@staticn0va/wigolo 0.4.0 → 0.6.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/SKILL.md +279 -68
- package/dist/agent/executor.d.ts +13 -0
- package/dist/agent/executor.d.ts.map +1 -0
- package/dist/agent/executor.js +128 -0
- package/dist/agent/executor.js.map +1 -0
- package/dist/agent/pipeline.d.ts +5 -0
- package/dist/agent/pipeline.d.ts.map +1 -0
- package/dist/agent/pipeline.js +198 -0
- package/dist/agent/pipeline.js.map +1 -0
- package/dist/agent/planner.d.ts +9 -0
- package/dist/agent/planner.d.ts.map +1 -0
- package/dist/agent/planner.js +190 -0
- package/dist/agent/planner.js.map +1 -0
- package/dist/cache/db.d.ts.map +1 -1
- package/dist/cache/db.js +32 -0
- package/dist/cache/db.js.map +1 -1
- package/dist/cache/store.d.ts +14 -0
- package/dist/cache/store.d.ts.map +1 -1
- package/dist/cache/store.js +69 -0
- package/dist/cache/store.js.map +1 -1
- package/dist/cli/warmup.d.ts +4 -0
- package/dist/cli/warmup.d.ts.map +1 -1
- package/dist/cli/warmup.js +58 -0
- package/dist/cli/warmup.js.map +1 -1
- package/dist/config.d.ts +8 -0
- package/dist/config.d.ts.map +1 -1
- package/dist/config.js +8 -0
- package/dist/config.js.map +1 -1
- package/dist/embedding/embed.d.ts +19 -0
- package/dist/embedding/embed.d.ts.map +1 -0
- package/dist/embedding/embed.js +131 -0
- package/dist/embedding/embed.js.map +1 -0
- package/dist/embedding/key-terms.d.ts +12 -0
- package/dist/embedding/key-terms.d.ts.map +1 -0
- package/dist/embedding/key-terms.js +138 -0
- package/dist/embedding/key-terms.js.map +1 -0
- package/dist/embedding/subprocess.d.ts +31 -0
- package/dist/embedding/subprocess.d.ts.map +1 -0
- package/dist/embedding/subprocess.js +213 -0
- package/dist/embedding/subprocess.js.map +1 -0
- package/dist/embedding/vector-index.d.ts +26 -0
- package/dist/embedding/vector-index.d.ts.map +1 -0
- package/dist/embedding/vector-index.js +78 -0
- package/dist/embedding/vector-index.js.map +1 -0
- package/dist/fetch/browser-pool.d.ts.map +1 -1
- package/dist/fetch/browser-pool.js +61 -0
- package/dist/fetch/browser-pool.js.map +1 -1
- package/dist/fetch/browser-types.js +1 -1
- package/dist/fetch/browser-types.js.map +1 -1
- package/dist/fetch/lightpanda.d.ts +28 -0
- package/dist/fetch/lightpanda.d.ts.map +1 -0
- package/dist/fetch/lightpanda.js +177 -0
- package/dist/fetch/lightpanda.js.map +1 -0
- package/dist/fetch/router.d.ts +1 -0
- package/dist/fetch/router.d.ts.map +1 -1
- package/dist/fetch/router.js.map +1 -1
- package/dist/instructions.d.ts +9 -6
- package/dist/instructions.d.ts.map +1 -1
- package/dist/instructions.js +111 -21
- package/dist/instructions.js.map +1 -1
- package/dist/logger.d.ts +1 -1
- package/dist/logger.d.ts.map +1 -1
- package/dist/repl/commands/agent.d.ts +5 -0
- package/dist/repl/commands/agent.d.ts.map +1 -0
- package/dist/repl/commands/agent.js +48 -0
- package/dist/repl/commands/agent.js.map +1 -0
- package/dist/repl/commands/find-similar.d.ts +5 -0
- package/dist/repl/commands/find-similar.d.ts.map +1 -0
- package/dist/repl/commands/find-similar.js +61 -0
- package/dist/repl/commands/find-similar.js.map +1 -0
- package/dist/repl/commands/research.d.ts +5 -0
- package/dist/repl/commands/research.d.ts.map +1 -0
- package/dist/repl/commands/research.js +50 -0
- package/dist/repl/commands/research.js.map +1 -0
- package/dist/repl/formatters.d.ts +4 -1
- package/dist/repl/formatters.d.ts.map +1 -1
- package/dist/repl/formatters.js +73 -0
- package/dist/repl/formatters.js.map +1 -1
- package/dist/repl/shell.d.ts.map +1 -1
- package/dist/repl/shell.js +22 -1
- package/dist/repl/shell.js.map +1 -1
- package/dist/research/decompose.d.ts +7 -0
- package/dist/research/decompose.d.ts.map +1 -0
- package/dist/research/decompose.js +195 -0
- package/dist/research/decompose.js.map +1 -0
- package/dist/research/pipeline.d.ts +5 -0
- package/dist/research/pipeline.d.ts.map +1 -0
- package/dist/research/pipeline.js +135 -0
- package/dist/research/pipeline.js.map +1 -0
- package/dist/research/synthesize.d.ts +10 -0
- package/dist/research/synthesize.d.ts.map +1 -0
- package/dist/research/synthesize.js +119 -0
- package/dist/research/synthesize.js.map +1 -0
- package/dist/search/answer-synthesis.d.ts +13 -0
- package/dist/search/answer-synthesis.d.ts.map +1 -0
- package/dist/search/answer-synthesis.js +120 -0
- package/dist/search/answer-synthesis.js.map +1 -0
- package/dist/search/find-similar.d.ts +5 -0
- package/dist/search/find-similar.d.ts.map +1 -0
- package/dist/search/find-similar.js +436 -0
- package/dist/search/find-similar.js.map +1 -0
- package/dist/search/multi-query.d.ts +22 -0
- package/dist/search/multi-query.d.ts.map +1 -0
- package/dist/search/multi-query.js +157 -0
- package/dist/search/multi-query.js.map +1 -0
- package/dist/search/rrf.d.ts +17 -0
- package/dist/search/rrf.d.ts.map +1 -0
- package/dist/search/rrf.js +48 -0
- package/dist/search/rrf.js.map +1 -0
- package/dist/search/sampling.d.ts +25 -0
- package/dist/search/sampling.d.ts.map +1 -0
- package/dist/search/sampling.js +52 -0
- package/dist/search/sampling.js.map +1 -0
- package/dist/server.d.ts.map +1 -1
- package/dist/server.js +210 -5
- package/dist/server.js.map +1 -1
- package/dist/tools/agent.d.ts +5 -0
- package/dist/tools/agent.d.ts.map +1 -0
- package/dist/tools/agent.js +67 -0
- package/dist/tools/agent.js.map +1 -0
- package/dist/tools/fetch.d.ts.map +1 -1
- package/dist/tools/fetch.js +16 -4
- package/dist/tools/fetch.js.map +1 -1
- package/dist/tools/find-similar.d.ts +5 -0
- package/dist/tools/find-similar.d.ts.map +1 -0
- package/dist/tools/find-similar.js +48 -0
- package/dist/tools/find-similar.js.map +1 -0
- package/dist/tools/research.d.ts +5 -0
- package/dist/tools/research.d.ts.map +1 -0
- package/dist/tools/research.js +50 -0
- package/dist/tools/research.js.map +1 -0
- package/dist/tools/search.d.ts +3 -2
- package/dist/tools/search.d.ts.map +1 -1
- package/dist/tools/search.js +217 -14
- package/dist/tools/search.js.map +1 -1
- package/dist/types.d.ts +108 -3
- package/dist/types.d.ts.map +1 -1
- package/package.json +10 -3
|
@@ -0,0 +1,436 @@
|
|
|
1
|
+
import { extractKeyTerms, buildFTS5Query } from '../embedding/key-terms.js';
|
|
2
|
+
import { reciprocalRankFusion, sortByRRFScore } from './rrf.js';
|
|
3
|
+
import { searchCache, getCachedContent, normalizeUrl } from '../cache/store.js';
|
|
4
|
+
import { filterByDomains } from './filters.js';
|
|
5
|
+
import { handleSearch } from '../tools/search.js';
|
|
6
|
+
import { extractContent } from '../extraction/pipeline.js';
|
|
7
|
+
import { getEmbeddingService } from '../embedding/embed.js';
|
|
8
|
+
import { createLogger } from '../logger.js';
|
|
9
|
+
const log = createLogger('search');
|
|
10
|
+
