vemora 0.1.0-alpha.10

package/dist/search/formatter.js

@@ -0,0 +1,208 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.formatTerse = formatTerse;
+exports.formatJson = formatJson;
+exports.formatMarkdown = formatMarkdown;
+const deps_1 = require("../indexer/deps");
+const signature_1 = require("./signature");
+// ─── Terse formatter ──────────────────────────────────────────────────────────
+/**
+ * Ultra-compact one-liner per result for small/local models with limited context windows.
+ * Format: file:startLine | symbol (type) | score | first-line-of-signature
+ *
+ * ~70-80% fewer tokens than the markdown format. Recommended for models with
+ * context windows ≤ 32K, or whenever token budget is the primary constraint.
+ */
+function formatTerse(results, options) {
+    const topK = options.topK ?? 10;
+    const seen = new Set();
+    const lines = [];
+    for (const { chunk, score, symbol } of results) {
+        if (seen.has(chunk.id))
+            continue;
+        seen.add(chunk.id);
+        const symbolPart = chunk.symbol
+            ? `${chunk.symbol} (${symbol?.type ?? "symbol"})`
+            : "(no symbol)";
+        const sig = (0, signature_1.extractSignature)(chunk.content).split("\n")[0].trim();
+        lines.push(`${chunk.file}:${chunk.start} | ${symbolPart} | ${score.toFixed(3)} | ${sig}`);
+        if (lines.length >= topK)
+            break;
+    }
+    return lines.join("\n");
+}
+function formatJson(query, results, depGraph, fileSummaries, options) {
+    const importedByMap = (0, deps_1.computeImportedBy)(depGraph);
+    const showCode = options.showCode ?? false;
+    const topK = options.topK ?? 10;
+    const seen = new Set();
+    const jsonResults = [];
+    for (const { chunk, score, symbol } of results) {
+        if (seen.has(chunk.id))
+            continue;
+        seen.add(chunk.id);
+        const rank = jsonResults.length + 1;
+        const tier = showCode ? "high" : (0, signature_1.getDisplayTier)(rank);
+        const fileDeps = depGraph[chunk.file];
+        const usedBy = importedByMap.get(chunk.file) ?? [];
+        let code = null;
+        let signature = null;
+        if (tier === "high") {
+            const codeLines = chunk.content.split("\n");
+            const limit = showCode ? codeLines.length : signature_1.HIGH_CODE_LINES;
+            code = codeLines.slice(0, limit).join("\n");
+        }
+        else if (tier === "med") {
+            signature = (0, signature_1.extractSignature)(chunk.content);
+        }
+        const fileSummary = fileSummaries[chunk.file];
+        jsonResults.push({
+            rank,
+            tier,
+            file: chunk.file,
+            symbol: chunk.symbol ?? null,
+            symbolType: symbol?.type ?? null,
+            lines: { start: chunk.start, end: chunk.end },
+            score: parseFloat(score.toFixed(4)),
+            code,
+            signature,
+            imports: fileDeps?.imports ?? [],
+            usedBy,
+            summary: fileSummary?.summary ?? null,
+        });
+        if (jsonResults.length >= topK)
+            break;
+    }
+    const output = {
+        query,
+        totalResults: jsonResults.length,
+        results: jsonResults,
+    };
+    return JSON.stringify(output, null, 2);
+}
+// ─── Markdown formatter ───────────────────────────────────────────────────────
+function formatMarkdown(query, results, depGraph, fileSummaries, callGraph, options) {
+    const importedByMap = (0, deps_1.computeImportedBy)(depGraph);
+    const showCode = options.showCode ?? false;
+    const topK = options.topK ?? 10;
+    const lines = [];
+    lines.push(`## Relevant code for: \`${query}\``);
+    lines.push("");
+    const seen = new Set();
+    let displayed = 0;
+    for (const { chunk, score, symbol } of results) {
+        if (seen.has(chunk.id))
+            continue;
+        seen.add(chunk.id);
+        const rank = displayed + 1;
+        const tier = showCode ? "high" : (0, signature_1.getDisplayTier)(rank);
+        // Header
+        lines.push(`### ${rank}. \`${chunk.file}\``);
+        if (chunk.symbol) {
+            const symType = symbol?.type ?? "symbol";
+            lines.push(`**${symType}** \`${chunk.symbol}\`  `);
+        }
+        lines.push(`Lines ${chunk.start}–${chunk.end} · Score: ${score.toFixed(4)} · Tier: ${tier}`);
+        lines.push("");
+        // Dependencies
+        if (tier !== "low") {
+            const fileDeps = depGraph[chunk.file];
+            const usedBy = importedByMap.get(chunk.file) ?? [];
+            if (fileDeps?.imports.length) {
+                const maxDeps = tier === "high" ? 6 : 3;
+                const shown = fileDeps.imports.slice(0, maxDeps);
+                const hidden = fileDeps.imports.length - shown.length;
+                lines.push("**Imports:**");
+                for (const imp of shown) {
+                    const syms = imp.symbols.length > 0
+                        ? ` — \`${imp.symbols.slice(0, 4).join(", ")}\``
+                        : "";
+                    lines.push(`- \`${imp.file}\`${syms}`);
+                }
+                if (hidden > 0)
+                    lines.push(`- _…and ${hidden} more_`);
+                lines.push("");
+            }
+            if (usedBy.length > 0) {
+                const maxUsed = tier === "high" ? 4 : 2;
+                const shown = usedBy.slice(0, maxUsed);
+                const hidden = usedBy.length - shown.length;
+                lines.push("**Used by (files):**");
+                for (const caller of shown) {
+                    lines.push(`- \`${caller}\``);
+                }
+                if (hidden > 0)
+                    lines.push(`- _…and ${hidden} more_`);
+                lines.push("");
+            }
+            // Call Graph context
+            const symbolId = chunk.symbol ? `${chunk.file}:${chunk.symbol}` : null;
+            const callInfo = symbolId ? callGraph[symbolId] : null;
+            if (callInfo) {
+                if (callInfo.calls.length > 0) {
+                    const maxCalls = tier === "high" ? 6 : 3;
+                    const shown = callInfo.calls.slice(0, maxCalls);
+                    const hidden = callInfo.calls.length - shown.length;
+                    lines.push("**Calls:**");
+                    for (const call of shown) {
+                        const loc = call.file ? ` (in \`${call.file}\`)` : "";
+                        lines.push(`- \`${call.name}\`${loc}`);
+                    }
+                    if (hidden > 0)
+                        lines.push(`- _…and ${hidden} more_`);
+                    lines.push("");
+                }
+                if (callInfo.calledBy.length > 0) {
+                    const maxCallers = tier === "high" ? 4 : 2;
+                    const shown = callInfo.calledBy.slice(0, maxCallers);
+                    const hidden = callInfo.calledBy.length - shown.length;
+                    lines.push("**Called by:**");
+                    for (const callerId of shown) {
+                        lines.push(`- \`${callerId}\``);
+                    }
+                    if (hidden > 0)
+                        lines.push(`- _…and ${hidden} more_`);
+                    lines.push("");
+                }
+            }
+        }
+        // Code / signature / summary
+        if (tier === "high") {
+            const codeLines = chunk.content.split("\n");
+            const limit = showCode ? codeLines.length : signature_1.HIGH_CODE_LINES;
+            const preview = codeLines.slice(0, limit).join("\n");
+            const ext = chunk.file.split(".").pop() ?? "";
+            lines.push(`\`\`\`${ext}`);
+            lines.push(preview);
+            if (codeLines.length > limit) {
+                lines.push(`// … (${codeLines.length - limit} more lines — use --show-code to expand)`);
+            }
+            lines.push("```");
+        }
+        else if (tier === "med") {
+            const sig = (0, signature_1.extractSignature)(chunk.content);
+            const ext = chunk.file.split(".").pop() ?? "";
+            lines.push(`\`\`\`${ext}`);
+            lines.push(sig);
+            lines.push("```");
+        }
+        else {
+            // LOW — summary if available
+            const fileSummary = fileSummaries[chunk.file];
+            if (fileSummary) {
+                lines.push(`> ${fileSummary.summary}`);
+            }
+        }
+        lines.push("");
+        lines.push("---");
+        lines.push("");
+        displayed++;
+        if (displayed >= topK)
+            break;
+    }
+    if (displayed === 0) {
+        lines.push("_No results found._");
+    }
+    return lines.join("\n");
+}
+//# sourceMappingURL=formatter.js.map

