,EAAE,kBAAkB,EAAE,GAAG,OAAO,CAAC;IAC9C,IAAI,SAAS,GAAG,KAAK,CAAC;IAEtB,IAAI,kBAAkB,GAAG,CAAC,EAAE,CAAC;QAC3B,IAAI,KAAK,CAAC,eAAe,IAAI,CAAC;YAAE,SAAS,GAAG,IAAI,CAAC;aAC5C,IAAI,KAAK,CAAC,eAAe,GAAG,kBAAkB;YAAE,OAAO,IAAI,CAAC;IACnE,CAAC;IAED,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;QACxB,IAAI,KAAK,CAAC,UAAU,KAAK,EAAE;YAAE,SAAS,GAAG,IAAI,CAAC;aACzC,IAAI,KAAK,CAAC,UAAU,GAAG,KAAK;YAAE,OAAO,IAAI,CAAC;IACjD,CAAC;IAED,OAAO,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC;AACvC,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,KAAoB;IAC/C,OAAO;QACL,OAAO,EAAE,KAAK,CAAC,EAAE;QACjB,KAAK,EAAE,KAAK,CAAC,KAAK;QAClB,GAAG,EAAE,KAAK,CAAC,GAAG;QACd,UAAU,EAAE,KAAK,CAAC,UAAU;QAC5B,eAAe,EAAE,KAAK,CAAC,QAAQ;QAC/B,KAAK,EAAE,SAAS;QAChB,UAAU,EAAE,CAAC;QACb,SAAS,EAAE,CAAC;KACb,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,QAAQ,CAAC,KAAsB;IAC7C,OAAO,EAAE,GAAG,KAAK,EAAE,KAAK,EAAE,UAAU,EAAE,UAAU,EAAE,CAAC,EAAE,SAAS,EAAE,CAAC,EAAE,CAAC;AACtE,CAAC;AAED,SAAS,OAAO,CAAC,KAAsB,EAAE,MAAc;IACrD,OAAO,EAAE,KAAK,EAAE,EAAE,GAAG,KAAK,EAAE,KAAK,EAAE,QAAQ,EAAE,KAAK,EAAE,MAAM,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC;AAC7E,CAAC"}

package/dist/utils/brain-lock.d.ts

@@ -0,0 +1,4 @@
+export { LOCK_STALE_MS } from './file-lock.js';
+/** Exclusive access to a brain while it is being built. See `withFileLock`. */
+export declare function withBuildLock<T>(channelId: string, build: () => Promise<T>): Promise<T>;
+//# sourceMappingURL=brain-lock.d.ts.map

package/dist/utils/brain-lock.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-lock.d.ts","sourceRoot":"","sources":["../../src/utils/brain-lock.ts"],"names":[],"mappings":"AAKA,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE/C,+EAA+E;AAC/E,wBAAsB,aAAa,CAAC,CAAC,EAAE,SAAS,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,OAAO,CAAC,CAAC,CAAC,GAAG,OAAO,CAAC,CAAC,CAAC,CAK7F"}

package/dist/utils/brain-lock.js

@@ -0,0 +1,17 @@
+import { ensureBrainDir } from './brain-storage.js';
+import { lockPath } from './brain-paths.js';
+import { YouTubeError } from './errors.js';
+import { withFileLock } from './file-lock.js';
+export { LOCK_STALE_MS } from './file-lock.js';
+/** Exclusive access to a brain while it is being built. See `withFileLock`. */
+export async function withBuildLock(channelId, build) {
+    await ensureBrainDir(channelId);
+    return withFileLock(lockPath(channelId), build, (existing) => alreadyRunning(channelId, existing));
+}
+function alreadyRunning(channelId, existing) {
+    const owner = existing === undefined ? '' : `Started at ${existing.startedAt} by process ${existing.pid}. `;
+    return new YouTubeError('INVALID_INPUT', `A build for ${channelId} is already running.`, {
+        nextStep: `${owner}Wait for it to finish, then call build_brain again to continue where it stopped.`,
+    });
+}
+//# sourceMappingURL=brain-lock.js.map

