@monoes/monomindcli 2.13.0 → 2.14.1

package/dist/src/knowledge/html-tags.js

@@ -0,0 +1,229 @@
+/**
+ * Tag and entity tables for the HTML extractor (RCL-01).
+ *
+ * Split out of `html-extract.ts` purely for size: these are data, the file
+ * next door is the algorithm.
+ *
+ * @module v1/cli/knowledge/html-tags
+ */
+export const VOID_TAGS = new Set([
+    'area',
+    'base',
+    'br',
+    'col',
+    'embed',
+    'hr',
+    'img',
+    'input',
+    'link',
+    'meta',
+    'param',
+    'source',
+    'track',
+    'wbr',
+]);
+/** Elements whose content is never prose. Dropped regardless of options. */
+export const NON_TEXT_TAGS = new Set([
+    'script',
+    'style',
+    'noscript',
+    'svg',
+    'math',
+    'template',
+    'canvas',
+    'iframe',
+    'object',
+    'applet',
+    'audio',
+    'video',
+]);
+/** Structural chrome. Dropped only when `stripBoilerplate` is on. */
+export const CHROME_TAGS = new Set([
+    'nav',
+    'header',
+    'footer',
+    'aside',
+    'menu',
+    'dialog',
+    'button',
+    'select',
+    'textarea',
+]);
+/** ARIA landmark roles that mark chrome rather than content. */
+export const CHROME_ROLES = new Set([
+    'navigation',
+    'banner',
+    'contentinfo',
+    'complementary',
+    'search',
+    'menubar',
+    'menu',
+    'toolbar',
+    'dialog',
+    'alertdialog',
+]);
+/**
+ * Class/id word list. Matched against `-`/`_`/space/camelCase-split tokens, so
+ * `site-header`, `cookieBanner` and `ad_slot` all hit and `content-wrapper`
+ * does not.
+ */
+export const CHROME_WORDS = new Set([
+    'cookie',
+    'cookies',
+    'consent',
+    'gdpr',
+    'ccpa',
+    'cmp',
+    'banner',
+    'promo',
+    'promotion',
+    'ad',
+    'ads',
+    'advert',
+    'advertisement',
+    'adsbygoogle',
+    'sponsored',
+    'newsletter',
+    'subscribe',
+    'signup',
+    'paywall',
+    'popup',
+    'modal',
+    'overlay',
+    'lightbox',
+    'nav',
+    'navbar',
+    'navigation',
+    'menu',
+    'breadcrumb',
+    'breadcrumbs',
+    'masthead',
+    'topbar',
+    'toolbar',
+    'sidebar',
+    'skiplink',
+    'pagination',
+    'paginate',
+    'pager',
+    'share',
+    'sharing',
+    'social',
+    'socials',
+    'footer',
+    'header',
+    'copyright',
+    'related',
+    'recommended',
+    'recirc',
+    'trending',
+    'comments',
+    'disqus',
+]);
+/** Elements that start a new output block. */
+export const BLOCK_TAGS = new Set([
+    'address',
+    'article',
+    'blockquote',
+    'br',
+    'dd',
+    'details',
+    'div',
+    'dl',
+    'dt',
+    'fieldset',
+    'figcaption',
+    'figure',
+    'footer',
+    'form',
+    'h1',
+    'h2',
+    'h3',
+    'h4',
+    'h5',
+    'h6',
+    'header',
+    'hr',
+    'li',
+    'main',
+    'nav',
+    'ol',
+    'p',
+    'pre',
+    'section',
+    'summary',
+    'table',
+    'tbody',
+    'tfoot',
+    'thead',
+    'tr',
+    'ul',
+]);
+/** Table cells are joined into their row with ` | ` rather than broken out
+ *  into blocks of their own, so a row survives as one readable line. */
+export const CELL_TAGS = new Set(['td', 'th']);
+/** Elements whose content is raw text — `<` inside them is not markup. */
+export const RAW_TEXT_TAGS = new Set(['script', 'style', 'textarea', 'title']);
+// ── Entities ───────────────────────────────────────────────────────
+const NAMED_ENTITIES = {
+    amp: '&',
+    lt: '<',
+    gt: '>',
+    quot: '"',
+    apos: "'",
+    nbsp: ' ',
+    ensp: ' ',
+    emsp: ' ',
+    thinsp: ' ',
+    shy: '',
+    mdash: '—',
+    ndash: '–',
+    hellip: '…',
+    lsquo: '‘',
+    rsquo: '’',
+    ldquo: '“',
+    rdquo: '”',
+    bull: '•',
+    middot: '·',
+    laquo: '«',
+    raquo: '»',
+    copy: '©',
+    reg: '®',
+    trade: '™',
+    deg: '°',
+    euro: '€',
+    pound: '£',
+    yen: '¥',
+    cent: '¢',
+    sect: '§',
+    para: '¶',
+    dagger: '†',
+    times: '×',
+    divide: '÷',
+    plusmn: '±',
+    frac12: '½',
+    prime: '′',
+    larr: '←',
+    rarr: '→',
+    harr: '↔',
+};
+export function decodeEntities(s) {
+    if (!s.includes('&'))
+        return s;
+    return s.replace(/&(#[xX]?[0-9a-fA-F]+|[a-zA-Z][a-zA-Z0-9]{1,31});/g, (whole, body) => {
+        if (body[0] === '#') {
+            const hex = body[1] === 'x' || body[1] === 'X';
+            const code = Number.parseInt(hex ? body.slice(2) : body.slice(1), hex ? 16 : 10);
+            if (!Number.isFinite(code) || code <= 0 || code > 0x10ffff)
+                return whole;