package/dist/search/formatter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"formatter.js","sourceRoot":"","sources":["../../src/search/formatter.ts"],"names":[],"mappings":";;AA0BA,kCAyBC;AAyBD,gCA8DC;AAID,wCAmJC;AA3RD,0CAAoD;AACpD,2CAAgF;AAUhF,iFAAiF;AAEjF;;;;;;GAMG;AACH,SAAgB,WAAW,CACzB,OAAuB,EACvB,OAAoC;IAEpC,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,EAAE,CAAC;IAChC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,KAAK,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,OAAO,EAAE,CAAC;QAC/C,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAAE,SAAS;QACjC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QAEnB,MAAM,UAAU,GAAG,KAAK,CAAC,MAAM;YAC7B,CAAC,CAAC,GAAG,KAAK,CAAC,MAAM,KAAK,MAAM,EAAE,IAAI,IAAI,QAAQ,GAAG;YACjD,CAAC,CAAC,aAAa,CAAC;QAClB,MAAM,GAAG,GAAG,IAAA,4BAAgB,EAAC,KAAK,CAAC,OAAO,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAElE,KAAK,CAAC,IAAI,CACR,GAAG,KAAK,CAAC,IAAI,IAAI,KAAK,CAAC,KAAK,MAAM,UAAU,MAAM,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,GAAG,EAAE,CAC9E,CAAC;QAEF,IAAI,KAAK,CAAC,MAAM,IAAI,IAAI;YAAE,MAAM;IAClC,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC;AAyBD,SAAgB,UAAU,CACxB,KAAa,EACb,OAAuB,EACvB,QAAyB,EACzB,aAA+B,EAC/B,OAAsB;IAEtB,MAAM,aAAa,GAAG,IAAA,wBAAiB,EAAC,QAAQ,CAAC,CAAC;IAClD,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,IAAI,KAAK,CAAC;IAC3C,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,EAAE,CAAC;IAEhC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,MAAM,WAAW,GAAiB,EAAE,CAAC;IAErC,KAAK,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,OAAO,EAAE,CAAC;QAC/C,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAAE,SAAS;QACjC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QAEnB,MAAM,IAAI,GAAG,WAAW,CAAC,MAAM,GAAG,CAAC,CAAC;QACpC,MAAM,IAAI,GAAG,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,IAAA,0BAAc,EAAC,IAAI,CAAC,CAAC;QAEtD,MAAM,QAAQ,GAAG,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACtC,MAAM,MAAM,GAAG,aAAa,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;QAEnD,IAAI,IAAI,GAAkB,IAAI,CAAC;QAC/B,IAAI,SAAS,GAAkB,IAAI,CAAC;QAEpC,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;YACpB,MAAM,SAAS,GAAG,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YAC5C,MAAM,KAAK,GAAG,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,2BAAe,CAAC;YAC5D,IAAI,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC9C,CAAC;aAAM,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;YAC1B,SAAS,GAAG,IAAA,4BAAgB,EAAC,KAAK,CAAC,OAAO,CAAC,CAAC;QAC9C,CAAC;QAED,MAAM,WAAW,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAE9C,WAAW,CAAC,IAAI,CAAC;YACf,IAAI;YACJ,IAAI;YACJ,IAAI,EAAE,KAAK,CAAC,IAAI;YAChB,MAAM,EAAE,KAAK,CAAC,MAAM,IAAI,IAAI;YAC5B,UAAU,EAAE,MAAM,EAAE,IAAI,IAAI,IAAI;YAChC,KAAK,EAAE,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,CAAC,GAAG,EAAE;YAC7C,KAAK,EAAE,UAAU,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;YACnC,IAAI;YACJ,SAAS;YACT,OAAO,EAAE,QAAQ,EAAE,OAAO,IAAI,EAAE;YAChC,MAAM;YACN,OAAO,EAAE,WAAW,EAAE,OAAO,IAAI,IAAI;SACtC,CAAC,CAAC;QAEH,IAAI,WAAW,CAAC,MAAM,IAAI,IAAI;YAAE,MAAM;IACxC,CAAC;IAED,MAAM,MAAM,GAAe;QACzB,KAAK;QACL,YAAY,EAAE,WAAW,CAAC,MAAM;QAChC,OAAO,EAAE,WAAW;KACrB,CAAC;IAEF,OAAO,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;AACzC,CAAC;AAED,iFAAiF;AAEjF,SAAgB,cAAc,CAC5B,KAAa,EACb,OAAuB,EACvB,QAAyB,EACzB,aAA+B,EAC/B,SAAoB,EACpB,OAAsB;IAEtB,MAAM,aAAa,GAAG,IAAA,wBAAiB,EAAC,QAAQ,CAAC,CAAC;IAClD,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,IAAI,KAAK,CAAC;IAC3C,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,EAAE,CAAC;IAEhC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,CAAC,IAAI,CAAC,2BAA2B,KAAK,IAAI,CAAC,CAAC;IACjD,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAEf,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,KAAK,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,OAAO,EAAE,CAAC;QAC/C,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YAAE,SAAS;QACjC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QAEnB,MAAM,IAAI,GAAG,SAAS,GAAG,CAAC,CAAC;QAC3B,MAAM,IAAI,GAAG,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,IAAA,0BAAc,EAAC,IAAI,CAAC,CAAC;QAEtD,SAAS;QACT,KAAK,CAAC,IAAI,CAAC,OAAO,IAAI,OAAO,KAAK,CAAC,IAAI,IAAI,CAAC,CAAC;QAC7C,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;YACjB,MAAM,OAAO,GAAG,MAAM,EAAE,IAAI,IAAI,QAAQ,CAAC;YACzC,KAAK,CAAC,IAAI,CAAC,KAAK,OAAO,QAAQ,KAAK,CAAC,MAAM,MAAM,CAAC,CAAC;QACrD,CAAC;QACD,KAAK,CAAC,IAAI,CACR,SAAS,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,GAAG,aAAa,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,YAAY,IAAI,EAAE,CACjF,CAAC;QACF,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAEf,eAAe;QACf,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;YACnB,MAAM,QAAQ,GAAG,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YACtC,MAAM,MAAM,GAAG,aAAa,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;YAEnD,IAAI,QAAQ,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC;gBAC7B,MAAM,OAAO,GAAG,IAAI,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;gBACxC,MAAM,KAAK,GAAG,QAAQ,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;gBACjD,MAAM,MAAM,GAAG,QAAQ,CAAC,OAAO,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC;gBAEtD,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC;gBAC3B,KAAK,MAAM,GAAG,IAAI,KAAK,EAAE,CAAC;oBACxB,MAAM,IAAI,GACR,GAAG,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC;wBACpB,CAAC,CAAC,QAAQ,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI;wBAChD,CAAC,CAAC,EAAE,CAAC;oBACT,KAAK,CAAC,IAAI,CAAC,OAAO,GAAG,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC,CAAC;gBACzC,CAAC;gBACD,IAAI,MAAM,GAAG,CAAC;oBAAE,KAAK,CAAC,IAAI,CAAC,WAAW,MAAM,QAAQ,CAAC,CAAC;gBACtD,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACjB,CAAC;YAED,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACtB,MAAM,OAAO,GAAG,IAAI,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;gBACxC,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;gBACvC,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC;gBAE5C,KAAK,CAAC,IAAI,CAAC,sBAAsB,CAAC,CAAC;gBACnC,KAAK,MAAM,MAAM,IAAI,KAAK,EAAE,CAAC;oBAC3B,KAAK,CAAC,IAAI,CAAC,OAAO,MAAM,IAAI,CAAC,CAAC;gBAChC,CAAC;gBACD,IAAI,MAAM,GAAG,CAAC;oBAAE,KAAK,CAAC,IAAI,CAAC,WAAW,MAAM,QAAQ,CAAC,CAAC;gBACtD,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACjB,CAAC;YAED,qBAAqB;YACrB,MAAM,QAAQ,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,IAAI,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;YACvE,MAAM,QAAQ,GAAG,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;YAEvD,IAAI,QAAQ,EAAE,CAAC;gBACb,IAAI,QAAQ,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAC9B,MAAM,QAAQ,GAAG,IAAI,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;oBACzC,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAC;oBAChD,MAAM,MAAM,GAAG,QAAQ,CAAC,KAAK,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC;oBAEpD,KAAK,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;oBACzB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;wBACzB,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,UAAU,IAAI,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;wBACtD,KAAK,CAAC,IAAI,CAAC,OAAO,IAAI,CAAC,IAAI,KAAK,GAAG,EAAE,CAAC,CAAC;oBACzC,CAAC;oBACD,IAAI,MAAM,GAAG,CAAC;wBAAE,KAAK,CAAC,IAAI,CAAC,WAAW,MAAM,QAAQ,CAAC,CAAC;oBACtD,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;gBACjB,CAAC;gBAED,IAAI,QAAQ,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBACjC,MAAM,UAAU,GAAG,IAAI,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;oBAC3C,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC;oBACrD,MAAM,MAAM,GAAG,QAAQ,CAAC,QAAQ,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC;oBAEvD,KAAK,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;oBAC7B,KAAK,MAAM,QAAQ,IAAI,KAAK,EAAE,CAAC;wBAC7B,KAAK,CAAC,IAAI,CAAC,OAAO,QAAQ,IAAI,CAAC,CAAC;oBAClC,CAAC;oBACD,IAAI,MAAM,GAAG,CAAC;wBAAE,KAAK,CAAC,IAAI,CAAC,WAAW,MAAM,QAAQ,CAAC,CAAC;oBACtD,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;gBACjB,CAAC;YACH,CAAC;QACH,CAAC;QAED,6BAA6B;QAC7B,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;YACpB,MAAM,SAAS,GAAG,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YAC5C,MAAM,KAAK,GAAG,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,2BAAe,CAAC;YAC5D,MAAM,OAAO,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACrD,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,CAAC;YAC9C,KAAK,CAAC,IAAI,CAAC,SAAS,GAAG,EAAE,CAAC,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YACpB,IAAI,SAAS,CAAC,MAAM,GAAG,KAAK,EAAE,CAAC;gBAC7B,KAAK,CAAC,IAAI,CACR,SAAS,SAAS,CAAC,MAAM,GAAG,KAAK,0CAA0C,CAC5E,CAAC;YACJ,CAAC;YACD,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACpB,CAAC;aAAM,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;YAC1B,MAAM,GAAG,GAAG,IAAA,4BAAgB,EAAC,KAAK,CAAC,OAAO,CAAC,CAAC;YAC5C,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,CAAC;YAC9C,KAAK,CAAC,IAAI,CAAC,SAAS,GAAG,EAAE,CAAC,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAChB,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACpB,CAAC;aAAM,CAAC;YACN,6BAA6B;YAC7B,MAAM,WAAW,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YAC9C,IAAI,WAAW,EAAE,CAAC;gBAChB,KAAK,CAAC,IAAI,CAAC,KAAK,WAAW,CAAC,OAAO,EAAE,CAAC,CAAC;YACzC,CAAC;QACH,CAAC;QAED,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACf,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAClB,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAEf,SAAS,EAAE,CAAC;QACZ,IAAI,SAAS,IAAI,IAAI;YAAE,MAAM;IAC/B,CAAC;IAED,IAAI,SAAS,KAAK,CAAC,EAAE,CAAC;QACpB,KAAK,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAC;IACpC,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC"}

