@amsterdamdatalabs/enact-docs-search 0.1.0

package/dist/search/search-service.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"search-service.js","sourceRoot":"","sources":["../../src/search/search-service.ts"],"names":[],"mappings":"AAgBA,OAAO,EAAE,iBAAiB,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AA8B/D,MAAM,KAAK,GAAG,EAAE,CAAC;AAEjB,MAAM,OAAO,aAAa;IAEd;IACA;IACA;IAHV,YACU,QAAuB,EACvB,WAAwB,EACxB,gBAAyC;QAFzC,aAAQ,GAAR,QAAQ,CAAe;QACvB,gBAAW,GAAX,WAAW,CAAa;QACxB,qBAAgB,GAAhB,gBAAgB,CAAyB;IAChD,CAAC;IAEJ,KAAK,CAAC,MAAM,CAAC,KAAa,EAAE,OAAsB;QAChD,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;QAC7B,IAAI,CAAC,OAAO;YAAE,OAAO,EAAE,CAAC;QAExB,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC;QAClC,MAAM,SAAS,GAAkB,OAAO,CAAC,SAAS,IAAI,OAAO,CAAC;QAC9D,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,CAAC;QAElC,QAAQ,OAAO,CAAC,IAAI,EAAE,CAAC;YACrB,KAAK,SAAS;gBACZ,OAAO,IAAI,CAAC,UAAU,CAAC,OAAO,EAAE,OAAO,CAAC,KAAK,EAAE,KAAK,EAAE,SAAS,EAAE,QAAQ,CAAC,CAAC;YAC7E,KAAK,UAAU;gBACb,OAAO,IAAI,CAAC,cAAc,CAAC,OAAO,EAAE,OAAO,CAAC,KAAK,EAAE,KAAK,EAAE,SAAS,EAAE,QAAQ,CAAC,CAAC;YACjF,KAAK,QAAQ;gBACX,OAAO,IAAI,CAAC,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,KAAK,EAAE,KAAK,EAAE,SAAS,EAAE,QAAQ,CAAC,CAAC;QACjF,CAAC;IACH,CAAC;IAEO,KAAK,CAAC,UAAU,CACtB,KAAa,EACb,KAAY,EACZ,KAAa,EACb,SAAwB,EACxB,QAAiB;QAEjB,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE,SAAS,EAAE,QAAQ,EAAE,CAAC,CAAC;QACzF,OAAO,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YACzB,EAAE,EAAE,CAAC,CAAC,EAAE;YACR,QAAQ,EAAE,CAAC,CAAC,QAAQ;YACpB,UAAU,EAAE,CAAC,CAAC,UAAU;YACxB,SAAS,EAAE,CAAC,CAAC,SAAS;YACtB,SAAS,EAAE,CAAC,CAAC,KAAK;YAClB,OAAO,EAAE,CAAC,CAAC,OAAO;YAClB,OAAO,EAAE,CAAC,CAAC,OAAO;YAClB,iBAAiB,EAAE,CAAC,CAAC,iBAAiB;YACtC,MAAM,EAAE,MAAe;SACxB,CAAC,CAAC,CAAC;IACN,CAAC;IAEO,KAAK,CAAC,cAAc,CAC1B,KAAa,EACb,KAAY,EACZ,KAAa,EACb,SAAwB,EACxB,QAAiB;QAEjB,IAAI,CAAC,IAAI,CAAC,gBAAgB,EAAE,CAAC;YAC3B,MAAM,IAAI,KAAK,CAAC,+CAA+C,CAAC,CAAC;QACnE,CAAC;QACD,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,gBAAgB,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QAC3D,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,SAAS,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC;QAEvE,MAAM,UAAU,GAAG,iBAAiB,CAAC,KAAK,CAAC,CAAC;QAE5C,MAAM,aAAa,GAAmB,EAAE,CAAC;QACzC,KAAK,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;YACxB,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YAClD,aAAa,CAAC,IAAI,CAAC;gBACjB,EAAE,EAAE,CAAC,CAAC,EAAE;gBACR,QAAQ,EAAE,CAAC,CAAC,QAAQ;gBACpB,UAAU,EAAE,CAAC,CAAC,UAAU;gBACxB,SAAS,EAAE,GAAG,EAAE,SAAS,IAAI,CAAC;gBAC9B,WAAW,EAAE,CAAC,CAAC,KAAK;gBACpB,OAAO,EAAE,GAAG,EAAE,OAAO;gBACrB,iBAAiB,EAAE,GAAG,EAAE,iBAAiB;gBACzC,OAAO,EAAE,GAAG;oBACV,CAAC,CAAC,MAAM,YAAY,CAChB,EAAE,QAAQ,EAAE,CAAC,CAAC,QAAQ,EAAE,OAAO,EAAE,GAAG,CAAC,OAAO,EAAE,SAAS,EAAE,GAAG,CAAC,SAAS,EAAE,EACxE,UAAU,EACV,SAAS,EACT,QAAQ,CACT;oBACH,CAAC,CAAC,SAAS;gBACb,MAAM,EAAE,QAAiB;aAC1B,CAAC,CAAC;QACL,CAAC;QACD,OAAO,aAAa,CAAC;IACvB,CAAC;IAEO,KAAK,CAAC,YAAY,CACxB,KAAa,EACb,KAAY,EACZ,KAAa,EACb,SAAwB,EACxB,QAAiB;QAEjB,MAAM,CAAC,WAAW,EAAE,aAAa,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;YACrD,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,EAAE,EAAE,KAAK,EAAE,KAAK,GAAG,CAAC,EAAE,KAAK,EAAE,SAAS,EAAE,QAAQ,EAAE,CAAC;YAC7E,IAAI,CAAC,gBAAgB;gBACnB,CAAC,CAAC,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,MAAM,IAAI,CAAC,gBAAgB,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,KAAK,GAAG,CAAC,EAAE,KAAK,CAAC;gBACrF,CAAC,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;SACxB,CAAC,CAAC;QAEH,MAAM,IAAI,GAAG,IAAI,GAAG,EAGjB,CAAC;QAEJ,WAAW,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE;YAC/B,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;gBACb,EAAE,EAAE,CAAC,CAAC,EAAE;gBACR,QAAQ,EAAE,CAAC,CAAC,QAAQ;gBACpB,UAAU,EAAE,CAAC,CAAC,UAAU;gBACxB,SAAS,EAAE,CAAC,CAAC,SAAS;gBACtB,SAAS,EAAE,CAAC,CAAC,KAAK;gBAClB,QAAQ,EAAE,KAAK,GAAG,CAAC;gBACnB,OAAO,EAAE,CAAC,CAAC,OAAO;gBAClB,OAAO,EAAE,CAAC,CAAC,OAAO;gBAClB,iBAAiB,EAAE,CAAC,CAAC,iBAAiB;gBACtC,MAAM,EAAE,MAAM;aACf,CAAC,CAAC;QACL,CAAC,CAAC,CAAC;QAEH,MAAM,UAAU,GAAG,iBAAiB,CAAC,KAAK,CAAC,CAAC;QAE5C,KAAK,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC,IAAI,aAAa,CAAC,OAAO,EAAE,EAAE,CAAC;YACjD,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YAChC,IAAI,QAAQ,EAAE,CAAC;gBACb,QAAQ,CAAC,WAAW,GAAG,CAAC,CAAC,KAAK,CAAC;gBAC/B,QAAQ,CAAC,UAAU,GAAG,KAAK,GAAG,CAAC,CAAC;gBAChC,QAAQ,CAAC,MAAM,GAAG,QAAQ,CAAC;YAC7B,CAAC;iBAAM,CAAC;gBACN,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;gBAClD,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;oBACb,EAAE,EAAE,CAAC,CAAC,EAAE;oBACR,QAAQ,EAAE,CAAC,CAAC,QAAQ;oBACpB,UAAU,EAAE,CAAC,CAAC,UAAU;oBACxB,SAAS,EAAE,GAAG,EAAE,SAAS,IAAI,CAAC;oBAC9B,WAAW,EAAE,CAAC,CAAC,KAAK;oBACpB,UAAU,EAAE,KAAK,GAAG,CAAC;oBACrB,OAAO,EAAE,GAAG,EAAE,OAAO;oBACrB,iBAAiB,EAAE,GAAG,EAAE,iBAAiB;oBACzC,OAAO,EAAE,GAAG;wBACV,CAAC,CAAC,MAAM,YAAY,CAChB,EAAE,QAAQ,EAAE,CAAC,CAAC,QAAQ,EAAE,OAAO,EAAE,GAAG,CAAC,OAAO,EAAE,SAAS,EAAE,GAAG,CAAC,SAAS,EAAE,EACxE,UAAU,EACV,SAAS,EACT,QAAQ,CACT;wBACH,CAAC,CAAC,SAAS;oBACb,MAAM,EAAE,QAAQ;iBACjB,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QAED,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;YAClD,IAAI,QAAQ,GAAG,CAAC,CAAC;YACjB,IAAI,CAAC,CAAC,QAAQ,EAAE,CAAC;gBACf,QAAQ,IAAI,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC;gBACrC,IAAI,CAAC,CAAC,QAAQ,KAAK,CAAC;oBAAE,QAAQ,IAAI,IAAI,CAAC;qBAClC,IAAI,CAAC,CAAC,QAAQ,IAAI,CAAC;oBAAE,QAAQ,IAAI,IAAI,CAAC;YAC7C,CAAC;YACD,IAAI,CAAC,CAAC,UAAU,EAAE,CAAC;gBACjB,QAAQ,IAAI,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC;gBACvC,IAAI,CAAC,CAAC,UAAU,KAAK,CAAC;oBAAE,QAAQ,IAAI,IAAI,CAAC;qBACpC,IAAI,CAAC,CAAC,UAAU,IAAI,CAAC;oBAAE,QAAQ,IAAI,IAAI,CAAC;YAC/C,CAAC;YACD,MAAM,EAAE,QAAQ,EAAE,UAAU,EAAE,GAAG,KAAK,EAAE,GAAG,CAAC,CAAC;YAC7C,OAAO,EAAE,GAAG,KAAK,EAAE,QAAQ,EAAE,CAAC;QAChC,CAAC,CAAC,CAAC;QAEH,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,QAAQ,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,IAAI,CAAC,CAAC,CAAC,CAAC;QAC9D,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;IACjC,CAAC;CACF"}

package/dist/search/snippet.d.ts

@@ -0,0 +1,65 @@
+/**
+ * Extract meaningful query terms for match/highlight purposes: lowercase,
+ * whitespace-split, drop short tokens (<3 chars) and stopwords. This is
+ * intentionally independent of whatever tokenization MiniSearch does
+ * internally for indexing/ranking — this list only drives snippet
+ * selection and highlighting, so it must not include stopwords regardless
+ * of what MiniSearch itself does under the hood.
+ */
+export declare function extractQueryTerms(query: string): string[];
+/** How matched terms should be marked in a generated snippet:
+ *  - "ansi": wrap with real ANSI bold escapes (\x1b[1m...\x1b[22m) — only
+ *    correct when writing to an interactive terminal.
+ *  - "plain": no markup at all. Used for --json output (a JSON consumer
+ *    wants raw data, not a display marker) and for non-TTY text output
+ *    (piped/redirected — a human doing that is likely about to grep/process
+ *    the output programmatically, so escape codes would just be noise, and
+ *    literal `**` clutter is exactly the bug being fixed). */
+export type HighlightMode = "ansi" | "plain";
+/** Context needed to resolve a chunk-local match position against the real,
+ * full source file rather than just the chunk's own text (fix #4 in the
+ * module doc-comment). `chunkStartLine` is the chunk's 1-based start line
+ * within the full file — already tracked per-chunk as
+ * `IndexedDocument.startLine`. */
+export interface FileContext {
+    fullContent: string;
+    chunkStartLine: number;
+}
+/**
+ * Build a human-readable snippet around the first occurrence of any of
+ * `terms` in `content`, then highlight those terms with word-boundary-safe
+ * bolding. Table rows are detected and handled specially so a match inside
+ * a markdown table doesn't produce a raw pipe-fragment slice.
+ *
+ * `fileContext`, when supplied, lets the rest of snippet assembly (most
+ * importantly table-header lookback) operate against the REAL file instead
+ * of just the chunk's own text — a chunk boundary can put a table's header
+ * outside the chunk that contains the matched row, even though the header
+ * is genuinely only a few lines away in the real file. The match itself is
+ * still located within `content` (the chunk actually matched by the
+ * search index); only the window used to build the surrounding snippet is
+ * widened to the full file, by converting the chunk-local match position
+ * to a file-absolute one via `chunkStartLine`.
+ */
+export declare function generateSnippet(content: string, terms: string[], highlightMode?: HighlightMode, fileContext?: FileContext): string;
+/** Minimal shape `buildSnippet` needs from an indexed chunk. */
+export interface SnippetDoc {
+    filePath: string;
+    content: string;
+    /** 1-based line number where this chunk starts in the source file. */
+    startLine: number;
+}
+/**
+ * Convenience wrapper around `generateSnippet` for the common call-site
+ * shape: an indexed chunk plus the repo root it lives under. Reads the full
+ * source file from disk so snippet generation can see the real document,
+ * not just the chunk (fix #4). Result counts per query are small (~10), so
+ * a handful of extra file reads at query time is the right tradeoff over
+ * silently degraded snippets.
