dsh-gitnexus-plus 0.0.0-stage → 0.1.1

package/lib/gitnexus/languages.d.ts

@@ -0,0 +1,73 @@
+/**
+ * Language classification for a file path, reusing GitNexus's own detector.
+ *
+ * `gitnexus-shared` is not published separately, but `analyze` bundles it into
+ * `gitnexus/dist/_shared/`. Importing that copy is deliberate: the detector that
+ * classified the files *in an existing index* is the one that must classify them
+ * in the report. Copying the map from source would be wrong — the installed
+ * 1.6.12 differs from the checkout (e.g. `.ipynb` resolves to `null` there, not
+ * to `python`), and such gaps are invisible until a percentage looks odd.
+ *
+ * Three classes, because "is this file in the graph as code?" has three answers:
+ *
+ *   parsed     — GitNexus has a LanguageProvider; symbols inside become graph nodes.
+ *   recognized — known format, no parser; the file exists as a `File` node only.
+ *   other      — not recognised at all.
+ *
+ * This is the distinction GitHub's language bar cannot express, and it is what
+ * makes the widget answer "how much of my repository does GitNexus actually
+ * understand?" rather than merely "what is this repo written in?".
+ */
+import type { ScannedFile } from './scan.js';
+/** Which of the three classes a path falls into. */
+export type LanguageClass = 'parsed' | 'recognized' | 'other';
+/**
+ * GitNexus's language detector, as this plugin uses it: the handle passed around
+ * by {@link classifyFile} and {@link aggregateLanguages}.
+ */
+export interface LanguageDetector {
+    getLanguageFromFilename: (filePath: string) => string | null;
+    getSyntaxLanguageFromFilename: (filePath: string) => string;
+}
+/** What {@link classifyFile} decides about one path. */
+export interface ClassifiedLanguage {
+    key: string;
+    label: string;
+    color: string;
+    klass: LanguageClass;
+}
+/** One row of the language breakdown: a {@link ClassifiedLanguage} with totals. */
+export interface LanguageRow extends ClassifiedLanguage {
+    bytes: number;
+    files: number;
+}
+/** The language breakdown itself, as returned by {@link aggregateLanguages}. */
+export interface LanguageAggregate {
+    rows: LanguageRow[];
+    totalFiles: number;
+    totalBytes: number;
+}
+/**
+ * Load GitNexus's bundled detector once per process.
+ * Returns `null` when the bundled copy is unavailable, in which case every file
+ * is classified `other` — a degraded but honest answer, never a fabricated one.
+ */
+export declare function loadLanguageDetector(options?: {
+    env?: NodeJS.ProcessEnv;
+}): Promise<LanguageDetector | null>;
+/**
+ * Classify one repo-relative path.
+ *
+ * @param filePath repo-relative, forward slashes
+ */
+export declare function classifyFile(filePath: string, det: LanguageDetector | null): ClassifiedLanguage;
+/**
+ * Aggregate files into a language breakdown.
+ *
+ * `klass` rank decides ordering ties so the three visual bands stay contiguous
+ * and the bar does not interleave "understood" and "not understood" segments.
+ *
+ * @param include extra filter, e.g. "is indexed"
+ */
+export declare function aggregateLanguages(files: ScannedFile[], det: LanguageDetector | null, include?: (filePath: string) => boolean): LanguageAggregate;
+//# sourceMappingURL=languages.d.ts.map

package/lib/gitnexus/languages.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"languages.d.ts","sourceRoot":"","sources":["../../src/gitnexus/languages.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAKH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,WAAW,CAAC;AAE7C,oDAAoD;AACpD,MAAM,MAAM,aAAa,GAAG,QAAQ,GAAG,YAAY,GAAG,OAAO,CAAC;AAiG9D;;;GAGG;AACH,MAAM,WAAW,gBAAgB;IAC/B,uBAAuB,EAAE,CAAC,QAAQ,EAAE,MAAM,KAAK,MAAM,GAAG,IAAI,CAAC;IAC7D,6BAA6B,EAAE,CAAC,QAAQ,EAAE,MAAM,KAAK,MAAM,CAAC;CAC7D;AAED,wDAAwD;AACxD,MAAM,WAAW,kBAAkB;IACjC,GAAG,EAAE,MAAM,CAAC;IACZ,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,aAAa,CAAC;CACtB;AAED,mFAAmF;AACnF,MAAM,WAAW,WAAY,SAAQ,kBAAkB;IACrD,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;CACf;AAED,gFAAgF;AAChF,MAAM,WAAW,iBAAiB;IAChC,IAAI,EAAE,WAAW,EAAE,CAAC;IACpB,UAAU,EAAE,MAAM,CAAC;IACnB,UAAU,EAAE,MAAM,CAAC;CACpB;AAID;;;;GAIG;AACH,wBAAsB,oBAAoB,CACxC,OAAO,GAAE;IAAE,GAAG,CAAC,EAAE,MAAM,CAAC,UAAU,CAAA;CAAO,GACxC,OAAO,CAAC,gBAAgB,GAAG,IAAI,CAAC,CAqBlC;AAED;;;;GAIG;AACH,wBAAgB,YAAY,CAAC,QAAQ,EAAE,MAAM,EAAE,GAAG,EAAE,gBAAgB,GAAG,IAAI,GAAG,kBAAkB,CA6C/F;AAED;;;;;;;GAOG;AACH,wBAAgB,kBAAkB,CAChC,KAAK,EAAE,WAAW,EAAE,EACpB,GAAG,EAAE,gBAAgB,GAAG,IAAI,EAC5B,OAAO,CAAC,EAAE,CAAC,QAAQ,EAAE,MAAM,KAAK,OAAO,GACtC,iBAAiB,CA4BnB"}