package/dist/search/hybrid.d.ts

@@ -0,0 +1,10 @@
+import type { Chunk, EmbeddingCache, SearchResult, SymbolIndex } from "../core/types";
+export interface HybridOptions {
+    alpha?: number;
+    topK?: number;
+}
+/**
+ * Combines Vector Search (semantic) and BM25 Search (keyword) for higher accuracy.
+ */
+export declare function hybridSearch(query: string, queryEmbedding: number[] | null, chunks: Chunk[], cache: EmbeddingCache, symbols: SymbolIndex, options?: HybridOptions): Promise<SearchResult[]>;
+//# sourceMappingURL=hybrid.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"hybrid.d.ts","sourceRoot":"","sources":["../../src/search/hybrid.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,KAAK,EACL,cAAc,EACd,YAAY,EAEZ,WAAW,EACZ,MAAM,eAAe,CAAC;AAIvB,MAAM,WAAW,aAAa;IAC5B,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,IAAI,CAAC,EAAE,MAAM,CAAC;CACf;AAED;;GAEG;AACH,wBAAsB,YAAY,CAChC,KAAK,EAAE,MAAM,EACb,cAAc,EAAE,MAAM,EAAE,GAAG,IAAI,EAC/B,MAAM,EAAE,KAAK,EAAE,EACf,KAAK,EAAE,cAAc,EACrB,OAAO,EAAE,WAAW,EACpB,OAAO,GAAE,aAAkB,GAC1B,OAAO,CAAC,YAAY,EAAE,CAAC,CAoEzB"}

