@amsterdamdatalabs/enact-docs-search 0.1.0

package/dist/indexing/chunker.d.ts

@@ -0,0 +1,52 @@
+/**
+ * Adapted from obsidian-qmd (MIT): https://github.com/thirteen37/obsidian-qmd
+ * Source: src/indexing/Chunker.ts — ported verbatim (no Obsidian dependency in
+ * the original); only the module doc-comment and export style were adjusted
+ * and `startLine` computation was added for path:line output.
+ *
+ * Document chunking for search indexing. Splits documents into overlapping
+ * chunks with awareness of markdown structure.
+ */
+export interface Chunk {
+    /** The text content of this chunk */
+    content: string;
+    /** Path to the source file */
+    filePath: string;
+    /** Zero-based index of this chunk within the document */
+    chunkIndex: number;
+    /** Start position in the original document (character offset) */
+    startPos: number;
+    /** End position in the original document (character offset) */
+    endPos: number;
+    /** 1-based line number in the original document where this chunk starts */
+    startLine: number;
+    /** Heading hierarchy at the start of this chunk (H1 to deepest level) */
+    headingBreadcrumb: string[];
+}
+export interface ChunkOptions {
+    /** Maximum chunk size in characters (default: 3200) */
+    maxChunkSize: number;
+    /** Overlap between chunks as a percentage (default: 0.15 = 15%) */
+    overlapPercent: number;
+}
+/**
+ * Chunker splits documents into overlapping chunks for search indexing.
+ * It attempts to break at natural boundaries (paragraphs, sentences, words)
+ * while respecting markdown structure (code blocks, headings).
+ */
+export declare class Chunker {
+    private options;
+    constructor(options?: Partial<ChunkOptions>);
+    /**
+     * Extract the heading hierarchy at a given position in the content.
+     */
+    getHeadingBreadcrumb(content: string, position: number): string[];
+    private findFirstNonHeadingContent;
+    /**
+     * Split a document into chunks.
+     */
+    chunk(content: string, filePath: string): Chunk[];
+    private findBreakPoint;
+    private findCodeBlockBoundary;
+}
+//# sourceMappingURL=chunker.d.ts.map

package/dist/indexing/chunker.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"chunker.d.ts","sourceRoot":"","sources":["../../src/indexing/chunker.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,MAAM,WAAW,KAAK;IACpB,qCAAqC;IACrC,OAAO,EAAE,MAAM,CAAC;IAChB,8BAA8B;IAC9B,QAAQ,EAAE,MAAM,CAAC;IACjB,yDAAyD;IACzD,UAAU,EAAE,MAAM,CAAC;IACnB,iEAAiE;IACjE,QAAQ,EAAE,MAAM,CAAC;IACjB,+DAA+D;IAC/D,MAAM,EAAE,MAAM,CAAC;IACf,2EAA2E;IAC3E,SAAS,EAAE,MAAM,CAAC;IAClB,yEAAyE;IACzE,iBAAiB,EAAE,MAAM,EAAE,CAAC;CAC7B;AAED,MAAM,WAAW,YAAY;IAC3B,uDAAuD;IACvD,YAAY,EAAE,MAAM,CAAC;IACrB,mEAAmE;IACnE,cAAc,EAAE,MAAM,CAAC;CACxB;AAOD;;;;GAIG;AACH,qBAAa,OAAO;IAClB,OAAO,CAAC,OAAO,CAAe;gBAElB,OAAO,GAAE,OAAO,CAAC,YAAY,CAAM;IAI/C;;OAEG;IACH,oBAAoB,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,MAAM,EAAE;IAqBjE,OAAO,CAAC,0BAA0B;IAiClC;;OAEG;IACH,KAAK,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,KAAK,EAAE;IAuEjD,OAAO,CAAC,cAAc;IA8CtB,OAAO,CAAC,qBAAqB;CAoB9B"}