package/lib/gitnexus/languages.js

@@ -0,0 +1,217 @@
+/**
+ * Language classification for a file path, reusing GitNexus's own detector.
+ *
+ * `gitnexus-shared` is not published separately, but `analyze` bundles it into
+ * `gitnexus/dist/_shared/`. Importing that copy is deliberate: the detector that
+ * classified the files *in an existing index* is the one that must classify them
+ * in the report. Copying the map from source would be wrong — the installed
+ * 1.6.12 differs from the checkout (e.g. `.ipynb` resolves to `null` there, not
+ * to `python`), and such gaps are invisible until a percentage looks odd.
+ *
+ * Three classes, because "is this file in the graph as code?" has three answers:
+ *
+ *   parsed     — GitNexus has a LanguageProvider; symbols inside become graph nodes.
+ *   recognized — known format, no parser; the file exists as a `File` node only.
+ *   other      — not recognised at all.
+ *
+ * This is the distinction GitHub's language bar cannot express, and it is what
+ * makes the widget answer "how much of my repository does GitNexus actually
+ * understand?" rather than merely "what is this repo written in?".
+ */
+import path from 'node:path';
+import { pathToFileURL } from 'node:url';
+import { findGitnexusRoot } from './root.js';
+/**
+ * Display names and stable colours. Colours follow GitHub linguist's palette for
+ * the languages it defines, so the bar reads the same as a repository page; the
+ * remainder are assigned from a fixed table rather than a hash so a colour never
+ * moves between renders.
+ */
+const DISPLAY = Object.freeze({
+    javascript: ['JavaScript', '#f1e05a'],
+    typescript: ['TypeScript', '#3178c6'],
+    python: ['Python', '#3572A5'],
+    java: ['Java', '#b07219'],
+    c: ['C', '#555555'],
+    cpp: ['C++', '#f34b7d'],
+    'objective-c': ['Objective-C', '#438eff'],
+    csharp: ['C#', '#178600'],
+    go: ['Go', '#00ADD8'],
+    ruby: ['Ruby', '#701516'],
+    rust: ['Rust', '#dea584'],
+    php: ['PHP', '#4F5D95'],
+    kotlin: ['Kotlin', '#A97BFF'],
+    swift: ['Swift', '#F05138'],
+    dart: ['Dart', '#00B4AB'],
+    vue: ['Vue', '#41b883'],
+    cobol: ['COBOL', '#6b6b6b'],
+    zig: ['Zig', '#ec915c'],
+    // Recognised formats: neutral colours, intentionally outside the code palette.
+    markdown: ['Markdown', '#8b949e'],
+    json: ['JSON', '#a8a8a8'],
+    yaml: ['YAML', '#7d9199'],
+    markup: ['HTML', '#e34c26'],
+    css: ['CSS', '#563d7c'],
+    bash: ['Shell', '#89e051'],
+    sql: ['SQL', '#9c9c9c'],
+    toml: ['TOML', '#9c4221'],
+    ini: ['INI', '#8a8a8a'],
+    docker: ['Dockerfile', '#384d54'],
+    makefile: ['Makefile', '#427819'],
+    // GitNexus's detector does not know these. They are added here so the bar reads
+    // like a repository page instead of dumping them into "other" — measured on the
+    // GitNexus repository itself, where 6 `.jinja` and 3 `.ps1` files were the only
+    // thing separating this breakdown from GitHub's.
+    jinja: ['Jinja', '#a52a22'],
+    powershell: ['PowerShell', '#012456'],
+    batch: ['Batchfile', '#C1F12E'],
+    diff: ['Diff', '#8b949e'],
+    plaintext: ['Text', '#6e7681'],
+    text: ['其他', '#484f58'],
+});
+/**
+ * Extensions GitNexus's detector has no entry for, mapped to the display keys
+ * above. Consulted only after that detector has declined the file, so it can
+ * never override a language GitNexus genuinely parses.
+ */
+const SUPPLEMENTARY_EXTENSIONS = Object.freeze({
+    '.jinja': 'jinja',
+    '.jinja2': 'jinja',
+    '.j2': 'jinja',
+    '.njk': 'jinja',
+    '.ps1': 'powershell',
+    '.psm1': 'powershell',
+    '.psd1': 'powershell',
+    '.ps1xml': 'powershell',
+    '.bat': 'batch',
+    '.cmd': 'batch',
+    '.patch': 'diff',
+    '.diff': 'diff',
+    '.txt': 'plaintext',
+    '.log': 'plaintext',
+});
+/**
+ * Cases no extension map can decide, resolved by filename.
+ * Kept explicit so an unlisted name falls to `other` rather than a guess.
+ */
+const BASENAME_LANGUAGES = Object.freeze({
+    dockerfile: 'docker',
+    makefile: 'makefile',
+    'cmakelists.txt': 'text',
+    '.gitignore': 'text',
+    '.gitattributes': 'text',
+    '.editorconfig': 'text',
+    '.npmrc': 'text',
+    '.nvmrc': 'text',
+});
+let detector = null;
+/**
+ * Load GitNexus's bundled detector once per process.
+ * Returns `null` when the bundled copy is unavailable, in which case every file
+ * is classified `other` — a degraded but honest answer, never a fabricated one.
+ */
+export async function loadLanguageDetector(options = {}) {
+    if (detector !== null)
+        return detector;
+    const root = findGitnexusRoot(options);
+    if (root === null)
+        return null;
+    try {
+        const mod = (await import(pathToFileURL(path.join(root, 'dist', '_shared', 'language-detection.js')).href));
+        if (typeof mod.getLanguageFromFilename === 'function') {
+            detector = {
+                getLanguageFromFilename: mod.getLanguageFromFilename,
+                getSyntaxLanguageFromFilename: typeof mod.getSyntaxLanguageFromFilename === 'function'
+                    ? mod.getSyntaxLanguageFromFilename
+                    : () => 'text',
+            };
+        }
+    }
+    catch {
+        /* fall through to the degraded path */
+    }
+    return detector;
+}
+/**
+ * Classify one repo-relative path.
+ *
+ * @param filePath repo-relative, forward slashes
+ */
+export function classifyFile(filePath, det) {
+    let key = null;
+    let klass = 'other';
+    if (det !== null) {
+        const parsed = det.getLanguageFromFilename(filePath);
+        if (parsed) {
+            key = parsed;
+            klass = 'parsed';
+        }
+        else {
+            const syntax = det.getSyntaxLanguageFromFilename(filePath);
+            if (syntax && syntax !== 'text') {
+                key = syntax;
+                klass = 'recognized';
+            }
+        }
+    }
+    if (key === null) {
+        const lastDot = filePath.lastIndexOf('.');
+        if (lastDot >= 0) {
+            const ext = filePath.slice(lastDot).toLowerCase();
+            const supplement = SUPPLEMENTARY_EXTENSIONS[ext];
+            if (supplement !== undefined)
+                key = supplement;
+        }
+    }
+    if (key === null) {
+        const base = filePath.split('/').pop()?.toLowerCase() ?? '';
+        const byName = BASENAME_LANGUAGES[base];
+        if (byName !== undefined) {
+            key = byName;
+            klass = byName === 'text' ? 'other' : 'recognized';
+        }
+    }
+    const entry = key !== null ? DISPLAY[key] : undefined;
+    // `key === null` already implies `entry === undefined`; stating both is what lets
+    // the checker narrow `key` to a string for the return below.
+    if (entry === undefined || key === null) {
+        return { key: 'text', label: '其他', color: DISPLAY.text[1], klass: 'other' };
+    }
+    // A supplementary hit is a known format GitNexus cannot parse — never `parsed`.
+    if (klass === 'other' && key !== 'text')
+        klass = 'recognized';
+    return { key, label: entry[0], color: entry[1], klass };
+}
+/**
+ * Aggregate files into a language breakdown.
+ *
+ * `klass` rank decides ordering ties so the three visual bands stay contiguous
+ * and the bar does not interleave "understood" and "not understood" segments.
+ *
+ * @param include extra filter, e.g. "is indexed"
+ */
+export function aggregateLanguages(files, det, include) {
+    const map = new Map();
+    let totalFiles = 0;
+    let totalBytes = 0;
+    for (const file of files) {
+        if (include && !include(file.path))
+            continue;
+        const { key, label, color, klass } = classifyFile(file.path, det);
+        let row = map.get(key);
+        if (row === undefined) {
+            row = { key, label, color, klass, files: 0, bytes: 0 };
+            map.set(key, row);
+        }
+        row.files += 1;
+        row.bytes += file.size;
+        totalFiles += 1;
+        totalBytes += file.size;
+    }
+    const klassRank = { parsed: 0, recognized: 1, other: 2 };
+    const rows = [...map.values()].sort((a, b) => klassRank[a.klass] - klassRank[b.klass] ||
+        b.bytes - a.bytes ||
+        (a.key < b.key ? -1 : 1));
+    return { rows, totalFiles, totalBytes };
+}
+//# sourceMappingURL=languages.js.map