+ *
+ * If `repoRoot` is omitted, or the file can't be read (moved/deleted since
+ * indexing, permission error, etc.), falls back to chunk-only snippet
+ * generation rather than failing the whole query.
+ */
+export declare function buildSnippet(doc: SnippetDoc, terms: string[], highlightMode: HighlightMode, repoRoot?: string): Promise<string>;
+//# sourceMappingURL=snippet.d.ts.map

package/dist/search/snippet.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"snippet.d.ts","sourceRoot":"","sources":["../../src/search/snippet.ts"],"names":[],"mappings":"AAwDA;;;;;;;GAOG;AACH,wBAAgB,iBAAiB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,EAAE,CAKzD;AAMD;;;;;;;8DAO8D;AAC9D,MAAM,MAAM,aAAa,GAAG,MAAM,GAAG,OAAO,CAAC;AAoO7C;;;;kCAIkC;AAClC,MAAM,WAAW,WAAW;IAC1B,WAAW,EAAE,MAAM,CAAC;IACpB,cAAc,EAAE,MAAM,CAAC;CACxB;AAED;;;;;;;;;;;;;;;GAeG;AACH,wBAAgB,eAAe,CAC7B,OAAO,EAAE,MAAM,EACf,KAAK,EAAE,MAAM,EAAE,EACf,aAAa,GAAE,aAAuB,EACtC,WAAW,CAAC,EAAE,WAAW,GACxB,MAAM,CA4BR;AAED,gEAAgE;AAChE,MAAM,WAAW,UAAU;IACzB,QAAQ,EAAE,MAAM,CAAC;IACjB,OAAO,EAAE,MAAM,CAAC;IAChB,sEAAsE;IACtE,SAAS,EAAE,MAAM,CAAC;CACnB;AAED;;;;;;;;;;;GAWG;AACH,wBAAsB,YAAY,CAChC,GAAG,EAAE,UAAU,EACf,KAAK,EAAE,MAAM,EAAE,EACf,aAAa,EAAE,aAAa,EAC5B,QAAQ,CAAC,EAAE,MAAM,GAChB,OAAO,CAAC,MAAM,CAAC,CAWjB"}

package/dist/search/snippet.js

@@ -0,0 +1,352 @@
+/**
+ * Shared query-term extraction and snippet/highlight logic used by both
+ * DocumentStore (BM25/keyword results) and SearchService (vector/hybrid
+ * results). Extracted so the two code paths can't drift and produce two
+ * different (and differently broken) renderings of the same kind of bug.
+ *
+ * Fixes several related readability bugs surfaced by real queries against
+ * this repo's own docs:
+ *
+ *  1. Highlighting used a bare substring match, so short common words like
+ *     "and" matched inside unrelated words ("H**and**le", "S**and**box").
+ *     Fixed by requiring a word boundary on whichever side of the term is
+ *     a word character.
+ *  2. Stopwords ("and", "to", ...) were treated as real query terms, so
+ *     they got matched/highlighted everywhere. Fixed by filtering a small
+ *     stopword list out of the extracted term list used for matching and
+ *     highlighting (MiniSearch's own internal tokenization for indexing/
+ *     ranking is untouched — this only affects the terms list used for
+ *     snippet selection and highlighting).
+ *  3. Snippets were raw character slices with no regard for word/sentence
+ *     boundaries or markdown table structure, so a match inside a table
+ *     row produced meaningless pipe-delimited fragments. Fixed by
+ *     expanding to word boundaries, preferring sentence boundaries within
+ *     a small extra budget, and special-casing table rows.
+ *  4. Table-header pairing (`findTableHeader`) only ever saw a single
+ *     CHUNK's text (`IndexedDocument.content`), not the full source file.
+ *     When a matched table row happened to sit near the start of its
+ *     chunk, the real header — genuinely just a handful of lines up in the
+ *     actual document — fell outside the chunk entirely, so the lookback
+ *     silently failed and every such row degraded to the raw `[table row]`
+ *     fallback even though a perfectly good header existed in the file.
+ *     Fixed by giving snippet generation an optional view of the FULL file
+ *     content plus the chunk's file-absolute start line, so the header
+ *     lookback (and the rest of snippet assembly, for consistency) walks
+ *     the real document rather than just the chunk. See `buildSnippet`.
+ *  5. Pre-existing markdown bold already present in the SOURCE content
+ *     (e.g. `**Schedule**` in a table cell) printed as literal `**`
+ *     clutter — `highlightTerms` only ever wrapped query-MATCHED terms,
+ *     never touching bold that was already there. Fixed by a universal
+ *     final pass (`applyMarkdownBold`) that runs LAST, after highlighting
+ *     and table formatting, and converts every remaining `**...**` pair in
+ *     the assembled snippet to real ANSI bold (TTY) or strips it entirely
+ *     (JSON / non-TTY).
+ */
+import { readFile } from "node:fs/promises";
+import { join } from "node:path";
+/** Small, deliberately conservative English stopword list — short, very
+ * common function words only. Not an aggressive/exhaustive list; the goal
+ * is just to stop them from being treated as meaningful match terms. */
+const STOPWORDS = new Set([
+    "a", "an", "and", "the", "to", "of", "in", "on", "for", "is", "are",
+    "was", "were", "be", "been", "this", "that", "with", "as", "at", "by",
+    "from", "or",
+]);
+/**
+ * Extract meaningful query terms for match/highlight purposes: lowercase,
+ * whitespace-split, drop short tokens (<3 chars) and stopwords. This is
+ * intentionally independent of whatever tokenization MiniSearch does
+ * internally for indexing/ranking — this list only drives snippet
+ * selection and highlighting, so it must not include stopwords regardless
+ * of what MiniSearch itself does under the hood.