package/dist/search/hybrid.js

@@ -0,0 +1,53 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.hybridSearch = hybridSearch;
+const bm25_1 = require("./bm25");
+const vector_1 = require("./vector");
+/**
+ * Combines Vector Search (semantic) and BM25 Search (keyword) for higher accuracy.
+ */
+async function hybridSearch(query, queryEmbedding, chunks, cache, symbols, options = {}) {
+    const alpha = options.alpha ?? 0.7;
+    const topK = options.topK ?? 10;
+    // 1. Get Vector Scores
+    let vectorResults = [];
+    const oversample = topK * 3;
+    if (queryEmbedding && queryEmbedding.length > 0) {
+        vectorResults = (0, vector_1.vectorSearch)(queryEmbedding, chunks, cache, symbols, oversample);
+    }
+    // 2. Get BM25 Scores
+    const bm25Results = (0, bm25_1.computeBM25Scores)(query, chunks, symbols, oversample);
+    // 3. Normalize and Combine
+    const combinedMap = new Map();
+    // Map for easy access and O(1) retrieval
+    const vMap = new Map();
+    vectorResults.forEach((r) => vMap.set(r.chunk.id, { score: r.score, result: r }));
+    const bMap = new Map();
+    // Normalize BM25 scores to [0, 1] relative to the max BM25 score found
+    const maxBM25 = bm25Results.length > 0 ? Math.max(...bm25Results.map((r) => r.score)) : 1;
+    bm25Results.forEach((r) => bMap.set(r.chunk.id, { score: r.score / maxBM25, result: r }));
+    // Iterate over all chunks that appeared in either result
+    const allIds = new Set([...vMap.keys(), ...bMap.keys()]);
+    for (const id of allIds) {
+        const vData = vMap.get(id);
+        const bData = bMap.get(id);
+        const vScore = vData?.score || 0;
+        const bScore = bData?.score || 0;
+        // Weighted combination
+        const finalScore = alpha * vScore + (1 - alpha) * bScore;
+        // Retrieve chunk and symbol in O(1)
+        const result = vData?.result || bData?.result;
+        if (result) {
+            combinedMap.set(id, {
+                chunk: result.chunk,
+                score: finalScore,
+                symbol: result.symbol,
+            });
+        }
+    }
+    // 4. Sort and Slice
+    return Array.from(combinedMap.values())
+        .sort((a, b) => b.score - a.score)
+        .slice(0, topK);
+}
+//# sourceMappingURL=hybrid.js.map

package/dist/search/hybrid.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"hybrid.js","sourceRoot":"","sources":["../../src/search/hybrid.ts"],"names":[],"mappings":";;AAkBA,oCA2EC;AAtFD,iCAA2C;AAC3C,qCAAwC;AAOxC;;GAEG;AACI,KAAK,UAAU,YAAY,CAChC,KAAa,EACb,cAA+B,EAC/B,MAAe,EACf,KAAqB,EACrB,OAAoB,EACpB,UAAyB,EAAE;IAE3B,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC;IACnC,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,EAAE,CAAC;IAEhC,uBAAuB;IACvB,IAAI,aAAa,GAAmB,EAAE,CAAC;IACvC,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,CAAC;IAC5B,IAAI,cAAc,IAAI,cAAc,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAChD,aAAa,GAAG,IAAA,qBAAY,EAC1B,cAAc,EACd,MAAM,EACN,KAAK,EACL,OAAO,EACP,UAAU,CACX,CAAC;IACJ,CAAC;IAED,qBAAqB;IACrB,MAAM,WAAW,GAAG,IAAA,wBAAiB,EAAC,KAAK,EAAE,MAAM,EAAE,OAAO,EAAE,UAAU,CAAC,CAAC;IAE1E,2BAA2B;IAC3B,MAAM,WAAW,GAAG,IAAI,GAAG,EAGxB,CAAC;IAEJ,yCAAyC;IACzC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAmD,CAAC;IACxE,aAAa,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAC1B,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC,CACpD,CAAC;IAEF,MAAM,IAAI,GAAG,IAAI,GAAG,EAAmD,CAAC;IACxE,uEAAuE;IACvE,MAAM,OAAO,GACX,WAAW,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAC5E,WAAW,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CACxB,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,GAAG,OAAO,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC,CAC9D,CAAC;IAEF,yDAAyD;IACzD,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,EAAE,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IAEzD,KAAK,MAAM,EAAE,IAAI,MAAM,EAAE,CAAC;QACxB,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAC3B,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAE3B,MAAM,MAAM,GAAG,KAAK,EAAE,KAAK,IAAI,CAAC,CAAC;QACjC,MAAM,MAAM,GAAG,KAAK,EAAE,KAAK,IAAI,CAAC,CAAC;QAEjC,uBAAuB;QACvB,MAAM,UAAU,GAAG,KAAK,GAAG,MAAM,GAAG,CAAC,CAAC,GAAG,KAAK,CAAC,GAAG,MAAM,CAAC;QAEzD,oCAAoC;QACpC,MAAM,MAAM,GAAG,KAAK,EAAE,MAAM,IAAI,KAAK,EAAE,MAAM,CAAC;QAC9C,IAAI,MAAM,EAAE,CAAC;YACX,WAAW,CAAC,GAAG,CAAC,EAAE,EAAE;gBAClB,KAAK,EAAE,MAAM,CAAC,KAAK;gBACnB,KAAK,EAAE,UAAU;gBACjB,MAAM,EAAE,MAAM,CAAC,MAAM;aACtB,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,oBAAoB;IACpB,OAAO,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,MAAM,EAAE,CAAC;SACpC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC;SACjC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;AACpB,CAAC"}