package/dist/indexing/chunker.js

@@ -0,0 +1,196 @@
+/**
+ * Adapted from obsidian-qmd (MIT): https://github.com/thirteen37/obsidian-qmd
+ * Source: src/indexing/Chunker.ts — ported verbatim (no Obsidian dependency in
+ * the original); only the module doc-comment and export style were adjusted
+ * and `startLine` computation was added for path:line output.
+ *
+ * Document chunking for search indexing. Splits documents into overlapping
+ * chunks with awareness of markdown structure.
+ */
+const DEFAULT_OPTIONS = {
+    maxChunkSize: 3200,
+    overlapPercent: 0.15,
+};
+/**
+ * Chunker splits documents into overlapping chunks for search indexing.
+ * It attempts to break at natural boundaries (paragraphs, sentences, words)
+ * while respecting markdown structure (code blocks, headings).
+ */
+export class Chunker {
+    options;
+    constructor(options = {}) {
+        this.options = { ...DEFAULT_OPTIONS, ...options };
+    }
+    /**
+     * Extract the heading hierarchy at a given position in the content.
+     */
+    getHeadingBreadcrumb(content, position) {
+        const scanEnd = this.findFirstNonHeadingContent(content, position);
+        const contentToScan = content.slice(0, scanEnd);
+        const headingRegex = /^(#{1,6})\s+(.+)$/gm;
+        const headings = [null, null, null, null, null, null];
+        let match;
+        while ((match = headingRegex.exec(contentToScan)) !== null) {
+            const level = match[1].length;
+            const text = match[2].trim();
+            headings[level - 1] = text;
+            for (let i = level; i < 6; i++) {
+                headings[i] = null;
+            }
+        }
+        return headings.filter((h) => h !== null);
+    }
+    findFirstNonHeadingContent(content, position) {
+        let i = position;
+        const len = content.length;
+        while (i < len) {
+            while (i < len && /\s/.test(content[i])) {
+                i++;
+            }
+            if (i >= len)
+                break;
+            const lineEnd = content.indexOf("\n", i);
+            const line = content.slice(i, lineEnd === -1 ? len : lineEnd);
+            if (/^#{1,6}\s+/.test(line)) {
+                i = lineEnd === -1 ? len : lineEnd + 1;
+            }
+            else if (line.trim() === "") {
+                i = lineEnd === -1 ? len : lineEnd + 1;
+            }
+            else if (line.startsWith("---") && i === 0) {
+                const frontmatterEnd = content.indexOf("\n---", i + 3);
+                if (frontmatterEnd !== -1) {
+                    i = frontmatterEnd + 4;
+                }
+                else {
+                    break;
+                }
+            }
+            else {
+                break;
+            }
+        }
+        return i;
+    }
+    /**
+     * Split a document into chunks.
+     */
+    chunk(content, filePath) {
+        const trimmed = content.trim();
+        if (!trimmed) {
+            return [];
+        }
+        const normalized = content.replace(/\r\n/g, "\n").replace(/\r/g, "\n");
+        if (normalized.length <= this.options.maxChunkSize) {
+            return [
+                {
+                    content: normalized,
+                    filePath,
+                    chunkIndex: 0,
+                    startPos: 0,
+                    endPos: normalized.length,
+                    startLine: 1,
+                    headingBreadcrumb: this.getHeadingBreadcrumb(normalized, 0),
+                },
+            ];
+        }
+        const chunks = [];
+        const overlapSize = Math.floor(this.options.maxChunkSize * this.options.overlapPercent);
+        const minChunkSize = Math.floor(this.options.maxChunkSize * 0.2);
+        let position = 0;
+        let chunkIndex = 0;
+        while (position < normalized.length) {
+            const remaining = normalized.length - position;
+            if (remaining < minChunkSize && chunks.length > 0) {
+                const lastChunk = chunks[chunks.length - 1];
+                lastChunk.content = normalized.slice(lastChunk.startPos);
+                lastChunk.endPos = normalized.length;
+                break;
+            }
+            const chunkEnd = Math.min(position + this.options.maxChunkSize, normalized.length);
+            let actualEnd;
+            if (remaining <= this.options.maxChunkSize) {
+                actualEnd = normalized.length;
+            }
+            else {
+                actualEnd = this.findBreakPoint(normalized, position, chunkEnd);
+            }
+            const chunkContent = normalized.slice(position, actualEnd);
+            chunks.push({
+                content: chunkContent,
+                filePath,
+                chunkIndex,
+                startPos: position,
+                endPos: actualEnd,
+                startLine: lineNumberAt(normalized, position),
+                headingBreadcrumb: this.getHeadingBreadcrumb(normalized, position),
+            });
+            const advance = actualEnd - position - overlapSize;
+            position += Math.max(advance, 1);
+            chunkIndex++;
+            if (chunkIndex > normalized.length) {
+                break;
+            }
+        }
+        return chunks;
+    }
+    findBreakPoint(content, start, maxEnd) {
+        const searchRange = content.slice(start, maxEnd);
+        const codeBlockBoundary = this.findCodeBlockBoundary(content, start, maxEnd);
+        if (codeBlockBoundary !== null) {
+            return codeBlockBoundary;
+        }
+        const searchStart = Math.floor(searchRange.length * 0.7);
+        const lateRange = searchRange.slice(searchStart);
+        const paragraphMatch = lateRange.lastIndexOf("\n\n");
+        if (paragraphMatch !== -1) {
+            return start + searchStart + paragraphMatch + 2;
+        }
+        const sentenceRegex = /[.!?]\s+/g;
+        let lastSentenceEnd = -1;
+        let match;
+        const fullSearchRange = searchRange.slice(searchStart);
+        while ((match = sentenceRegex.exec(fullSearchRange)) !== null) {
+            lastSentenceEnd = match.index + match[0].length;
+        }
+        if (lastSentenceEnd !== -1) {
+            return start + searchStart + lastSentenceEnd;
+        }
+        const wordSearchStart = Math.floor(searchRange.length * 0.8);
+        const wordRange = searchRange.slice(wordSearchStart);
+        let lastSpace = Math.max(wordRange.lastIndexOf(" "), wordRange.lastIndexOf("\n"));
+        if (lastSpace !== -1) {
+            return start + wordSearchStart + lastSpace + 1;
+        }
+        lastSpace = Math.max(searchRange.lastIndexOf(" "), searchRange.lastIndexOf("\n"), searchRange.lastIndexOf("\t"));
+        if (lastSpace !== -1) {
+            return start + lastSpace + 1;
+        }
+        return maxEnd;
+    }
+    findCodeBlockBoundary(content, start, maxEnd) {
+        const beforeChunk = content.slice(0, start);
+        const chunk = content.slice(start, maxEnd);
+        const openMarkersBefore = (beforeChunk.match(/```/g) || []).length;
+        const markersInChunk = (chunk.match(/```/g) || []).length;
+        const startedInsideBlock = openMarkersBefore % 2 === 1;
+        if (startedInsideBlock || markersInChunk % 2 === 1) {
+            const afterMaxEnd = content.slice(maxEnd);
+            const closingMarker = afterMaxEnd.indexOf("```");
+            if (closingMarker !== -1 && closingMarker < this.options.maxChunkSize) {
+                return maxEnd + closingMarker + 3;
+            }
+        }
+        return null;
+    }
+}
+/** 1-based line number of a character offset within `content`. */
+function lineNumberAt(content, offset) {
+    let line = 1;
+    for (let i = 0; i < offset; i++) {
+        if (content[i] === "\n")
+            line++;
+    }
+    return line;
+}
+//# sourceMappingURL=chunker.js.map