+            try {
+                return String.fromCodePoint(code);
+            }
+            catch {
+                return whole;
+            }
+        }
+        const hit = NAMED_ENTITIES[body] ?? NAMED_ENTITIES[body.toLowerCase()];
+        return hit === undefined ? whole : hit;
+    });
+}
+//# sourceMappingURL=html-tags.js.map

package/dist/src/knowledge/html-tags.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-tags.js","sourceRoot":"","sources":["../../../src/knowledge/html-tags.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,MAAM,CAAC,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC;IAC/B,MAAM;IACN,MAAM;IACN,IAAI;IACJ,KAAK;IACL,OAAO;IACP,IAAI;IACJ,KAAK;IACL,OAAO;IACP,MAAM;IACN,MAAM;IACN,OAAO;IACP,QAAQ;IACR,OAAO;IACP,KAAK;CACN,CAAC,CAAC;AAEH,4EAA4E;AAC5E,MAAM,CAAC,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC;IACnC,QAAQ;IACR,OAAO;IACP,UAAU;IACV,KAAK;IACL,MAAM;IACN,UAAU;IACV,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,OAAO;IACP,OAAO;CACR,CAAC,CAAC;AAEH,qEAAqE;AACrE,MAAM,CAAC,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC;IACjC,KAAK;IACL,QAAQ;IACR,QAAQ;IACR,OAAO;IACP,MAAM;IACN,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,UAAU;CACX,CAAC,CAAC;AAEH,gEAAgE;AAChE,MAAM,CAAC,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC;IAClC,YAAY;IACZ,QAAQ;IACR,aAAa;IACb,eAAe;IACf,QAAQ;IACR,SAAS;IACT,MAAM;IACN,SAAS;IACT,QAAQ;IACR,aAAa;CACd,CAAC,CAAC;AAEH;;;;GAIG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC;IAClC,QAAQ;IACR,SAAS;IACT,SAAS;IACT,MAAM;IACN,MAAM;IACN,KAAK;IACL,QAAQ;IACR,OAAO;IACP,WAAW;IACX,IAAI;IACJ,KAAK;IACL,QAAQ;IACR,eAAe;IACf,aAAa;IACb,WAAW;IACX,YAAY;IACZ,WAAW;IACX,QAAQ;IACR,SAAS;IACT,OAAO;IACP,OAAO;IACP,SAAS;IACT,UAAU;IACV,KAAK;IACL,QAAQ;IACR,YAAY;IACZ,MAAM;IACN,YAAY;IACZ,aAAa;IACb,UAAU;IACV,QAAQ;IACR,SAAS;IACT,SAAS;IACT,UAAU;IACV,YAAY;IACZ,UAAU;IACV,OAAO;IACP,OAAO;IACP,SAAS;IACT,QAAQ;IACR,SAAS;IACT,QAAQ;IACR,QAAQ;IACR,WAAW;IACX,SAAS;IACT,aAAa;IACb,QAAQ;IACR,UAAU;IACV,UAAU;IACV,QAAQ;CACT,CAAC,CAAC;AAEH,8CAA8C;AAC9C,MAAM,CAAC,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC;IAChC,SAAS;IACT,SAAS;IACT,YAAY;IACZ,IAAI;IACJ,IAAI;IACJ,SAAS;IACT,KAAK;IACL,IAAI;IACJ,IAAI;IACJ,UAAU;IACV,YAAY;IACZ,QAAQ;IACR,QAAQ;IACR,MAAM;IACN,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,QAAQ;IACR,IAAI;IACJ,IAAI;IACJ,MAAM;IACN,KAAK;IACL,IAAI;IACJ,GAAG;IACH,KAAK;IACL,SAAS;IACT,SAAS;IACT,OAAO;IACP,OAAO;IACP,OAAO;IACP,OAAO;IACP,IAAI;IACJ,IAAI;CACL,CAAC,CAAC;AAEH;wEACwE;AACxE,MAAM,CAAC,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;AAE/C,0EAA0E;AAC1E,MAAM,CAAC,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC,CAAC,QAAQ,EAAE,OAAO,EAAE,UAAU,EAAE,OAAO,CAAC,CAAC,CAAC;AAE/E,sEAAsE;AAEtE,MAAM,cAAc,GAA2B;IAC7C,GAAG,EAAE,GAAG;IACR,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,GAAG;IACP,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,MAAM,EAAE,GAAG;IACX,GAAG,EAAE,EAAE;IACP,KAAK,EAAE,GAAG;IACV,KAAK,EAAE,GAAG;IACV,MAAM,EAAE,GAAG;IACX,KAAK,EAAE,GAAG;IACV,KAAK,EAAE,GAAG;IACV,KAAK,EAAE,GAAG;IACV,KAAK,EAAE,GAAG;IACV,IAAI,EAAE,GAAG;IACT,MAAM,EAAE,GAAG;IACX,KAAK,EAAE,GAAG;IACV,KAAK,EAAE,GAAG;IACV,IAAI,EAAE,GAAG;IACT,GAAG,EAAE,GAAG;IACR,KAAK,EAAE,GAAG;IACV,GAAG,EAAE,GAAG;IACR,IAAI,EAAE,GAAG;IACT,KAAK,EAAE,GAAG;IACV,GAAG,EAAE,GAAG;IACR,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,MAAM,EAAE,GAAG;IACX,KAAK,EAAE,GAAG;IACV,MAAM,EAAE,GAAG;IACX,MAAM,EAAE,GAAG;IACX,MAAM,EAAE,GAAG;IACX,KAAK,EAAE,GAAG;IACV,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;CACV,CAAC;AAEF,MAAM,UAAU,cAAc,CAAC,CAAS;IACtC,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,CAAC,CAAC;IAC/B,OAAO,CAAC,CAAC,OAAO,CAAC,mDAAmD,EAAE,CAAC,KAAK,EAAE,IAAY,EAAE,EAAE;QAC5F,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,GAAG,EAAE,CAAC;YACpB,MAAM,GAAG,GAAG,IAAI,CAAC,CAAC,CAAC,KAAK,GAAG,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC;YAC/C,MAAM,IAAI,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YACjF,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,IAAI,IAAI,GAAG,QAAQ;gBAAE,OAAO,KAAK,CAAC;YACzE,IAAI,CAAC;gBACH,OAAO,MAAM,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC;YACpC,CAAC;YAAC,MAAM,CAAC;gBACP,OAAO,KAAK,CAAC;YACf,CAAC;QACH,CAAC;QACD,MAAM,GAAG,GAAG,cAAc,CAAC,IAAI,CAAC,IAAI,cAAc,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;QACvE,OAAO,GAAG,KAAK,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC;IACzC,CAAC,CAAC,CAAC;AACL,CAAC"}

package/dist/src/knowledge/library.d.ts

@@ -0,0 +1,106 @@
+/**
+ * The capture library (RCL-09) — browsing what has been captured, rather than
+ * searching it.