+ */
+export function extractQueryTerms(query) {
+    return query
+        .toLowerCase()
+        .split(/\s+/)
+        .filter((t) => t.length >= 3 && !STOPWORDS.has(t));
+}
+function escapeRegex(str) {
+    return str.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
+}
+const ANSI_BOLD_ON = "\x1b[1m";
+const ANSI_BOLD_OFF = "\x1b[22m";
+/** Universal final pass: convert every remaining `**...**` pair in a fully
+ * assembled snippet — whether it's markdown bold that was already present
+ * in the source content, or bold left untouched by `highlightTerms` (which
+ * only ever wraps query-MATCHED terms) — according to the same TTY/non-TTY
+ * rule as everything else. Must run LAST, after highlighting and table
+ * formatting, so nothing downstream re-introduces or misses a `**` pair. */
+function applyMarkdownBold(text, mode) {
+    const paired = mode === "ansi"
+        ? text.replace(/\*\*(.+?)\*\*/g, `${ANSI_BOLD_ON}$1${ANSI_BOLD_OFF}`)
+        : text.replace(/\*\*(.+?)\*\*/g, "$1");
+    // A snippet window can cut through a bold span, leaving one lone `**`
+    // marker with no partner (see e.g. the 220-char prose window truncating
+    // mid-`**word**`). The paired pass above only touches complete pairs, so
+    // strip whatever's left over rather than leave literal asterisks in
+    // either mode — there's no meaningful bolding left to apply to it.
+    return paired.replace(/\*\*/g, "");
+}
+/** Highlight `terms` inside `text`, requiring a word boundary on whichever
+ * side of a term is itself a word character (so "and" won't match inside
+ * "handle", but a term like "project_id" that starts/ends with `_` still
+ * matches sensibly). In "plain" mode this is a no-op (no markers added). */
+function highlightTerms(text, terms, mode) {
+    if (mode !== "ansi")
+        return text;
+    let result = text;
+    for (const term of terms) {
+        if (!term)
+            continue;
+        const escaped = escapeRegex(term);
+        const leadingBoundary = /^\w/.test(term) ? "\\b" : "";
+        const trailingBoundary = /\w$/.test(term) ? "\\b" : "";
+        const regex = new RegExp(`${leadingBoundary}(${escaped})${trailingBoundary}`, "gi");
+        result = result.replace(regex, `${ANSI_BOLD_ON}$1${ANSI_BOLD_OFF}`);
+    }
+    return result;
+}
+const SNIPPET_LENGTH = 220;
+const SENTENCE_EXPAND_BUDGET = 80;
+function findFirstMatchIndex(content, terms) {
+    let firstMatch = -1;
+    const lower = content.toLowerCase();
+    for (const term of terms) {
+        const index = lower.indexOf(term.toLowerCase());
+        if (index !== -1 && (firstMatch === -1 || index < firstMatch))
+            firstMatch = index;
+    }
+    return firstMatch;
+}
+function lineBounds(content, pos) {
+    const lineStart = content.lastIndexOf("\n", Math.max(0, pos - 1)) + 1;
+    let lineEnd = content.indexOf("\n", pos);
+    if (lineEnd === -1)
+        lineEnd = content.length;
+    return { lineStart, lineEnd };
+}
+/** A line "is a table row" if it has 2+ pipe characters — enough to
+ * produce multiple cells when split, which is what makes a naive
+ * character-slice truncation unreadable. */
+function isTableLine(line) {
+    return (line.match(/\|/g)?.length ?? 0) >= 2;
+}
+function splitTableRow(line) {
+    let l = line.trim();
+    if (l.startsWith("|"))
+        l = l.slice(1);
+    if (l.endsWith("|"))
+        l = l.slice(0, -1);
+    return l.split("|").map((c) => c.trim());
+}
+const TABLE_SEPARATOR_RE = /^\|?\s*:?-{2,}:?\s*(\|\s*:?-{2,}:?\s*)+\|?$/;
+/** Walk upward from a table row looking for the `|---|---|` separator
+ * line, and return the header row immediately above it, if found within
+ * a reasonable number of lines (bounded so we don't walk the whole file). */
+function findTableHeader(content, lineStart) {
+    const before = content.slice(0, Math.max(0, lineStart - 1)).split("\n");
+    const scanFrom = before.length - 1;
+    for (let i = scanFrom; i >= 0 && i >= scanFrom - 30; i--) {
+        const l = before[i].trim();
+        if (TABLE_SEPARATOR_RE.test(l)) {
+            const headerLine = before[i - 1];
+            return headerLine ? splitTableRow(headerLine) : null;
+        }
+    }
+    return null;
+}
+/** Look a bounded number of lines before/after a table block for a
+ * non-table, non-heading, non-empty line to use as a fallback snippet.
+ * Rejects lines that are just leftover table punctuation — a lone "|"
+ * left over at a chunk boundary (which can have too few pipes to trip
+ * `isTableLine`), a run of "-"/":" characters, etc. — via both a
+ * minimum-word-content check AND a blanket "no pipe at all" check, since
+ * a chunk boundary can split a table row so only its last fragment (e.g.
+ * `enabled. |`) lands in this chunk, and that fragment alone has just one
+ * pipe but is exactly as meaningless as a full broken row. */
+function looksLikeRealProse(line) {
+    return /[A-Za-z]{3,}/.test(line) && !line.includes("|");
+}
+function findNearestProseLine(content, lineStart, lineEnd) {
+    const before = content.slice(0, lineStart).split("\n");
+    for (let i = before.length - 1; i >= 0 && i >= before.length - 15; i--) {
+        const l = before[i].trim();
+        if (l && looksLikeRealProse(l) && !TABLE_SEPARATOR_RE.test(l) && !/^#{1,6}\s/.test(l)) {
+            return l;
+        }
+    }
+    const after = content.slice(lineEnd).split("\n");
+    for (let i = 0; i < after.length && i < 15; i++) {
+        const l = after[i].trim();
+        if (l && looksLikeRealProse(l) && !TABLE_SEPARATOR_RE.test(l) && !/^#{1,6}\s/.test(l)) {
+            return l;
+        }
+    }
+    return null;
+}
+function tableSnippet(content, lineStart, lineEnd) {
+    const line = content.slice(lineStart, lineEnd);
+    const cells = splitTableRow(line);
+    const header = findTableHeader(content, lineStart);
+    if (header && header.length === cells.length) {
+        const paired = cells
+            .map((cell, i) => {
+            const h = (header[i] ?? "").trim();
+            return h ? `${h}: ${cell}` : cell;
+        })
+            .filter((p) => p.trim() !== "");
+        if (paired.length > 0)
+            return paired.join(" | ");
+    }
+    const prose = findNearestProseLine(content, lineStart, lineEnd);
+    if (prose)
+        return prose.length > 200 ? prose.slice(0, 200) + "..." : prose;
+    // Last resort: label it plainly rather than presenting broken fragments
+    // as if they were prose, and collapse repeated pipe padding.