package/dist/indexing/chunker.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"chunker.js","sourceRoot":"","sources":["../../src/indexing/chunker.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AA0BH,MAAM,eAAe,GAAiB;IACpC,YAAY,EAAE,IAAI;IAClB,cAAc,EAAE,IAAI;CACrB,CAAC;AAEF;;;;GAIG;AACH,MAAM,OAAO,OAAO;IACV,OAAO,CAAe;IAE9B,YAAY,UAAiC,EAAE;QAC7C,IAAI,CAAC,OAAO,GAAG,EAAE,GAAG,eAAe,EAAE,GAAG,OAAO,EAAE,CAAC;IACpD,CAAC;IAED;;OAEG;IACH,oBAAoB,CAAC,OAAe,EAAE,QAAgB;QACpD,MAAM,OAAO,GAAG,IAAI,CAAC,0BAA0B,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;QACnE,MAAM,aAAa,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;QAChD,MAAM,YAAY,GAAG,qBAAqB,CAAC;QAE3C,MAAM,QAAQ,GAAsB,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;QAEzE,IAAI,KAAK,CAAC;QACV,OAAO,CAAC,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YAC3D,MAAM,KAAK,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,MAAM,CAAC;YAC/B,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;YAE9B,QAAQ,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC;YAC3B,KAAK,IAAI,CAAC,GAAG,KAAK,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC/B,QAAQ,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC;YACrB,CAAC;QACH,CAAC;QAED,OAAO,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAe,EAAE,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC;IACzD,CAAC;IAEO,0BAA0B,CAAC,OAAe,EAAE,QAAgB;QAClE,IAAI,CAAC,GAAG,QAAQ,CAAC;QACjB,MAAM,GAAG,GAAG,OAAO,CAAC,MAAM,CAAC;QAE3B,OAAO,CAAC,GAAG,GAAG,EAAE,CAAC;YACf,OAAO,CAAC,GAAG,GAAG,IAAI,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAE,CAAC,EAAE,CAAC;gBACzC,CAAC,EAAE,CAAC;YACN,CAAC;YAED,IAAI,CAAC,IAAI,GAAG;gBAAE,MAAM;YAEpB,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;YACzC,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;YAE9D,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC5B,CAAC,GAAG,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC;YACzC,CAAC;iBAAM,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;gBAC9B,CAAC,GAAG,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC;YACzC,CAAC;iBAAM,IAAI,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC7C,MAAM,cAAc,GAAG,OAAO,CAAC,OAAO,CAAC,OAAO,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;gBACvD,IAAI,cAAc,KAAK,CAAC,CAAC,EAAE,CAAC;oBAC1B,CAAC,GAAG,cAAc,GAAG,CAAC,CAAC;gBACzB,CAAC;qBAAM,CAAC;oBACN,MAAM;gBACR,CAAC;YACH,CAAC;iBAAM,CAAC;gBACN,MAAM;YACR,CAAC;QACH,CAAC;QAED,OAAO,CAAC,CAAC;IACX,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,OAAe,EAAE,QAAgB;QACrC,MAAM,OAAO,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC;QAC/B,IAAI,CAAC,OAAO,EAAE,CAAC;YACb,OAAO,EAAE,CAAC;QACZ,CAAC;QAED,MAAM,UAAU,GAAG,OAAO,CAAC,OAAO,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;QAEvE,IAAI,UAAU,CAAC,MAAM,IAAI,IAAI,CAAC,OAAO,CAAC,YAAY,EAAE,CAAC;YACnD,OAAO;gBACL;oBACE,OAAO,EAAE,UAAU;oBACnB,QAAQ;oBACR,UAAU,EAAE,CAAC;oBACb,QAAQ,EAAE,CAAC;oBACX,MAAM,EAAE,UAAU,CAAC,MAAM;oBACzB,SAAS,EAAE,CAAC;oBACZ,iBAAiB,EAAE,IAAI,CAAC,oBAAoB,CAAC,UAAU,EAAE,CAAC,CAAC;iBAC5D;aACF,CAAC;QACJ,CAAC;QAED,MAAM,MAAM,GAAY,EAAE,CAAC;QAC3B,MAAM,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,YAAY,GAAG,IAAI,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC;QACxF,MAAM,YAAY,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,YAAY,GAAG,GAAG,CAAC,CAAC;QACjE,IAAI,QAAQ,GAAG,CAAC,CAAC;QACjB,IAAI,UAAU,GAAG,CAAC,CAAC;QAEnB,OAAO,QAAQ,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC;YACpC,MAAM,SAAS,GAAG,UAAU,CAAC,MAAM,GAAG,QAAQ,CAAC;YAE/C,IAAI,SAAS,GAAG,YAAY,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAClD,MAAM,SAAS,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAE,CAAC;gBAC7C,SAAS,CAAC,OAAO,GAAG,UAAU,CAAC,KAAK,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC;gBACzD,SAAS,CAAC,MAAM,GAAG,UAAU,CAAC,MAAM,CAAC;gBACrC,MAAM;YACR,CAAC;YAED,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,YAAY,EAAE,UAAU,CAAC,MAAM,CAAC,CAAC;YAEnF,IAAI,SAAiB,CAAC;YACtB,IAAI,SAAS,IAAI,IAAI,CAAC,OAAO,CAAC,YAAY,EAAE,CAAC;gBAC3C,SAAS,GAAG,UAAU,CAAC,MAAM,CAAC;YAChC,CAAC;iBAAM,CAAC;gBACN,SAAS,GAAG,IAAI,CAAC,cAAc,CAAC,UAAU,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAC;YAClE,CAAC;YAED,MAAM,YAAY,GAAG,UAAU,CAAC,KAAK,CAAC,QAAQ,EAAE,SAAS,CAAC,CAAC;YAE3D,MAAM,CAAC,IAAI,CAAC;gBACV,OAAO,EAAE,YAAY;gBACrB,QAAQ;gBACR,UAAU;gBACV,QAAQ,EAAE,QAAQ;gBAClB,MAAM,EAAE,SAAS;gBACjB,SAAS,EAAE,YAAY,CAAC,UAAU,EAAE,QAAQ,CAAC;gBAC7C,iBAAiB,EAAE,IAAI,CAAC,oBAAoB,CAAC,UAAU,EAAE,QAAQ,CAAC;aACnE,CAAC,CAAC;YAEH,MAAM,OAAO,GAAG,SAAS,GAAG,QAAQ,GAAG,WAAW,CAAC;YACnD,QAAQ,IAAI,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;YACjC,UAAU,EAAE,CAAC;YAEb,IAAI,UAAU,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC;gBACnC,MAAM;YACR,CAAC;QACH,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAEO,cAAc,CAAC,OAAe,EAAE,KAAa,EAAE,MAAc;QACnE,MAAM,WAAW,GAAG,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;QAEjD,MAAM,iBAAiB,GAAG,IAAI,CAAC,qBAAqB,CAAC,OAAO,EAAE,KAAK,EAAE,MAAM,CAAC,CAAC;QAC7E,IAAI,iBAAiB,KAAK,IAAI,EAAE,CAAC;YAC/B,OAAO,iBAAiB,CAAC;QAC3B,CAAC;QAED,MAAM,WAAW,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC;QACzD,MAAM,SAAS,GAAG,WAAW,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC;QAEjD,MAAM,cAAc,GAAG,SAAS,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC;QACrD,IAAI,cAAc,KAAK,CAAC,CAAC,EAAE,CAAC;YAC1B,OAAO,KAAK,GAAG,WAAW,GAAG,cAAc,GAAG,CAAC,CAAC;QAClD,CAAC;QAED,MAAM,aAAa,GAAG,WAAW,CAAC;QAClC,IAAI,eAAe,GAAG,CAAC,CAAC,CAAC;QACzB,IAAI,KAAK,CAAC;QACV,MAAM,eAAe,GAAG,WAAW,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC;QACvD,OAAO,CAAC,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YAC9D,eAAe,GAAG,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;QAClD,CAAC;QACD,IAAI,eAAe,KAAK,CAAC,CAAC,EAAE,CAAC;YAC3B,OAAO,KAAK,GAAG,WAAW,GAAG,eAAe,CAAC;QAC/C,CAAC;QAED,MAAM,eAAe,GAAG,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC;QAC7D,MAAM,SAAS,GAAG,WAAW,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC;QACrD,IAAI,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,WAAW,CAAC,GAAG,CAAC,EAAE,SAAS,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC;QAClF,IAAI,SAAS,KAAK,CAAC,CAAC,EAAE,CAAC;YACrB,OAAO,KAAK,GAAG,eAAe,GAAG,SAAS,GAAG,CAAC,CAAC;QACjD,CAAC;QAED,SAAS,GAAG,IAAI,CAAC,GAAG,CAClB,WAAW,CAAC,WAAW,CAAC,GAAG,CAAC,EAC5B,WAAW,CAAC,WAAW,CAAC,IAAI,CAAC,EAC7B,WAAW,CAAC,WAAW,CAAC,IAAI,CAAC,CAC9B,CAAC;QACF,IAAI,SAAS,KAAK,CAAC,CAAC,EAAE,CAAC;YACrB,OAAO,KAAK,GAAG,SAAS,GAAG,CAAC,CAAC;QAC/B,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAEO,qBAAqB,CAAC,OAAe,EAAE,KAAa,EAAE,MAAc;QAC1E,MAAM,WAAW,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;QAC5C,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;QAE3C,MAAM,iBAAiB,GAAG,CAAC,WAAW,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;QACnE,MAAM,cAAc,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;QAE1D,MAAM,kBAAkB,GAAG,iBAAiB,GAAG,CAAC,KAAK,CAAC,CAAC;QAEvD,IAAI,kBAAkB,IAAI,cAAc,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YACnD,MAAM,WAAW,GAAG,OAAO,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;YAC1C,MAAM,aAAa,GAAG,WAAW,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;YAEjD,IAAI,aAAa,KAAK,CAAC,CAAC,IAAI,aAAa,GAAG,IAAI,CAAC,OAAO,CAAC,YAAY,EAAE,CAAC;gBACtE,OAAO,MAAM,GAAG,aAAa,GAAG,CAAC,CAAC;YACpC,CAAC;QACH,CAAC;QAED,OAAO,IAAI,CAAC;IACd,CAAC;CACF;AAED,kEAAkE;AAClE,SAAS,YAAY,CAAC,OAAe,EAAE,MAAc;IACnD,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAChC,IAAI,OAAO,CAAC,CAAC,CAAC,KAAK,IAAI;YAAE,IAAI,EAAE,CAAC;IAClC,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC"}