package/lib/gitnexus/languages.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"languages.js","sourceRoot":"","sources":["../../src/gitnexus/languages.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAAE,gBAAgB,EAAE,MAAM,WAAW,CAAC;AAM7C;;;;;GAKG;AACH,MAAM,OAAO,GAAsE,MAAM,CAAC,MAAM,CAAC;IAC/F,UAAU,EAAE,CAAC,YAAY,EAAE,SAAS,CAAC;IACrC,UAAU,EAAE,CAAC,YAAY,EAAE,SAAS,CAAC;IACrC,MAAM,EAAE,CAAC,QAAQ,EAAE,SAAS,CAAC;IAC7B,IAAI,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;IACzB,CAAC,EAAE,CAAC,GAAG,EAAE,SAAS,CAAC;IACnB,GAAG,EAAE,CAAC,KAAK,EAAE,SAAS,CAAC;IACvB,aAAa,EAAE,CAAC,aAAa,EAAE,SAAS,CAAC;IACzC,MAAM,EAAE,CAAC,IAAI,EAAE,SAAS,CAAC;IACzB,EAAE,EAAE,CAAC,IAAI,EAAE,SAAS,CAAC;IACrB,IAAI,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;IACzB,IAAI,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;IACzB,GAAG,EAAE,CAAC,KAAK,EAAE,SAAS,CAAC;IACvB,MAAM,EAAE,CAAC,QAAQ,EAAE,SAAS,CAAC;IAC7B,KAAK,EAAE,CAAC,OAAO,EAAE,SAAS,CAAC;IAC3B,IAAI,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;IACzB,GAAG,EAAE,CAAC,KAAK,EAAE,SAAS,CAAC;IACvB,KAAK,EAAE,CAAC,OAAO,EAAE,SAAS,CAAC;IAC3B,GAAG,EAAE,CAAC,KAAK,EAAE,SAAS,CAAC;IACvB,+EAA+E;IAC/E,QAAQ,EAAE,CAAC,UAAU,EAAE,SAAS,CAAC;IACjC,IAAI,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;IACzB,IAAI,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;IACzB,MAAM,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;IAC3B,GAAG,EAAE,CAAC,KAAK,EAAE,SAAS,CAAC;IACvB,IAAI,EAAE,CAAC,OAAO,EAAE,SAAS,CAAC;IAC1B,GAAG,EAAE,CAAC,KAAK,EAAE,SAAS,CAAC;IACvB,IAAI,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;IACzB,GAAG,EAAE,CAAC,KAAK,EAAE,SAAS,CAAC;IACvB,MAAM,EAAE,CAAC,YAAY,EAAE,SAAS,CAAC;IACjC,QAAQ,EAAE,CAAC,UAAU,EAAE,SAAS,CAAC;IACjC,gFAAgF;IAChF,gFAAgF;IAChF,gFAAgF;IAChF,iDAAiD;IACjD,KAAK,EAAE,CAAC,OAAO,EAAE,SAAS,CAAC;IAC3B,UAAU,EAAE,CAAC,YAAY,EAAE,SAAS,CAAC;IACrC,KAAK,EAAE,CAAC,WAAW,EAAE,SAAS,CAAC;IAC/B,IAAI,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;IACzB,SAAS,EAAE,CAAC,MAAM,EAAE,SAAS,CAAC;IAC9B,IAAI,EAAE,CAAC,IAAI,EAAE,SAAS,CAAC;CACxB,CAAC,CAAC;AAEH;;;;GAIG;AACH,MAAM,wBAAwB,GAAqC,MAAM,CAAC,MAAM,CAAC;IAC/E,QAAQ,EAAE,OAAO;IACjB,SAAS,EAAE,OAAO;IAClB,KAAK,EAAE,OAAO;IACd,MAAM,EAAE,OAAO;IACf,MAAM,EAAE,YAAY;IACpB,OAAO,EAAE,YAAY;IACrB,OAAO,EAAE,YAAY;IACrB,SAAS,EAAE,YAAY;IACvB,MAAM,EAAE,OAAO;IACf,MAAM,EAAE,OAAO;IACf,QAAQ,EAAE,MAAM;IAChB,OAAO,EAAE,MAAM;IACf,MAAM,EAAE,WAAW;IACnB,MAAM,EAAE,WAAW;CACpB,CAAC,CAAC;AAEH;;;GAGG;AACH,MAAM,kBAAkB,GAAqC,MAAM,CAAC,MAAM,CAAC;IACzE,UAAU,EAAE,QAAQ;IACpB,QAAQ,EAAE,UAAU;IACpB,gBAAgB,EAAE,MAAM;IACxB,YAAY,EAAE,MAAM;IACpB,gBAAgB,EAAE,MAAM;IACxB,eAAe,EAAE,MAAM;IACvB,QAAQ,EAAE,MAAM;IAChB,QAAQ,EAAE,MAAM;CACjB,CAAC,CAAC;AAyCH,IAAI,QAAQ,GAA4B,IAAI,CAAC;AAE7C;;;;GAIG;AACH,MAAM,CAAC,KAAK,UAAU,oBAAoB,CACxC,UAAuC,EAAE;IAEzC,IAAI,QAAQ,KAAK,IAAI;QAAE,OAAO,QAAQ,CAAC;IACvC,MAAM,IAAI,GAAG,gBAAgB,CAAC,OAAO,CAAC,CAAC;IACvC,IAAI,IAAI,KAAK,IAAI;QAAE,OAAO,IAAI,CAAC;IAC/B,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,CAAC,MAAM,MAAM,CACvB,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,MAAM,EAAE,SAAS,EAAE,uBAAuB,CAAC,CAAC,CAAC,IAAI,CAChF,CAAmB,CAAC;QACrB,IAAI,OAAO,GAAG,CAAC,uBAAuB,KAAK,UAAU,EAAE,CAAC;YACtD,QAAQ,GAAG;gBACT,uBAAuB,EAAE,GAAG,CAAC,uBAAuB;gBACpD,6BAA6B,EAC3B,OAAO,GAAG,CAAC,6BAA6B,KAAK,UAAU;oBACrD,CAAC,CAAC,GAAG,CAAC,6BAA6B;oBACnC,CAAC,CAAC,