package/dist/search/merge.d.ts

@@ -0,0 +1,33 @@
+import type { EmbeddingCache, SearchResult } from "../core/types";
+/**
+ * Merges adjacent or overlapping chunks from the same file into a single chunk.
+ *
+ * Two chunks are merged when the next chunk's start line is within `gapThreshold`
+ * lines of the current chunk's end line. Overlapping lines (from sliding-window
+ * chunks) are deduplicated by skipping the already-covered lines when appending.
+ *
+ * Properties of the merged result:
+ *  - `start` / `end` span the full combined line range
+ *  - `content` is reconstructed without duplication
+ *  - `score` is the maximum score among merged chunks
+ *  - `symbol` is preserved only when all merged chunks share the same symbol name
+ *
+ * @param results      Search results (any order).
+ * @param gapThreshold Max line gap between chunks to still merge them (default: 3).
+ */
+export declare function mergeAdjacentChunks(results: SearchResult[], gapThreshold?: number): SearchResult[];
+/**
+ * Removes near-duplicate chunks before token budget is applied.
+ *
+ * Two chunks are considered near-duplicates when:
+ *  - Embeddings available: cosineSimilarity(a, b) > threshold (default 0.92)
+ *  - No embeddings:        Jaccard similarity of word-token sets > 0.80
+ *
+ * Processes results in score order (highest first); earlier chunks win.
+ * The first result is always kept regardless of similarity.
+ *
+ * @param cache     Embedding cache (may be null — falls back to Jaccard)
+ * @param threshold Cosine similarity above which two chunks are near-duplicates
+ */
+export declare function deduplicateBySimilarity(results: SearchResult[], cache: EmbeddingCache | null, threshold?: number): SearchResult[];
+//# sourceMappingURL=merge.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"merge.d.ts","sourceRoot":"","sources":["../../src/search/merge.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,cAAc,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAGlE;;;;;;;;;;;;;;;GAeG;AACH,wBAAgB,mBAAmB,CACjC,OAAO,EAAE,YAAY,EAAE,EACvB,YAAY,SAAI,GACf,YAAY,EAAE,CAwEhB;AAID;;;;;;;;;;;;GAYG;AACH,wBAAgB,uBAAuB,CACrC,OAAO,EAAE,YAAY,EAAE,EACvB,KAAK,EAAE,cAAc,GAAG,IAAI,EAC5B,SAAS,SAAO,GACf,YAAY,EAAE,CAgDhB"}

package/dist/search/merge.js

@@ -0,0 +1,158 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.mergeAdjacentChunks = mergeAdjacentChunks;
+exports.deduplicateBySimilarity = deduplicateBySimilarity;
+const vector_1 = require("./vector");
+/**
+ * Merges adjacent or overlapping chunks from the same file into a single chunk.
+ *
+ * Two chunks are merged when the next chunk's start line is within `gapThreshold`
+ * lines of the current chunk's end line. Overlapping lines (from sliding-window
+ * chunks) are deduplicated by skipping the already-covered lines when appending.
+ *
+ * Properties of the merged result:
+ *  - `start` / `end` span the full combined line range
+ *  - `content` is reconstructed without duplication
+ *  - `score` is the maximum score among merged chunks
+ *  - `symbol` is preserved only when all merged chunks share the same symbol name
+ *
+ * @param results      Search results (any order).
+ * @param gapThreshold Max line gap between chunks to still merge them (default: 3).
+ */
+function mergeAdjacentChunks(results, gapThreshold = 3) {
+    if (results.length <= 1)
+        return results;
+    // Group results by file
+    const byFile = new Map();
+    for (const result of results) {
+        const list = byFile.get(result.chunk.file) ?? [];
+        list.push(result);
+        byFile.set(result.chunk.file, list);
+    }
+    const merged = [];
+    for (const fileResults of byFile.values()) {
+        // Sort ascending by start line so we can merge left-to-right
+        const sorted = [...fileResults].sort((a, b) => a.chunk.start - b.chunk.start);
+        let cur = sorted[0];
+        for (let i = 1; i < sorted.length; i++) {
+            const next = sorted[i];
+            if (next.chunk.start <= cur.chunk.end + gapThreshold) {
+                // Deduplicate overlapping lines: compute how many lines of `next` are
+                // already covered by `cur` based on line numbers, but cap at the actual
+                // content length to handle reconstructed (previously merged) chunks.
+                const overlapByLineNum = Math.max(0, cur.chunk.end - next.chunk.start + 1);
+                const nextLines = next.chunk.content.split("\n");
+                const overlapLines = Math.min(overlapByLineNum, nextLines.length);
+                const appendLines = nextLines.slice(overlapLines);
+                const mergedContent = appendLines.length > 0
+                    ? cur.chunk.content + "\n" + appendLines.join("\n")
+                    : cur.chunk.content;
+                const mergedSymbol = cur.chunk.symbol !== undefined &&
+                    cur.chunk.symbol === next.chunk.symbol
+                    ? cur.chunk.symbol
+                    : undefined;
+                cur = {
+                    chunk: {
+                        id: cur.chunk.id + "+" + next.chunk.id,
+                        file: cur.chunk.file,
+                        start: cur.chunk.start,
+                        end: Math.max(cur.chunk.end, next.chunk.end),
+                        symbol: mergedSymbol,
+                        content: mergedContent,
+                    },
+                    score: Math.max(cur.score, next.score),
+                    symbol: mergedSymbol !== undefined
+                        ? (cur.symbol ?? next.symbol)
+                        : undefined,
+                };
+            }
+            else {
+                merged.push(cur);
+                cur = next;
+            }
+        }
+        merged.push(cur);
+    }
+    // Restore score-descending order
+    return merged.sort((a, b) => b.score - a.score);
+}
+// ─── Semantic deduplication ───────────────────────────────────────────────────
+/**
+ * Removes near-duplicate chunks before token budget is applied.
+ *
+ * Two chunks are considered near-duplicates when:
+ *  - Embeddings available: cosineSimilarity(a, b) > threshold (default 0.92)
+ *  - No embeddings:        Jaccard similarity of word-token sets > 0.80
+ *
+ * Processes results in score order (highest first); earlier chunks win.
+ * The first result is always kept regardless of similarity.
+ *
+ * @param cache     Embedding cache (may be null — falls back to Jaccard)
+ * @param threshold Cosine similarity above which two chunks are near-duplicates
+ */
+function deduplicateBySimilarity(results, cache, threshold = 0.92) {
+    if (results.length <= 1)
+        return results;
+    // Precompute a mapping from chunk ID to embedding for O(1) retrieval during deduplication.
+    let idToIndex;
+    if (cache && cache.chunkIds) {
+        idToIndex = new Map(cache.chunkIds.map((id, i) => [id, i]));
+    }
+    const embeddingOf = (id) => {
+        if (!cache)
+            return null;
+        // Dense binary format (chunkIds + vectors buffer)
+        if (cache.chunkIds && cache.vectors && idToIndex) {
+            const idx = idToIndex.get(id);
+            if (idx !== undefined) {
+                const dims = cache.dimensions;
+                return Array.from(cache.vectors.subarray(idx * dims, idx * dims + dims));
+            }
+        }
+        // Legacy map format
+        return cache.embeddings?.[id] ?? null;
+    };
+    const selected = [];
+    const selectedEmbeddings = [];
+    for (const candidate of results) {
+        const candidateEmb = embeddingOf(candidate.chunk.id);
+        let tooSimilar = false;
+        for (let i = 0; i < selected.length; i++) {
+            const selEmb = selectedEmbeddings[i];
+            if (candidateEmb && selEmb) {
+                tooSimilar = (0, vector_1.cosineSimilarity)(candidateEmb, selEmb) > threshold;
+            }
+            else {
+                tooSimilar =
+                    jaccardSimilarity(candidate.chunk.content, selected[i].chunk.content) >
+                        0.65;
+            }
+            if (tooSimilar)
+                break;
+        }
+        if (!tooSimilar) {
+            selected.push(candidate);
+            selectedEmbeddings.push(candidateEmb);
+        }
+    }
+    return selected;
+}
+function jaccardSimilarity(a, b) {
+    const tokA = wordTokens(a);
+    const tokB = wordTokens(b);
+    if (tokA.size === 0 && tokB.size === 0)
+        return 1;
+    let intersection = 0;
+    for (const t of tokA)
+        if (tokB.has(t))
+            intersection++;
+    const union = tokA.size + tokB.size - intersection;
+    return union === 0 ? 0 : intersection / union;
+}
+function wordTokens(text) {
+    return new Set(text
+        .toLowerCase()
+        .split(/[\s\W]+/)
+        .filter((t) => t.length >= 2));
+}
+//# sourceMappingURL=merge.js.map