package/dist/indexing/indexing-service.d.ts

@@ -0,0 +1,29 @@
+import type { Scope } from "../types.js";
+export interface IndexProgress {
+    phase: "scanning" | "text-indexing" | "embedding" | "done";
+    filesProcessed: number;
+    totalFiles: number;
+    currentFile?: string;
+}
+export type ProgressCallback = (progress: IndexProgress) => void;
+export interface IndexRunResult {
+    scope: Scope;
+    filesScanned: number;
+    filesChanged: number;
+    filesRemoved: number;
+    chunksIndexed: number;
+    embeddingsGenerated: number;
+    totalDurationMs: number;
+}
+export interface IndexOptions {
+    repoPath: string;
+    scope: Scope;
+    /** Skip vector embedding entirely (keyword-only index, much faster). */
+    skipEmbeddings?: boolean;
+    /** Ignore the manifest and rebuild every file in scope from scratch. */
+    force?: boolean;
+    exclude?: string[];
+    onProgress?: ProgressCallback;
+}
+export declare function runIndex(options: IndexOptions): Promise<IndexRunResult>;
+//# sourceMappingURL=indexing-service.d.ts.map

package/dist/indexing/indexing-service.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"indexing-service.d.ts","sourceRoot":"","sources":["../../src/indexing/indexing-service.ts"],"names":[],"mappings":"AAgCA,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAEzC,MAAM,WAAW,aAAa;IAC5B,KAAK,EAAE,UAAU,GAAG,eAAe,GAAG,WAAW,GAAG,MAAM,CAAC;IAC3D,cAAc,EAAE,MAAM,CAAC;IACvB,UAAU,EAAE,MAAM,CAAC;IACnB,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAED,MAAM,MAAM,gBAAgB,GAAG,CAAC,QAAQ,EAAE,aAAa,KAAK,IAAI,CAAC;AAEjE,MAAM,WAAW,cAAc;IAC7B,KAAK,EAAE,KAAK,CAAC;IACb,YAAY,EAAE,MAAM,CAAC;IACrB,YAAY,EAAE,MAAM,CAAC;IACrB,YAAY,EAAE,MAAM,CAAC;IACrB,aAAa,EAAE,MAAM,CAAC;IACtB,mBAAmB,EAAE,MAAM,CAAC;IAC5B,eAAe,EAAE,MAAM,CAAC;CACzB;AAED,MAAM,WAAW,YAAY;IAC3B,QAAQ,EAAE,MAAM,CAAC;IACjB,KAAK,EAAE,KAAK,CAAC;IACb,wEAAwE;IACxE,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB,wEAAwE;IACxE,KAAK,CAAC,EAAE,OAAO,CAAC;IAChB,OAAO,CAAC,EAAE,MAAM,EAAE,CAAC;IACnB,UAAU,CAAC,EAAE,gBAAgB,CAAC;CAC/B;AAKD,wBAAsB,QAAQ,CAAC,OAAO,EAAE,YAAY,GAAG,OAAO,CAAC,cAAc,CAAC,CAoJ7E"}