+ *
+ * `listDocuments` answers "what is indexed". A library answers a different
+ * question: everything from this site, captured last week, tagged `reading`,
+ * newest first. The facets come from the capture envelope's `meta.json`
+ * (RCL-07), which is already on every record — this module is the query layer
+ * over them, and the data layer the desktop GUI renders.
+ *
+ * Rules, once, here:
+ *  - within a facet values are OR'd, across facets AND'd (`--site a --site b
+ *    --tag x` = "from a or b, tagged x");
+ *  - `--site example.com` matches `www.example.com` and `docs.example.com`,
+ *    because nobody browsing their own library means the bare host only;
+ *  - a document with no provenance (an ordinary file on disk) sorts by its
+ *    index time and matches no capture facet — it is not a capture.
+ *
+ * @module v1/cli/knowledge/library
+ */
+import type { CaptureProvenance } from './capture-envelope.js';
+import type { DocumentMeta } from './document-pipeline.js';
+export interface LibraryFilter {
+    /** Host, or a parent domain of it. */
+    site?: string[];
+    tag?: string[];
+    collection?: string[];
+    /** `extension` | `monobrowse` | `crawl`. */
+    source?: string[];
+    /** ISO date/time, or a relative age: `7d`, `24h`, `30m`. */
+    since?: string;
+    until?: string;
+    /** Only documents that came from a capture envelope. */
+    capturedOnly?: boolean;
+    scope?: string;
+    /** Substring over title, URL and path. */
+    text?: string;
+}
+export type LibrarySort = 'captured' | 'indexed' | 'title' | 'path' | 'size';
+export interface LibraryEntry {
+    filePath: string;
+    scope: string;
+    title: string;
+    url?: string;
+    site?: string;
+    capturedAt: string;
+    indexedAt: string;
+    source?: string;
+    collection?: string;
+    tags: string[];
+    chunkCount: number;
+    size: number;
+    version?: number;
+    /** False for ordinary files on disk: no capture envelope backed this one. */
+    captured: boolean;
+}
+/** The host a document was captured from, or undefined for a local file. */
+export declare function documentSite(meta: {
+    canonicalUrl?: string;
+    provenance?: CaptureProvenance;
+}): string | undefined;
+/** When the page was captured; falls back to when it was indexed. */
+export declare function documentCapturedAt(meta: {
+    indexedAt?: string;
+    provenance?: CaptureProvenance;
+}): string;
+/**
+ * A point in time from `2026-09-01`, a full ISO stamp, or an AGE like `7d`
+ * (meaning "7 days ago"). Returns null for anything unparseable so a typo
+ * cannot silently filter the whole library away.