GAAG,EAAE,CAAC,MAAM;aACnB,CAAC;QACJ,CAAC;IACH,CAAC;IAAC,MAAM,CAAC;QACP,uCAAuC;IACzC,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,YAAY,CAAC,QAAgB,EAAE,GAA4B;IACzE,IAAI,GAAG,GAAkB,IAAI,CAAC;IAC9B,IAAI,KAAK,GAAkB,OAAO,CAAC;IAEnC,IAAI,GAAG,KAAK,IAAI,EAAE,CAAC;QACjB,MAAM,MAAM,GAAG,GAAG,CAAC,uBAAuB,CAAC,QAAQ,CAAC,CAAC;QACrD,IAAI,MAAM,EAAE,CAAC;YACX,GAAG,GAAG,MAAM,CAAC;YACb,KAAK,GAAG,QAAQ,CAAC;QACnB,CAAC;aAAM,CAAC;YACN,MAAM,MAAM,GAAG,GAAG,CAAC,6BAA6B,CAAC,QAAQ,CAAC,CAAC;YAC3D,IAAI,MAAM,IAAI,MAAM,KAAK,MAAM,EAAE,CAAC;gBAChC,GAAG,GAAG,MAAM,CAAC;gBACb,KAAK,GAAG,YAAY,CAAC;YACvB,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,GAAG,KAAK,IAAI,EAAE,CAAC;QACjB,MAAM,OAAO,GAAG,QAAQ,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;QAC1C,IAAI,OAAO,IAAI,CAAC,EAAE,CAAC;YACjB,MAAM,GAAG,GAAG,QAAQ,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,WAAW,EAAE,CAAC;YAClD,MAAM,UAAU,GAAG,wBAAwB,CAAC,GAAG,CAAC,CAAC;YACjD,IAAI,UAAU,KAAK,SAAS;gBAAE,GAAG,GAAG,UAAU,CAAC;QACjD,CAAC;IACH,CAAC;IAED,IAAI,GAAG,KAAK,IAAI,EAAE,CAAC;QACjB,MAAM,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC;QAC5D,MAAM,MAAM,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAC;QACxC,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;YACzB,GAAG,GAAG,MAAM,CAAC;YACb,KAAK,GAAG,MAAM,KAAK,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,YAAY,CAAC;QACrD,CAAC;IACH,CAAC;IAED,MAAM,KAAK,GAAG,GAAG,KAAK,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IACtD,kFAAkF;IAClF,6DAA6D;IAC7D,IAAI,KAAK,KAAK,SAAS,IAAI,GAAG,KAAK,IAAI,EAAE,CAAC;QACxC,OAAO,EAAE,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,CAAC;IAC9E,CAAC;IACD,gFAAgF;IAChF,IAAI,KAAK,KAAK,OAAO,IAAI,GAAG,KAAK,MAAM;QAAE,KAAK,GAAG,YAAY,CAAC;IAC9D,OAAO,EAAE,GAAG,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC;AAC1D,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,kBAAkB,CAChC,KAAoB,EACpB,GAA4B,EAC5B,OAAuC;IAEvC,MAAM,GAAG,GAAG,IAAI,GAAG,EAAuB,CAAC;IAC3C,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,IAAI,UAAU,GAAG,CAAC,CAAC;IAEnB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,OAAO,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC;YAAE,SAAS;QAC7C,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,GAAG,YAAY,CAAC,IAAI,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;QAClE,IAAI,GAAG,GAAG,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACvB,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;YACtB,GAAG,GAAG,EAAE,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC;YACvD,GAAG,CAAC,GAAG,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC;QACpB,CAAC;QACD,GAAG,CAAC,KAAK,IAAI,CAAC,CAAC;QACf,GAAG,CAAC,KAAK,IAAI,IAAI,CAAC,IAAI,CAAC;QACvB,UAAU,IAAI,CAAC,CAAC;QAChB,UAAU,IAAI,IAAI,CAAC,IAAI,CAAC;IAC1B,CAAC;IAED,MAAM,SAAS,GAAG,EAAE,MAAM,EAAE,CAAC,EAAE,UAAU,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC;IACzD,MAAM,IAAI,GAAG,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,CAAC,IAAI,CACjC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACP,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,GAAG,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC;QACvC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK;QACjB,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAC3B,CAAC;IAEF,OAAO,EAAE,IAAI,EAAE,UAAU,EAAE,UAAU,EAAE,CAAC;AAC1C,CAAC"}