package/dist/indexing/indexing-service.js

@@ -0,0 +1,151 @@
+/**
+ * indexing-service.ts — orchestrates indexing for one target repo.
+ *
+ * Loosely modelled on obsidian-qmd's IndexingService.ts (scan -> chunk ->
+ * index text -> embed), but rewritten: the original drives itself off
+ * in-memory Obsidian vault state and doesn't persist between runs (Obsidian
+ * keeps the plugin process alive). This CLI process exits after every
+ * command, so this version is built around a manifest-driven incremental
+ * diff (new/changed/removed files by mtime) against a store that is loaded
+ * from and saved back to disk on every run, scoped per target repo.
+ */
+import { DocumentStore } from "../storage/document-store.js";
+import { VectorStore } from "../storage/vector-store.js";
+import { EmbeddingService, EMBEDDING_DIMENSIONS } from "../embeddings/embedding-service.js";
+import { Chunker } from "./chunker.js";
+import { listMarkdownFiles, readRepoFile } from "../adapter/repo-walker.js";
+import { ensureCacheDir, readManifest, writeManifest, emptyManifest, readDocumentStoreExport, writeDocumentStoreExport, readVectorSidecar, writeVectorSidecar, vectorIndexExists, repoCacheDir, paths, EMBEDDING_MODEL_ID, } from "../storage/persistence.js";
+import { scopesToRoots } from "../types.js";
+const CHUNK_SIZE = 3200;
+const CHUNK_OVERLAP = 0.15;
+export async function runIndex(options) {
+    const start = Date.now();
+    const { repoPath, scope, onProgress } = options;
+    const cacheDir = repoCacheDir(repoPath);
+    await ensureCacheDir(cacheDir);
+    const manifest = (options.force ? null : await readManifest(cacheDir)) ?? emptyManifest();
+    onProgress?.({ phase: "scanning", filesProcessed: 0, totalFiles: 0 });
+    const roots = scopesToRoots(scope);
+    const files = await listMarkdownFiles(repoPath, roots, { exclude: options.exclude });
+    const docStore = new DocumentStore();
+    await docStore.initialize((await readDocumentStoreExport(cacheDir)) ?? undefined);
+    const vectorStore = new VectorStore({ dimensions: EMBEDDING_DIMENSIONS });
+    const hasExistingVectors = await vectorIndexExists(cacheDir);
+    if (hasExistingVectors && !options.force) {
+        const sidecar = await readVectorSidecar(cacheDir);
+        if (sidecar) {
+            await vectorStore.loadFromDisk(paths(cacheDir).vectorIndex, sidecar);
+        }
+        else {
+            await vectorStore.initialize();
+        }
+    }
+    else {
+        await vectorStore.initialize();
+    }
+    const chunker = new Chunker({ maxChunkSize: CHUNK_SIZE, overlapPercent: CHUNK_OVERLAP });
+    const currentPaths = new Set(files.map((f) => f.path));
+    const changed = [];
+    for (const f of files) {
+        const entry = manifest.files[f.path];
+        if (!entry || entry.mtimeMs !== f.mtimeMs)
+            changed.push(f);
+    }
+    // Files that were indexed under this manifest for this scope but no longer exist on disk.
+    const removed = Object.keys(manifest.files).filter((p) => {
+        const inScope = roots.some((r) => p === r || p.startsWith(r + "/"));
+        return inScope && !currentPaths.has(p);
+    });
+    onProgress?.({
+        phase: "text-indexing",
+        filesProcessed: 0,
+        totalFiles: changed.length,
+    });
+    let chunksIndexed = 0;
+    const changedDocsForEmbedding = [];
+    for (const path of removed) {
+        await docStore.deleteByFilePath(path);
+        await vectorStore.deleteByFilePath(path);
+        delete manifest.files[path];
+    }
+    for (let i = 0; i < changed.length; i++) {
+        const file = changed[i];
+        onProgress?.({
+            phase: "text-indexing",
+            filesProcessed: i,
+            totalFiles: changed.length,
+            currentFile: file.path,
+        });
+        const content = await readRepoFile(file.absPath);
+        await docStore.deleteByFilePath(file.path);
+        await vectorStore.deleteByFilePath(file.path);
+        const chunks = chunker.chunk(content, file.path);
+        const docs = chunks.map((chunk, idx) => ({
+            id: `${file.path}:${idx}`,
+            filePath: file.path,
+            content: chunk.content,
+            chunkIndex: idx,
+            mtimeMs: file.mtimeMs,
+            startLine: chunk.startLine,
+            headingBreadcrumb: chunk.headingBreadcrumb,
+        }));
+        await docStore.indexDocuments(docs);
+        chunksIndexed += docs.length;
+        changedDocsForEmbedding.push(...docs);
+        manifest.files[file.path] = { mtimeMs: file.mtimeMs };
+    }
+    let embeddingsGenerated = 0;
+    if (!options.skipEmbeddings && changedDocsForEmbedding.length > 0) {
+        const embeddingService = new EmbeddingService();
+        await embeddingService.initialize();
+        onProgress?.({
+            phase: "embedding",
+            filesProcessed: 0,
+            totalFiles: changedDocsForEmbedding.length,
+        });
+        const BATCH = 16;
+        for (let i = 0; i < changedDocsForEmbedding.length; i += BATCH) {
+            const batch = changedDocsForEmbedding.slice(i, i + BATCH);
+            const embeddings = await embeddingService.embedBatch(batch.map((d) => d.content));
+            for (let j = 0; j < batch.length; j++) {
+                const doc = batch[j];
+                await vectorStore.add({
+                    id: doc.id,
+                    filePath: doc.filePath,
+                    chunkIndex: doc.chunkIndex,
+                    embedding: embeddings[j],
+                });
+                embeddingsGenerated++;
+            }
+            onProgress?.({
+                phase: "embedding",
+                filesProcessed: Math.min(i + BATCH, changedDocsForEmbedding.length),
+                totalFiles: changedDocsForEmbedding.length,
+            });
+        }
+    }
+    // Persist.
+    await writeDocumentStoreExport(cacheDir, docStore.export());
+    if (!options.skipEmbeddings) {
+        const sidecar = await vectorStore.saveToDisk(paths(cacheDir).vectorIndex);
+        await writeVectorSidecar(cacheDir, sidecar);
+    }
+    const updatedManifest = {
+        ...manifest,
+        embeddingModel: EMBEDDING_MODEL_ID,
+        indexedScopes: Array.from(new Set([...manifest.indexedScopes, scope])),
+        lastIndexedAt: Date.now(),
+    };
+    await writeManifest(cacheDir, updatedManifest);
+    onProgress?.({ phase: "done", filesProcessed: files.length, totalFiles: files.length });
+    return {
+        scope,
+        filesScanned: files.length,
+        filesChanged: changed.length,
+        filesRemoved: removed.length,