package/dist/search/merge.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"merge.js","sourceRoot":"","sources":["../../src/search/merge.ts"],"names":[],"mappings":";;AAmBA,kDA2EC;AAiBD,0DAoDC;AAlKD,qCAA4C;AAE5C;;;;;;;;;;;;;;;GAeG;AACH,SAAgB,mBAAmB,CACjC,OAAuB,EACvB,YAAY,GAAG,CAAC;IAEhB,IAAI,OAAO,CAAC,MAAM,IAAI,CAAC;QAAE,OAAO,OAAO,CAAC;IAExC,wBAAwB;IACxB,MAAM,MAAM,GAAG,IAAI,GAAG,EAA0B,CAAC;IACjD,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;QACjD,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAClB,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;IACtC,CAAC;IAED,MAAM,MAAM,GAAmB,EAAE,CAAC;IAElC,KAAK,MAAM,WAAW,IAAI,MAAM,CAAC,MAAM,EAAE,EAAE,CAAC;QAC1C,6DAA6D;QAC7D,MAAM,MAAM,GAAG,CAAC,GAAG,WAAW,CAAC,CAAC,IAAI,CAClC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,KAAK,CACxC,CAAC;QACF,IAAI,GAAG,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;QAEpB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACvC,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;YAEvB,IAAI,IAAI,CAAC,KAAK,CAAC,KAAK,IAAI,GAAG,CAAC,KAAK,CAAC,GAAG,GAAG,YAAY,EAAE,CAAC;gBACrD,sEAAsE;gBACtE,wEAAwE;gBACxE,qEAAqE;gBACrE,MAAM,gBAAgB,GAAG,IAAI,CAAC,GAAG,CAC/B,CAAC,EACD,GAAG,CAAC,KAAK,CAAC,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CACrC,CAAC;gBACF,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;gBACjD,MAAM,YAAY,GAAG,IAAI,CAAC,GAAG,CAAC,gBAAgB,EAAE,SAAS,CAAC,MAAM,CAAC,CAAC;gBAClE,MAAM,WAAW,GAAG,SAAS,CAAC,KAAK,CAAC,YAAY,CAAC,CAAC;gBAElD,MAAM,aAAa,GACjB,WAAW,CAAC,MAAM,GAAG,CAAC;oBACpB,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,GAAG,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC;oBACnD,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC;gBAExB,MAAM,YAAY,GAChB,GAAG,CAAC,KAAK,CAAC,MAAM,KAAK,SAAS;oBAC9B,GAAG,CAAC,KAAK,CAAC,MAAM,KAAK,IAAI,CAAC,KAAK,CAAC,MAAM;oBACpC,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM;oBAClB,CAAC,CAAC,SAAS,CAAC;gBAEhB,GAAG,GAAG;oBACJ,KAAK,EAAE;wBACL,EAAE,EAAE,GAAG,CAAC,KAAK,CAAC,EAAE,GAAG,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE;wBACtC,IAAI,EAAE,GAAG,CAAC,KAAK,CAAC,IAAI;wBACpB,KAAK,EAAE,GAAG,CAAC,KAAK,CAAC,KAAK;wBACtB,GAAG,EAAE,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC;wBAC5C,MAAM,EAAE,YAAY;wBACpB,OAAO,EAAE,aAAa;qBACvB;oBACD,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC;oBACtC,MAAM,EACJ,YAAY,KAAK,SAAS;wBACxB,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC;wBAC7B,CAAC,CAAC,SAAS;iBAChB,CAAC;YACJ,CAAC;iBAAM,CAAC;gBACN,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBACjB,GAAG,GAAG,IAAI,CAAC;YACb,CAAC;QACH,CAAC;QAED,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACnB,CAAC;IAED,iCAAiC;IACjC,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;AAClD,CAAC;AAED,iFAAiF;AAEjF;;;;;;;;;;;;GAYG;AACH,SAAgB,uBAAuB,CACrC,OAAuB,EACvB,KAA4B,EAC5B,SAAS,GAAG,IAAI;IAEhB,IAAI,OAAO,CAAC,MAAM,IAAI,CAAC;QAAE,OAAO,OAAO,CAAC;IAExC,2FAA2F;IAC3F,IAAI,SAA0C,CAAC;IAC/C,IAAI,KAAK,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;QAC5B,SAAS,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAC9D,CAAC;IACD,MAAM,WAAW,GAAG,CAAC,EAAU,EAAmB,EAAE;QAClD,IAAI,CAAC,KAAK;YAAE,OAAO,IAAI,CAAC;QACxB,kDAAkD;QAClD,IAAI,KAAK,CAAC,QAAQ,IAAI,KAAK,CAAC,OAAO,IAAI,SAAS,EAAE,CAAC;YACjD,MAAM,GAAG,GAAG,SAAS,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YAC9B,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;gBACtB,MAAM,IAAI,GAAG,KAAK,CAAC,UAAU,CAAC;gBAC9B,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,QAAQ,CAAC,GAAG,GAAG,IAAI,EAAE,GAAG,GAAG,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC;YAC3E,CAAC;QACH,CAAC;QACD,oBAAoB;QACpB,OAAO,KAAK,CAAC,UAAU,EAAE,CAAC,EAAE,CAAC,IAAI,IAAI,CAAC;IACxC,CAAC,CAAC;IAEF,MAAM,QAAQ,GAAmB,EAAE,CAAC;IACpC,MAAM,kBAAkB,GAA2B,EAAE,CAAC;IAEtD,KAAK,MAAM,SAAS,IAAI,OAAO,EAAE,CAAC;QAChC,MAAM,YAAY,GAAG,WAAW,CAAC,SAAS,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QACrD,IAAI,UAAU,GAAG,KAAK,CAAC;QAEvB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACzC,MAAM,MAAM,GAAG,kBAAkB,CAAC,CAAC,CAAC,CAAC;YACrC,IAAI,YAAY,IAAI,MAAM,EAAE,CAAC;gBAC3B,UAAU,GAAG,IAAA,yBAAgB,EAAC,YAAY,EAAE,MAAM,CAAC,GAAG,SAAS,CAAC;YAClE,CAAC;iBAAM,CAAC;gBACN,UAAU;oBACR,iBAAiB,CAAC,SAAS,CAAC,KAAK,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC;wBACrE,IAAI,CAAC;YACT,CAAC;YACD,IAAI,UAAU;gBAAE,MAAM;QACxB,CAAC;QAED,IAAI,CAAC,UAAU,EAAE,CAAC;YAChB,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;YACzB,kBAAkB,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;QACxC,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,SAAS,iBAAiB,CAAC,CAAS,EAAE,CAAS;IAC7C,MAAM,IAAI,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;IAC3B,MAAM,IAAI,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;IAC3B,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC,IAAI,IAAI,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAEjD,IAAI,YAAY,GAAG,CAAC,CAAC;IACrB,KAAK,MAAM,CAAC,IAAI,IAAI;QAAE,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;YAAE,YAAY,EAAE,CAAC;IACtD,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,GAAG,YAAY,CAAC;IACnD,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,YAAY,GAAG,KAAK,CAAC;AAChD,CAAC;AAED,SAAS,UAAU,CAAC,IAAY;IAC9B,OAAO,IAAI,GAAG,CACZ,IAAI;SACD,WAAW,EAAE;SACb,KAAK,CAAC,SAAS,CAAC;SAChB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC,CAChC,CAAC;AACJ,CAAC"}