const DEFAULT_MAX_RESULTS = 10;
|
|
11
|
+
const MAX_FTS5_CANDIDATES = 20;
|
|
12
|
+
const MAX_EMBEDDING_CANDIDATES = 20;
|
|
13
|
+
const WEB_SEARCH_QUERY_COUNT = 3;
|
|
14
|
+
export async function findSimilar(input, engines, router, backendStatus) {
|
|
15
|
+
const start = Date.now();
|
|
16
|
+
// Probe embedding availability once up front for the whole request
|
|
17
|
+
const embeddingAvailable = checkEmbeddingAvailable();
|
|
18
|
+
try {
|
|
19
|
+
const url = input.url?.trim();
|
|
20
|
+
const concept = input.concept?.trim();
|
|
21
|
+
if (!url && !concept) {
|
|
22
|
+
return {
|
|
23
|
+
results: [],
|
|
24
|
+
method: 'fts5',
|
|
25
|
+
cache_hits: 0,
|
|
26
|
+
search_hits: 0,
|
|
27
|
+
embedding_available: embeddingAvailable,
|
|
28
|
+
error: 'Either url or concept must be provided',
|
|
29
|
+
total_time_ms: Date.now() - start,
|
|
30
|
+
};
|
|
31
|
+
}
|
|
32
|
+
const maxResults = input.max_results ?? DEFAULT_MAX_RESULTS;
|
|
33
|
+
const includeCache = input.include_cache ?? true;
|
|
34
|
+
const includeWeb = input.include_web ?? true;
|
|
35
|
+
const signal = await prepareSignal(url, concept, router);
|
|
36
|
+
if (signal.terms.length === 0 && !signal.queryText) {
|
|
37
|
+
log.warn('no key terms or query text extracted, falling back to web search');
|
|
38
|
+
if (!includeWeb) {
|
|
39
|
+
return {
|
|
40
|
+
results: [],
|
|
41
|
+
method: 'fts5',
|
|
42
|
+
cache_hits: 0,
|
|
43
|
+
search_hits: 0,
|
|
44
|
+
embedding_available: embeddingAvailable,
|
|
45
|
+
error: 'Could not extract key terms from input and web search is disabled',
|
|
46
|
+
total_time_ms: Date.now() - start,
|
|
47
|
+
};
|
|
48
|
+
}
|
|
49
|
+
}
|
|
50
|
+
// Phase 1: FTS5 + embedding in parallel (both hit local state, cheap)
|
|
51
|
+
let cacheResults = [];
|
|
52
|
+
const fts5RankMap = new Map();
|
|
53
|
+
let embeddingResults = [];
|
|
54
|
+
const embeddingRankMap = new Map();
|
|
55
|
+
await Promise.all([
|
|
56
|
+
(async () => {
|
|
57
|
+
if (includeCache && signal.terms.length > 0) {
|
|
58
|
+
cacheResults = runFTS5Search(signal.terms, signal.inputNormalizedUrl, input.include_domains, input.exclude_domains, MAX_FTS5_CANDIDATES, fts5RankMap);
|
|
59
|
+
log.debug('FTS5 search complete', { hits: cacheResults.length });
|
|
60
|
+
}
|
|
61
|
+
})(),
|
|
62
|
+
(async () => {
|
|
63
|
+
if (includeCache && embeddingAvailable && signal.queryText) {
|
|
64
|
+
embeddingResults = await runEmbeddingSearch(signal.queryText, signal.inputNormalizedUrl, input.include_domains, input.exclude_domains, MAX_EMBEDDING_CANDIDATES, embeddingRankMap);
|
|
65
|
+
log.debug('embedding search complete', { hits: embeddingResults.length });
|
|
66
|
+
}
|
|
67
|
+
})(),
|
|
68
|
+
]);
|
|
69
|
+
// Phase 2: Web search fallback (only if combined unique local hits < maxResults)
|
|
70
|
+
let searchResults = [];
|
|
71
|
+
const searchRankMap = new Map();
|
|
72
|
+
const combinedLocalHits = new Set();
|
|
73
|
+
for (const r of cacheResults)
|
|
74
|
+
combinedLocalHits.add(safeNormalize(r.url));
|
|
75
|
+
for (const r of embeddingResults)
|
|
76
|
+
combinedLocalHits.add(safeNormalize(r.url));
|
|
77
|
+
if (combinedLocalHits.size < maxResults && includeWeb) {
|
|
78
|
+
searchResults = await runWebSearchFallback(signal, engines, router, backendStatus, maxResults, signal.inputNormalizedUrl, input.include_domains, input.exclude_domains, searchRankMap);
|
|
79
|
+
log.debug('web search fallback complete', { hits: searchResults.length });
|
|
80
|
+
}
|
|
81
|
+
// Phase 3: 3-way RRF fusion
|
|
82
|
+
const rankedLists = [];
|
|
83
|
+
if (fts5RankMap.size > 0)
|
|
84
|
+
rankedLists.push(fts5RankMap);
|
|
85
|
+
if (embeddingRankMap.size > 0)
|
|
86
|
+
rankedLists.push(embeddingRankMap);
|
|
87
|
+
if (searchRankMap.size > 0)
|
|
88
|
+
rankedLists.push(searchRankMap);
|
|
89
|
+
const allResults = mergeResults(cacheResults, embeddingResults, searchResults);
|
|
90
|
+
let finalResults;
|
|
91
|
+
if (rankedLists.length >= 1) {
|
|
92
|
+
finalResults = fuseResults(rankedLists, allResults, maxResults);
|
|
93
|
+
}
|
|
94
|
+
else {
|
|
95
|
+
finalResults = allResults
|
|
96
|
+
.sort((a, b) => b.relevance_score - a.relevance_score)
|
|
97
|
+
.slice(0, maxResults);
|
|
98
|
+
}
|
|
99
|
+
const method = determineMethod(cacheResults.length > 0, embeddingResults.length > 0, searchResults.length > 0);
|
|
100
|
+
const cacheHits = finalResults.filter(r => r.source === 'cache').length;
|
|
101
|
+
const searchHits = finalResults.filter(r => r.source === 'search').length;
|
|
102
|
+
return {
|
|
103
|
+
results: finalResults,
|
|
104
|
+
method,
|
|
105
|
+
cache_hits: cacheHits,
|
|
106
|
+
search_hits: searchHits,
|
|
107
|
+
embedding_available: embeddingAvailable,
|
|
108
|
+
total_time_ms: Date.now() - start,
|
|
109
|
+
};
|
|
110
|
+
}
|
|
111
|
+
catch (err) {
|
|
112
|
+
log.error('findSimilar failed', { error: String(err) });
|
|
113
|
+
return {
|
|
114
|
+
results: [],
|
|
115
|
+
method: 'fts5',
|
|
116
|
+
cache_hits: 0,
|
|
117
|
+
search_hits: 0,
|
|
118
|
+
embedding_available: embeddingAvailable,
|
|
119
|
+
error: `find_similar failed: ${err instanceof Error ? err.message : String(err)}`,
|
|
120
|
+
total_time_ms: Date.now() - start,
|
|
121
|
+
};
|
|
122
|
+
}
|
|
123
|
+
}
|
|
124
|
+
function checkEmbeddingAvailable() {
|
|
125
|
+
try {
|
|
126
|
+
const svc = getEmbeddingService();
|
|
127
|
+
return svc.isAvailable() && svc.getIndex().size() > 0;
|
|
128
|
+
}
|
|
129
|
+
catch {
|
|
130
|
+
return false;
|
|
131
|
+
}
|
|
132
|
+
}
|
|
133
|
+
function safeNormalize(url) {
|
|
134
|
+
try {
|
|
135
|
+
return normalizeUrl(url);
|
|
136
|
+
}
|
|
137
|
+
catch {
|
|
138
|
+
return url;
|
|
139
|
+
}
|
|
140
|
+
}
|
|
141
|
+
function mergeResults(...lists) {
|
|
142
|
+
const seen = new Map();
|
|
143
|
+
for (const list of lists) {
|
|
144
|
+
for (const r of list) {
|
|
145
|
+
const key = safeNormalize(r.url);
|
|
146
|
+
const existing = seen.get(key);
|
|
147
|
+
if (!existing) {
|
|
148
|
+
seen.set(key, r);
|
|
149
|
+
}
|
|
150
|
+
else {
|
|
151
|
+
// Merge match_signals so fused result records the most-specific source info
|
|
152
|
+
existing.match_signals = {
|
|
153
|
+
...existing.match_signals,
|
|
154
|
+
...r.match_signals,
|
|
155
|
+
fused_score: existing.match_signals.fused_score,
|
|
156
|
+
};
|
|
157
|
+
}
|
|
158
|
+
}
|
|
159
|
+
}
|
|
160
|
+
return [...seen.values()];
|
|
161
|
+
}
|
|
162
|
+
async function prepareSignal(url, concept, router) {
|
|
163
|
+
if (url) {
|
|
164
|
+
return await prepareSignalFromUrl(url, router);
|
|
165
|
+
}
|
|
166
|
+
if (concept) {
|
|
167
|
+
const terms = extractKeyTerms(concept, '');
|
|
168
|
+