+    const compact = line.trim().replace(/\s*\|\s*/g, " | ").replace(/^\s*\|\s*/, "").trim();
+    return `[table row] ${compact}`;
+}
+/** Move `start` left until the character immediately before it is
+ * whitespace (or we hit the beginning of the string) — never cuts a word
+ * in half. */
+function expandStartToWordBoundary(content, start) {
+    let s = start;
+    while (s > 0 && !/\s/.test(content[s - 1]))
+        s--;
+    return s;
+}
+/** Move `end` right until the character at it is whitespace (or we hit
+ * the end of the string) — never cuts a word in half. */
+function expandEndToWordBoundary(content, end) {
+    let e = end;
+    while (e < content.length && !/\s/.test(content[e]))
+        e++;
+    return e;
+}
+/** Try to extend `end` forward, within `budget` extra characters, to the
+ * next sentence-terminating punctuation followed by whitespace/EOF. */
+function extendToSentenceEnd(content, end, budget) {
+    const zone = content.slice(end, Math.min(content.length, end + budget));
+    const match = zone.match(/[.!?](\s|$)/);
+    if (match && match.index !== undefined)
+        return end + match.index + 1;
+    return end;
+}
+/** Try to pull `start` backward, within `budget` extra characters, to just
+ * after the nearest preceding sentence-terminating punctuation. */
+function extendToSentenceStart(content, start, budget) {
+    const zoneStart = Math.max(0, start - budget);
+    const zone = content.slice(zoneStart, start);
+    const matches = [...zone.matchAll(/[.!?]\s+/g)];
+    if (matches.length > 0) {
+        const last = matches[matches.length - 1];
+        return zoneStart + last.index + last[0].length;
+    }
+    return start;
+}
+function proseSnippet(content, matchPos) {
+    let start = Math.max(0, matchPos - Math.floor(SNIPPET_LENGTH / 3));
+    let end = Math.min(content.length, start + SNIPPET_LENGTH);
+    start = expandStartToWordBoundary(content, start);
+    end = expandEndToWordBoundary(content, end);
+    end = extendToSentenceEnd(content, end, SENTENCE_EXPAND_BUDGET);
+    start = extendToSentenceStart(content, start, SENTENCE_EXPAND_BUDGET);
+    let snippet = content.slice(start, end).replace(/\s+/g, " ").trim();
+    if (start > 0)
+        snippet = "..." + snippet;
+    if (end < content.length)
+        snippet = snippet + "...";
+    return snippet;
+}
+/** 0-based index of the line containing character offset `pos` within
+ * `content` (i.e. how many newlines precede `pos`). */
+function lineIndexAt(content, pos) {
+    let count = 0;
+    const bound = Math.min(pos, content.length);
+    for (let i = 0; i < bound; i++) {
+        if (content[i] === "\n")
+            count++;
+    }
+    return count;
+}
+/** Character offset where 0-based line `lineIndex` starts within `content`. */
+function offsetOfLineIndex(content, lineIndex) {
+    if (lineIndex <= 0)
+        return 0;
+    let seen = 0;
+    for (let i = 0; i < content.length; i++) {
+        if (content[i] === "\n") {
+            seen++;
+            if (seen === lineIndex)
+                return i + 1;
+        }
+    }
+    return content.length;
+}
+/**
+ * Build a human-readable snippet around the first occurrence of any of
+ * `terms` in `content`, then highlight those terms with word-boundary-safe
+ * bolding. Table rows are detected and handled specially so a match inside
+ * a markdown table doesn't produce a raw pipe-fragment slice.
+ *
+ * `fileContext`, when supplied, lets the rest of snippet assembly (most
+ * importantly table-header lookback) operate against the REAL file instead
+ * of just the chunk's own text — a chunk boundary can put a table's header
+ * outside the chunk that contains the matched row, even though the header
+ * is genuinely only a few lines away in the real file. The match itself is
+ * still located within `content` (the chunk actually matched by the
+ * search index); only the window used to build the surrounding snippet is
+ * widened to the full file, by converting the chunk-local match position
+ * to a file-absolute one via `chunkStartLine`.
+ */
+export function generateSnippet(content, terms, highlightMode = "plain", fileContext) {
+    const firstMatch = findFirstMatchIndex(content, terms);
+    const matchPosInChunk = firstMatch === -1 ? 0 : firstMatch;
+    let workingContent = content;
+    let matchPos = matchPosInChunk;
+    if (fileContext) {
+        const chunkLineIndex = lineIndexAt(content, matchPosInChunk);
+        const columnInLine = matchPosInChunk - lineBounds(content, matchPosInChunk).lineStart;
+        const fileLineIndex = fileContext.chunkStartLine - 1 + chunkLineIndex;
+        const fileLineOffset = offsetOfLineIndex(fileContext.fullContent, fileLineIndex);
+        workingContent = fileContext.fullContent;
+        matchPos = Math.min(fileLineOffset + columnInLine, workingContent.length);
+    }
+    const { lineStart, lineEnd } = lineBounds(workingContent, matchPos);
+    const line = workingContent.slice(lineStart, lineEnd);
+    const snippet = isTableLine(line)
+        ? tableSnippet(workingContent, lineStart, lineEnd)
+        : proseSnippet(workingContent, matchPos);
+    const highlighted = highlightTerms(snippet, terms, highlightMode);
+    // Universal final pass — must run LAST, after highlighting and table
+    // formatting, so it catches pre-existing source markdown bold as well as
+    // anything table formatting left untouched (fix #5).
+    return applyMarkdownBold(highlighted, highlightMode);
+}
+/**
+ * Convenience wrapper around `generateSnippet` for the common call-site
+ * shape: an indexed chunk plus the repo root it lives under. Reads the full
+ * source file from disk so snippet generation can see the real document,
+ * not just the chunk (fix #4). Result counts per query are small (~10), so
+ * a handful of extra file reads at query time is the right tradeoff over
+ * silently degraded snippets.
+ *
+ * If `repoRoot` is omitted, or the file can't be read (moved/deleted since
+ * indexing, permission error, etc.), falls back to chunk-only snippet
+ * generation rather than failing the whole query.