package/dist/utils/brain-lock.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-lock.js","sourceRoot":"","sources":["../../src/utils/brain-lock.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAC;AACpD,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAC3C,OAAO,EAAE,YAAY,EAAmB,MAAM,gBAAgB,CAAC;AAE/D,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAE/C,+EAA+E;AAC/E,MAAM,CAAC,KAAK,UAAU,aAAa,CAAI,SAAiB,EAAE,KAAuB;IAC/E,MAAM,cAAc,CAAC,SAAS,CAAC,CAAC;IAChC,OAAO,YAAY,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,KAAK,EAAE,CAAC,QAAQ,EAAE,EAAE,CAC3D,cAAc,CAAC,SAAS,EAAE,QAAQ,CAAC,CACpC,CAAC;AACJ,CAAC;AAED,SAAS,cAAc,CAAC,SAAiB,EAAE,QAAgC;IACzE,MAAM,KAAK,GACT,QAAQ,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,cAAc,QAAQ,CAAC,SAAS,eAAe,QAAQ,CAAC,GAAG,IAAI,CAAC;IAEhG,OAAO,IAAI,YAAY,CAAC,eAAe,EAAE,eAAe,SAAS,sBAAsB,EAAE;QACvF,QAAQ,EAAE,GAAG,KAAK,kFAAkF;KACrG,CAAC,CAAC;AACL,CAAC"}

package/dist/utils/brain-lookup.d.ts

@@ -0,0 +1,11 @@
+import type { BrainManifest } from '../brain-schemas.js';
+/**
+ * Find a brain from whatever the caller called the channel.
+ *
+ * Every brain tool except `build_brain` answers from disk, so none of them
+ * should need the network to work out which brain is meant. The matching
+ * itself is shared with every other channel-keyed store; what is specific here
+ * is what to say when there is no brain, or more than one.
+ */
+export declare function resolveBrain(channel: string): Promise<BrainManifest>;
+//# sourceMappingURL=brain-lookup.d.ts.map

package/dist/utils/brain-lookup.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-lookup.d.ts","sourceRoot":"","sources":["../../src/utils/brain-lookup.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AAKzD;;;;;;;GAOG;AACH,wBAAsB,YAAY,CAAC,OAAO,EAAE,MAAM,GAAG,OAAO,CAAC,aAAa,CAAC,CAU1E"}

package/dist/utils/brain-lookup.js

@@ -0,0 +1,36 @@
+import { listManifests } from './brain-storage.js';
+import { findByChannel } from './channel-lookup.js';
+import { YouTubeError } from './errors.js';
+/**
+ * Find a brain from whatever the caller called the channel.
+ *
+ * Every brain tool except `build_brain` answers from disk, so none of them
+ * should need the network to work out which brain is meant. The matching
+ * itself is shared with every other channel-keyed store; what is specific here
+ * is what to say when there is no brain, or more than one.
+ */
+export async function resolveBrain(channel) {
+    const manifests = await listManifests();
+    const matches = findByChannel(manifests, channel);
+    const [first, second] = matches;
+    if (first === undefined)
+        throw noBrain(channel, manifests);
+    if (second !== undefined)
+        throw ambiguous(channel, matches);
+    return first;
+}
+function noBrain(channel, manifests) {
+    const built = manifests.map((manifest) => manifest.channel.handle || manifest.channel.name);
+    return new YouTubeError('NOT_FOUND', `No brain has been built for "${channel}".`, {
+        nextStep: built.length === 0
+            ? 'Call build_brain with the channel URL or handle to build one.'
+            : `Call build_brain to build it. Brains that exist: ${built.join(', ')}.`,
+    });
+}
+function ambiguous(channel, matches) {
+    const names = matches.map((manifest) => manifest.channel.channelId);
+    return new YouTubeError('INVALID_INPUT', `"${channel}" matches more than one brain.`, {
+        nextStep: `Use the channel id instead: ${names.join(', ')}.`,
+    });
+}
+//# sourceMappingURL=brain-lookup.js.map