package/lib/gitnexus/root.d.ts

@@ -0,0 +1,68 @@
+/**
+ * Locate the installed GitNexus package and reuse its own scan primitives.
+ *
+ * WHY REUSE INSTEAD OF REIMPLEMENT
+ * --------------------------------
+ * "Which files did GitNexus skip?" has exactly one correct answer: the set
+ * `walkRepositoryPaths` produced during `analyze`. Reimplementing that decision
+ * here would drift from it silently — the plugin would report a large-file list
+ * that no longer matches the index, and nothing would fail loudly.
+ *
+ * `gitnexus/package.json` declares no `exports` map, so its `dist/**` files are
+ * deep-importable. That is what makes reuse possible; it is also why every entry
+ * point below is guarded. If a future release closes the exports map or renames a
+ * module, `loadScanPrimitives` returns `null` and the caller falls back to the
+ * local scan (`./scan-fallback.js`) rather than throwing.
+ */
+/** What a caller may say about where to look for the installed package. */
+export interface RootDiscoveryOptions {
+    env?: NodeJS.ProcessEnv;
+    moduleUrl?: string;
+    discover?: boolean;
+}
+/**
+ * Resolve the GitNexus package root, or `null` when no usable install is found.
+ */
+export declare function findGitnexusRoot(options?: RootDiscoveryOptions): string | null;
+/**
+ * Read the installed GitNexus version. Returned for provenance so a report can
+ * state which analyzer produced the index it is judging.
+ */
+export declare function readGitnexusVersion(root: string): string | null;
+/** `walkRepositoryPaths(repoPath, onProgress?, options?)` — GitNexus's own walk. */
+export type WalkRepositoryPaths = (repoPath: string, onProgress?: unknown, options?: object) => Promise<Array<{
+    path: string;
+    size: number;
+}>>;
+/** `createIgnoreFilter(repoPath, options?)` — the matcher `glob` is handed. */
+export type CreateIgnoreFilter = (repoPath: string, options?: object) => Promise<unknown>;
+/** The slice of the `glob` package this plugin calls. */
+export type GlobFunction = (pattern: string, options: object) => Promise<string[]>;
+/**
+ * The scan primitives this plugin reuses, all sourced from one GitNexus install.
+ */
+export interface ScanPrimitives {
+    gitnexusRoot: string;
+    gitnexusVersion: string | null;
+    glob: GlobFunction;
+    createIgnoreFilter: CreateIgnoreFilter;
+    walkRepositoryPaths: WalkRepositoryPaths;
+    getMaxFileSizeBytes: () => number;
+    DEFAULT_MAX_FILE_SIZE_BYTES: number;
+    MAX_FILE_SIZE_UPPER_BOUND_BYTES: number;
+    source: 'gitnexus' | 'fallback';
+}
+/** What a caller may override when loading the primitives. */
+export interface LoadScanPrimitivesOptions extends RootDiscoveryOptions {
+    root?: string;
+}
+/**
+ * Load GitNexus's own scan primitives by deep-importing its dist modules.
+ *
+ * Every exported member is validated before use: a partial or reshaped module
+ * degrades to the fallback rather than producing a wrong answer.
+ *
+ * @returns `null` when reuse is not possible.
+ */
+export declare function loadScanPrimitives(options?: LoadScanPrimitivesOptions): Promise<ScanPrimitives | null>;
+//# sourceMappingURL=root.d.ts.map

package/lib/gitnexus/root.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"root.d.ts","sourceRoot":"","sources":["../../src/gitnexus/root.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAcH,2EAA2E;AAC3E,MAAM,WAAW,oBAAoB;IACnC,GAAG,CAAC,EAAE,MAAM,CAAC,UAAU,CAAC;IACxB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,QAAQ,CAAC,EAAE,OAAO,CAAC;CACpB;AAqDD;;GAEG;AACH,wBAAgB,gBAAgB,CAAC,OAAO,GAAE,oBAAyB,GAAG,MAAM,GAAG,IAAI,CAWlF;AAED;;;GAGG;AACH,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,CAS/D;AAKD,oFAAoF;AACpF,MAAM,MAAM,mBAAmB,GAAG,CAChC,QAAQ,EAAE,MAAM,EAChB,UAAU,CAAC,EAAE,OAAO,EACpB,OAAO,CAAC,EAAE,MAAM,KACb,OAAO,CAAC,KAAK,CAAC;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAC,CAAC,CAAC;AAEpD,+EAA+E;AAC/E,MAAM,MAAM,kBAAkB,GAAG,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,MAAM,KAAK,OAAO,CAAC,OAAO,CAAC,CAAC;AAE1F,yDAAyD;AACzD,MAAM,MAAM,YAAY,GAAG,CAAC,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,KAAK,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;AA6BnF;;GAEG;AACH,MAAM,WAAW,cAAc;IAC7B,YAAY,EAAE,MAAM,CAAC;IACrB,eAAe,EAAE,MAAM,GAAG,IAAI,CAAC;IAC/B,IAAI,EAAE,YAAY,CAAC;IACnB,kBAAkB,EAAE,kBAAkB,CAAC;IACvC,mBAAmB,EAAE,mBAAmB,CAAC;IACzC,mBAAmB,EAAE,MAAM,MAAM,CAAC;IAClC,2BAA2B,EAAE,MAAM,CAAC;IACpC,+BAA+B,EAAE,MAAM,CAAC;IACxC,MAAM,EAAE,UAAU,GAAG,UAAU,CAAC;CACjC;AAED,8DAA8D;AAC9D,MAAM,WAAW,yBAA0B,SAAQ,oBAAoB;IACrE,IAAI,CAAC,EAAE,MAAM,CAAC;CACf;AAED;;;;;;;GAOG;AACH,wBAAsB,kBAAkB,CACtC,OAAO,GAAE,yBAA8B,GACtC,OAAO,CAAC,cAAc,GAAG,IAAI,CAAC,CAiDhC"}