return { terms, title: concept, queryText: concept };
|
|
169
|
+
}
|
|
170
|
+
return { terms: [], title: '' };
|
|
171
|
+
}
|
|
172
|
+
async function prepareSignalFromUrl(url, router) {
|
|
173
|
+
let normalizedInputUrl;
|
|
174
|
+
try {
|
|
175
|
+
normalizedInputUrl = normalizeUrl(url);
|
|
176
|
+
}
|
|
177
|
+
catch {
|
|
178
|
+
normalizedInputUrl = url;
|
|
179
|
+
}
|
|
180
|
+
const cached = getCachedContent(url);
|
|
181
|
+
if (cached) {
|
|
182
|
+
const terms = extractKeyTerms(cached.markdown, cached.title);
|
|
183
|
+
return {
|
|
184
|
+
terms,
|
|
185
|
+
title: cached.title,
|
|
186
|
+
inputUrl: url,
|
|
187
|
+
inputNormalizedUrl: normalizedInputUrl,
|
|
188
|
+
queryText: cached.markdown,
|
|
189
|
+
};
|
|
190
|
+
}
|
|
191
|
+
try {
|
|
192
|
+
log.info('fetching URL for signal extraction', { url });
|
|
193
|
+
const raw = await router.fetch(url, { renderJs: 'auto' });
|
|
194
|
+
const extraction = await extractContent(raw.html, raw.finalUrl, {
|
|
195
|
+
contentType: raw.contentType,
|
|
196
|
+
});
|
|
197
|
+
const terms = extractKeyTerms(extraction.markdown, extraction.title);
|
|
198
|
+
return {
|
|
199
|
+
terms,
|
|
200
|
+
title: extraction.title,
|
|
201
|
+
inputUrl: url,
|
|
202
|
+
inputNormalizedUrl: normalizedInputUrl,
|
|
203
|
+
queryText: extraction.markdown,
|
|
204
|
+
};
|
|
205
|
+
}
|
|
206
|
+
catch (err) {
|
|
207
|
+
log.warn('failed to fetch URL for signal extraction', { url, error: String(err) });
|
|
208
|
+
const urlTerms = extractKeyTerms('', url);
|
|
209
|
+
return {
|
|
210
|
+
terms: urlTerms,
|
|
211
|
+
title: url,
|
|
212
|
+
inputUrl: url,
|
|
213
|
+
inputNormalizedUrl: normalizedInputUrl,
|
|
214
|
+
};
|
|
215
|
+
}
|
|
216
|
+
}
|
|
217
|
+
async function runEmbeddingSearch(queryText, excludeNormalizedUrl, includeDomains, excludeDomains, topK, rankMap) {
|
|
218
|
+
try {
|
|
219
|
+
const service = getEmbeddingService();
|
|
220
|
+
if (!service.isAvailable() || service.getIndex().size() === 0)
|
|
221
|
+
return [];
|
|
222
|
+
const excludeUrls = excludeNormalizedUrl ? new Set([excludeNormalizedUrl]) : undefined;
|
|
223
|
+
const similar = await service.findSimilar(queryText, topK, excludeUrls);
|
|
224
|
+
if (similar.length === 0)
|
|
225
|
+
return [];
|
|
226
|
+
// Hydrate with cached content and apply domain filters on the hydrated pool
|
|
227
|
+
const hydrated = [];
|
|
228
|
+
for (const { url: nUrl, score } of similar) {
|
|
229
|
+
const cached = getCachedContent(nUrl);
|
|
230
|
+
hydrated.push({ entry: cached, url: nUrl, score });
|
|
231
|
+
}
|
|
232
|
+
const filterableInputs = hydrated.map(h => ({
|
|
233
|
+
url: h.entry?.url ?? h.url,
|
|
234
|
+
}));
|
|
235
|
+
const filtered = filterByDomains(filterableInputs, includeDomains, excludeDomains);
|
|
236
|
+
const allowedUrls = new Set(filtered.map(f => f.url));
|
|
237
|
+
const results = [];
|
|
238
|
+
let rank = 0;
|
|
239
|
+
for (const h of hydrated) {
|
|
240
|
+
const displayUrl = h.entry?.url ?? h.url;
|
|
241
|
+
if (!allowedUrls.has(displayUrl))
|
|
242
|
+
continue;
|
|
243
|
+
rank++;
|
|
244
|
+
rankMap.set(safeNormalize(displayUrl), rank);
|
|
245
|
+
results.push({
|
|
246
|
+
url: displayUrl,
|
|
247
|
+
title: h.entry?.title ?? displayUrl,
|
|
248
|
+
markdown: (h.entry?.markdown ?? '').slice(0, 5000),
|
|
249
|
+
relevance_score: h.score,
|
|
250
|
+
source: 'cache',
|
|
251
|
+
match_signals: {
|
|
252
|
+
embedding_rank: rank,
|
|
253
|
+
fused_score: 0,
|
|
254
|
+
},
|
|
255
|
+
});
|
|
256
|
+
}
|
|
257
|
+
return results;
|
|
258
|
+
}
|
|
259
|
+
catch (err) {
|
|
260
|
+
log.warn('embedding search failed', { error: String(err) });
|
|
261
|
+
return [];
|
|
262
|
+
}
|
|
263
|
+
}
|
|
264
|
+
function runFTS5Search(terms, excludeNormalizedUrl, includeDomains, excludeDomains, maxCandidates, rankMap) {
|
|
265
|
+
try {
|
|
266
|
+
const fts5Query = buildFTS5Query(terms);
|
|
267
|
+
if (!fts5Query)
|
|
268
|
+
return [];
|
|
269
|
+
let cached = searchCache(fts5Query);
|
|
270
|
+
if (excludeNormalizedUrl) {
|
|
271
|
+
cached = cached.filter(c => {
|
|
272
|
+
try {
|
|
273
|
+
return normalizeUrl(c.url) !== excludeNormalizedUrl;
|
|
274
|
+
}
|
|
275
|
+
catch {
|
|
276
|
+
return c.url !== excludeNormalizedUrl;
|
|
277
|
+
}
|
|
278
|
+
});
|
|
279
|
+
}
|
|
280
|
+
cached = filterByDomains(cached, includeDomains, excludeDomains);
|
|
281
|
+
cached = cached.slice(0, maxCandidates);
|
|
282
|
+
const results = [];
|
|
283
|
+
for (let i = 0; i < cached.length; i++) {
|
|
284
|
+
const entry = cached[i];
|
|
285
|
+
let nUrl;
|
|
286
|
+
try {
|
|
287
|
+
nUrl = normalizeUrl(entry.url);
|
|
288
|
+
}
|
|
289
|
+
catch {
|
|
290
|
+
nUrl = entry.url;
|
|
291
|
+
}
|
|
292
|
+
rankMap.set(nUrl, i + 1);
|
|
293
|
+
results.push({
|
|
294
|
+
url: entry.url,
|
|
295
|
+
title: entry.title,
|
|
296
|
+
markdown: entry.markdown.slice(0, 5000),
|
|
297
|
+
relevance_score: 0,
|
|
298
|
+
source: 'cache',
|
|
299
|
+
match_signals: {
|
|
300
|
+
fts5_rank: i + 1,
|
|
301
|
+
fused_score: 0,
|
|
302
|
+
},
|
|
303
|
+
});
|
|
304
|
+
}
|
|
305
|
+
return results;
|
|
306
|
+
}
|
|
307
|
+
catch (err) {
|
|
308
|
+
log.error('FTS5 search failed', { error: String(err) });
|
|
309
|
+
return [];
|
|
310
|
+
}
|
|
311
|
+
}
|
|
312
|
+
async function runWebSearchFallback(signal, engines, router, backendStatus, maxResults, excludeNormalizedUrl, includeDomains, excludeDomains, rankMap) {
|
|
313
|
+
try {
|
|
314
|
+
const queries = generateSearchQueries(signal.terms, signal.title);
|
|
315
|
+
if (queries.length === 0)
|
|
316
|
+
return [];
|
|
317
|
+
const allResults = [];
|
|
318
|
+
const seenUrls = new Set();
|
|
319
|
+
if (excludeNormalizedUrl) {
|
|
320
|
+
seenUrls.add(excludeNormalizedUrl);
|
|
321
|
+
}
|
|
322
|
+
for (const query of queries) {
|
|
323
|
+
try {
|
|
324
|
+
const searchOutput = await handleSearch({
|
|
325
|
+
query,
|
|
326
|
+
max_results: maxResults,
|
|
327
|
+
include_content: true,
|
|
328
|
+
include_domains: includeDomains,
|
|
329
|
+
exclude_domains: excludeDomains,
|
|
330
|
+
}, engines, router, backendStatus);
|
|
331
|
+
for (const item of searchOutput.results) {
|
|
332
|
+
let nUrl;
|
|
333
|
+
try {
|
|
334
|
+
nUrl = normalizeUrl(item.url);
|
|
335
|
+
}
|
|
336
|
+
catch {
|
|
337
|
+
nUrl = item.url;
|
|
338
|
+
}
|
|
339
|
+
if (seenUrls.has(nUrl))
|
|
340
|
+
continue;
|
|
341
|
+
seenUrls.add(nUrl);
|
|
342
|
+
const rank = allResults.length + 1;
|
|
343
|
+
rankMap.set(nUrl, rank);
|
|
344
|
+
allResults.push({
|
|
345
|
+
url: item.url,
|
|
346
|
+
title: item.title,
|
|
347
|
+
markdown: (item.markdown_content ?? item.snippet).slice(0, 5000),
|
|
348
|
+
relevance_score: item.relevance_score,