+ */
+export declare function parseWhen(value: string | undefined, now?: number): number | null;
+/** True when a document belongs in this view. */
+export declare function matchesLibraryFilter(meta: DocumentMeta, filter?: LibraryFilter, now?: number): boolean;
+export declare function filterDocuments(docs: DocumentMeta[], filter?: LibraryFilter, now?: number): DocumentMeta[];
+export declare function sortDocuments(docs: DocumentMeta[], sort?: LibrarySort, order?: 'asc' | 'desc'): DocumentMeta[];
+/** The flat, JSON-friendly row the GUI and `--json` consumers read. */
+export declare function libraryEntry(meta: DocumentMeta): LibraryEntry;
+export interface ListLibraryOptions extends LibraryFilter {
+    sort?: LibrarySort;
+    order?: 'asc' | 'desc';
+    limit?: number;
+    now?: number;
+}
+/** Browse the library: filter, sort, cap. Never throws on an empty store. */
+export declare function listLibrary(docs: DocumentMeta[], opts?: ListLibraryOptions): LibraryEntry[];
+/** Facet counts for the current view — what a sidebar renders. */
+export declare function libraryFacets(docs: DocumentMeta[]): {
+    sites: Array<{
+        value: string;
+        count: number;
+    }>;
+    tags: Array<{
+        value: string;
+        count: number;
+    }>;
+    collections: Array<{
+        value: string;
+        count: number;
+    }>;
+    sources: Array<{
+        value: string;
+        count: number;
+    }>;
+};
+//# sourceMappingURL=library.d.ts.map

package/dist/src/knowledge/library.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"library.d.ts","sourceRoot":"","sources":["../../../src/knowledge/library.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAGH,OAAO,KAAK,EAAE,iBAAiB,EAAE,MAAM,uBAAuB,CAAC;AAC/D,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,wBAAwB,CAAC;AAE3D,MAAM,WAAW,aAAa;IAC5B,sCAAsC;IACtC,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC;IAChB,GAAG,CAAC,EAAE,MAAM,EAAE,CAAC;IACf,UAAU,CAAC,EAAE,MAAM,EAAE,CAAC;IACtB,4CAA4C;IAC5C,MAAM,CAAC,EAAE,MAAM,EAAE,CAAC;IAClB,4DAA4D;IAC5D,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,wDAAwD;IACxD,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,0CAA0C;IAC1C,IAAI,CAAC,EAAE,MAAM,CAAC;CACf;AAED,MAAM,MAAM,WAAW,GAAG,UAAU,GAAG,SAAS,GAAG,OAAO,GAAG,MAAM,GAAG,MAAM,CAAC;AAE7E,MAAM,WAAW,YAAY;IAC3B,QAAQ,EAAE,MAAM,CAAC;IACjB,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,GAAG,CAAC,EAAE,MAAM,CAAC;IACb,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,IAAI,EAAE,MAAM,EAAE,CAAC;IACf,UAAU,EAAE,MAAM,CAAC;IACnB,IAAI,EAAE,MAAM,CAAC;IACb,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,6EAA6E;IAC7E,QAAQ,EAAE,OAAO,CAAC;CACnB;AAID,4EAA4E;AAC5E,wBAAgB,YAAY,CAAC,IAAI,EAAE;IACjC,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,UAAU,CAAC,EAAE,iBAAiB,CAAC;CAChC,GAAG,MAAM,GAAG,SAAS,CAQrB;AAED,qEAAqE;AACrE,wBAAgB,kBAAkB,CAAC,IAAI,EAAE;IACvC,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,UAAU,CAAC,EAAE,iBAAiB,CAAC;CAChC,GAAG,MAAM,CAET;AAUD;;;;GAIG;AACH,wBAAgB,SAAS,CAAC,KAAK,EAAE,MAAM,GAAG,SAAS,EAAE,GAAG,SAAa,GAAG,MAAM,GAAG,IAAI,CAOpF;AAkBD,iDAAiD;AACjD,wBAAgB,oBAAoB,CAClC,IAAI,EAAE,YAAY,EAClB,MAAM,GAAE,aAAkB,EAC1B,GAAG,SAAa,GACf,OAAO,CAqCT;AAED,wBAAgB,eAAe,CAC7B,IAAI,EAAE,YAAY,EAAE,EACpB,MAAM,GAAE,aAAkB,EAC1B,GAAG,SAAa,GACf,YAAY,EAAE,CAEhB;AAED,wBAAgB,aAAa,CAC3B,IAAI,EAAE,YAAY,EAAE,EACpB,IAAI,GAAE,WAAwB,EAC9B,KAAK,GAAE,KAAK,GAAG,MAAe,GAC7B,YAAY,EAAE,CAsBhB;AAED,uEAAuE;AACvE,wBAAgB,YAAY,CAAC,IAAI,EAAE,YAAY,GAAG,YAAY,CAmB7D;AAED,MAAM,WAAW,kBAAmB,SAAQ,aAAa;IACvD,IAAI,CAAC,EAAE,WAAW,CAAC;IACnB,KAAK,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC;IACvB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,GAAG,CAAC,EAAE,MAAM,CAAC;CACd;AAED,6EAA6E;AAC7E,wBAAgB,WAAW,CAAC,IAAI,EAAE,YAAY,EAAE,EAAE,IAAI,GAAE,kBAAuB,GAAG,YAAY,EAAE,CAM/F;AAED,kEAAkE;AAClE,wBAAgB,aAAa,CAAC,IAAI,EAAE,YAAY,EAAE,GAAG;IACnD,KAAK,EAAE,KAAK,CAAC;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAC/C,IAAI,EAAE,KAAK,CAAC;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAC9C,WAAW,EAAE,KAAK,CAAC;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IACrD,OAAO,EAAE,KAAK,CAAC;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;CAClD,CAiBA"}

package/dist/src/knowledge/library.js

@@ -0,0 +1,195 @@
+/**
+ * The capture library (RCL-09) — browsing what has been captured, rather than
+ * searching it.
+ *
+ * `listDocuments` answers "what is indexed". A library answers a different
+ * question: everything from this site, captured last week, tagged `reading`,
+ * newest first. The facets come from the capture envelope's `meta.json`
+ * (RCL-07), which is already on every record — this module is the query layer
+ * over them, and the data layer the desktop GUI renders.
+ *
+ * Rules, once, here:
+ *  - within a facet values are OR'd, across facets AND'd (`--site a --site b
+ *    --tag x` = "from a or b, tagged x");
+ *  - `--site example.com` matches `www.example.com` and `docs.example.com`,
+ *    because nobody browsing their own library means the bare host only;
+ *  - a document with no provenance (an ordinary file on disk) sorts by its
+ *    index time and matches no capture facet — it is not a capture.