package/dist/search/mmr.d.ts

@@ -0,0 +1,23 @@
+import type { EmbeddingCache, SearchResult } from "../core/types";
+/**
+ * Maximal Marginal Relevance (MMR) reranking.
+ *
+ * Selects up to `topK` results from `candidates` by iteratively picking
+ * the candidate that maximises:
+ *
+ *   MMR(d) = lambda * relevance(d) - (1 - lambda) * max_sim(d, selected)
+ *
+ * where:
+ *   - relevance(d) is the original retrieval score (normalised to [0, 1])
+ *   - max_sim(d, selected) is the maximum cosine similarity between d and
+ *     any already-selected result (requires embeddings in the cache)
+ *
+ * lambda=1.0 → pure relevance ordering (identical to original ranking)
+ * lambda=0.5 → balanced relevance / diversity (default)
+ * lambda=0.0 → maximum diversity, ignores relevance
+ *
+ * Falls back to returning `candidates.slice(0, topK)` when embeddings
+ * are unavailable (keyword-only mode).
+ */
+export declare function applyMMR(candidates: SearchResult[], cache: EmbeddingCache | null, topK: number, lambda?: number): SearchResult[];
+//# sourceMappingURL=mmr.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"mmr.d.ts","sourceRoot":"","sources":["../../src/search/mmr.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,cAAc,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAGlE;;;;;;;;;;;;;;;;;;;GAmBG;AACH,wBAAgB,QAAQ,CACtB,UAAU,EAAE,YAAY,EAAE,EAC1B,KAAK,EAAE,cAAc,GAAG,IAAI,EAC5B,IAAI,EAAE,MAAM,EACZ,MAAM,SAAM,GACX,YAAY,EAAE,CAmFhB"}

package/dist/search/mmr.js

@@ -0,0 +1,95 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.applyMMR = applyMMR;
+const vector_1 = require("./vector");
+/**
+ * Maximal Marginal Relevance (MMR) reranking.
+ *
+ * Selects up to `topK` results from `candidates` by iteratively picking
+ * the candidate that maximises:
+ *
+ *   MMR(d) = lambda * relevance(d) - (1 - lambda) * max_sim(d, selected)
+ *
+ * where:
+ *   - relevance(d) is the original retrieval score (normalised to [0, 1])
+ *   - max_sim(d, selected) is the maximum cosine similarity between d and
+ *     any already-selected result (requires embeddings in the cache)
+ *
+ * lambda=1.0 → pure relevance ordering (identical to original ranking)
+ * lambda=0.5 → balanced relevance / diversity (default)
+ * lambda=0.0 → maximum diversity, ignores relevance
+ *
+ * Falls back to returning `candidates.slice(0, topK)` when embeddings
+ * are unavailable (keyword-only mode).
+ */
+function applyMMR(candidates, cache, topK, lambda = 0.5) {
+    if (candidates.length <= 1)
+        return candidates.slice(0, topK);
+    // Without vector data MMR cannot compute inter-result similarity.
+    if (!cache?.vectors || !cache.chunkIds || !cache.dimensions) {
+        return candidates.slice(0, topK);
+    }
+    const { vectors, chunkIds, dimensions } = cache;
+    // Build a fast chunk-id → flat-buffer-index map.
+    const idToIdx = new Map();
+    chunkIds.forEach((id, i) => idToIdx.set(id, i));
+    // Normalise relevance scores to [0, 1] so lambda is meaningful regardless
+    // of whether scores come from cosine similarity (already ~[0,1]) or BM25.
+    // Use reduce instead of Math.max(...spread) to avoid call-stack overflow on large arrays.
+    let maxScore = -Infinity;
+    let minScore = Infinity;
+    for (const r of candidates) {
+        if (r.score > maxScore)
+            maxScore = r.score;
+        if (r.score < minScore)
+            minScore = r.score;
+    }
+    const range = maxScore - minScore || 1;
+    const relNorm = candidates.map((r) => (r.score - minScore) / range);
+    // Pre-extract each candidate's vector as number[] so the inner loop avoids
+    // repeated Array.from() allocations (O(n² * d) → O(n * d) allocations).
+    const candidateVecs = candidates.map((r) => {
+        const idx = idToIdx.get(r.chunk.id);
+        if (idx === undefined)
+            return null;
+        const offset = idx * dimensions;
+        return Array.from(vectors.subarray(offset, offset + dimensions));
+    });
+    const selected = [];
+    // Flat-buffer indices of already-selected results (for similarity queries).
+    const selectedVecIdx = [];
+    // Track remaining candidates as index positions into `candidates`.
+    const remaining = candidates.map((_, i) => i);
+    while (selected.length < topK && remaining.length > 0) {
+        let bestPos = -1; // position in `remaining`
+        let bestMMR = -Infinity;
+        for (let pos = 0; pos < remaining.length; pos++) {
+            const candIdx = remaining[pos];
+            const relevance = relNorm[candIdx];
+            let maxSim = 0;
+            if (selectedVecIdx.length > 0) {
+                const vecA = candidateVecs[candIdx];
+                if (vecA !== null) {
+                    for (const selIdx of selectedVecIdx) {
+                        const sim = (0, vector_1.cosineSimilarityBinary)(vecA, vectors, selIdx * dimensions, dimensions);
+                        if (sim > maxSim)
+                            maxSim = sim;
+                    }
+                }
+            }
+            const mmrScore = lambda * relevance - (1 - lambda) * maxSim;
+            if (mmrScore > bestMMR) {
+                bestMMR = mmrScore;
+                bestPos = pos;
+            }
+        }
+        const chosenIdx = remaining[bestPos];
+        selected.push(candidates[chosenIdx]);
+        const vecIdx = idToIdx.get(candidates[chosenIdx].chunk.id);
+        if (vecIdx !== undefined)
+            selectedVecIdx.push(vecIdx);
+        remaining.splice(bestPos, 1);
+    }
+    return selected;
+}
+//# sourceMappingURL=mmr.js.map