|
|
349
|
+
source: 'search',
|
|
350
|
+
match_signals: {
|
|
351
|
+
fused_score: 0,
|
|
352
|
+
},
|
|
353
|
+
});
|
|
354
|
+
}
|
|
355
|
+
}
|
|
356
|
+
catch (err) {
|
|
357
|
+
log.warn('web search query failed', { query, error: String(err) });
|
|
358
|
+
}
|
|
359
|
+
}
|
|
360
|
+
try {
|
|
361
|
+
const embeddingService = getEmbeddingService();
|
|
362
|
+
if (embeddingService.isAvailable()) {
|
|
363
|
+
for (const result of allResults) {
|
|
364
|
+
if (result.markdown) {
|
|
365
|
+
embeddingService.embedAsync(result.url, result.markdown);
|
|
366
|
+
}
|
|
367
|
+
}
|
|
368
|
+
}
|
|
369
|
+
}
|
|
370
|
+
catch (err) {
|
|
371
|
+
log.debug('embedding hook skipped for find_similar results', { error: String(err) });
|
|
372
|
+
}
|
|
373
|
+
return allResults;
|
|
374
|
+
}
|
|
375
|
+
catch (err) {
|
|
376
|
+
log.error('web search fallback failed', { error: String(err) });
|
|
377
|
+
return [];
|
|
378
|
+
}
|
|
379
|
+
}
|
|
380
|
+
function generateSearchQueries(terms, title) {
|
|
381
|
+
if (terms.length === 0 && !title)
|
|
382
|
+
return [];
|
|
383
|
+
const queries = [];
|
|
384
|
+
if (title && title.length > 3) {
|
|
385
|
+
queries.push(title.slice(0, 150));
|
|
386
|
+
}
|
|
387
|
+
if (terms.length >= 3) {
|
|
388
|
+
queries.push(terms.slice(0, 5).join(' '));
|
|
389
|
+
}
|
|
390
|
+
if (terms.length >= 2) {
|
|
391
|
+
queries.push(`${terms.slice(0, 3).join(' ')} tutorial guide`);
|
|
392
|
+
}
|
|
393
|
+
const unique = [...new Set(queries)];
|
|
394
|
+
return unique.slice(0, WEB_SEARCH_QUERY_COUNT);
|
|
395
|
+
}
|
|
396
|
+
function fuseResults(rankedLists, allResults, maxResults) {
|
|
397
|
+
const scores = reciprocalRankFusion(rankedLists);
|
|
398
|
+
const sorted = sortByRRFScore(scores);
|
|
399
|
+
const resultsByNormalizedUrl = new Map();
|
|
400
|
+
for (const r of allResults) {
|
|
401
|
+
const key = safeNormalize(r.url);
|
|
402
|
+
if (!resultsByNormalizedUrl.has(key)) {
|
|
403
|
+
resultsByNormalizedUrl.set(key, r);
|
|
404
|
+
}
|
|
405
|
+
}
|
|
406
|
+
const fused = [];
|
|
407
|
+
for (const [nUrl, score] of sorted) {
|
|
408
|
+
if (fused.length >= maxResults)
|
|
409
|
+
break;
|
|
410
|
+
const result = resultsByNormalizedUrl.get(nUrl);
|
|
411
|
+
if (!result)
|
|
412
|
+
continue;
|
|
413
|
+
fused.push({
|
|
414
|
+
...result,
|
|
415
|
+
relevance_score: score,
|
|
416
|
+
match_signals: {
|
|
417
|
+
...result.match_signals,
|
|
418
|
+
fused_score: score,
|
|
419
|
+
},
|
|
420
|
+
});
|
|
421
|
+
}
|
|
422
|
+
return fused;
|
|
423
|
+
}
|
|
424
|
+
function determineMethod(hasCache, hasEmbedding, hasSearch) {
|
|
425
|
+
const sources = [hasCache, hasEmbedding, hasSearch].filter(Boolean).length;
|
|
426
|
+
if (sources >= 2)
|
|
427
|
+
return 'hybrid';
|
|
428
|
+
if (hasEmbedding)
|
|
429
|
+
return 'embedding';
|
|
430
|
+
if (hasCache)
|
|
431
|
+
return 'fts5';
|
|
432
|
+
if (hasSearch)
|
|
433
|
+
return 'search';
|
|
434
|
+
return 'fts5';
|
|
435
|
+
}
|
|
436
|
+
//# sourceMappingURL=find-similar.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"find-similar.js","sourceRoot":"","sources":["../../src/search/find-similar.ts"],"names":[],"mappings":"AASA,OAAO,EAAE,eAAe,EAAE,cAAc,EAAE,MAAM,2BAA2B,CAAC;AAC5E,OAAO,EAAE,oBAAoB,EAAgB,cAAc,EAAE,MAAM,UAAU,CAAC;AAC9E,OAAO,EAAE,WAAW,EAAE,gBAAgB,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AAChF,OAAO,EAAE,eAAe,EAAE,MAAM,cAAc,CAAC;AAC/C,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAClD,OAAO,EAAE,cAAc,EAAE,MAAM,2BAA2B,CAAC;AAC3D,OAAO,EAAE,mBAAmB,EAAE,MAAM,uBAAuB,CAAC;AAC5D,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AAE5C,MAAM,GAAG,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;AAEnC,MAAM,mBAAmB,GAAG,EAAE,CAAC;AAC/B,MAAM,mBAAmB,GAAG,EAAE,CAAC;AAC/B,MAAM,wBAAwB,GAAG,EAAE,CAAC;AACpC,MAAM,sBAAsB,GAAG,CAAC,CAAC;AAUjC,MAAM,CAAC,KAAK,UAAU,WAAW,CAC/B,KAAuB,EACvB,OAAuB,EACvB,MAAmB,EACnB,aAA6B;IAE7B,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAEzB,mEAAmE;IACnE,MAAM,kBAAkB,GAAG,uBAAuB,EAAE,CAAC;IAErD,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,KAAK,CAAC,GAAG,EAAE,IAAI,EAAE,CAAC;QAC9B,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,EAAE,IAAI,EAAE,CAAC;QAEtC,IAAI,CAAC,GAAG,IAAI,CAAC,OAAO,EAAE,CAAC;YACrB,OAAO;gBACL,OAAO,EAAE,EAAE;gBACX,MAAM,EAAE,MAAM;gBACd,UAAU,EAAE,CAAC;gBACb,WAAW,EAAE,CAAC;gBACd,mBAAmB,EAAE,kBAAkB;gBACvC,KAAK,EAAE,wCAAwC;gBAC/C,aAAa,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;aAClC,CAAC;QACJ,CAAC;QAED,MAAM,UAAU,GAAG,KAAK,CAAC,WAAW,IAAI,mBAAmB,CAAC;QAC5D,MAAM,YAAY,GAAG,KAAK,CAAC,aAAa,IAAI,IAAI,CAAC;QACjD,MAAM,UAAU,GAAG,KAAK,CAAC,WAAW,IAAI,IAAI,CAAC;QAE7C,MAAM,MAAM,GAAG,MAAM,aAAa,CAAC,GAAG,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;QAEzD,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,CAAC;YACnD,GAAG,CAAC,IAAI,CAAC,kEAAkE,CAAC,CAAC;YAE7E,IAAI,CAAC,UAAU,EAAE,CAAC;gBAChB,OAAO;oBACL,OAAO,EAAE,EAAE;oBACX,MAAM,EAAE,MAAM;oBACd,UAAU,EAAE,CAAC;oBACb,WAAW,EAAE,CAAC;oBACd,mBAAmB,EAAE,kBAAkB;oBACvC,KAAK,EAAE,mEAAmE;oBAC1E,aAAa,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;iBAClC,CAAC;YACJ,CAAC;QACH,CAAC;QAED,sEAAsE;QACtE,IAAI,YAAY,GAAwB,EAAE,CAAC;QAC3C,MAAM,WAAW,GAAG,IAAI,GAAG,EAAkB,CAAC;QAC9C,IAAI,gBAAgB,GAAwB,EAAE,CAAC;QAC/C,MAAM,gBAAgB,GAAG,IAAI,GAAG,EAAkB,CAAC;QAEnD,MAAM,OAAO,CAAC,GAAG,CAAC;YAChB,CAAC,KAAK,IAAI,EAAE;gBACV,IAAI,YAAY,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAC5C,YAAY,GAAG,aAAa,CAC1B,MAAM,CAAC,KAAK,EACZ,MAAM,CAAC,kBAAkB,EACzB,KAAK,CAAC,eAAe,EACrB,KAAK,CAAC,eAAe,EACrB,mBAAmB,EACnB,WAAW,CACZ,CAAC;oBACF,GAAG,CAAC,KAAK,CAAC,sBAAsB,EAAE,EAAE,IAAI,EAAE,YAAY,CAAC,MAAM,EAAE,CAAC,CAAC;gBACnE,CAAC;YACH,CAAC,CAAC,EAAE;YACJ,CAAC,KAAK,IAAI,EAAE;gBACV,IAAI,YAAY,IAAI,kBAAkB,IAAI,MAAM,CAAC,SAAS,EAAE,CAAC;oBAC3D,gBAAgB,GAAG,MAAM,kBAAkB,CACzC,MAAM,CAAC,SAAS,EAChB,MAAM,CAAC,kBAAkB,EACzB,KAAK,CAAC,eAAe,EACrB,KAAK,CAAC,eAAe,EACrB,wBAAwB,EACxB,gBAAgB,CACjB,CAAC;oBACF,GAAG,CAAC,KAAK,CAAC,2BAA2B,EAAE,EAAE,IAAI,EAAE,gBAAgB,CAAC,MAAM,EAAE,CAAC,CAAC;gBAC5E,CAAC;YACH,CAAC,CAAC,EAAE;SACL,CAAC,CAAC;QAEH,iFAAiF;QACjF,IAAI,aAAa,GAAwB,EAAE,CAAC;QAC5C,MAAM,aAAa,GAAG,IAAI,GAAG,EAAkB,CAAC;QAEhD,MAAM,iBAAiB,GAAG,IAAI,GAAG,EAAU,CAAC;QAC5C,KAAK,MAAM,CAAC,IAAI,YAAY;YAAE,iBAAiB,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QAC1E,KAAK,MAAM,CAAC,IAAI,gBAAgB;YAAE,iBAAiB,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QAE9E,IAAI,iBAAiB,CAAC,IAAI,GAAG,UAAU,IAAI,UAAU,EAAE,CAAC;YACtD,aAAa,GAAG,MAAM,oBAAoB,CACxC,MAAM,EACN,OAAO,EACP,MAAM,EACN,aAAa,EACb,UAAU,EACV,MAAM,CAAC,kBAAkB,EACzB,KAAK,CAAC,eAAe,EACrB,KAAK,CAAC,eAAe,EACrB,aAAa,CACd,CAAC;YACF,GAAG,CAAC,KAAK,CAAC,8BAA8B,EAAE,EAAE,IAAI,EAAE,aAAa,CAAC,MAAM,EAAE,CAAC,CAAC;QAC5E,CAAC;QAED,4BAA4B;QAC5B,MAAM,WAAW,GAA0B,EAAE,CAAC;QAC9C,IAAI,WAAW,CAAC,IAAI,GAAG,CAAC;YAAE,WAAW,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;QACxD,IAAI,gBAAgB,CAAC,IAAI,GAAG,CAAC;YAAE,WAAW,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;QAClE,IAAI,aAAa,CAAC,IAAI,GAAG,CAAC;YAAE,WAAW,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;QAE5D,MAAM,UAAU,GAAG,YAAY,CAAC,YAAY,EAAE,gBAAgB,EAAE,aAAa,CAAC,CAAC;QAE/E,IAAI,YAAiC,CAAC;QAEtC,IAAI,WAAW,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;YAC5B,YAAY,GAAG,WAAW,CAAC,WAAW,EAAE,UAAU,EAAE,UAAU,CAAC,CAAC;QAClE,CAAC;aAAM,CAAC;YACN,YAAY,GAAG,UAAU;iBACtB,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,eAAe,GAAG,CAAC,CAAC,eAAe,CAAC;iBACrD,KAAK,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC;QAC1B,CAAC;QAED,MAAM,MAAM,GAAG,eAAe,CAC5B,YAAY,CAAC,MAAM,GAAG,CAAC,EACvB,gBAAgB,CAAC,MAAM,GAAG,CAAC,EAC3B,aAAa,CAAC,MAAM,GAAG,CAAC,CACzB,CAAC;QAEF,MAAM,SAAS,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,OAAO,CAAC,CAAC,MAAM,CAAC;QACxE,MAAM,UAAU,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC,MAAM,CAAC;QAE1E,OAAO;YACL,OAAO,EAAE,YAAY;YACrB,MAAM;YACN,UAAU,EAAE,SAAS;YACrB,WAAW,EAAE,UAAU;YACvB,mBAAmB,EAAE,kBAAkB;YACvC,aAAa,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;SAClC,CAAC;IACJ,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,GAAG,CAAC,KAAK,CAAC,oBAAoB,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACxD,OAAO;YACL,OAAO,EAAE,EAAE;YACX,MAAM,EAAE,MAAM;YACd,UAAU,EAAE,CAAC;YACb,WAAW,EAAE,CAAC;YACd,mBAAmB,EAAE,kBAAkB;YACvC,KAAK,EAAE,wBAAwB,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE;YACjF,aAAa,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;SAClC,CAAC;IACJ,CAAC;AACH,CAAC;AAED,SAAS,uBAAuB;IAC9B,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,mBAAmB,EAAE,CAAC;QAClC,OAAO,GAAG,CAAC,WAAW,EAAE,IAAI,GAAG,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IACxD,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,KAAK,CAAC;IACf,CAAC;AACH,CAAC;AAED,SAAS,aAAa,CAAC,GAAW;IAChC,IAAI,CAAC;QACH,OAAO,YAAY,CAAC,GAAG,CAAC,CAAC;IAC3B,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,GAAG,CAAC;IACb,CAAC;AACH,CAAC;AAED,SAAS,YAAY,CAAC,GAAG,KAA4B;IACnD,MAAM,IAAI,GAAG,IAAI,GAAG,EAA6B,CAAC;IAClD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,KAAK,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC;YACrB,MAAM,GAAG,GAAG,aAAa,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;YACjC,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YAC/B,IAAI,CAAC,QAAQ,EAAE,CAAC;gBACd,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC;YACnB,CAAC;iBAAM,CAAC;gBACN,4EAA4E;gBAC5E,QAAQ,CAAC,aAAa,GAAG;oBACvB,GAAG,QAAQ,CAAC,aAAa;oBACzB,GAAG,CAAC,CAAC,aAAa;oBAClB,WAAW,EAAE,QAAQ,CAAC,aAAa,CAAC,WAAW;iBAChD,CAAC;YACJ,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC;AAC5B,CAAC;AAED,KAAK,UAAU,aAAa,CAC1B,GAAuB,EACvB,OAA2B,EAC3B,MAAmB;IAEnB,IAAI,GAAG,EAAE,CAAC;QACR,OAAO,MAAM,oBAAoB,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;IACjD,CAAC;IAED,IAAI,OAAO,EAAE,CAAC;QACZ,MAAM,KAAK,GAAG,eAAe,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;QAC3C,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC;IACvD,CAAC;IAED,OAAO,EAAE,KAAK,EAAE,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC;AAClC,CAAC;AAED,KAAK,UAAU,oBAAoB,CACjC,GAAW,EACX,MAAmB;IAEnB,IAAI,kBAA0B,CAAC;IAC/B,IAAI,CAAC;QACH,kBAAkB,GAAG,YAAY,CAAC,GAAG,CAAC,CAAC;IACzC,CAAC;IAAC,MAAM,CAAC;QACP,kBAAkB,GAAG,GAAG,CAAC;IAC3B,CAAC;IAED,MAAM,MAAM,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC;IACrC,IAAI,MAAM,EAAE,CAAC;QACX,MAAM,KAAK,GAAG,eAAe,CAAC,MAAM,CAAC,QAAQ,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC;QAC7D,OAAO;YACL,KAAK;YACL,KAAK,EAAE,MAAM,CAAC,KAAK;YACnB,QAAQ,EAAE,GAAG;YACb,kBAAkB,EAAE,kBAAkB;YACtC,SAAS,EAAE,MAAM,CAAC,QAAQ;SAC3B,CAAC;IACJ,CAAC;IAED,IAAI,CAAC;QACH,GAAG,CAAC,IAAI,CAAC,oCAAoC,EAAE,EAAE,GAAG,EAAE,CAAC,CAAC;QACxD,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,KAAK,CAAC,GAAG,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC,CAAC;QAC1D,MAAM,UAAU,GAAG,MAAM,cAAc,CAAC,GAAG,CAAC,IAAI,EAAE,GAAG,CAAC,QAAQ,EAAE;YAC9D,WAAW,EAAE,GAAG,CAAC,WAAW;SAC7B,CAAC,CAAC;QACH,MAAM,KAAK,GAAG,eAAe,CAAC,UAAU,CAAC,QAAQ,EAAE,UAAU,CAAC,KAAK,CAAC,CAAC;QACrE,OAAO;YACL,KAAK;YACL,KAAK,EAAE,UAAU,CAAC,KAAK;YACvB,QAAQ,EAAE,GAAG;YACb,kBAAkB,EAAE,kBAAkB;YACtC,SAAS,EAAE,UAAU,CAAC,QAAQ;SAC/B,CAAC;IACJ,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,GAAG,CAAC,IAAI,CAAC,2CAA2C,EAAE,EAAE,GAAG,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACnF,MAAM,QAAQ,GAAG,eAAe,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC;QAC1C,OAAO;YACL,KAAK,EAAE,QAAQ;YACf,KAAK,EAAE,GAAG;YACV,QAAQ,EAAE,GAAG;YACb,kBAAkB,EAAE,kBAAkB;SACvC,CAAC;IACJ,CAAC;AACH,CAAC;AAED,KAAK,UAAU,kBAAkB,CAC/B,SAAiB,EACjB,oBAAwC,EACxC,cAAoC,EACpC,cAAoC,EACpC,IAAY,EACZ,OAA4B;IAE5B,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,mBAAmB,EAAE,CAAC;QACtC,IAAI,CAAC,OAAO,CAAC,WAAW,EAAE,IAAI,OAAO,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEzE,MAAM,WAAW,GAAG,oBAAoB,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,oBAAoB,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;QACvF,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,WAAW,CAAC,SAAS,EAAE,IAAI,EAAE,WAAW,CAAC,CAAC;QACxE,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEpC,4EAA4E;QAC5E,MAAM,QAAQ,GAAuE,EAAE,CAAC;QACxF,KAAK,MAAM,EAAE,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,OAAO,EAAE,CAAC;YAC3C,MAAM,MAAM,GAAG,gBAAgB,CAAC,IAAI,CAAC,CAAC;YACtC,QAAQ,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;QACrD,CAAC;QAED,MAAM,gBAAgB,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YAC1C,GAAG,EAAE,CAAC,CAAC,KAAK,EAAE,GAAG,IAAI,CAAC,CAAC,GAAG;SAC3B,CAAC,CAA+B,CAAC;QAClC,MAAM,QAAQ,GAAG,eAAe,CAAC,gBAAgB,EAAE,cAAc,EAAE,cAAc,CAE/E,CAAC;QACH,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;QAEtD,MAAM,OAAO,GAAwB,EAAE,CAAC;QACxC,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,KAAK,MAAM,CAAC,IAAI,QAAQ,EAAE,CAAC;YACzB,MAAM,UAAU,GAAG,CAAC,CAAC,KAAK,EAAE,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC;YACzC,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,UAAU,CAAC;gBAAE,SAAS;YAE3C,IAAI,EAAE,CAAC;YACP,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,UAAU,CAAC,EAAE,IAAI,CAAC,CAAC;YAE7C,OAAO,CAAC,IAAI,CAAC;gBACX,GAAG,EAAE,UAAU;gBACf,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,KAAK,IAAI,UAAU;gBACnC,QAAQ,EAAE,CAAC,CAAC,CAAC,KAAK,EAAE,QAAQ,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC;gBAClD,eAAe,EAAE,CAAC,CAAC,KAAK;gBACxB,MAAM,EAAE,OAAO;gBACf,aAAa,EAAE;oBACb,cAAc,EAAE,IAAI;oBACpB,WAAW,EAAE,CAAC;iBACf;aACF,CAAC,CAAC;QACL,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,GAAG,CAAC,IAAI,CAAC,yBAAyB,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAC5D,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED,SAAS,aAAa,CACpB,KAAe,EACf,oBAAwC,EACxC,cAAoC,EACpC,cAAoC,EACpC,aAAqB,EACrB,OAA4B;IAE5B,IAAI,CAAC;QACH,MAAM,SAAS,GAAG,cAAc,CAAC,KAAK,CAAC,CAAC;QACxC,IAAI,CAAC,SAAS;YAAE,OAAO,EAAE,CAAC;QAE1B,IAAI,MAAM,GAAG,WAAW,CAAC,SAAS,CAAC,CAAC;QAEpC,IAAI,oBAAoB,EAAE,CAAC;YACzB,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE;gBACzB,IAAI,CAAC;oBACH,OAAO,YAAY,CAAC,CAAC,CAAC,GAAG,CAAC,KAAK,oBAAoB,CAAC;gBACtD,CAAC;gBAAC,MAAM,CAAC;oBACP,OAAO,CAAC,CAAC,GAAG,KAAK,oBAAoB,CAAC;gBACxC,CAAC;YACH,CAAC,CAAC,CAAC;QACL,CAAC;QAED,MAAM,GAAG,eAAe,CAAC,MAAM,EAAE,cAAc,EAAE,cAAc,CAAoB,CAAC;QACpF,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,aAAa,CAAC,CAAC;QAExC,MAAM,OAAO,GAAwB,EAAE,CAAC;QACxC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACvC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;YACxB,IAAI,IAAY,CAAC;YACjB,IAAI,CAAC;gBACH,IAAI,GAAG,YAAY,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;YACjC,CAAC;YAAC,MAAM,CAAC;gBACP,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC;YACnB,CAAC;YAED,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;YAEzB,OAAO,CAAC,IAAI,CAAC;gBACX,GAAG,EAAE,KAAK,CAAC,GAAG;gBACd,KAAK,EAAE,KAAK,CAAC,KAAK;gBAClB,QAAQ,EAAE,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC;gBACvC,eAAe,EAAE,CAAC;gBAClB,MAAM,EAAE,OAAO;gBACf,aAAa,EAAE;oBACb,SAAS,EAAE,CAAC,GAAG,CAAC;oBAChB,WAAW,EAAE,CAAC;iBACf;aACF,CAAC,CAAC;QACL,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,GAAG,CAAC,KAAK,CAAC,oBAAoB,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACxD,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED,KAAK,UAAU,oBAAoB,CACjC,MAAsB,EACtB,OAAuB,EACvB,MAAmB,EACnB,aAAwC,EACxC,UAAkB,EAClB,oBAAwC,EACxC,cAAoC,EACpC,cAAoC,EACpC,OAA4B;IAE5B,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,qBAAqB,CAAC,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC;QAClE,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAEpC,MAAM,UAAU,GAAwB,EAAE,CAAC;QAC3C,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAC;QAEnC,IAAI,oBAAoB,EAAE,CAAC;YACzB,QAAQ,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAC;QACrC,CAAC;QAED,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,IAAI,CAAC;gBACH,MAAM,YAAY,GAAG,MAAM,YAAY,CACrC;oBACE,KAAK;oBACL,WAAW,EAAE,UAAU;oBACvB,eAAe,EAAE,IAAI;oBACrB,eAAe,EAAE,cAAc;oBAC/B,eAAe,EAAE,cAAc;iBAChC,EACD,OAAO,EACP,MAAM,EACN,aAAa,CACd,CAAC;gBAEF,KAAK,MAAM,IAAI,IAAI,YAAY,CAAC,OAAO,EAAE,CAAC;oBACxC,IAAI,IAAY,CAAC;oBACjB,IAAI,CAAC;wBACH,IAAI,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;oBAChC,CAAC;oBAAC,MAAM,CAAC;wBACP,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC;oBAClB,CAAC;oBAED,IAAI,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC;wBAAE,SAAS;oBACjC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;oBAEnB,MAAM,IAAI,GAAG,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC;oBACnC,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;oBAExB,UAAU,CAAC,IAAI,CAAC;wBACd,GAAG,EAAE,IAAI,CAAC,GAAG;wBACb,KAAK,EAAE,IAAI,CAAC,KAAK;wBACjB,QAAQ,EAAE,CAAC,IAAI,CAAC,gBAAgB,IAAI,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC;wBAChE,eAAe,EAAE,IAAI,CAAC,eAAe;wBACrC,MAAM,EAAE,QAAQ;wBAChB,aAAa,EAAE;4BACb,WAAW,EAAE,CAAC;yBACf;qBACF,CAAC,CAAC;gBACL,CAAC;YACH,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,GAAG,CAAC,IAAI,CAAC,yBAAyB,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YACrE,CAAC;QACH,CAAC;QAED,IAAI,CAAC;YACH,MAAM,gBAAgB,GAAG,mBAAmB,EAAE,CAAC;YAC/C,IAAI,gBAAgB,CAAC,WAAW,EAAE,EAAE,CAAC;gBACnC,KAAK,MAAM,MAAM,IAAI,UAAU,EAAE,CAAC;oBAChC,IAAI,MAAM,CAAC,QAAQ,EAAE,CAAC;wBACpB,gBAAgB,CAAC,UAAU,CAAC,MAAM,CAAC,GAAG,EAAE,MAAM,CAAC,QAAQ,CAAC,CAAC;oBAC3D,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,GAAG,CAAC,KAAK,CAAC,iDAAiD,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACvF,CAAC;QAED,OAAO,UAAU,CAAC;IACpB,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,GAAG,CAAC,KAAK,CAAC,4BAA4B,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAChE,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED,SAAS,qBAAqB,CAAC,KAAe,EAAE,KAAa;IAC3D,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IAE5C,MAAM,OAAO,GAAa,EAAE,CAAC;IAE7B,IAAI,KAAK,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC9B,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;IACpC,CAAC;IAED,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;QACtB,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IAC5C,CAAC;IAED,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;QACtB,OAAO,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,iBAAiB,CAAC,CAAC;IAChE,CAAC;IAED,MAAM,MAAM,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC;IACrC,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,sBAAsB,CAAC,CAAC;AACjD,CAAC;AAED,SAAS,WAAW,CAClB,WAAkC,EAClC,UAA+B,EAC/B,UAAkB;IAElB,MAAM,MAAM,GAAG,oBAAoB,CAAC,WAAW,CAAC,CAAC;IACjD,MAAM,MAAM,GAAG,cAAc,CAAC,MAAM,CAAC,CAAC;IAEtC,MAAM,sBAAsB,GAAG,IAAI,GAAG,EAA6B,CAAC;IACpE,KAAK,MAAM,CAAC,IAAI,UAAU,EAAE,CAAC;QAC3B,MAAM,GAAG,GAAG,aAAa,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;QACjC,IAAI,CAAC,sBAAsB,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YACrC,sBAAsB,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC;QACrC,CAAC;IACH,CAAC;IAED,MAAM,KAAK,GAAwB,EAAE,CAAC;IACtC,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,EAAE,CAAC;QACnC,IAAI,KAAK,CAAC,MAAM,IAAI,UAAU;YAAE,MAAM;QAEtC,MAAM,MAAM,GAAG,sBAAsB,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAChD,IAAI,CAAC,MAAM;YAAE,SAAS;QAEtB,KAAK,CAAC,IAAI,CAAC;YACT,GAAG,MAAM;YACT,eAAe,EAAE,KAAK;YACtB,aAAa,EAAE;gBACb,GAAG,MAAM,CAAC,aAAa;gBACvB,WAAW,EAAE,KAAK;aACnB;SACF,CAAC,CAAC;IACL,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,eAAe,CACtB,QAAiB,EACjB,YAAqB,EACrB,SAAkB;IAElB,MAAM,OAAO,GAAG,CAAC,QAAQ,EAAE,YAAY,EAAE,SAAS,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC;IAC3E,IAAI,OAAO,IAAI,CAAC;QAAE,OAAO,QAAQ,CAAC;IAClC,IAAI,YAAY;QAAE,OAAO,WAAW,CAAC;IACrC,IAAI,QAAQ;QAAE,OAAO,MAAM,CAAC;IAC5B,IAAI,SAAS;QAAE,OAAO,QAAQ,CAAC;IAC/B,OAAO,MAAM,CAAC;AAChB,CAAC"}
|
|
@@ -0,0 +1,22 @@
|
|
|
1
|
+
import type { RawSearchResult, SearchEngine } from '../types.js';
|
|
2
|
+
import type { MergedSearchResult } from './dedup.js';
|
|
3
|
+
export declare function normalizeQueries(queries: string[]): string[];
|
|
4
|
+
export interface FanOutOptions {
|
|
5
|
+
maxResults: number;
|
|
6
|
+
timeRange?: string;
|
|
7
|
+
language?: string;
|
|
8
|
+
includeDomains?: string[];
|
|
9
|
+
excludeDomains?: string[];
|
|
10
|
+
fromDate?: string;
|
|
11
|
+
toDate?: string;
|
|
12
|
+
category?: 'general' | 'news' | 'code' | 'docs' | 'papers' | 'images';
|
|
13
|
+
}
|
|
14
|
+
export interface FanOutResult {
|
|
15
|
+
results: RawSearchResult[];
|
|
16
|
+
enginesUsed: string[];
|
|
17
|
+
errors: string[];
|
|
18
|
+
}
|
|
19
|
+