+ *
+ * @module v1/cli/knowledge/library
+ */
+import * as path from 'node:path';
+// ── Facet extraction ───────────────────────────────────────────────
+/** The host a document was captured from, or undefined for a local file. */
+export function documentSite(meta) {
+    const url = meta.canonicalUrl ?? meta.provenance?.canonicalUrl ?? meta.provenance?.url;
+    if (!url)
+        return undefined;
+    try {
+        return new URL(url).hostname.toLowerCase().replace(/^www\./, '');
+    }
+    catch {
+        return undefined;
+    }
+}
+/** When the page was captured; falls back to when it was indexed. */
+export function documentCapturedAt(meta) {
+    return meta.provenance?.capturedAt ?? meta.indexedAt ?? '';
+}
+const RELATIVE_RE = /^(\d+(?:\.\d+)?)\s*(m|h|d|w)$/i;
+const UNIT_MS = {
+    m: 60_000,
+    h: 3_600_000,
+    d: 86_400_000,
+    w: 604_800_000,
+};
+/**
+ * A point in time from `2026-09-01`, a full ISO stamp, or an AGE like `7d`
+ * (meaning "7 days ago"). Returns null for anything unparseable so a typo
+ * cannot silently filter the whole library away.
+ */
+export function parseWhen(value, now = Date.now()) {
+    const raw = (value ?? '').trim();
+    if (!raw)
+        return null;
+    const rel = RELATIVE_RE.exec(raw);
+    if (rel)
+        return now - Number(rel[1]) * UNIT_MS[rel[2].toLowerCase()];
+    const parsed = Date.parse(raw.length === 10 ? `${raw}T00:00:00.000Z` : raw);
+    return Number.isNaN(parsed) ? null : parsed;
+}
+function siteMatches(host, wanted) {
+    if (!host)
+        return false;
+    return wanted.some((w) => {
+        const want = w
+            .trim()
+            .toLowerCase()
+            .replace(/^www\./, '');
+        return !!want && (host === want || host.endsWith(`.${want}`));
+    });
+}
+function anyOf(values, wanted) {
+    const have = (values ?? []).map((v) => v.toLowerCase());
+    return wanted.some((w) => have.includes(w.trim().toLowerCase()));
+}
+/** True when a document belongs in this view. */
+export function matchesLibraryFilter(meta, filter = {}, now = Date.now()) {
+    if (filter.scope && meta.scope !== filter.scope)
+        return false;
+    const prov = meta.provenance;
+    const isCapture = !!prov || !!meta.canonicalUrl;
+    if (filter.capturedOnly && !isCapture)
+        return false;
+    if (filter.site?.length && !siteMatches(documentSite(meta), filter.site))
+        return false;
+    if (filter.tag?.length && !anyOf(prov?.tags, filter.tag))
+        return false;
+    if (filter.collection?.length &&
+        !anyOf(prov?.collection ? [prov.collection] : [], filter.collection))
+        return false;
+    if (filter.source?.length && !anyOf(prov?.source ? [prov.source] : [], filter.source))
+        return false;
+    if (filter.since || filter.until) {
+        const at = Date.parse(documentCapturedAt(meta));
+        const since = parseWhen(filter.since, now);
+        const until = parseWhen(filter.until, now);
+        if (Number.isNaN(at))
+            return false;
+        if (since !== null && at < since)
+            return false;
+        if (until !== null && at > until)
+            return false;
+    }
+    if (filter.text) {
+        const needle = filter.text.toLowerCase();
+        const hay = [
+            prov?.title ?? '',
+            meta.canonicalUrl ?? prov?.url ?? '',
+            meta.filePath,
+            prov?.note ?? '',
+        ]
+            .join(' ')
+            .toLowerCase();
+        if (!hay.includes(needle))
+            return false;
+    }
+    return true;
+}
+export function filterDocuments(docs, filter = {}, now = Date.now()) {
+    return docs.filter((d) => matchesLibraryFilter(d, filter, now));
+}
+export function sortDocuments(docs, sort = 'captured', order = 'desc') {
+    const dir = order === 'asc' ? 1 : -1;
+    const key = (d) => {
+        switch (sort) {
+            case 'indexed':
+                return d.indexedAt ?? '';
+            case 'title':
+                return (d.provenance?.title ?? path.basename(d.filePath)).toLowerCase();
+            case 'path':
+                return d.filePath;
+            case 'size':
+                return d.size ?? 0;
+            default:
+                return documentCapturedAt(d);
+        }
+    };
+    return [...docs].sort((a, b) => {
+        const ka = key(a);
+        const kb = key(b);
+        if (ka === kb)
+            return a.filePath.localeCompare(b.filePath);
+        return ka < kb ? -dir : dir;
+    });
+}
+/** The flat, JSON-friendly row the GUI and `--json` consumers read. */
+export function libraryEntry(meta) {
+    const prov = meta.provenance;
+    const url = meta.canonicalUrl ?? prov?.canonicalUrl ?? prov?.url;
+    return {
+        filePath: meta.filePath,