package/dist/utils/brain-lookup.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-lookup.js","sourceRoot":"","sources":["../../src/utils/brain-lookup.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAC;AACnD,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAE3C;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,YAAY,CAAC,OAAe;IAChD,MAAM,SAAS,GAAG,MAAM,aAAa,EAAE,CAAC;IACxC,MAAM,OAAO,GAAG,aAAa,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;IAElD,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,GAAG,OAAO,CAAC;IAEhC,IAAI,KAAK,KAAK,SAAS;QAAE,MAAM,OAAO,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC;IAC3D,IAAI,MAAM,KAAK,SAAS;QAAE,MAAM,SAAS,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;IAE5D,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,OAAO,CAAC,OAAe,EAAE,SAA0B;IAC1D,MAAM,KAAK,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,MAAM,IAAI,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IAE5F,OAAO,IAAI,YAAY,CAAC,WAAW,EAAE,gCAAgC,OAAO,IAAI,EAAE;QAChF,QAAQ,EACN,KAAK,CAAC,MAAM,KAAK,CAAC;YAChB,CAAC,CAAC,+DAA+D;YACjE,CAAC,CAAC,oDAAoD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG;KAC9E,CAAC,CAAC;AACL,CAAC;AAED,SAAS,SAAS,CAAC,OAAe,EAAE,OAAwB;IAC1D,MAAM,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;IAEpE,OAAO,IAAI,YAAY,CAAC,eAAe,EAAE,IAAI,OAAO,gCAAgC,EAAE;QACpF,QAAQ,EAAE,+BAA+B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG;KAC7D,CAAC,CAAC;AACL,CAAC"}

package/dist/utils/brain-paths.d.ts

@@ -0,0 +1,19 @@
+/**
+ * Every file a brain is made of.
+ *
+ * Separate from the modules that read and write them so that the manifest layer
+ * and the index layer can each know where things live without importing each
+ * other.
+ *
+ * A brain is keyed on the channel id rather than the handle: handles get
+ * renamed, and a renamed handle would orphan the brain built under the old one.
+ * The id is also what reaches the filesystem, so it is validated here — once,
+ * at the only place that turns one into a path.
+ */
+export declare function brainsDir(): string;
+export declare function brainDir(channelId: string): string;
+export declare function manifestPath(channelId: string): string;
+export declare function chunksPath(channelId: string): string;
+export declare function profilePath(channelId: string): string;
+export declare function lockPath(channelId: string): string;
+//# sourceMappingURL=brain-paths.d.ts.map

package/dist/utils/brain-paths.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-paths.d.ts","sourceRoot":"","sources":["../../src/utils/brain-paths.ts"],"names":[],"mappings":"AAIA;;;;;;;;;;;GAWG;AAEH,wBAAgB,SAAS,IAAI,MAAM,CAElC;AAED,wBAAgB,QAAQ,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAElD;AAED,wBAAgB,YAAY,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAEtD;AAED,wBAAgB,UAAU,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAEpD;AAED,wBAAgB,WAAW,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAErD;AAED,wBAAgB,QAAQ,CAAC,SAAS,EAAE,MAAM,GAAG,MAAM,CAElD"}

package/dist/utils/brain-paths.js

@@ -0,0 +1,34 @@
+import { join } from 'node:path';
+import { dataDir } from './paths.js';
+import { assertChannelId } from './validate.js';
+/**
+ * Every file a brain is made of.
+ *
+ * Separate from the modules that read and write them so that the manifest layer
+ * and the index layer can each know where things live without importing each
+ * other.
+ *
+ * A brain is keyed on the channel id rather than the handle: handles get
+ * renamed, and a renamed handle would orphan the brain built under the old one.
+ * The id is also what reaches the filesystem, so it is validated here — once,
+ * at the only place that turns one into a path.
+ */
+export function brainsDir() {
+    return dataDir('brains');
+}
+export function brainDir(channelId) {
+    return join(brainsDir(), assertChannelId(channelId));
+}
+export function manifestPath(channelId) {
+    return join(brainDir(channelId), 'manifest.json');
+}
+export function chunksPath(channelId) {
+    return join(brainDir(channelId), 'chunks.json');
+}
+export function profilePath(channelId) {
+    return join(brainDir(channelId), 'profile.md');
+}
+export function lockPath(channelId) {
+    return join(brainDir(channelId), 'build.lock');
+}
+//# sourceMappingURL=brain-paths.js.map

package/dist/utils/brain-paths.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-paths.js","sourceRoot":"","sources":["../../src/utils/brain-paths.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AACjC,OAAO,EAAE,OAAO,EAAE,MAAM,YAAY,CAAC;AACrC,OAAO,EAAE,eAAe,EAAE,MAAM,eAAe,CAAC;AAEhD;;;;;;;;;;;GAWG;AAEH,MAAM,UAAU,SAAS;IACvB,OAAO,OAAO,CAAC,QAAQ,CAAC,CAAC;AAC3B,CAAC;AAED,MAAM,UAAU,QAAQ,CAAC,SAAiB;IACxC,OAAO,IAAI,CAAC,SAAS,EAAE,EAAE,eAAe,CAAC,SAAS,CAAC,CAAC,CAAC;AACvD,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,SAAiB;IAC5C,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,eAAe,CAAC,CAAC;AACpD,CAAC;AAED,MAAM,UAAU,UAAU,CAAC,SAAiB;IAC1C,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,aAAa,CAAC,CAAC;AAClD,CAAC;AAED,MAAM,UAAU,WAAW,CAAC,SAAiB;IAC3C,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,YAAY,CAAC,CAAC;AACjD,CAAC;AAED,MAAM,UAAU,QAAQ,CAAC,SAAiB;IACxC,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,YAAY,CAAC,CAAC;AACjD,CAAC"}