package/lib/gitnexus/root.js

@@ -0,0 +1,162 @@
+/**
+ * Locate the installed GitNexus package and reuse its own scan primitives.
+ *
+ * WHY REUSE INSTEAD OF REIMPLEMENT
+ * --------------------------------
+ * "Which files did GitNexus skip?" has exactly one correct answer: the set
+ * `walkRepositoryPaths` produced during `analyze`. Reimplementing that decision
+ * here would drift from it silently — the plugin would report a large-file list
+ * that no longer matches the index, and nothing would fail loudly.
+ *
+ * `gitnexus/package.json` declares no `exports` map, so its `dist/**` files are
+ * deep-importable. That is what makes reuse possible; it is also why every entry
+ * point below is guarded. If a future release closes the exports map or renames a
+ * module, `loadScanPrimitives` returns `null` and the caller falls back to the
+ * local scan (`./scan-fallback.js`) rather than throwing.
+ */
+import { createRequire } from 'node:module';
+import { pathToFileURL } from 'node:url';
+import { existsSync, readFileSync } from 'node:fs';
+import path from 'node:path';
+/** Deep paths inside the `gitnexus` package that this plugin depends on. */
+const DIST = Object.freeze({
+    walker: 'dist/core/ingestion/filesystem-walker.js',
+    maxFileSize: 'dist/core/ingestion/utils/max-file-size.js',
+    ignore: 'dist/config/ignore-service.js',
+});
+/**
+ * Candidate directories that may contain an installed `gitnexus` package.
+ * Ordered most-specific first: an explicit override beats PATH discovery, which
+ * beats a plain resolution from this package.
+ *
+ * Discovery always consults the real process environment for `PATH`, even when a
+ * caller passes its own `env`. The `env` option governs *configuration*
+ * (`GITNEXUS_MAX_FILE_SIZE`, `GITNEXUS_HOME`); which GitNexus is installed on this
+ * machine is a fact about the machine, not a per-call setting, and letting a
+ * sanitized env blank it out would silently downgrade every caller to the
+ * fallback scan.
+ *
+ * @returns absolute candidate directories (not yet validated)
+ */
+function candidateRoots(options = {}) {
+    if (options.discover === false)
+        return [];
+    const env = { ...process.env, ...(options.env ?? {}) };
+    const roots = [];
+    if (env.GITNEXUS_PACKAGE_ROOT)
+        roots.push(env.GITNEXUS_PACKAGE_ROOT);
+    // Probe every PATH entry for a sibling `node_modules/gitnexus` directly.
+    //
+    // Deliberately NOT "find the gitnexus shim, then derive its prefix": the shim
+    // is `.ps1` under nvm4w, and `PATHEXT` on Windows does not list `.ps1`, so
+    // shim-based discovery silently finds nothing on the very machines where the
+    // CLI is most likely to be installed. Testing the directory is
+    // extension-independent and needs no PATHEXT reasoning.
+    for (const dir of (env.PATH ?? '').split(path.delimiter)) {
+        if (dir === '')
+            continue;
+        roots.push(path.join(dir, 'node_modules', 'gitnexus'));
+        roots.push(path.join(dir, '..', 'node_modules', 'gitnexus'));
+        roots.push(path.join(dir, '..', '..', 'node_modules', 'gitnexus'));
+    }
+    // nvm-for-windows keeps the real tree off PATH behind a junction.
+    if (env.NVM_SYMLINK)
+        roots.push(path.join(env.NVM_SYMLINK, 'node_modules', 'gitnexus'));
+    if (env.NVM_HOME)
+        roots.push(path.join(env.NVM_HOME, 'node_modules', 'gitnexus'));
+    // Resolve from this package (a real dependency edge, if declared).
+    try {
+        const here = options.moduleUrl ? path.dirname(new URL(options.moduleUrl).pathname.replace(/^\/([A-Za-z]:)/, '$1')) : undefined;
+        const req = createRequire(here ? path.join(here, 'noop.js') : import.meta.url);
+        roots.push(path.dirname(req.resolve('gitnexus/package.json')));
+    }
+    catch {
+        /* not resolvable from here — the other candidates may still work */
+    }
+    return [...new Set(roots.map((r) => path.resolve(r)))];
+}
+/**
+ * Resolve the GitNexus package root, or `null` when no usable install is found.
+ */
+export function findGitnexusRoot(options = {}) {
+    for (const root of candidateRoots(options)) {
+        try {
+            if (existsSync(path.join(root, 'package.json')) && existsSync(path.join(root, DIST.walker))) {
+                return root;
+            }
+        }
+        catch {
+            /* unreadable candidate — keep looking */
+        }
+    }
+    return null;
+}
+/**
+ * Read the installed GitNexus version. Returned for provenance so a report can
+ * state which analyzer produced the index it is judging.
+ */
+export function readGitnexusVersion(root) {
+    try {
+        const pkg = JSON.parse(readFileSync(path.join(root, 'package.json'), 'utf8'));
+        return typeof pkg.version === 'string' ? pkg.version : null;
+    }
+    catch {
+        return null;
+    }
+}
+/** The URL a deep-imported `dist` module is loaded from. */
+const distUrl = (root, relative) => pathToFileURL(path.join(root, relative)).href;
+/**
+ * Load GitNexus's own scan primitives by deep-importing its dist modules.
+ *
+ * Every exported member is validated before use: a partial or reshaped module
+ * degrades to the fallback rather than producing a wrong answer.
+ *
+ * @returns `null` when reuse is not possible.
+ */
+export async function loadScanPrimitives(options = {}) {
+    const root = options.root ?? findGitnexusRoot(options);
+    if (root === null)
+        return null;
+    try {
+        const [walker, maxSize, ignore] = await Promise.all([
+            import(distUrl(root, DIST.walker)),
+            import(distUrl(root, DIST.maxFileSize)),
+            import(distUrl(root, DIST.ignore)),
+        ]);
+        const createIgnoreFilter = ignore.createIgnoreFilter;
+        const walkRepositoryPaths = walker.walkRepositoryPaths;
+        const getMaxFileSizeBytes = maxSize.getMaxFileSizeBytes;
+        const DEFAULT_MAX_FILE_SIZE_BYTES = maxSize.DEFAULT_MAX_FILE_SIZE_BYTES;
+        if (typeof createIgnoreFilter !== 'function' ||
+            typeof walkRepositoryPaths !== 'function' ||
+            typeof getMaxFileSizeBytes !== 'function' ||
+            typeof DEFAULT_MAX_FILE_SIZE_BYTES !== 'number') {
+            return null;
+        }
+        // `glob` is resolved from GitNexus's own dependency tree so the version that
+        // built the ignore filter is the version that consumes it.
+        const req = createRequire(path.join(root, 'package.json'));
+        const globModule = (await import(pathToFileURL(req.resolve('glob')).href));
+        const glob = globModule.glob ?? globModule.default?.glob ?? globModule.default;
+        if (typeof glob !== 'function')
+            return null;
+        return {
+            gitnexusRoot: root,
+            gitnexusVersion: readGitnexusVersion(root),
+            glob,
+            createIgnoreFilter,
+            walkRepositoryPaths,
+            getMaxFileSizeBytes,
+            DEFAULT_MAX_FILE_SIZE_BYTES,
+            MAX_FILE_SIZE_UPPER_BOUND_BYTES: typeof maxSize.MAX_FILE_SIZE_UPPER_BOUND_BYTES === 'number'
+                ? maxSize.MAX_FILE_SIZE_UPPER_BOUND_BYTES
+                : DEFAULT_MAX_FILE_SIZE_BYTES,
+            source: 'gitnexus',
+        };
+    }
+    catch {
+        return null;
+    }
+}
+//# sourceMappingURL=root.js.map

package/lib/gitnexus/root.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"root.js","sourceRoot":"","sources":["../../src/gitnexus/root.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAEH,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAC5C,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAAE,UAAU,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AACnD,OAAO,IAAI,MAAM,WAAW,CAAC;AAE7B,4EAA4E;AAC5E,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC;IACzB,MAAM,EAAE,0CAA0C;IAClD,WAAW,EAAE,4CAA4C;IACzD,MAAM,EAAE,+BAA+B;CACxC,CAAC,CAAC;AASH;;;;;;;;;;;;;GAaG;AACH,SAAS,cAAc,CAAC,UAAgC,EAAE;IACxD,IAAI,OAAO,CAAC,QAAQ,KAAK,KAAK;QAAE,OAAO,EAAE,CAAC;IAC1C,MAAM,GAAG,GAAsB,EAAE,GAAG,OAAO,CAAC,GAAG,EAAE,GAAG,CAAC,OAAO,CAAC,GAAG,IAAI,EAAE,CAAC,EAAE,CAAC;IAC1E,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,IAAI,GAAG,CAAC,qBAAqB;QAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,qBAAqB,CAAC,CAAC;IAErE,yEAAyE;IACzE,EAAE;IACF,8EAA8E;IAC9E,2EAA2E;IAC3E,6EAA6E;IAC7E,+DAA+D;IAC/D,wDAAwD;IACxD,KAAK,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;QACzD,IAAI,GAAG,KAAK,EAAE;YAAE,SAAS;QACzB,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,cAAc,EAAE,UAAU,CAAC,CAAC,CAAC;QACvD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,EAAE,cAAc,EAAE,UAAU,CAAC,CAAC,CAAC;QAC7D,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,cAAc,EAAE,UAAU,CAAC,CAAC,CAAC;IACrE,CAAC;IAED,kEAAkE;IAClE,IAAI,GAAG,CAAC,WAAW;QAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,WAAW,EAAE,cAAc,EAAE,UAAU,CAAC,CAAC,CAAC;IACxF,IAAI,GAAG,CAAC,QAAQ;QAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,cAAc,EAAE,UAAU,CAAC,CAAC,CAAC;IAElF,mEAAmE;IACnE,IAAI,CAAC;QACH,MAAM,IAAI,GAAG,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,gBAAgB,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;QAC/H,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAC/E,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,uBAAuB,CAAC,CAAC,CAAC,CAAC;IACjE,CAAC;IAAC,MAAM,CAAC;QACP,oEAAoE;IACtE,CAAC;IAED,OAAO,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AACzD,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,gBAAgB,CAAC,UAAgC,EAAE;IACjE,KAAK,MAAM,IAAI,IAAI,cAAc,CAAC,OAAO,CAAC,EAAE,CAAC;QAC3C,IAAI,CAAC;YACH,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,cAAc,CAAC,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC,EAAE,CAAC;gBAC5F,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,yCAAyC;QAC3C,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,mBAAmB,CAAC,IAAY;IAC9C,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,cAAc,CAAC,EAAE,MAAM,CAAC,CAE3E,CAAC;QACF,OAAO,OAAO,GAAG,CAAC,OAAO,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC;IAC9D,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED,4DAA4D;AAC5D,MAAM,OAAO,GAAG,CAAC,IAAY,EAAE,QAAgB,EAAU,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC;AA8D1G;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,kBAAkB,CACtC,UAAqC,EAAE;IAEvC,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,gBAAgB,CAAC,OAAO,CAAC,CAAC;IACvD,IAAI,IAAI,KAAK,IAAI;QAAE,OAAO,IAAI,CAAC;IAE/B,IAAI,CAAC;QACH,MAAM,CAAC,MAAM,EAAE,OAAO,EAAE,MAAM,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;YAClD,MAAM,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAoC;YACrE,MAAM,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,WAAW,CAAC,CAA+B;YACrE,MAAM,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAiC;SACnE,CAAC,CAAC;QAEH,MAAM,kBAAkB,GAAG,MAAM,CAAC,kBAAkB,CAAC;QACrD,MAAM,mBAAmB,GAAG,MAAM,CAAC,mBAAmB,CAAC;QACvD,MAAM,mBAAmB,GAAG,OAAO,CAAC,mBAAmB,CAAC;QACxD,MAAM,2BAA2B,GAAG,OAAO,CAAC,2BAA2B,CAAC;QAExE,IACE,OAAO,kBAAkB,KAAK,UAAU;YACxC,OAAO,mBAAmB,KAAK,UAAU;YACzC,OAAO,mBAAmB,KAAK,UAAU;YACzC,OAAO,2BAA2B,KAAK,QAAQ,EAC/C,CAAC;YACD,OAAO,IAAI,CAAC;QACd,CAAC;QAED,6EAA6E;QAC7E,2DAA2D;QAC3D,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,cAAc,CAAC,CAAC,CAAC;QAC3D,MAAM,UAAU,GAAG,CAAC,MAAM,MAAM,CAAC,aAAa,CAAC,GAAG,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAe,CAAC;QACzF,MAAM,IAAI,GAAG,UAAU,CAAC,IAAI,IAAI,UAAU,CAAC,OAAO,EAAE,IAAI,IAAI,UAAU,CAAC,OAAO,CAAC;QAC/E,IAAI,OAAO,IAAI,KAAK,UAAU;YAAE,OAAO,IAAI,CAAC;QAE5C,OAAO;YACL,YAAY,EAAE,IAAI;YAClB,eAAe,EAAE,mBAAmB,CAAC,IAAI,CAAC;YAC1C,IAAI;YACJ,kBAAkB;YAClB,mBAAmB;YACnB,mBAAmB;YACnB,2BAA2B;YAC3B,+BAA+B,EAC7B,OAAO,OAAO,CAAC,+BAA+B,KAAK,QAAQ;gBACzD,CAAC,CAAC,OAAO,CAAC,+BAA+B;gBACzC,CAAC,CAAC,2BAA2B;YACjC,MAAM,EAAE,UAAU;SACnB,CAAC;IACJ,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC"}