export declare function fanOutSearch(queries: string[], engines: SearchEngine[], options: FanOutOptions): Promise<FanOutResult>;
|
|
20
|
+
export declare function synthesizeIntent(queries: string[]): string;
|
|
21
|
+
export declare function mergeWithRRF(rankedLists: MergedSearchResult[][]): MergedSearchResult[];
|
|
22
|
+
//# sourceMappingURL=multi-query.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"multi-query.d.ts","sourceRoot":"","sources":["../../src/search/multi-query.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,YAAY,EAAuB,MAAM,aAAa,CAAC;AACtF,OAAO,KAAK,EAAE,kBAAkB,EAAE,MAAM,YAAY,CAAC;AASrD,wBAAgB,gBAAgB,CAAC,OAAO,EAAE,MAAM,EAAE,GAAG,MAAM,EAAE,CA2B5D;AAED,MAAM,WAAW,aAAa;IAC5B,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,cAAc,CAAC,EAAE,MAAM,EAAE,CAAC;IAC1B,cAAc,CAAC,EAAE,MAAM,EAAE,CAAC;IAC1B,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,QAAQ,CAAC,EAAE,SAAS,GAAG,MAAM,GAAG,MAAM,GAAG,MAAM,GAAG,QAAQ,GAAG,QAAQ,CAAC;CACvE;AAED,MAAM,WAAW,YAAY;IAC3B,OAAO,EAAE,eAAe,EAAE,CAAC;IAC3B,WAAW,EAAE,MAAM,EAAE,CAAC;IACtB,MAAM,EAAE,MAAM,EAAE,CAAC;CAClB;AAED,wBAAsB,YAAY,CAChC,OAAO,EAAE,MAAM,EAAE,EACjB,OAAO,EAAE,YAAY,EAAE,EACvB,OAAO,EAAE,aAAa,GACrB,OAAO,CAAC,YAAY,CAAC,CAsEvB;AAED,wBAAgB,gBAAgB,CAAC,OAAO,EAAE,MAAM,EAAE,GAAG,MAAM,CAO1D;AAED,wBAAgB,YAAY,CAAC,WAAW,EAAE,kBAAkB,EAAE,EAAE,GAAG,kBAAkB,EAAE,CAmDtF"}
|
|
@@ -0,0 +1,157 @@
|
|
|
1
|
+
import { normalizeUrl } from '../cache/store.js';
|
|
2
|
+
import { getConfig } from '../config.js';
|
|
3
|
+
import { createLogger } from '../logger.js';
|
|
4
|
+
const log = createLogger('search');
|
|
5
|
+
const RRF_K = 60;
|
|
6
|
+
export function normalizeQueries(queries) {
|
|
7
|
+
try {
|
|
8
|
+
const config = getConfig();
|
|
9
|
+
const seen = new Set();
|
|
10
|
+
const normalized = [];
|
|
11
|
+
for (const raw of queries) {
|
|
12
|
+
const q = raw.toLowerCase().trim().replace(/\s+/g, ' ');
|
|
13
|
+
if (q.length === 0)
|
|
14
|
+
continue;
|
|
15
|
+
if (seen.has(q))
|
|
16
|
+
continue;
|
|
17
|
+
seen.add(q);
|
|
18
|
+
normalized.push(q);
|
|
19
|
+
}
|
|
20
|
+
if (normalized.length > config.multiQueryMax) {
|
|
21
|
+
log.warn('multi-query array exceeds max, truncating', {
|
|
22
|
+
provided: normalized.length,
|
|
23
|
+
max: config.multiQueryMax,
|
|
24
|
+
});
|
|
25
|
+
return normalized.slice(0, config.multiQueryMax);
|
|
26
|
+
}
|
|
27
|
+
return normalized;
|
|
28
|
+
}
|
|
29
|
+
catch (err) {
|
|
30
|
+
log.error('normalizeQueries failed', { error: String(err) });
|
|
31
|
+
return [];
|
|
32
|
+
}
|
|
33
|
+
}
|
|
34
|
+
export async function fanOutSearch(queries, engines, options) {
|
|
35
|
+
const allResults = [];
|
|
36
|
+
const enginesUsed = new Set();
|
|
37
|
+
const errors = [];
|
|
38
|
+
if (queries.length === 0 || engines.length === 0) {
|
|
39
|
+
return { results: [], enginesUsed: [], errors: [] };
|
|
40
|
+
}
|
|
41
|
+
try {
|
|
42
|
+
const config = getConfig();
|
|
43
|
+
const concurrency = config.multiQueryConcurrency;
|
|
44
|
+
const hasFilterAttrition = !!(options.includeDomains?.length || options.excludeDomains?.length);
|
|
45
|
+
const overfetchFactor = hasFilterAttrition ? 3 : 2;
|
|
46
|
+
const engineOptions = {
|
|
47
|
+
maxResults: options.maxResults * overfetchFactor,
|
|
48
|
+
timeRange: options.timeRange,
|
|
49
|
+
language: options.language,
|
|
50
|
+
includeDomains: options.includeDomains,
|
|
51
|
+
excludeDomains: options.excludeDomains,
|
|
52
|
+
fromDate: options.fromDate,
|
|
53
|
+
toDate: options.toDate,
|
|
54
|
+
category: options.category,
|
|
55
|
+
};
|
|
56
|
+
const tasks = [];
|
|
57
|
+
for (const engine of engines) {
|
|
58
|
+
for (const query of queries) {
|
|
59
|
+
tasks.push({ engine, query });
|
|
60
|
+
}
|
|
61
|
+
}
|
|
62
|
+
for (let i = 0; i < tasks.length; i += concurrency) {
|
|
63
|
+
const batch = tasks.slice(i, i + concurrency);
|
|
64
|
+
const promises = batch.map(async ({ engine, query }) => {
|
|
65
|
+
try {
|
|
66
|
+
const results = await engine.search(query, engineOptions);
|
|
67
|
+
for (const r of results) {
|
|
68
|
+
allResults.push(r);
|
|
69
|
+
enginesUsed.add(engine.name);
|
|
70
|
+
}
|
|
71
|
+
}
|
|
72
|
+
catch (err) {
|
|
73
|
+
const msg = err instanceof Error ? err.message : String(err);
|
|
74
|
+
log.warn('multi-query engine search failed', {
|
|
75
|
+
engine: engine.name,
|
|
76
|
+
query,
|
|
77
|
+
error: msg,
|
|
78
|
+
});
|
|
79
|
+
errors.push(`${engine.name}(${query}): ${msg}`);
|
|
80
|
+
}
|
|
81
|
+
});
|
|
82
|
+
await Promise.allSettled(promises);
|
|
83
|
+
}
|
|
84
|
+
return {
|
|
85
|
+
results: allResults,
|
|
86
|
+
enginesUsed: [...enginesUsed],
|
|
87
|
+
errors,
|
|
88
|
+
};
|
|
89
|
+
}
|
|
90
|
+
catch (err) {
|
|
91
|
+
log.error('fanOutSearch failed', { error: String(err) });
|
|
92
|
+
return {
|
|
93
|
+
results: allResults,
|
|
94
|
+
enginesUsed: [...enginesUsed],
|
|
95
|
+
errors: [...errors, `fanOutSearch: ${String(err)}`],
|
|
96
|
+
};
|
|
97
|
+
}
|
|
98
|
+
}
|
|
99
|
+
export function synthesizeIntent(queries) {
|
|
100
|
+
try {
|
|
101
|
+
return queries.map(q => q.trim()).filter(Boolean).join('; ');
|
|
102
|
+
}
|
|
103
|
+
catch (err) {
|
|
104
|
+
log.error('synthesizeIntent failed', { error: String(err) });
|
|
105
|
+
return '';
|
|
106
|
+
}
|
|
107
|
+
}
|
|
108
|
+
export function mergeWithRRF(rankedLists) {
|
|
109
|
+
try {
|
|
110
|
+
if (rankedLists.length === 0)
|
|
111
|
+
return [];
|
|
112
|
+
const nonEmpty = rankedLists.filter(l => l.length > 0);
|
|
113
|
+
if (nonEmpty.length === 0)
|
|
114
|
+
return [];
|
|
115
|
+
const rrfScores = new Map();
|
|
116
|
+
const bestAppearance = new Map();
|
|
117
|
+
for (const list of nonEmpty) {
|
|
118
|
+
for (let rank = 0; rank < list.length; rank++) {
|
|
119
|
+
const item = list[rank];
|
|
120
|
+
let normalizedUrlStr;
|
|
121
|
+
try {
|
|
122
|
+
normalizedUrlStr = normalizeUrl(item.url);
|
|
123
|
+
}
|
|
124
|
+
catch {
|
|
125
|
+
normalizedUrlStr = item.url;
|
|
126
|
+
}
|
|
127
|
+
const rrfContribution = 1 / (RRF_K + rank + 1);
|
|
128
|
+
const current = rrfScores.get(normalizedUrlStr) ?? 0;
|
|
129
|
+
rrfScores.set(normalizedUrlStr, current + rrfContribution);
|
|
130
|
+
const existing = bestAppearance.get(normalizedUrlStr);
|
|
131
|
+
if (!existing || rank < existing.bestRank) {
|
|
132
|
+
bestAppearance.set(normalizedUrlStr, { result: item, bestRank: rank });
|
|
133
|
+
}
|
|
134
|
+
}
|
|
135
|
+
}
|
|
136
|
+
let maxScore = 0;
|
|
137
|
+
for (const score of rrfScores.values()) {
|
|
138
|
+
if (score > maxScore)
|
|
139
|
+
maxScore = score;
|
|
140
|
+
}
|
|
141
|
+
const merged = [];
|
|
142
|
+
for (const [normalizedUrlStr, score] of rrfScores.entries()) {
|
|
143
|
+