+ */
+export async function buildSnippet(doc, terms, highlightMode, repoRoot) {
+    let fileContext;
+    if (repoRoot) {
+        try {
+            const fullContent = await readFile(join(repoRoot, doc.filePath), "utf-8");
+            fileContext = { fullContent, chunkStartLine: doc.startLine };
+        }
+        catch {
+            fileContext = undefined;
+        }
+    }
+    return generateSnippet(doc.content, terms, highlightMode, fileContext);
+}
+//# sourceMappingURL=snippet.js.map

package/dist/search/snippet.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"snippet.js","sourceRoot":"","sources":["../../src/search/snippet.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2CG;AACH,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAEjC;;wEAEwE;AACxE,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC;IACxB,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK;IACnE,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI;IACrE,MAAM,EAAE,IAAI;CACb,CAAC,CAAC;AAEH;;;;;;;GAOG;AACH,MAAM,UAAU,iBAAiB,CAAC,KAAa;IAC7C,OAAO,KAAK;SACT,WAAW,EAAE;SACb,KAAK,CAAC,KAAK,CAAC;SACZ,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;AACvD,CAAC;AAED,SAAS,WAAW,CAAC,GAAW;IAC9B,OAAO,GAAG,CAAC,OAAO,CAAC,qBAAqB,EAAE,MAAM,CAAC,CAAC;AACpD,CAAC;AAYD,MAAM,YAAY,GAAG,SAAS,CAAC;AAC/B,MAAM,aAAa,GAAG,UAAU,CAAC;AAEjC;;;;;4EAK4E;AAC5E,SAAS,iBAAiB,CAAC,IAAY,EAAE,IAAmB;IAC1D,MAAM,MAAM,GACV,IAAI,KAAK,MAAM;QACb,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,gBAAgB,EAAE,GAAG,YAAY,KAAK,aAAa,EAAE,CAAC;QACrE,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,gBAAgB,EAAE,IAAI,CAAC,CAAC;IAC3C,sEAAsE;IACtE,wEAAwE;IACxE,yEAAyE;IACzE,oEAAoE;IACpE,mEAAmE;IACnE,OAAO,MAAM,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;AACrC,CAAC;AAED;;;4EAG4E;AAC5E,SAAS,cAAc,CAAC,IAAY,EAAE,KAAe,EAAE,IAAmB;IACxE,IAAI,IAAI,KAAK,MAAM;QAAE,OAAO,IAAI,CAAC;IACjC,IAAI,MAAM,GAAG,IAAI,CAAC;IAClB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,MAAM,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC;QAClC,MAAM,eAAe,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;QACtD,MAAM,gBAAgB,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;QACvD,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,GAAG,eAAe,IAAI,OAAO,IAAI,gBAAgB,EAAE,EAAE,IAAI,CAAC,CAAC;QACpF,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,YAAY,KAAK,aAAa,EAAE,CAAC,CAAC;IACtE,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,MAAM,cAAc,GAAG,GAAG,CAAC;AAC3B,MAAM,sBAAsB,GAAG,EAAE,CAAC;AAElC,SAAS,mBAAmB,CAAC,OAAe,EAAE,KAAe;IAC3D,IAAI,UAAU,GAAG,CAAC,CAAC,CAAC;IACpB,MAAM,KAAK,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC;IACpC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;QAChD,IAAI,KAAK,KAAK,CAAC,CAAC,IAAI,CAAC,UAAU,KAAK,CAAC,CAAC,IAAI,KAAK,GAAG,UAAU,CAAC;YAAE,UAAU,GAAG,KAAK,CAAC;IACpF,CAAC;IACD,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,SAAS,UAAU,CAAC,OAAe,EAAE,GAAW;IAC9C,MAAM,SAAS,GAAG,OAAO,CAAC,WAAW,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;IACtE,IAAI,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IACzC,IAAI,OAAO,KAAK,CAAC,CAAC;QAAE,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC;IAC7C,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC;AAChC,CAAC;AAED;;4CAE4C;AAC5C,SAAS,WAAW,CAAC,IAAY;IAC/B,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,MAAM,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;AAC/C,CAAC;AAED,SAAS,aAAa,CAAC,IAAY;IACjC,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IACpB,IAAI,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACtC,IAAI,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACxC,OAAO,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;AAC3C,CAAC;AAED,MAAM,kBAAkB,GAAG,6CAA6C,CAAC;AAEzE;;6EAE6E;AAC7E,SAAS,eAAe,CAAC,OAAe,EAAE,SAAiB;IACzD,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACxE,MAAM,QAAQ,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC;IACnC,KAAK,IAAI,CAAC,GAAG,QAAQ,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,QAAQ,GAAG,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC;QACzD,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAC5B,IAAI,kBAAkB,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;YAC/B,MAAM,UAAU,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YACjC,OAAO,UAAU,CAAC,CAAC,CAAC,aAAa,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;QACvD,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;;;;;;;8DAQ8D;AAC9D,SAAS,kBAAkB,CAAC,IAAY;IACtC,OAAO,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC;AAC1D,CAAC;AAED,SAAS,oBAAoB,CAAC,OAAe,EAAE,SAAiB,EAAE,OAAe;IAC/E,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACvD,KAAK,IAAI,CAAC,GAAG,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,MAAM,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC;QACvE,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAC5B,IAAI,CAAC,IAAI,kBAAkB,CAAC,CAAC,CAAC,IAAI,CAAC,kBAAkB,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;YACtF,OAAO,CAAC,CAAC;QACX,CAAC;IACH,CAAC;IACD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACjD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,IAAI,CAAC,GAAG,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC;QAChD,MAAM,CAAC,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAC3B,IAAI,CAAC,IAAI,kBAAkB,CAAC,CAAC,CAAC,IAAI,CAAC,kBAAkB,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;YACtF,OAAO,CAAC,CAAC;QACX,CAAC;IACH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,YAAY,CAAC,OAAe,EAAE,SAAiB,EAAE,OAAe;IACvE,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;IAC/C,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,MAAM,GAAG,eAAe,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC;IAEnD,IAAI,MAAM,IAAI,MAAM,CAAC,MAAM,KAAK,KAAK,CAAC,MAAM,EAAE,CAAC;QAC7C,MAAM,MAAM,GAAG,KAAK;aACjB,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,EAAE;YACf