package/lib/gitnexus/scan.d.ts

@@ -0,0 +1,67 @@
+/**
+ * Reproduce GitNexus's scan decision and expose what it discards.
+ *
+ * `walkRepositoryPaths` (gitnexus/dist/core/ingestion/filesystem-walker.js)
+ * makes three cuts and only returns the survivors:
+ *
+ *   1. `glob(recursive, { nodir: true, dot: false, ignore: createIgnoreFilter(repo) })`
+ *      — drops dotfiles (`dot: false`) and anything the ignore rules match.
+ *   2. `stat.size > maxFileSizeBytes` — the large-file cut. **Counted and logged
+ *      but never returned**, so an external tool cannot see it.
+ *   3. `.d.ts` / `.d.mts` / `.d.cts` whose same-stem implementation was also
+ *      scanned — suppressed as redundant.
+ *
+ * This module applies cuts 1 and 3 exactly as upstream does (reusing upstream's
+ * own `createIgnoreFilter` and glob), and turns cut 2 — the one that is invisible
+ * from outside — into data.
+ *
+ * Deliberate duplication: the single `glob(...)` call below mirrors
+ * `filesystem-walker.js:62-67`. It is four lines and must stay in lockstep with
+ * upstream. If it ever drifts, `test/run.mjs` catches it by cross-checking the
+ * reported large-file count against what `analyze` itself prints.
+ */
+import { type ScanPrimitives } from './root.js';
+/** One file the scan walked. */
+export interface ScannedFile {
+    /** repo-relative, forward slashes */
+    path: string;
+    /** bytes */
+    size: number;
+}
+/** Where the effective size cap came from. */
+export type CapSource = 'env' | 'default' | 'index';
+/** What {@link scanRepository} reports: the survivors plus all three cuts as data. */
+export interface ScanReport {
+    /** files GitNexus would hand to the pipeline */
+    selected: ScannedFile[];
+    /** files dropped by the size cap, largest first */
+    large: ScannedFile[];
+    /** `.d.ts` dropped as redundant */
+    suppressedDeclarations: ScannedFile[];
+    /** files that survived the ignore/dotfile cut */
+    scannedCount: number;
+    /** the effective size cap */
+    capBytes: number;
+    /** the shipped default */
+    defaultCapBytes: number;
+    /** where the effective cap came from */
+    capSource: CapSource;
+    /** hard tree-sitter ceiling */
+    maxCapBytes: number;
+    gitnexusVersion: string;
+    /** reused upstream primitives or a local glob */
+    primitives: 'gitnexus' | 'fallback';
+}
+/**
+ * Scan a repository the way GitNexus does, and report the size cut as data.
+ *
+ * @param repoPath absolute path to the repository root
+ * @param options `capBytes` overrides the cap (used when the index records its own coverage
+ *   policy); otherwise `GITNEXUS_MAX_FILE_SIZE` decides, as it does for `analyze`.
+ */
+export declare function scanRepository(repoPath: string, options?: {
+    capBytes?: number;
+    env?: NodeJS.ProcessEnv;
+    primitives?: ScanPrimitives | null;
+}): Promise<ScanReport>;
+//# sourceMappingURL=scan.d.ts.map

package/lib/gitnexus/scan.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"scan.d.ts","sourceRoot":"","sources":["../../src/gitnexus/scan.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AAKH,OAAO,EAAsB,KAAK,cAAc,EAAE,MAAM,WAAW,CAAC;AAkBpE,gCAAgC;AAChC,MAAM,WAAW,WAAW;IAC1B,qCAAqC;IACrC,IAAI,EAAE,MAAM,CAAC;IACb,YAAY;IACZ,IAAI,EAAE,MAAM,CAAC;CACd;AAED,8CAA8C;AAC9C,MAAM,MAAM,SAAS,GAAG,KAAK,GAAG,SAAS,GAAG,OAAO,CAAC;AAEpD,sFAAsF;AACtF,MAAM,WAAW,UAAU;IACzB,gDAAgD;IAChD,QAAQ,EAAE,WAAW,EAAE,CAAC;IACxB,mDAAmD;IACnD,KAAK,EAAE,WAAW,EAAE,CAAC;IACrB,mCAAmC;IACnC,sBAAsB,EAAE,WAAW,EAAE,CAAC;IACtC,iDAAiD;IACjD,YAAY,EAAE,MAAM,CAAC;IACrB,6BAA6B;IAC7B,QAAQ,EAAE,MAAM,CAAC;IACjB,0BAA0B;IAC1B,eAAe,EAAE,MAAM,CAAC;IACxB,wCAAwC;IACxC,SAAS,EAAE,SAAS,CAAC;IACrB,+BAA+B;IAC/B,WAAW,EAAE,MAAM,CAAC;IACpB,eAAe,EAAE,MAAM,CAAC;IACxB,iDAAiD;IACjD,UAAU,EAAE,UAAU,GAAG,UAAU,CAAC;CACrC;AA6CD;;;;;;GAMG;AACH,wBAAsB,cAAc,CAClC,QAAQ,EAAE,MAAM,EAChB,OAAO,GAAE;IACP,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,GAAG,CAAC,EAAE,MAAM,CAAC,UAAU,CAAC;IACxB,UAAU,CAAC,EAAE,cAAc,GAAG,IAAI,CAAC;CAC/B,GACL,OAAO,CAAC,UAAU,CAAC,CAgErB"}