package/dist/search/mmr.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"mmr.js","sourceRoot":"","sources":["../../src/search/mmr.ts"],"names":[],"mappings":";;AAuBA,4BAwFC;AA9GD,qCAAkD;AAElD;;;;;;;;;;;;;;;;;;;GAmBG;AACH,SAAgB,QAAQ,CACtB,UAA0B,EAC1B,KAA4B,EAC5B,IAAY,EACZ,MAAM,GAAG,GAAG;IAEZ,IAAI,UAAU,CAAC,MAAM,IAAI,CAAC;QAAE,OAAO,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;IAE7D,kEAAkE;IAClE,IAAI,CAAC,KAAK,EAAE,OAAO,IAAI,CAAC,KAAK,CAAC,QAAQ,IAAI,CAAC,KAAK,CAAC,UAAU,EAAE,CAAC;QAC5D,OAAO,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;IACnC,CAAC;IAED,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,UAAU,EAAE,GAAG,KAAK,CAAC;IAEhD,iDAAiD;IACjD,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC1C,QAAQ,CAAC,OAAO,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC;IAEhD,0EAA0E;IAC1E,0EAA0E;IAC1E,0FAA0F;IAC1F,IAAI,QAAQ,GAAG,CAAC,QAAQ,CAAC;IACzB,IAAI,QAAQ,GAAG,QAAQ,CAAC;IACxB,KAAK,MAAM,CAAC,IAAI,UAAU,EAAE,CAAC;QAC3B,IAAI,CAAC,CAAC,KAAK,GAAG,QAAQ;YAAE,QAAQ,GAAG,CAAC,CAAC,KAAK,CAAC;QAC3C,IAAI,CAAC,CAAC,KAAK,GAAG,QAAQ;YAAE,QAAQ,GAAG,CAAC,CAAC,KAAK,CAAC;IAC7C,CAAC;IACD,MAAM,KAAK,GAAG,QAAQ,GAAG,QAAQ,IAAI,CAAC,CAAC;IACvC,MAAM,OAAO,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,QAAQ,CAAC,GAAG,KAAK,CAAC,CAAC;IAEpE,2EAA2E;IAC3E,wEAAwE;IACxE,MAAM,aAAa,GAA2B,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;QACjE,MAAM,GAAG,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QACpC,IAAI,GAAG,KAAK,SAAS;YAAE,OAAO,IAAI,CAAC;QACnC,MAAM,MAAM,GAAG,GAAG,GAAG,UAAU,CAAC;QAChC,OAAO,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,MAAM,EAAE,MAAM,GAAG,UAAU,CAAC,CAAC,CAAC;IACnE,CAAC,CAAC,CAAC;IAEH,MAAM,QAAQ,GAAmB,EAAE,CAAC;IACpC,4EAA4E;IAC5E,MAAM,cAAc,GAAa,EAAE,CAAC;IAEpC,mEAAmE;IACnE,MAAM,SAAS,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC;IAE9C,OAAO,QAAQ,CAAC,MAAM,GAAG,IAAI,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtD,IAAI,OAAO,GAAG,CAAC,CAAC,CAAC,CAAC,0BAA0B;QAC5C,IAAI,OAAO,GAAG,CAAC,QAAQ,CAAC;QAExB,KAAK,IAAI,GAAG,GAAG,CAAC,EAAE,GAAG,GAAG,SAAS,CAAC,MAAM,EAAE,GAAG,EAAE,EAAE,CAAC;YAChD,MAAM,OAAO,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC;YAC/B,MAAM,SAAS,GAAG,OAAO,CAAC,OAAO,CAAC,CAAC;YAEnC,IAAI,MAAM,GAAG,CAAC,CAAC;YACf,IAAI,cAAc,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC9B,MAAM,IAAI,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC;gBACpC,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;oBAClB,KAAK,MAAM,MAAM,IAAI,cAAc,EAAE,CAAC;wBACpC,MAAM,GAAG,GAAG,IAAA,+BAAsB,EAChC,IAAI,EACJ,OAAO,EACP,MAAM,GAAG,UAAU,EACnB,UAAU,CACX,CAAC;wBACF,IAAI,GAAG,GAAG,MAAM;4BAAE,MAAM,GAAG,GAAG,CAAC;oBACjC,CAAC;gBACH,CAAC;YACH,CAAC;YAED,MAAM,QAAQ,GAAG,MAAM,GAAG,SAAS,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,GAAG,MAAM,CAAC;YAC5D,IAAI,QAAQ,GAAG,OAAO,EAAE,CAAC;gBACvB,OAAO,GAAG,QAAQ,CAAC;gBACnB,OAAO,GAAG,GAAG,CAAC;YAChB,CAAC;QACH,CAAC;QAED,MAAM,SAAS,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC;QACrC,QAAQ,CAAC,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,CAAC,CAAC;QAErC,MAAM,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC,UAAU,CAAC,SAAS,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;QAC3D,IAAI,MAAM,KAAK,SAAS;YAAE,cAAc,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAEtD,SAAS,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;IAC/B,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC"}

package/dist/search/rerank.d.ts

@@ -0,0 +1,12 @@
+import type { RerankConfig, SearchResult } from "../core/types";
+/**
+ * Re-scores search results using the configured reranker.
+ *
+ * @param query       The user's natural language query
+ * @param results     Initial search results (from vector or keyword search)
+ * @param topK        Number of results to return after reranking
+ * @param config      Reranker config (default: xenova cross-encoder)
+ * @param fallbackModel  Model name to use when config.model is not set (ollama only)
+ */
+export declare function rerankResults(query: string, results: SearchResult[], topK?: number, config?: RerankConfig, fallbackModel?: string): Promise<SearchResult[]>;
+//# sourceMappingURL=rerank.d.ts.map