,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YACnC,OAAO,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;QACpC,CAAC,CAAC;aACD,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;QAClC,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC;YAAE,OAAO,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACnD,CAAC;IAED,MAAM,KAAK,GAAG,oBAAoB,CAAC,OAAO,EAAE,SAAS,EAAE,OAAO,CAAC,CAAC;IAChE,IAAI,KAAK;QAAE,OAAO,KAAK,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC;IAE3E,wEAAwE;IACxE,6DAA6D;IAC7D,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,WAAW,EAAE,KAAK,CAAC,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IACxF,OAAO,eAAe,OAAO,EAAE,CAAC;AAClC,CAAC;AAED;;cAEc;AACd,SAAS,yBAAyB,CAAC,OAAe,EAAE,KAAa;IAC/D,IAAI,CAAC,GAAG,KAAK,CAAC;IACd,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC;QAAE,CAAC,EAAE,CAAC;IACjD,OAAO,CAAC,CAAC;AACX,CAAC;AAED;yDACyD;AACzD,SAAS,uBAAuB,CAAC,OAAe,EAAE,GAAW;IAC3D,IAAI,CAAC,GAAG,GAAG,CAAC;IACZ,OAAO,CAAC,GAAG,OAAO,CAAC,MAAM,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAE,CAAC;QAAE,CAAC,EAAE,CAAC;IAC1D,OAAO,CAAC,CAAC;AACX,CAAC;AAED;uEACuE;AACvE,SAAS,mBAAmB,CAAC,OAAe,EAAE,GAAW,EAAE,MAAc;IACvE,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,CAAC,GAAG,EAAE,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC;IACxE,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,aAAa,CAAC,CAAC;IACxC,IAAI,KAAK,IAAI,KAAK,CAAC,KAAK,KAAK,SAAS;QAAE,OAAO,GAAG,GAAG,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC;IACrE,OAAO,GAAG,CAAC;AACb,CAAC;AAED;mEACmE;AACnE,SAAS,qBAAqB,CAAC,OAAe,EAAE,KAAa,EAAE,MAAc;IAC3E,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC,CAAC;IAC9C,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC;IAC7C,MAAM,OAAO,GAAG,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC,CAAC,CAAC;IAChD,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACvB,MAAM,IAAI,GAAG,OAAO,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAE,CAAC;QAC1C,OAAO,SAAS,GAAG,IAAI,CAAC,KAAM,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAClD,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,YAAY,CAAC,OAAe,EAAE,QAAgB;IACrD,IAAI,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,GAAG,IAAI,CAAC,KAAK,CAAC,cAAc,GAAG,CAAC,CAAC,CAAC,CAAC;IACnE,IAAI,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,MAAM,EAAE,KAAK,GAAG,cAAc,CAAC,CAAC;IAE3D,KAAK,GAAG,yBAAyB,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;IAClD,GAAG,GAAG,uBAAuB,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC;IAE5C,GAAG,GAAG,mBAAmB,CAAC,OAAO,EAAE,GAAG,EAAE,sBAAsB,CAAC,CAAC;IAChE,KAAK,GAAG,qBAAqB,CAAC,OAAO,EAAE,KAAK,EAAE,sBAAsB,CAAC,CAAC;IAEtE,IAAI,OAAO,GAAG,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IACpE,IAAI,KAAK,GAAG,CAAC;QAAE,OAAO,GAAG,KAAK,GAAG,OAAO,CAAC;IACzC,IAAI,GAAG,GAAG,OAAO,CAAC,MAAM;QAAE,OAAO,GAAG,OAAO,GAAG,KAAK,CAAC;IACpD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;uDACuD;AACvD,SAAS,WAAW,CAAC,OAAe,EAAE,GAAW;IAC/C,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC;IAC5C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,EAAE,CAAC,EAAE,EAAE,CAAC;QAC/B,IAAI,OAAO,CAAC,CAAC,CAAC,KAAK,IAAI;YAAE,KAAK,EAAE,CAAC;IACnC,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,+EAA+E;AAC/E,SAAS,iBAAiB,CAAC,OAAe,EAAE,SAAiB;IAC3D,IAAI,SAAS,IAAI,CAAC;QAAE,OAAO,CAAC,CAAC;IAC7B,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,IAAI,OAAO,CAAC,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YACxB,IAAI,EAAE,CAAC;YACP,IAAI,IAAI,KAAK,SAAS;gBAAE,OAAO,CAAC,GAAG,CAAC,CAAC;QACvC,CAAC;IACH,CAAC;IACD,OAAO,OAAO,CAAC,MAAM,CAAC;AACxB,CAAC;AAYD;;;;;;;;;;;;;;;GAeG;AACH,MAAM,UAAU,eAAe,CAC7B,OAAe,EACf,KAAe,EACf,gBAA+B,OAAO,EACtC,WAAyB;IAEzB,MAAM,UAAU,GAAG,mBAAmB,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;IACvD,MAAM,eAAe,GAAG,UAAU,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC;IAE3D,IAAI,cAAc,GAAG,OAAO,CAAC;IAC7B,IAAI,QAAQ,GAAG,eAAe,CAAC;IAE/B,IAAI,WAAW,EAAE,CAAC;QAChB,MAAM,cAAc,GAAG,WAAW,CAAC,OAAO,EAAE,eAAe,CAAC,CAAC;QAC7D,MAAM,YAAY,GAAG,eAAe,GAAG,UAAU,CAAC,OAAO,EAAE,eAAe,CAAC,CAAC,SAAS,CAAC;QACtF,MAAM,aAAa,GAAG,WAAW,CAAC,cAAc,GAAG,CAAC,GAAG,cAAc,CAAC;QACtE,MAAM,cAAc,GAAG,iBAAiB,CAAC,WAAW,CAAC,WAAW,EAAE,aAAa,CAAC,CAAC;QACjF,cAAc,GAAG,WAAW,CAAC,WAAW,CAAC;QACzC,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,cAAc,GAAG,YAAY,EAAE,cAAc,CAAC,MAAM,CAAC,CAAC;IAC5E,CAAC;IAED,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,GAAG,UAAU,CAAC,cAAc,EAAE,QAAQ,CAAC,CAAC;IACpE,MAAM,IAAI,GAAG,cAAc,CAAC,KAAK,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;IAEtD,MAAM,OAAO,GAAG,WAAW,CAAC,IAAI,CAAC;QAC/B,CAAC,CAAC,YAAY,CAAC,cAAc,EAAE,SAAS,EAAE,OAAO,CAAC;QAClD,CAAC,CAAC,YAAY,CAAC,cAAc,EAAE,QAAQ,CAAC,CAAC;IAE3C,MAAM,WAAW,GAAG,cAAc,CAAC,OAAO,EAAE,KAAK,EAAE,aAAa,CAAC,CAAC;IAClE,qEAAqE;IACrE,yEAAyE;IACzE,qDAAqD;IACrD,OAAO,iBAAiB,CAAC,WAAW,EAAE,aAAa,CAAC,CAAC;AACvD,CAAC;AAUD;;;;;;;;;;;GAWG;AACH,MAAM,CAAC,KAAK,UAAU,YAAY,CAChC,GAAe,EACf,KAAe,EACf,aAA4B,EAC5B,QAAiB;IAEjB,IAAI,WAAoC,CAAC;IACzC,IAAI,QAAQ,EAAE,CAAC;QACb,IAAI,CAAC;YACH,MAAM,WAAW,GAAG,MAAM,QAAQ,CAAC,IAAI,CAAC,QAAQ,EAAE,GAAG,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC,CAAC;YAC1E,WAAW,GAAG,EAAE,WAAW,EAAE,cAAc,EAAE,GAAG,CAAC,SAAS,EAAE,CAAC;QAC/D,CAAC;QAAC,MAAM,CAAC;YACP,WAAW,GAAG,SAAS,CAAC;QAC1B,CAAC;IACH,CAAC;IACD,OAAO,eAAe,CAAC,GAAG,CAAC,OAAO,EAAE,KAAK,EAAE,aAAa,EAAE,WAAW,CAAC,CAAC;AACzE,CAAC"}

package/dist/storage/document-store.d.ts

@@ -0,0 +1,72 @@
+/**
+ * Adapted from obsidian-qmd (MIT): https://github.com/thirteen37/obsidian-qmd
+ * Source: src/storage/DocumentStore.ts — the original had no Obsidian
+ * dependency at all (pure MiniSearch wrapper). Adjustments: added
+ * `startLine` to the indexed schema for path:line output, added a `scope`
+ * pre-filter to `search()`, and dropped the Obsidian-flavoured
+ * `pluginBasePath` option that was already unused upstream.