package/dist/utils/brain-phrases.d.ts

@@ -0,0 +1,31 @@
+import type { BrainChunk } from '../brain-schemas.js';
+/**
+ * Phrases a creator repeats across videos.
+ *
+ * This is the one statistic here that is not a plain count, so it is worth
+ * saying what it does and does not claim. A phrase appearing in several
+ * different videos is repeated; that is a fact about the corpus. Whether it is
+ * a catchphrase, a verbal tic or a segment title is a judgement, and belongs to
+ * whoever reads the list.
+ *
+ * Counting every 3-to-6-word window at once would hold millions of keys in
+ * memory. Instead each length is counted only where the length below it
+ * survived — a phrase cannot recur more often than its own prefix does — so all
+ * but the first pass work on a small candidate set.
+ *
+ * It is still the most expensive thing done to a corpus: 12.6s over 100,000
+ * passages, which is why a build runs it once at the end rather than at every
+ * checkpoint.
+ */
+export declare const PHRASE_MIN_WORDS = 3;
+export declare const PHRASE_MAX_WORDS = 6;
+/** Repetition inside one video is a habit of that video, not of the channel. */
+export declare const PHRASE_MIN_VIDEOS = 3;
+export declare const PHRASE_LIMIT = 20;
+export interface RecurringPhrase {
+    phrase: string;
+    videoCount: number;
+    occurrences: number;
+}
+export declare function findRecurringPhrases(chunks: BrainChunk[]): RecurringPhrase[];
+//# sourceMappingURL=brain-phrases.d.ts.map

package/dist/utils/brain-phrases.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-phrases.d.ts","sourceRoot":"","sources":["../../src/utils/brain-phrases.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,qBAAqB,CAAC;AAGtD;;;;;;;;;;;;;;;;;GAiBG;AAEH,eAAO,MAAM,gBAAgB,IAAI,CAAC;AAClC,eAAO,MAAM,gBAAgB,IAAI,CAAC;AAClC,gFAAgF;AAChF,eAAO,MAAM,iBAAiB,IAAI,CAAC;AACnC,eAAO,MAAM,YAAY,KAAK,CAAC;AAE/B,MAAM,WAAW,eAAe;IAC9B,MAAM,EAAE,MAAM,CAAC;IACf,UAAU,EAAE,MAAM,CAAC;IACnB,WAAW,EAAE,MAAM,CAAC;CACrB;AAaD,wBAAgB,oBAAoB,CAAC,MAAM,EAAE,UAAU,EAAE,GAAG,eAAe,EAAE,CAuB5E"}