+        chunksIndexed,
+        embeddingsGenerated,
+        totalDurationMs: Date.now() - start,
+    };
+}
+//# sourceMappingURL=indexing-service.js.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"indexing-service.js","sourceRoot":"","sources":["../../src/indexing/indexing-service.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AACH,OAAO,EAAE,aAAa,EAAwB,MAAM,8BAA8B,CAAC;AACnF,OAAO,EAAE,WAAW,EAAE,MAAM,4BAA4B,CAAC;AACzD,OAAO,EAAE,gBAAgB,EAAE,oBAAoB,EAAE,MAAM,oCAAoC,CAAC;AAC5F,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAC;AACvC,OAAO,EAAE,iBAAiB,EAAE,YAAY,EAAiB,MAAM,2BAA2B,CAAC;AAC3F,OAAO,EACL,cAAc,EACd,YAAY,EACZ,aAAa,EACb,aAAa,EACb,uBAAuB,EACvB,wBAAwB,EACxB,iBAAiB,EACjB,kBAAkB,EAClB,iBAAiB,EACjB,YAAY,EACZ,KAAK,EACL,kBAAkB,GAEnB,MAAM,2BAA2B,CAAC;AACnC,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAiC5C,MAAM,UAAU,GAAG,IAAI,CAAC;AACxB,MAAM,aAAa,GAAG,IAAI,CAAC;AAE3B,MAAM,CAAC,KAAK,UAAU,QAAQ,CAAC,OAAqB;IAClD,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IACzB,MAAM,EAAE,QAAQ,EAAE,KAAK,EAAE,UAAU,EAAE,GAAG,OAAO,CAAC;IAChD,MAAM,QAAQ,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACxC,MAAM,cAAc,CAAC,QAAQ,CAAC,CAAC;IAE/B,MAAM,QAAQ,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,YAAY,CAAC,QAAQ,CAAC,CAAC,IAAI,aAAa,EAAE,CAAC;IAE1F,UAAU,EAAE,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,cAAc,EAAE,CAAC,EAAE,UAAU,EAAE,CAAC,EAAE,CAAC,CAAC;IACtE,MAAM,KAAK,GAAG,aAAa,CAAC,KAAK,CAAC,CAAC;IACnC,MAAM,KAAK,GAAG,MAAM,iBAAiB,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE,OAAO,EAAE,OAAO,CAAC,OAAO,EAAE,CAAC,CAAC;IAErF,MAAM,QAAQ,GAAG,IAAI,aAAa,EAAE,CAAC;IACrC,MAAM,QAAQ,CAAC,UAAU,CAAC,CAAC,MAAM,uBAAuB,CAAC,QAAQ,CAAC,CAAC,IAAI,SAAS,CAAC,CAAC;IAElF,MAAM,WAAW,GAAG,IAAI,WAAW,CAAC,EAAE,UAAU,EAAE,oBAAoB,EAAE,CAAC,CAAC;IAC1E,MAAM,kBAAkB,GAAG,MAAM,iBAAiB,CAAC,QAAQ,CAAC,CAAC;IAC7D,IAAI,kBAAkB,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC;QACzC,MAAM,OAAO,GAAG,MAAM,iBAAiB,CAAC,QAAQ,CAAC,CAAC;QAClD,IAAI,OAAO,EAAE,CAAC;YACZ,MAAM,WAAW,CAAC,YAAY,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,OAAO,CAAC,CAAC;QACvE,CAAC;aAAM,CAAC;YACN,MAAM,WAAW,CAAC,UAAU,EAAE,CAAC;QACjC,CAAC;IACH,CAAC;SAAM,CAAC;QACN,MAAM,WAAW,CAAC,UAAU,EAAE,CAAC;IACjC,CAAC;IAED,MAAM,OAAO,GAAG,IAAI,OAAO,CAAC,EAAE,YAAY,EAAE,UAAU,EAAE,cAAc,EAAE,aAAa,EAAE,CAAC,CAAC;IAEzF,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IACvD,MAAM,OAAO,GAAe,EAAE,CAAC;IAC/B,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;QACtB,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QACrC,IAAI,CAAC,KAAK,IAAI,KAAK,CAAC,OAAO,KAAK,CAAC,CAAC,OAAO;YAAE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC7D,CAAC;IAED,0FAA0F;IAC1F,MAAM,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;QACvD,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;QACpE,OAAO,OAAO,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;IACzC,CAAC,CAAC,CAAC;IAEH,UAAU,EAAE,CAAC;QACX,KAAK,EAAE,eAAe;QACtB,cAAc,EAAE,CAAC;QACjB,UAAU,EAAE,OAAO,CAAC,MAAM;KAC3B,CAAC,CAAC;IAEH,IAAI,aAAa,GAAG,CAAC,CAAC;IACtB,MAAM,uBAAuB,GAAsB,EAAE,CAAC;IAEtD,KAAK,MAAM,IAAI,IAAI,OAAO,EAAE,CAAC;QAC3B,MAAM,QAAQ,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC;QACtC,MAAM,WAAW,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC;QACzC,OAAO,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC9B,CAAC;IAED,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,CAAE,CAAC;QACzB,UAAU,EAAE,CAAC;YACX,KAAK,EAAE,eAAe;YACtB,cAAc,EAAE,CAAC;YACjB,UAAU,EAAE,OAAO,CAAC,MAAM;YAC1B,WAAW,EAAE,IAAI,CAAC,IAAI;SACvB,CAAC,CAAC;QAEH,MAAM,OAAO,GAAG,MAAM,YAAY,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACjD,MAAM,QAAQ,CAAC,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC3C,MAAM,WAAW,CAAC,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAE9C,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC;QACjD,MAAM,IAAI,GAAsB,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,GAAG,EAAE,EAAE,CAAC,CAAC;YAC1D,EAAE,EAAE,GAAG,IAAI,CAAC,IAAI,IAAI,GAAG,EAAE;YACzB,QAAQ,EAAE,IAAI,CAAC,IAAI;YACnB,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,UAAU,EAAE,GAAG;YACf,OAAO,EAAE,IAAI,CAAC,OAAO;YACrB,SAAS,EAAE,KAAK,CAAC,SAAS;YAC1B,iBAAiB,EAAE,KAAK,CAAC,iBAAiB;SAC3C,CAAC,CAAC,CAAC;QAEJ,MAAM,QAAQ,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;QACpC,aAAa,IAAI,IAAI,CAAC,MAAM,CAAC;QAC7B,uBAAuB,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,CAAC;QAEtC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,CAAC;IACxD,CAAC;IAED,IAAI,mBAAmB,GAAG,CAAC,CAAC;IAC5B,IAAI,CAAC,OAAO,CAAC,cAAc,IAAI,uBAAuB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClE,MAAM,gBAAgB,GAAG,IAAI,gBAAgB,EAAE,CAAC;QAChD,MAAM,gBAAgB,CAAC,UAAU,EAAE,CAAC;QAEpC,UAAU,EAAE,CAAC;YACX,KAAK,EAAE,WAAW;YAClB,cAAc,EAAE,CAAC;YACjB,UAAU,EAAE,uBAAuB,CAAC,MAAM;SAC3C,CAAC,CAAC;QAEH,MAAM,KAAK,GAAG,EAAE,CAAC;QACjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,uBAAuB,CAAC,MAAM,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC;YAC/D,MAAM,KAAK,GAAG,uBAAuB,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,CAAC;YAC1D,MAAM,UAAU,GAAG,MAAM,gBAAgB,CAAC,UAAU,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;YAClF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACtC,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;gBACtB,MAAM,WAAW,CAAC,GAAG,CAAC;oBACpB,EAAE,EAAE,GAAG,CAAC,EAAE;oBACV,QAAQ,EAAE,GAAG,CAAC,QAAQ;oBACtB,UAAU,EAAE,GAAG,CAAC,UAAU;oBAC1B,SAAS,EAAE,UAAU,CAAC,CAAC,CAAE;iBAC1B,CAAC,CAAC;gBACH,mBAAmB,EAAE,CAAC;YACxB,CAAC;YACD,UAAU,EAAE,CAAC;gBACX,KAAK,EAAE,WAAW;gBAClB,cAAc,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,KAAK,EAAE,uBAAuB,CAAC,MAAM,CAAC;gBACnE,UAAU,EAAE,uBAAuB,CAAC,MAAM;aAC3C,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,WAAW;IACX,MAAM,wBAAwB,CAAC,QAAQ,EAAE,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC;IAC5D,IAAI,CAAC,OAAO,CAAC,cAAc,EAAE,CAAC;QAC5B,MAAM,OAAO,GAAG,MAAM,WAAW,CAAC,UAAU,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,WAAW,CAAC,CAAC;QAC1E,MAAM,kBAAkB,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;IAC9C,CAAC;IAED,MAAM,eAAe,GAAa;QAChC,GAAG,QAAQ;QACX,cAAc,EAAE,kBAAkB;QAClC,aAAa,EAAE,KAAK,CAAC,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC,GAAG,QAAQ,CAAC,aAAa,EAAE,KAAK,CAAC,CAAC,CAAC;QACtE,aAAa,EAAE,IAAI,CAAC,GAAG,EAAE;KAC1B,CAAC;IACF,MAAM,aAAa,CAAC,QAAQ,EAAE,eAAe,CAAC,CAAC;IAE/C,UAAU,EAAE,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,cAAc,EAAE,KAAK,CAAC,MAAM,EAAE,UAAU,EAAE,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC;IAExF,OAAO;QACL,KAAK;QACL,YAAY,EAAE,KAAK,CAAC,MAAM;QAC1B,YAAY,EAAE,OAAO,CAAC,MAAM;QAC5B,YAAY,EAAE,OAAO,CAAC,MAAM;QAC5B,aAAa;QACb,mBAAmB;QACnB,eAAe,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;KACpC,CAAC;AACJ,CAAC"}