+ *
+ * DocumentStore provides full-text search using MiniSearch.
+ * Implements proper BM25 ranking with inverted index.
+ */
+import MiniSearch from "minisearch";
+import type { Scope } from "../types.js";
+import type { HighlightMode } from "../search/snippet.js";
+export interface IndexedDocument {
+    /** Unique identifier for this document/chunk */
+    id: string;
+    /** Path to the source file, relative to the repo root */
+    filePath: string;
+    /** The text content */
+    content: string;
+    /** Chunk index within the file (0 for single-chunk files) */
+    chunkIndex: number;
+    /** Last modification time of the source file (epoch ms) */
+    mtimeMs: number;
+    /** 1-based line number where this chunk starts in the source file */
+    startLine: number;
+    /** Heading hierarchy at the start of this chunk */
+    headingBreadcrumb?: string[];
+}
+export interface SearchResult {
+    id: string;
+    filePath: string;
+    chunkIndex: number;
+    startLine: number;
+    score: number;
+    snippet: string;
+    headingBreadcrumb?: string[];
+    content: string;
+}
+export interface SearchOptions {
+    limit?: number;
+    scope?: Scope;
+    /** How matched terms should be marked in generated snippets. Defaults to
+     * "plain" (no markup) when omitted. */
+    highlight?: HighlightMode;
+    /** Repo root used to read each result's full source file from disk for
+     * snippet generation (so e.g. table-header pairing can see the real
+     * document, not just the matched chunk). When omitted, snippets fall
+     * back to chunk-only generation. */
+    repoRoot?: string;
+}
+export interface ExportedDocumentStore {
+    documents: IndexedDocument[];
+    index: ReturnType<MiniSearch<IndexedDocument>["toJSON"]>;
+}
+export declare class DocumentStore {
+    private documents;
+    private searchIndex;
+    private ready;
+    initialize(existingData?: ExportedDocumentStore): Promise<void>;
+    isReady(): boolean;
+    indexDocuments(docs: IndexedDocument[]): Promise<void>;
+    getDocument(id: string): Promise<IndexedDocument | null>;
+    deleteByFilePath(filePath: string): Promise<void>;
+    /** Search for documents using MiniSearch with BM25 scoring. */
+    search(query: string, options?: SearchOptions): Promise<SearchResult[]>;
+    getDocumentCount(): Promise<number>;
+    getAllFilePaths(): Promise<string[]>;
+    export(): ExportedDocumentStore;
+}
+//# sourceMappingURL=document-store.d.ts.map

package/dist/storage/document-store.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"document-store.d.ts","sourceRoot":"","sources":["../../src/storage/document-store.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AACH,OAAO,UAAU,MAAM,YAAY,CAAC;AAEpC,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAEzC,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,sBAAsB,CAAC;AAE1D,MAAM,WAAW,eAAe;IAC9B,gDAAgD;IAChD,EAAE,EAAE,MAAM,CAAC;IACX,yDAAyD;IACzD,QAAQ,EAAE,MAAM,CAAC;IACjB,uBAAuB;IACvB,OAAO,EAAE,MAAM,CAAC;IAChB,6DAA6D;IAC7D,UAAU,EAAE,MAAM,CAAC;IACnB,2DAA2D;IAC3D,OAAO,EAAE,MAAM,CAAC;IAChB,qEAAqE;IACrE,SAAS,EAAE,MAAM,CAAC;IAClB,mDAAmD;IACnD,iBAAiB,CAAC,EAAE,MAAM,EAAE,CAAC;CAC9B;AAED,MAAM,WAAW,YAAY;IAC3B,EAAE,EAAE,MAAM,CAAC;IACX,QAAQ,EAAE,MAAM,CAAC;IACjB,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,CAAC;IAClB,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,MAAM,CAAC;IAChB,iBAAiB,CAAC,EAAE,MAAM,EAAE,CAAC;IAC7B,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,MAAM,WAAW,aAAa;IAC5B,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,KAAK,CAAC,EAAE,KAAK,CAAC;IACd;2CACuC;IACvC,SAAS,CAAC,EAAE,aAAa,CAAC;IAC1B;;;wCAGoC;IACpC,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,WAAW,qBAAqB;IACpC,SAAS,EAAE,eAAe,EAAE,CAAC;IAC7B,KAAK,EAAE,UAAU,CAAC,UAAU,CAAC,eAAe,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC;CAC1D;AAYD,qBAAa,aAAa;IACxB,OAAO,CAAC,SAAS,CAA2C;IAC5D,OAAO,CAAC,WAAW,CAA4C;IAC/D,OAAO,CAAC,KAAK,CAAS;IAEhB,UAAU,CAAC,YAAY,CAAC,EAAE,qBAAqB,GAAG,OAAO,CAAC,IAAI,CAAC;IAyBrE,OAAO,IAAI,OAAO;IAIZ,cAAc,CAAC,IAAI,EAAE,eAAe,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC;IAiBtD,WAAW,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAC,eAAe,GAAG,IAAI,CAAC;IAIxD,gBAAgB,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;IAgBvD,+DAA+D;IACzD,MAAM,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,GAAE,aAAkB,GAAG,OAAO,CAAC,YAAY,EAAE,CAAC;IA0C3E,gBAAgB,IAAI,OAAO,CAAC,MAAM,CAAC;IAInC,eAAe,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;IAQ1C,MAAM,IAAI,qBAAqB;CAOhC"}