+        scope: meta.scope,
+        title: prov?.title ?? path.basename(meta.filePath),
+        ...(url ? { url } : {}),
+        ...(documentSite(meta) ? { site: documentSite(meta) } : {}),
+        capturedAt: documentCapturedAt(meta),
+        indexedAt: meta.indexedAt,
+        ...(prov?.source ? { source: prov.source } : {}),
+        ...(prov?.collection ? { collection: prov.collection } : {}),
+        tags: prov?.tags ?? [],
+        chunkCount: meta.chunkCount,
+        size: meta.size,
+        ...(meta.version ? { version: meta.version } : {}),
+        captured: !!prov || !!meta.canonicalUrl,
+    };
+}
+/** Browse the library: filter, sort, cap. Never throws on an empty store. */
+export function listLibrary(docs, opts = {}) {
+    const now = opts.now ?? Date.now();
+    const filtered = filterDocuments(docs, opts, now);
+    const sorted = sortDocuments(filtered, opts.sort ?? 'captured', opts.order ?? 'desc');
+    const capped = opts.limit && opts.limit > 0 ? sorted.slice(0, opts.limit) : sorted;
+    return capped.map(libraryEntry);
+}
+/** Facet counts for the current view — what a sidebar renders. */
+export function libraryFacets(docs) {
+    const tally = (values) => {
+        const counts = new Map();
+        for (const v of values) {
+            if (!v)
+                continue;
+            counts.set(v, (counts.get(v) ?? 0) + 1);
+        }
+        return [...counts.entries()]
+            .map(([value, count]) => ({ value, count }))
+            .sort((a, b) => b.count - a.count || a.value.localeCompare(b.value));
+    };
+    return {
+        sites: tally(docs.map(documentSite)),
+        tags: tally(docs.flatMap((d) => d.provenance?.tags ?? [])),
+        collections: tally(docs.map((d) => d.provenance?.collection)),
+        sources: tally(docs.map((d) => d.provenance?.source)),
+    };
+}
+//# sourceMappingURL=library.js.map

package/dist/src/knowledge/library.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"library.js","sourceRoot":"","sources":["../../../src/knowledge/library.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH,OAAO,KAAK,IAAI,MAAM,WAAW,CAAC;AAyClC,sEAAsE;AAEtE,4EAA4E;AAC5E,MAAM,UAAU,YAAY,CAAC,IAG5B;IACC,MAAM,GAAG,GAAG,IAAI,CAAC,YAAY,IAAI,IAAI,CAAC,UAAU,EAAE,YAAY,IAAI,IAAI,CAAC,UAAU,EAAE,GAAG,CAAC;IACvF,IAAI,CAAC,GAAG;QAAE,OAAO,SAAS,CAAC;IAC3B,IAAI,CAAC;QACH,OAAO,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;IACnE,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;AACH,CAAC;AAED,qEAAqE;AACrE,MAAM,UAAU,kBAAkB,CAAC,IAGlC;IACC,OAAO,IAAI,CAAC,UAAU,EAAE,UAAU,IAAI,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC;AAC7D,CAAC;AAED,MAAM,WAAW,GAAG,gCAAgC,CAAC;AACrD,MAAM,OAAO,GAA2B;IACtC,CAAC,EAAE,MAAM;IACT,CAAC,EAAE,SAAS;IACZ,CAAC,EAAE,UAAU;IACb,CAAC,EAAE,WAAW;CACf,CAAC;AAEF;;;;GAIG;AACH,MAAM,UAAU,SAAS,CAAC,KAAyB,EAAE,GAAG,GAAG,IAAI,CAAC,GAAG,EAAE;IACnE,MAAM,GAAG,GAAG,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IACjC,IAAI,CAAC,GAAG;QAAE,OAAO,IAAI,CAAC;IACtB,MAAM,GAAG,GAAG,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAClC,IAAI,GAAG;QAAE,OAAO,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;IACrE,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,KAAK,EAAE,CAAC,CAAC,CAAC,GAAG,GAAG,gBAAgB,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;IAC5E,OAAO,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC;AAC9C,CAAC;AAED,SAAS,WAAW,CAAC,IAAwB,EAAE,MAAgB;IAC7D,IAAI,CAAC,IAAI;QAAE,OAAO,KAAK,CAAC;IACxB,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE;QACvB,MAAM,IAAI,GAAG,CAAC;aACX,IAAI,EAAE;aACN,WAAW,EAAE;aACb,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;QACzB,OAAO,CAAC,CAAC,IAAI,IAAI,CAAC,IAAI,KAAK,IAAI,IAAI,IAAI,CAAC,QAAQ,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,CAAC;IAChE,CAAC,CAAC,CAAC;AACL,CAAC;AAED,SAAS,KAAK,CAAC,MAA4B,EAAE,MAAgB;IAC3D,MAAM,IAAI,GAAG,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;IACxD,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;AACnE,CAAC;AAED,iDAAiD;AACjD,MAAM,UAAU,oBAAoB,CAClC,IAAkB,EAClB,MAAM,GAAkB,EAAE,EAC1B,GAAG,GAAG,IAAI,CAAC,GAAG,EAAE;IAEhB,IAAI,MAAM,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,KAAK,MAAM,CAAC,KAAK;QAAE,OAAO,KAAK,CAAC;IAC9D,MAAM,IAAI,GAAG,IAAI,CAAC,UAAU,CAAC;IAC7B,MAAM,SAAS,GAAG,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC;IAChD,IAAI,MAAM,CAAC,YAAY,IAAI,CAAC,SAAS;QAAE,OAAO,KAAK,CAAC;IACpD,IAAI,MAAM,CAAC,IAAI,EAAE,MAAM,IAAI,CAAC,WAAW,CAAC,YAAY,CAAC,IAAI,CAAC,EAAE,MAAM,CAAC,IAAI,CAAC;QAAE,OAAO,KAAK,CAAC;IACvF,IAAI,MAAM,CAAC,GAAG,EAAE,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,EAAE,MAAM,CAAC,GAAG,CAAC;QAAE,OAAO,KAAK,CAAC;IACvE,IACE,MAAM,CAAC,UAAU,EAAE,MAAM;QACzB,CAAC,KAAK,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,MAAM,CAAC,UAAU,CAAC;QAEpE,OAAO,KAAK,CAAC;IACf,IAAI,MAAM,CAAC,MAAM,EAAE,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,MAAM,CAAC,MAAM,CAAC;QACnF,OAAO,KAAK,CAAC;IAEf,IAAI,MAAM,CAAC,KAAK,IAAI,MAAM,CAAC,KAAK,EAAE,CAAC;QACjC,MAAM,EAAE,GAAG,IAAI,CAAC,KAAK,CAAC,kBAAkB,CAAC,IAAI,CAAC,CAAC,CAAC;QAChD,MAAM,KAAK,GAAG,SAAS,CAAC,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;QAC3C,MAAM,KAAK,GAAG,SAAS,CAAC,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;QAC3C,IAAI,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC;YAAE,OAAO,KAAK,CAAC;QACnC,IAAI,KAAK,KAAK,IAAI,IAAI,EAAE,GAAG,KAAK;YAAE,OAAO,KAAK,CAAC;QAC/C,IAAI,KAAK,KAAK,IAAI,IAAI,EAAE,GAAG,KAAK;YAAE,OAAO,KAAK,CAAC;IACjD,CAAC;IAED,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;QAChB,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;QACzC,MAAM,GAAG,GAAG;YACV,IAAI,EAAE,KAAK,IAAI,EAAE;YACjB,IAAI,CAAC,YAAY,IAAI,IAAI,EAAE,GAAG,IAAI,EAAE;YACpC,IAAI,CAAC,QAAQ;YACb,IAAI,EAAE,IAAI,IAAI,EAAE;SACjB;aACE,IAAI,CAAC,GAAG,CAAC;aACT,WAAW,EAAE,CAAC;QACjB,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,MAAM,CAAC;YAAE,OAAO,KAAK,CAAC;IAC1C,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,MAAM,UAAU,eAAe,CAC7B,IAAoB,EACpB,MAAM,GAAkB,EAAE,EAC1B,GAAG,GAAG,IAAI,CAAC,GAAG,EAAE;IAEhB,OAAO,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,oBAAoB,CAAC,CAAC,EAAE,MAAM,EAAE,GAAG,CAAC,CAAC,CAAC;AAClE,CAAC;AAED,MAAM,UAAU,aAAa,CAC3B,IAAoB,EACpB,IAAI,GAAgB,UAAU,EAC9B,KAAK,GAAmB,MAAM;IAE9B,MAAM,GAAG,GAAG,KAAK,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACrC,MAAM,GAAG,GAAG,CAAC,CAAe,EAAmB,EAAE;QAC/C,QAAQ,IAAI,EAAE,CAAC;YACb,KAAK,SAAS;gBACZ,OAAO,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC;YAC3B,KAAK,OAAO;gBACV,OAAO,CAAC,CAAC,CAAC,UAAU,EAAE,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC;YAC1E,KAAK,MAAM;gBACT,OAAO,CAAC,CAAC,QAAQ,CAAC;YACpB,KAAK,MAAM;gBACT,OAAO,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC;YACrB;gBACE,OAAO,kBAAkB,CAAC,CAAC,CAAC,CAAC;QACjC,CAAC;IACH,CAAC,CAAC;IACF,OAAO,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;QAC7B,MAAM,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;QAClB,MAAM,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;QAClB,IAAI,EAAE,KAAK,EAAE;YAAE,OAAO,CAAC,CAAC,QAAQ,CAAC,aAAa,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC;QAC3D,OAAO,EAAE,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC;IAC9B,CAAC,CAAC,CAAC;AACL,CAAC;AAED,uEAAuE;AACvE,MAAM,UAAU,YAAY,CAAC,IAAkB;IAC7C,MAAM,IAAI,GAAG,IAAI,CAAC,UAAU,CAAC;IAC7B,MAAM,GAAG,GAAG,IAAI,CAAC,YAAY,IAAI,IAAI,EAAE,YAAY,IAAI,IAAI,EAAE,GAAG,CAAC;IACjE,OAAO;QACL,QAAQ,EAAE,IAAI,CAAC,QAAQ;QACvB,KAAK,EAAE,IAAI,CAAC,KAAK;QACjB,KAAK,EAAE,IAAI,EAAE,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC;QAClD,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACvB,GAAG,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3D,UAAU,EAAE,kBAAkB,CAAC,IAAI,CAAC;QACpC,SAAS,EAAE,IAAI,CAAC,SAAS;QACzB,GAAG,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAChD,GAAG,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC5D,IAAI,EAAE,IAAI,EAAE,IAAI,IAAI,EAAE;QACtB,UAAU,EAAE,IAAI,CAAC,UAAU;QAC3B,IAAI,EAAE,IAAI,CAAC,IAAI;QACf,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAClD,QAAQ,EAAE,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,IAAI,CAAC,YAAY;KACxC,CAAC;AACJ,CAAC;AASD,6EAA6E;AAC7E,MAAM,UAAU,WAAW,CAAC,IAAoB,EAAE,IAAI,GAAuB,EAAE;IAC7E,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC;IACnC,MAAM,QAAQ,GAAG,eAAe,CAAC,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC;IAClD,MAAM,MAAM,GAAG,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,IAAI,IAAI,UAAU,EAAE,IAAI,CAAC,KAAK,IAAI,MAAM,CAAC,CAAC;IACtF,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IACnF,OAAO,MAAM,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;AAClC,CAAC;AAED,kEAAkE;AAClE,MAAM,UAAU,aAAa,CAAC,IAAoB;IAMhD,MAAM,KAAK,GAAG,CAAC,MAAiC,EAAE,EAAE;QAClD,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;QACzC,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;YACvB,IAAI,CAAC,CAAC;gBAAE,SAAS;YACjB,MAAM,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QAC1C,CAAC;QACD,OAAO,CAAC,GAAG,MAAM,CAAC,OAAO,EAAE,CAAC;aACzB,GAAG,CAAC,CAAC,CAAC,KAAK,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC;aAC3C,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,KAAK,CAAC,aAAa,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC;IACzE,CAAC,CAAC;IACF,OAAO;QACL,KAAK,EAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;QACpC,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,EAAE,IAAI,IAAI,EAAE,CAAC,CAAC;QAC1D,WAAW,EAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,EAAE,UAAU,CAAC,CAAC;QAC7D,OAAO,EAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;KACtD,CAAC;AACJ,CAAC"}

package/dist/src/knowledge/mhtml.d.ts

@@ -0,0 +1,60 @@
+/**
+ * MHTML / MHT container parsing (RCL-01).