package/dist/search/search-service.d.ts

@@ -0,0 +1,53 @@
+/**
+ * Adapted from obsidian-qmd (MIT): https://github.com/thirteen37/obsidian-qmd
+ * Source: src/search/SearchService.ts. Carried over: the core idea of
+ * Reciprocal Rank Fusion over independently-ranked BM25 and vector result
+ * lists, with small positional bonuses for top ranks (RRF_K=60, same
+ * constant as upstream). Dropped: `QueryExpansionService` and
+ * `RerankingService` integration — both call out to an LLM, which is out of
+ * scope for this offline CLI (the brief's three dependencies are minisearch,
+ * @xenova/transformers, and hnswlib-node; no LLM client). Also dropped:
+ * `groupByFile`/section dedup UI conveniences not needed by a line-oriented
+ * CLI, since callers get raw scored chunks that already carry a line number.
+ */
+import { DocumentStore } from "../storage/document-store.js";
+import { VectorStore } from "../storage/vector-store.js";
+import { EmbeddingService } from "../embeddings/embedding-service.js";
+import type { Scope, SearchMode } from "../types.js";
+import type { HighlightMode } from "./snippet.js";
+export interface HybridResult {
+    id: string;
+    filePath: string;
+    chunkIndex: number;
+    startLine: number;
+    textScore?: number;
+    vectorScore?: number;
+    rrfScore?: number;
+    snippet?: string;
+    content?: string;
+    headingBreadcrumb?: string[];
+    source: "text" | "vector" | "hybrid";
+}
+export interface SearchOptions {
+    mode: SearchMode;
+    scope: Scope;
+    limit?: number;
+    /** 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. When omitted, snippets fall back to chunk-only
+     * generation. */
+    repoRoot?: string;
+}
+export declare class SearchService {
+    private docStore;
+    private vectorStore;
+    private embeddingService;
+    constructor(docStore: DocumentStore, vectorStore: VectorStore, embeddingService: EmbeddingService | null);
+    search(query: string, options: SearchOptions): Promise<HybridResult[]>;
+    private textSearch;
+    private semanticSearch;
+    private hybridSearch;
+}
+//# sourceMappingURL=search-service.d.ts.map