package/dist/utils/brain-phrases.js

@@ -0,0 +1,116 @@
+import { isStopWord } from './search-index.js';
+/**
+ * Phrases a creator repeats across videos.
+ *
+ * This is the one statistic here that is not a plain count, so it is worth
+ * saying what it does and does not claim. A phrase appearing in several
+ * different videos is repeated; that is a fact about the corpus. Whether it is
+ * a catchphrase, a verbal tic or a segment title is a judgement, and belongs to
+ * whoever reads the list.
+ *
+ * Counting every 3-to-6-word window at once would hold millions of keys in
+ * memory. Instead each length is counted only where the length below it
+ * survived — a phrase cannot recur more often than its own prefix does — so all
+ * but the first pass work on a small candidate set.
+ *
+ * It is still the most expensive thing done to a corpus: 12.6s over 100,000
+ * passages, which is why a build runs it once at the end rather than at every
+ * checkpoint.
+ */
+export const PHRASE_MIN_WORDS = 3;
+export const PHRASE_MAX_WORDS = 6;
+/** Repetition inside one video is a habit of that video, not of the channel. */
+export const PHRASE_MIN_VIDEOS = 3;
+export const PHRASE_LIMIT = 20;
+export function findRecurringPhrases(chunks) {
+    const corpus = wordsByVideo(chunks);
+    const kept = [];
+    let candidates = new Set();
+    for (let length = PHRASE_MIN_WORDS; length <= PHRASE_MAX_WORDS; length++) {
+        const tallies = countPhrases(corpus, length, length === PHRASE_MIN_WORDS ? undefined : candidates);
+        candidates = new Set();
+        for (const [phrase, tally] of tallies) {
+            if (tally.videoCount < PHRASE_MIN_VIDEOS)
+                continue;
+            candidates.add(phrase);
+            kept.push({ phrase, videoCount: tally.videoCount, occurrences: tally.occurrences });
+        }
+        if (candidates.size === 0)
+            break;
+    }
+    return rank(kept);
+}
+/**
+ * The corpus as one word list per video, in a stable order.
+ *
+ * Grouping is what lets distinct-video counts be kept as a single field per
+ * phrase rather than a set of video ids per phrase, which for a large channel
+ * is the difference between tens of megabytes and hundreds.
+ */
+function wordsByVideo(chunks) {
+    const byVideo = new Map();
+    for (const chunk of chunks) {
+        const words = byVideo.get(chunk.videoId) ?? [];
+        words.push(...phraseWords(chunk.text));
+        byVideo.set(chunk.videoId, words);
+    }
+    return [...byVideo.values()];
+}
+/** Lowercased words, apostrophes kept because "don't" is one word. */
+function phraseWords(text) {
+    return text
+        .toLowerCase()
+        .split(/[^\p{L}\p{N}']+/u)
+        .filter(Boolean);
+}
+function countPhrases(corpus, length, candidatePrefixes) {
+    const tallies = new Map();
+    for (const [videoIndex, words] of corpus.entries()) {
+        for (let start = 0; start + length <= words.length; start++) {
+            const window = words.slice(start, start + length);
+            // A longer phrase only recurs where its prefix already did. A phrase
+            // whose opening words are all filler is therefore never reached, which is
+            // the intended trade: it costs "and then i went on" to drop "and then i".
+            if (candidatePrefixes !== undefined) {
+                if (!candidatePrefixes.has(window.slice(0, length - 1).join(' ')))
+                    continue;
+            }
+            else if (window.every(isStopWord)) {
+                continue;
+            }
+            const phrase = window.join(' ');
+            const tally = tallies.get(phrase);
+            if (tally === undefined) {
+                tallies.set(phrase, { occurrences: 1, videoCount: 1, lastVideoIndex: videoIndex });
+                continue;
+            }
+            tally.occurrences++;
+            if (tally.lastVideoIndex !== videoIndex) {
+                tally.videoCount++;
+                tally.lastVideoIndex = videoIndex;
+            }
+        }
+    }
+    return tallies;
+}
+/**
+ * Drop the truncations, then order by reach.
+ *
+ * A phrase contained in a longer kept phrase that is just as widespread says
+ * nothing the longer one does not — otherwise every long catchphrase drags in
+ * three truncations of itself. Subsumption has to consider the longest first,
+ * which is not the order worth reporting, so the two happen in separate passes.
+ */
+function rank(phrases) {
+    const longestFirst = [...phrases].sort((a, b) => b.phrase.split(' ').length - a.phrase.split(' ').length || b.videoCount - a.videoCount);
+    const distinct = [];
+    for (const candidate of longestFirst) {
+        const subsumed = distinct.some((existing) => existing.phrase.includes(candidate.phrase) && existing.videoCount >= candidate.videoCount);
+        if (!subsumed)
+            distinct.push(candidate);
+    }
+    return distinct
+        .sort((a, b) => b.videoCount - a.videoCount || b.occurrences - a.occurrences)
+        .slice(0, PHRASE_LIMIT);
+}
+//# sourceMappingURL=brain-phrases.js.map