const appearance = bestAppearance.get(normalizedUrlStr);
|
|
144
|
+
merged.push({
|
|
145
|
+
...appearance.result,
|
|
146
|
+
relevance_score: maxScore > 0 ? score / maxScore : 0,
|
|
147
|
+
});
|
|
148
|
+
}
|
|
149
|
+
merged.sort((a, b) => b.relevance_score - a.relevance_score);
|
|
150
|
+
return merged;
|
|
151
|
+
}
|
|
152
|
+
catch (err) {
|
|
153
|
+
log.error('mergeWithRRF failed', { error: String(err) });
|
|
154
|
+
return [];
|
|
155
|
+
}
|
|
156
|
+
}
|
|
157
|
+
//# sourceMappingURL=multi-query.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"multi-query.js","sourceRoot":"","sources":["../../src/search/multi-query.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AACjD,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AAE5C,MAAM,GAAG,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;AAEnC,MAAM,KAAK,GAAG,EAAE,CAAC;AAEjB,MAAM,UAAU,gBAAgB,CAAC,OAAiB;IAChD,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,SAAS,EAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;QAC/B,MAAM,UAAU,GAAa,EAAE,CAAC;QAEhC,KAAK,MAAM,GAAG,IAAI,OAAO,EAAE,CAAC;YAC1B,MAAM,CAAC,GAAG,GAAG,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;YACxD,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC;gBAAE,SAAS;YAC7B,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;gBAAE,SAAS;YAC1B,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YACZ,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACrB,CAAC;QAED,IAAI,UAAU,CAAC,MAAM,GAAG,MAAM,CAAC,aAAa,EAAE,CAAC;YAC7C,GAAG,CAAC,IAAI,CAAC,2CAA2C,EAAE;gBACpD,QAAQ,EAAE,UAAU,CAAC,MAAM;gBAC3B,GAAG,EAAE,MAAM,CAAC,aAAa;aAC1B,CAAC,CAAC;YACH,OAAO,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,MAAM,CAAC,aAAa,CAAC,CAAC;QACnD,CAAC;QAED,OAAO,UAAU,CAAC;IACpB,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,GAAG,CAAC,KAAK,CAAC,yBAAyB,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAC7D,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAmBD,MAAM,CAAC,KAAK,UAAU,YAAY,CAChC,OAAiB,EACjB,OAAuB,EACvB,OAAsB;IAEtB,MAAM,UAAU,GAAsB,EAAE,CAAC;IACzC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;IACtC,MAAM,MAAM,GAAa,EAAE,CAAC;IAE5B,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACjD,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,WAAW,EAAE,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC;IACtD,CAAC;IAED,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,SAAS,EAAE,CAAC;QAC3B,MAAM,WAAW,GAAG,MAAM,CAAC,qBAAqB,CAAC;QAEjD,MAAM,kBAAkB,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,cAAc,EAAE,MAAM,IAAI,OAAO,CAAC,cAAc,EAAE,MAAM,CAAC,CAAC;QAChG,MAAM,eAAe,GAAG,kBAAkB,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAEnD,MAAM,aAAa,GAAwB;YACzC,UAAU,EAAE,OAAO,CAAC,UAAU,GAAG,eAAe;YAChD,SAAS,EAAE,OAAO,CAAC,SAAS;YAC5B,QAAQ,EAAE,OAAO,CAAC,QAAQ;YAC1B,cAAc,EAAE,OAAO,CAAC,cAAc;YACtC,cAAc,EAAE,OAAO,CAAC,cAAc;YACtC,QAAQ,EAAE,OAAO,CAAC,QAAQ;YAC1B,MAAM,EAAE,OAAO,CAAC,MAAM;YACtB,QAAQ,EAAE,OAAO,CAAC,QAAQ;SAC3B,CAAC;QAEF,MAAM,KAAK,GAAmD,EAAE,CAAC;QACjE,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;gBAC5B,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;YAChC,CAAC;QACH,CAAC;QAED,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,WAAW,EAAE,CAAC;YACnD,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,WAAW,CAAC,CAAC;YAC9C,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE,EAAE,EAAE;gBACrD,IAAI,CAAC;oBACH,MAAM,OAAO,GAAG,MAAM,MAAM,CAAC,MAAM,CAAC,KAAK,EAAE,aAAa,CAAC,CAAC;oBAC1D,KAAK,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;wBACxB,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;wBACnB,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;oBAC/B,CAAC;gBACH,CAAC;gBAAC,OAAO,GAAG,EAAE,CAAC;oBACb,MAAM,GAAG,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;oBAC7D,GAAG,CAAC,IAAI,CAAC,kCAAkC,EAAE;wBAC3C,MAAM,EAAE,MAAM,CAAC,IAAI;wBACnB,KAAK;wBACL,KAAK,EAAE,GAAG;qBACX,CAAC,CAAC;oBACH,MAAM,CAAC,IAAI,CAAC,GAAG,MAAM,CAAC,IAAI,IAAI,KAAK,MAAM,GAAG,EAAE,CAAC,CAAC;gBAClD,CAAC;YACH,CAAC,CAAC,CAAC;YAEH,MAAM,OAAO,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC;QACrC,CAAC;QAED,OAAO;YACL,OAAO,EAAE,UAAU;YACnB,WAAW,EAAE,CAAC,GAAG,WAAW,CAAC;YAC7B,MAAM;SACP,CAAC;IACJ,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,GAAG,CAAC,KAAK,CAAC,qBAAqB,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACzD,OAAO;YACL,OAAO,EAAE,UAAU;YACnB,WAAW,EAAE,CAAC,GAAG,WAAW,CAAC;YAC7B,MAAM,EAAE,CAAC,GAAG,MAAM,EAAE,iBAAiB,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;SACpD,CAAC;IACJ,CAAC;AACH,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,OAAiB;IAChD,IAAI,CAAC;QACH,OAAO,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC/D,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,GAAG,CAAC,KAAK,CAAC,yBAAyB,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAC7D,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,WAAmC;IAC9D,IAAI,CAAC;QACH,IAAI,WAAW,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAExC,MAAM,QAAQ,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QACvD,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QAErC,MAAM,SAAS,GAAG,IAAI,GAAG,EAAkB,CAAC;QAC5C,MAAM,cAAc,GAAG,IAAI,GAAG,EAA4D,CAAC;QAE3F,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;YAC5B,KAAK,IAAI,IAAI,GAAG,CAAC,EAAE,IAAI,GAAG,IAAI,CAAC,MAAM,EAAE,IAAI,EAAE,EAAE,CAAC;gBAC9C,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC;gBACxB,IAAI,gBAAwB,CAAC;gBAC7B,IAAI,CAAC;oBACH,gBAAgB,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBAC5C,CAAC;gBAAC,MAAM,CAAC;oBACP,gBAAgB,GAAG,IAAI,CAAC,GAAG,CAAC;gBAC9B,CAAC;gBAED,MAAM,eAAe,GAAG,CAAC,GAAG,CAAC,KAAK,GAAG,IAAI,GAAG,CAAC,CAAC,CAAC;gBAC/C,MAAM,OAAO,GAAG,SAAS,CAAC,GAAG,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC;gBACrD,SAAS,CAAC,GAAG,CAAC,gBAAgB,EAAE,OAAO,GAAG,eAAe,CAAC,CAAC;gBAE3D,MAAM,QAAQ,GAAG,cAAc,CAAC,GAAG,CAAC,gBAAgB,CAAC,CAAC;gBACtD,IAAI,CAAC,QAAQ,IAAI,IAAI,GAAG,QAAQ,CAAC,QAAQ,EAAE,CAAC;oBAC1C,cAAc,CAAC,GAAG,CAAC,gBAAgB,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC;gBACzE,CAAC;YACH,CAAC;QACH,CAAC;QAED,IAAI,QAAQ,GAAG,CAAC,CAAC;QACjB,KAAK,MAAM,KAAK,IAAI,SAAS,CAAC,MAAM,EAAE,EAAE,CAAC;YACvC,IAAI,KAAK,GAAG,QAAQ;gBAAE,QAAQ,GAAG,KAAK,CAAC;QACzC,CAAC;QAED,MAAM,MAAM,GAAyB,EAAE,CAAC;QACxC,KAAK,MAAM,CAAC,gBAAgB,EAAE,KAAK,CAAC,IAAI,SAAS,CAAC,OAAO,EAAE,EAAE,CAAC;YAC5D,MAAM,UAAU,GAAG,cAAc,CAAC,GAAG,CAAC,gBAAgB,CAAE,CAAC;YACzD,MAAM,CAAC,IAAI,CAAC;gBACV,GAAG,UAAU,CAAC,MAAM;gBACpB,eAAe,EAAE,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;aACrD,CAAC,CAAC;QACL,CAAC;QAED,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,eAAe,GAAG,CAAC,CAAC,eAAe,CAAC,CAAC;QAC7D,OAAO,MAAM,CAAC;IAChB,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,GAAG,CAAC,KAAK,CAAC,qBAAqB,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACzD,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC"}
|