package/dist/search/search-service.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"search-service.d.ts","sourceRoot":"","sources":["../../src/search/search-service.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AACH,OAAO,EAAE,aAAa,EAAE,MAAM,8BAA8B,CAAC;AAC7D,OAAO,EAAE,WAAW,EAAE,MAAM,4BAA4B,CAAC;AACzD,OAAO,EAAE,gBAAgB,EAAE,MAAM,oCAAoC,CAAC;AACtE,OAAO,KAAK,EAAE,KAAK,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAErD,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,cAAc,CAAC;AAElD,MAAM,WAAW,YAAY;IAC3B,EAAE,EAAE,MAAM,CAAC;IACX,QAAQ,EAAE,MAAM,CAAC;IACjB,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,iBAAiB,CAAC,EAAE,MAAM,EAAE,CAAC;IAC7B,MAAM,EAAE,MAAM,GAAG,QAAQ,GAAG,QAAQ,CAAC;CACtC;AAED,MAAM,WAAW,aAAa;IAC5B,IAAI,EAAE,UAAU,CAAC;IACjB,KAAK,EAAE,KAAK,CAAC;IACb,KAAK,CAAC,EAAE,MAAM,CAAC;IACf;2CACuC;IACvC,SAAS,CAAC,EAAE,aAAa,CAAC;IAC1B;;qBAEiB;IACjB,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB;AAID,qBAAa,aAAa;IAEtB,OAAO,CAAC,QAAQ;IAChB,OAAO,CAAC,WAAW;IACnB,OAAO,CAAC,gBAAgB;gBAFhB,QAAQ,EAAE,aAAa,EACvB,WAAW,EAAE,WAAW,EACxB,gBAAgB,EAAE,gBAAgB,GAAG,IAAI;IAG7C,MAAM,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,aAAa,GAAG,OAAO,CAAC,YAAY,EAAE,CAAC;YAkB9D,UAAU;YAqBV,cAAc;YAwCd,YAAY;CAqF3B"}

package/dist/search/search-service.js

@@ -0,0 +1,139 @@
+import { extractQueryTerms, buildSnippet } from "./snippet.js";
+const RRF_K = 60;
+export class SearchService {
+    docStore;
+    vectorStore;
+    embeddingService;
+    constructor(docStore, vectorStore, embeddingService) {
+        this.docStore = docStore;
+        this.vectorStore = vectorStore;
+        this.embeddingService = embeddingService;
+    }
+    async search(query, options) {
+        const trimmed = query.trim();
+        if (!trimmed)
+            return [];
+        const limit = options.limit ?? 10;
+        const highlight = options.highlight ?? "plain";
+        const repoRoot = options.repoRoot;
+        switch (options.mode) {
+            case "keyword":
+                return this.textSearch(trimmed, options.scope, limit, highlight, repoRoot);
+            case "semantic":
+                return this.semanticSearch(trimmed, options.scope, limit, highlight, repoRoot);
+            case "hybrid":
+                return this.hybridSearch(trimmed, options.scope, limit, highlight, repoRoot);
+        }
+    }
+    async textSearch(query, scope, limit, highlight, repoRoot) {
+        const results = await this.docStore.search(query, { limit, scope, highlight, repoRoot });
+        return results.map((r) => ({
+            id: r.id,
+            filePath: r.filePath,
+            chunkIndex: r.chunkIndex,
+            startLine: r.startLine,
+            textScore: r.score,
+            snippet: r.snippet,
+            content: r.content,
+            headingBreadcrumb: r.headingBreadcrumb,
+            source: "text",
+        }));
+    }
+    async semanticSearch(query, scope, limit, highlight, repoRoot) {
+        if (!this.embeddingService) {
+            throw new Error("Semantic search requires an embedding service");
+        }
+        const embedding = await this.embeddingService.embed(query);
+        const results = await this.vectorStore.search(embedding, limit, scope);
+        const queryTerms = extractQueryTerms(query);
+        const hybridResults = [];
+        for (const r of results) {
+            const doc = await this.docStore.getDocument(r.id);
+            hybridResults.push({
+                id: r.id,
+                filePath: r.filePath,
+                chunkIndex: r.chunkIndex,
+                startLine: doc?.startLine ?? 0,
+                vectorScore: r.score,
+                content: doc?.content,
+                headingBreadcrumb: doc?.headingBreadcrumb,
+                snippet: doc
+                    ? await buildSnippet({ filePath: r.filePath, content: doc.content, startLine: doc.startLine }, queryTerms, highlight, repoRoot)
+                    : undefined,
+                source: "vector",
+            });
+        }
+        return hybridResults;
+    }
+    async hybridSearch(query, scope, limit, highlight, repoRoot) {
+        const [textResults, vectorResults] = await Promise.all([
+            this.docStore.search(query, { limit: limit * 2, scope, highlight, repoRoot }),
+            this.embeddingService
+                ? this.vectorStore.search(await this.embeddingService.embed(query), limit * 2, scope)
+                : Promise.resolve([]),
+        ]);
+        const byId = new Map();
+        textResults.forEach((r, index) => {
+            byId.set(r.id, {
+                id: r.id,
+                filePath: r.filePath,
+                chunkIndex: r.chunkIndex,
+                startLine: r.startLine,
+                textScore: r.score,
+                textRank: index + 1,
+                snippet: r.snippet,
+                content: r.content,
+                headingBreadcrumb: r.headingBreadcrumb,
+                source: "text",
+            });
+        });
+        const queryTerms = extractQueryTerms(query);
+        for (const [index, r] of vectorResults.entries()) {
+            const existing = byId.get(r.id);
+            if (existing) {
+                existing.vectorScore = r.score;
+                existing.vectorRank = index + 1;
+                existing.source = "hybrid";
+            }
+            else {
+                const doc = await this.docStore.getDocument(r.id);
+                byId.set(r.id, {
+                    id: r.id,
+                    filePath: r.filePath,
+                    chunkIndex: r.chunkIndex,
+                    startLine: doc?.startLine ?? 0,
+                    vectorScore: r.score,
+                    vectorRank: index + 1,
+                    content: doc?.content,
+                    headingBreadcrumb: doc?.headingBreadcrumb,
+                    snippet: doc
+                        ? await buildSnippet({ filePath: r.filePath, content: doc.content, startLine: doc.startLine }, queryTerms, highlight, repoRoot)
+                        : undefined,
+                    source: "vector",
+                });
+            }
+        }
+        const results = Array.from(byId.values()).map((r) => {
+            let rrfScore = 0;
+            if (r.textRank) {
+                rrfScore += 1 / (RRF_K + r.textRank);
+                if (r.textRank === 1)
+                    rrfScore += 0.05;
+                else if (r.textRank <= 3)
+                    rrfScore += 0.02;
+            }
+            if (r.vectorRank) {
+                rrfScore += 1 / (RRF_K + r.vectorRank);
+                if (r.vectorRank === 1)
+                    rrfScore += 0.05;
+                else if (r.vectorRank <= 3)
+                    rrfScore += 0.02;
+            }
+            const { textRank, vectorRank, ...clean } = r;
+            return { ...clean, rrfScore };
+        });
+        results.sort((a, b) => (b.rrfScore ?? 0) - (a.rrfScore ?? 0));
+        return results.slice(0, limit);
+    }
+}
+//# sourceMappingURL=search-service.js.map