package/dist/utils/brain-phrases.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-phrases.js","sourceRoot":"","sources":["../../src/utils/brain-phrases.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAE/C;;;;;;;;;;;;;;;;;GAiBG;AAEH,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,CAAC;AAClC,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,CAAC;AAClC,gFAAgF;AAChF,MAAM,CAAC,MAAM,iBAAiB,GAAG,CAAC,CAAC;AACnC,MAAM,CAAC,MAAM,YAAY,GAAG,EAAE,CAAC;AAmB/B,MAAM,UAAU,oBAAoB,CAAC,MAAoB;IACvD,MAAM,MAAM,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC;IACpC,MAAM,IAAI,GAAsB,EAAE,CAAC;IACnC,IAAI,UAAU,GAAG,IAAI,GAAG,EAAU,CAAC;IAEnC,KAAK,IAAI,MAAM,GAAG,gBAAgB,EAAE,MAAM,IAAI,gBAAgB,EAAE,MAAM,EAAE,EAAE,CAAC;QACzE,MAAM,OAAO,GAAG,YAAY,CAC1B,MAAM,EACN,MAAM,EACN,MAAM,KAAK,gBAAgB,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,UAAU,CACrD,CAAC;QAEF,UAAU,GAAG,IAAI,GAAG,EAAE,CAAC;QACvB,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,OAAO,EAAE,CAAC;YACtC,IAAI,KAAK,CAAC,UAAU,GAAG,iBAAiB;gBAAE,SAAS;YACnD,UAAU,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;YACvB,IAAI,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,UAAU,EAAE,KAAK,CAAC,UAAU,EAAE,WAAW,EAAE,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;QACtF,CAAC;QAED,IAAI,UAAU,CAAC,IAAI,KAAK,CAAC;YAAE,MAAM;IACnC,CAAC;IAED,OAAO,IAAI,CAAC,IAAI,CAAC,CAAC;AACpB,CAAC;AAED;;;;;;GAMG;AACH,SAAS,YAAY,CAAC,MAAoB;IACxC,MAAM,OAAO,GAAG,IAAI,GAAG,EAAoB,CAAC;IAE5C,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;QAC/C,KAAK,CAAC,IAAI,CAAC,GAAG,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;QACvC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;IACpC,CAAC;IAED,OAAO,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;AAC/B,CAAC;AAED,sEAAsE;AACtE,SAAS,WAAW,CAAC,IAAY;IAC/B,OAAO,IAAI;SACR,WAAW,EAAE;SACb,KAAK,CAAC,kBAAkB,CAAC;SACzB,MAAM,CAAC,OAAO,CAAC,CAAC;AACrB,CAAC;AAED,SAAS,YAAY,CACnB,MAAkB,EAClB,MAAc,EACd,iBAA0C;IAE1C,MAAM,OAAO,GAAG,IAAI,GAAG,EAAiB,CAAC;IAEzC,KAAK,MAAM,CAAC,UAAU,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,EAAE,EAAE,CAAC;QACnD,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,MAAM,IAAI,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;YAC5D,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,GAAG,MAAM,CAAC,CAAC;YAElD,qEAAqE;YACrE,0EAA0E;YAC1E,0EAA0E;YAC1E,IAAI,iBAAiB,KAAK,SAAS,EAAE,CAAC;gBACpC,IAAI,CAAC,iBAAiB,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,MAAM,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;oBAAE,SAAS;YAC9E,CAAC;iBAAM,IAAI,MAAM,CAAC,KAAK,CAAC,UAAU,CAAC,EAAE,CAAC;gBACpC,SAAS;YACX,CAAC;YAED,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAChC,MAAM,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;YAElC,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;gBACxB,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,EAAE,WAAW,EAAE,CAAC,EAAE,UAAU,EAAE,CAAC,EAAE,cAAc,EAAE,UAAU,EAAE,CAAC,CAAC;gBACnF,SAAS;YACX,CAAC;YAED,KAAK,CAAC,WAAW,EAAE,CAAC;YACpB,IAAI,KAAK,CAAC,cAAc,KAAK,UAAU,EAAE,CAAC;gBACxC,KAAK,CAAC,UAAU,EAAE,CAAC;gBACnB,KAAK,CAAC,cAAc,GAAG,UAAU,CAAC;YACpC,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,IAAI,CAAC,OAA0B;IACtC,MAAM,YAAY,GAAG,CAAC,GAAG,OAAO,CAAC,CAAC,IAAI,CACpC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC,UAAU,GAAG,CAAC,CAAC,UAAU,CACjG,CAAC;IAEF,MAAM,QAAQ,GAAsB,EAAE,CAAC;IAEvC,KAAK,MAAM,SAAS,IAAI,YAAY,EAAE,CAAC;QACrC,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,CAC5B,CAAC,QAAQ,EAAE,EAAE,CACX,QAAQ,CAAC,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,MAAM,CAAC,IAAI,QAAQ,CAAC,UAAU,IAAI,SAAS,CAAC,UAAU,CAC5F,CAAC;QACF,IAAI,CAAC,QAAQ;YAAE,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;IAC1C,CAAC;IAED,OAAO,QAAQ;SACZ,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,GAAG,CAAC,CAAC,UAAU,IAAI,CAAC,CAAC,WAAW,GAAG,CAAC,CAAC,WAAW,CAAC;SAC5E,KAAK,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC;AAC5B,CAAC"}