+ *
+ * An MHTML archive is an RFC 2557 `multipart/related` MIME message: the page
+ * plus every subresource it referenced, each part carrying its own transfer
+ * encoding and charset. Chrome's `Page.captureSnapshot` writes them
+ * quoted-printable; Internet Explorer and most "save as web archive" tools
+ * write base64.
+ *
+ * Only the ROOT frame is prose. Picking the first `text/html` part instead is
+ * how an ad iframe ends up indexed as the article, so the root is resolved
+ * from the container's `Snapshot-Content-Location` / `Content-Location` when
+ * one is present.
+ *
+ * Read the file as `latin1` (or pass a Buffer): byte fidelity matters before
+ * transfer decoding, and the real charset is only known per part.
+ *
+ * @module v1/cli/knowledge/mhtml
+ */
+export interface MhtmlPart {
+    /** Lowercased MIME type without parameters, e.g. `text/html`. */
+    contentType: string;
+    /** Charset from the part's Content-Type, lowercased, or null. */
+    charset: string | null;
+    /** Content-Location, or null. */
+    location: string | null;
+    /** Lowercased Content-Transfer-Encoding; `binary` when unstated. */
+    encoding: string;
+    /** Raw bytes after transfer decoding. */
+    body: Buffer;
+    headers: Record<string, string>;
+}
+export interface ParsedMhtml {
+    /** The part holding the page itself, or null when there is no HTML part. */
+    root: MhtmlPart | null;
+    parts: MhtmlPart[];
+    /** Container `Snapshot-Content-Location`/`Content-Location` header. */
+    snapshotLocation: string | null;
+    /** Container `Subject` header — Chrome writes the page title here. */
+    subject: string | null;
+}
+/** Decode a quoted-printable body to bytes: soft line breaks vanish, `=XX`
+ *  becomes one byte, everything else is its own latin1 byte. */
+export declare function decodeQuotedPrintable(input: string): Buffer;
+/** Decode bytes under a MIME charset, falling back to UTF-8 (then latin1)
+ *  rather than failing an ingest over an exotic label. */
+export declare function decodeCharset(bytes: Buffer, charset: string | null): string;
+/**
+ * Parse an MHTML container. Anything that is not a well-formed multipart
+ * message is treated as a single part, so a plain `.html` file renamed to
+ * `.mhtml` still ingests.
+ */
+export declare function parseMhtml(raw: string | Buffer): ParsedMhtml;
+/** Parse a container and return its root frame decoded to a string. */
+export declare function mhtmlToHtml(raw: string | Buffer): {
+    html: string;
+    location: string | null;
+    subject: string | null;
+};
+//# sourceMappingURL=mhtml.d.ts.map

package/dist/src/knowledge/mhtml.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"mhtml.d.ts","sourceRoot":"","sources":["../../../src/knowledge/mhtml.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAEH,MAAM,WAAW,SAAS;IACxB,iEAAiE;IACjE,WAAW,EAAE,MAAM,CAAC;IACpB,iEAAiE;IACjE,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;IACvB,iCAAiC;IACjC,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAC;IACxB,oEAAoE;IACpE,QAAQ,EAAE,MAAM,CAAC;IACjB,yCAAyC;IACzC,IAAI,EAAE,MAAM,CAAC;IACb,OAAO,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CACjC;AAED,MAAM,WAAW,WAAW;IAC1B,4EAA4E;IAC5E,IAAI,EAAE,SAAS,GAAG,IAAI,CAAC;IACvB,KAAK,EAAE,SAAS,EAAE,CAAC;IACnB,uEAAuE;IACvE,gBAAgB,EAAE,MAAM,GAAG,IAAI,CAAC;IAChC,sEAAsE;IACtE,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;CACxB;AAED;gEACgE;AAChE,wBAAgB,qBAAqB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAyB3D;AAaD;0DAC0D;AAC1D,wBAAgB,aAAa,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,GAAG,IAAI,GAAG,MAAM,CAU3E;AA4CD;;;;GAIG;AACH,wBAAgB,UAAU,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,GAAG,WAAW,CAiC5D;AAED,uEAAuE;AACvE,wBAAgB,WAAW,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,GAAG;IACjD,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAC;IACxB,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;CACxB,CAOA"}