package/dist/utils/brain-stats.d.ts

@@ -0,0 +1,25 @@
+import type { BrainChunk, BrainStats, BrainVideoState } from '../brain-schemas.js';
+/**
+ * What the server can say about a channel without inferring anything.
+ *
+ * Everything here is counted or measured from the corpus. Deliberately absent:
+ * named entities, because auto-generated captions arrive lowercase and
+ * unpunctuated so capitalisation heuristics produce noise; and distinctive
+ * vocabulary by TF-IDF, because there is no honest background corpus to compare
+ * a channel against and inventing one would produce confident nonsense.
+ *
+ * Anything the numbers cannot support belongs in the profile, which a model
+ * writes from passages it can cite.
+ */
+export interface StatsOptions {
+    /**
+     * Whether to run the phrase pass, which reads every word of the corpus. It is
+     * measured in seconds at the sizes a brain is allowed to reach — 12.6s over
+     * 100,000 passages on the machine this was written on — so a build that
+     * checkpoints every ten videos must not pay for it fifty times over. Omitted
+     * rather than guessed at when it has not been run.
+     */
+    phrases: boolean;
+}
+export declare function computeStats(all: BrainVideoState[], chunks: BrainChunk[], options: StatsOptions): BrainStats;
+//# sourceMappingURL=brain-stats.d.ts.map

package/dist/utils/brain-stats.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-stats.d.ts","sourceRoot":"","sources":["../../src/utils/brain-stats.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,UAAU,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AAGnF;;;;;;;;;;;GAWG;AAEH,MAAM,WAAW,YAAY;IAC3B;;;;;;OAMG;IACH,OAAO,EAAE,OAAO,CAAC;CAClB;AAED,wBAAgB,YAAY,CAC1B,GAAG,EAAE,eAAe,EAAE,EACtB,MAAM,EAAE,UAAU,EAAE,EACpB,OAAO,EAAE,YAAY,GACpB,UAAU,CAyBZ"}

package/dist/utils/brain-stats.js

@@ -0,0 +1,64 @@
+import { findRecurringPhrases } from './brain-phrases.js';
+export function computeStats(all, chunks, options) {
+    // A video the filters ruled out is not part of this brain, so it does not get
+    // to move the brain's numbers — its date must not stretch the upload range,
+    // and its length must not enter the speaking rate. It is reported as its own
+    // count instead, because "we chose not to read 40 videos" is worth knowing.
+    const videos = all.filter((video) => video.state !== 'excluded');
+    const uploadDates = videos
+        .map((video) => video.uploadDate)
+        .filter(Boolean)
+        .sort();
+    return {
+        videoCount: videos.length,
+        excludedCount: all.length - videos.length,
+        indexedCount: countState(videos, 'indexed'),
+        noCaptionsCount: countState(videos, 'no-captions'),
+        failedCount: countState(videos, 'failed'),
+        pendingCount: countState(videos, 'pending'),
+        chunkCount: chunks.length,
+        totalWords: videos.reduce((total, video) => total + video.wordCount, 0),
+        medianWordsPerMinute: medianWordsPerMinute(videos),
+        ...firstAndLast(uploadDates),
+        uploadsPerMonth: uploadsPerMonth(uploadDates),
+        ...(options.phrases ? { recurringPhrases: findRecurringPhrases(chunks) } : {}),
+    };
+}
+function countState(videos, state) {
+    return videos.filter((video) => video.state === state).length;
+}
+/**
+ * The median rather than the mean: one mis-timed transcript, or a ten-hour
+ * ambience upload with forty words in it, would drag an average anywhere.
+ */
+function medianWordsPerMinute(videos) {
+    const rates = videos
+        .filter((video) => video.durationSeconds > 0 && video.wordCount > 0)
+        .map((video) => (video.wordCount / video.durationSeconds) * 60)
+        .sort((a, b) => a - b);
+    if (rates.length === 0)
+        return 0;
+    const middle = Math.floor(rates.length / 2);
+    const median = rates.length % 2 === 0
+        ? ((rates[middle - 1] ?? 0) + (rates[middle] ?? 0)) / 2
+        : (rates[middle] ?? 0);
+    return Math.round(median);
+}
+function firstAndLast(sortedDates) {
+    const first = sortedDates[0];
+    const last = sortedDates[sortedDates.length - 1];
+    return {
+        ...(first === undefined ? {} : { firstUpload: first }),
+        ...(last === undefined ? {} : { lastUpload: last }),
+    };
+}
+/** Months with no uploads are absent rather than zero: this is a tally, not a series. */
+function uploadsPerMonth(sortedDates) {
+    const months = new Map();
+    for (const date of sortedDates) {
+        const month = date.slice(0, 7);
+        months.set(month, (months.get(month) ?? 0) + 1);
+    }
+    return [...months.entries()].map(([month, videos]) => ({ month, videos }));
+}
+//# sourceMappingURL=brain-stats.js.map

package/dist/utils/brain-stats.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"brain-stats.js","sourceRoot":"","sources":["../../src/utils/brain-stats.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,oBAAoB,EAAE,MAAM,oBAAoB,CAAC;AA0B1D,MAAM,UAAU,YAAY,CAC1B,GAAsB,EACtB,MAAoB,EACpB,OAAqB;IAErB,8EAA8E;IAC9E,4EAA4E;IAC5E,6EAA6E;IAC7E,4EAA4E;IAC5E,MAAM,MAAM,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,UAAU,CAAC,CAAC;IACjE,MAAM,WAAW,GAAG,MAAM;SACvB,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,UAAU,CAAC;SAChC,MAAM,CAAC,OAAO,CAAC;SACf,IAAI,EAAE,CAAC;IAEV,OAAO;QACL,UAAU,EAAE,MAAM,CAAC,MAAM;QACzB,aAAa,EAAE,GAAG,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM;QACzC,YAAY,EAAE,UAAU,CAAC,MAAM,EAAE,SAAS,CAAC;QAC3C,eAAe,EAAE,UAAU,CAAC,MAAM,EAAE,aAAa,CAAC;QAClD,WAAW,EAAE,UAAU,CAAC,MAAM,EAAE,QAAQ,CAAC;QACzC,YAAY,EAAE,UAAU,CAAC,MAAM,EAAE,SAAS,CAAC;QAC3C,UAAU,EAAE,MAAM,CAAC,MAAM;QACzB,UAAU,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,GAAG,KAAK,CAAC,SAAS,EAAE,CAAC,CAAC;QACvE,oBAAoB,EAAE,oBAAoB,CAAC,MAAM,CAAC;QAClD,GAAG,YAAY,CAAC,WAAW,CAAC;QAC5B,eAAe,EAAE,eAAe,CAAC,WAAW,CAAC;QAC7C,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,gBAAgB,EAAE,oBAAoB,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAC/E,CAAC;AACJ,CAAC;AAED,SAAS,UAAU,CAAC,MAAyB,EAAE,KAA+B;IAC5E,OAAO,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC,MAAM,CAAC;AAChE,CAAC;AAED;;;GAGG;AACH,SAAS,oBAAoB,CAAC,MAAyB;IACrD,MAAM,KAAK,GAAG,MAAM;SACjB,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,eAAe,GAAG,CAAC,IAAI,KAAK,CAAC,SAAS,GAAG,CAAC,CAAC;SACnE,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,SAAS,GAAG,KAAK,CAAC,eAAe,CAAC,GAAG,EAAE,CAAC;SAC9D,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAEzB,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAEjC,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC5C,MAAM,MAAM,GACV,KAAK,CAAC,MAAM,GAAG,CAAC,KAAK,CAAC;QACpB,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC;QACvD,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC;IAE3B,OAAO,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;AAC5B,CAAC;AAED,SAAS,YAAY,CAAC,WAAqB;IACzC,MAAM,KAAK,GAAG,WAAW,CAAC,CAAC,CAAC,CAAC;IAC7B,MAAM,IAAI,GAAG,WAAW,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAEjD,OAAO;QACL,GAAG,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,KAAK,EAAE,CAAC;QACtD,GAAG,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC;KACpD,CAAC;AACJ,CAAC;AAED,yFAAyF;AACzF,SAAS,eAAe,CAAC,WAAqB;IAC5C,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;IAEzC,KAAK,MAAM,IAAI,IAAI,WAAW,EAAE,CAAC;QAC/B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;QAC/B,MAAM,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAClD,CAAC;IAED,OAAO,CAAC,GAAG,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,EAAE,MAAM,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC;AAC7E,CAAC"}