opencode-mem 1.0.0

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.
Files changed (84) hide show
  1. package/README.md +588 -0
  2. package/dist/config.d.ts +33 -0
  3. package/dist/config.d.ts.map +1 -0
  4. package/dist/config.js +258 -0
  5. package/dist/index.d.ts +3 -0
  6. package/dist/index.d.ts.map +1 -0
  7. package/dist/index.js +618 -0
  8. package/dist/plugin.d.ts +5 -0
  9. package/dist/plugin.d.ts.map +1 -0
  10. package/dist/plugin.js +15 -0
  11. package/dist/services/api-handlers.d.ts +102 -0
  12. package/dist/services/api-handlers.d.ts.map +1 -0
  13. package/dist/services/api-handlers.js +494 -0
  14. package/dist/services/auto-capture.d.ts +32 -0
  15. package/dist/services/auto-capture.d.ts.map +1 -0
  16. package/dist/services/auto-capture.js +451 -0
  17. package/dist/services/cleanup-service.d.ts +20 -0
  18. package/dist/services/cleanup-service.d.ts.map +1 -0
  19. package/dist/services/cleanup-service.js +88 -0
  20. package/dist/services/client.d.ts +104 -0
  21. package/dist/services/client.d.ts.map +1 -0
  22. package/dist/services/client.js +251 -0
  23. package/dist/services/compaction.d.ts +92 -0
  24. package/dist/services/compaction.d.ts.map +1 -0
  25. package/dist/services/compaction.js +421 -0
  26. package/dist/services/context.d.ts +17 -0
  27. package/dist/services/context.d.ts.map +1 -0
  28. package/dist/services/context.js +41 -0
  29. package/dist/services/deduplication-service.d.ts +30 -0
  30. package/dist/services/deduplication-service.d.ts.map +1 -0
  31. package/dist/services/deduplication-service.js +131 -0
  32. package/dist/services/embedding.d.ts +10 -0
  33. package/dist/services/embedding.d.ts.map +1 -0
  34. package/dist/services/embedding.js +77 -0
  35. package/dist/services/jsonc.d.ts +7 -0
  36. package/dist/services/jsonc.d.ts.map +1 -0
  37. package/dist/services/jsonc.js +76 -0
  38. package/dist/services/logger.d.ts +2 -0
  39. package/dist/services/logger.d.ts.map +1 -0
  40. package/dist/services/logger.js +16 -0
  41. package/dist/services/migration-service.d.ts +42 -0
  42. package/dist/services/migration-service.d.ts.map +1 -0
  43. package/dist/services/migration-service.js +258 -0
  44. package/dist/services/privacy.d.ts +4 -0
  45. package/dist/services/privacy.d.ts.map +1 -0
  46. package/dist/services/privacy.js +10 -0
  47. package/dist/services/sqlite/connection-manager.d.ts +10 -0
  48. package/dist/services/sqlite/connection-manager.d.ts.map +1 -0
  49. package/dist/services/sqlite/connection-manager.js +45 -0
  50. package/dist/services/sqlite/shard-manager.d.ts +20 -0
  51. package/dist/services/sqlite/shard-manager.d.ts.map +1 -0
  52. package/dist/services/sqlite/shard-manager.js +221 -0
  53. package/dist/services/sqlite/types.d.ts +39 -0
  54. package/dist/services/sqlite/types.d.ts.map +1 -0
  55. package/dist/services/sqlite/types.js +1 -0
  56. package/dist/services/sqlite/vector-search.d.ts +18 -0
  57. package/dist/services/sqlite/vector-search.d.ts.map +1 -0
  58. package/dist/services/sqlite/vector-search.js +129 -0
  59. package/dist/services/sqlite-client.d.ts +116 -0
  60. package/dist/services/sqlite-client.d.ts.map +1 -0
  61. package/dist/services/sqlite-client.js +284 -0
  62. package/dist/services/tags.d.ts +20 -0
  63. package/dist/services/tags.d.ts.map +1 -0
  64. package/dist/services/tags.js +76 -0
  65. package/dist/services/web-server-lock.d.ts +12 -0
  66. package/dist/services/web-server-lock.d.ts.map +1 -0
  67. package/dist/services/web-server-lock.js +157 -0
  68. package/dist/services/web-server-worker.d.ts +2 -0
  69. package/dist/services/web-server-worker.d.ts.map +1 -0
  70. package/dist/services/web-server-worker.js +221 -0
  71. package/dist/services/web-server.d.ts +22 -0
  72. package/dist/services/web-server.d.ts.map +1 -0
  73. package/dist/services/web-server.js +134 -0
  74. package/dist/types/index.d.ts +48 -0
  75. package/dist/types/index.d.ts.map +1 -0
  76. package/dist/types/index.js +1 -0
  77. package/dist/web/app.d.ts +2 -0
  78. package/dist/web/app.d.ts.map +1 -0
  79. package/dist/web/app.js +691 -0
  80. package/dist/web/favicon.ico +0 -0
  81. package/dist/web/favicon.svg +14 -0
  82. package/dist/web/index.html +202 -0
  83. package/dist/web/styles.css +851 -0
  84. package/package.json +52 -0
@@ -0,0 +1,421 @@
1
+ import { existsSync, mkdirSync, readdirSync, readFileSync, writeFileSync } from "node:fs";
2
+ import { join } from "node:path";
3
+ import { homedir } from "node:os";
4
+ import { memoryClient } from "./client.js";
5
+ import { log } from "./logger.js";
6
+ import { CONFIG } from "../config.js";
7
+ const MESSAGE_STORAGE = join(homedir(), ".opencode", "messages");
8
+ const PART_STORAGE = join(homedir(), ".opencode", "parts");
9
+ const DEFAULT_THRESHOLD = 0.80;
10
+ const MIN_TOKENS_FOR_COMPACTION = 50_000;
11
+ const COMPACTION_COOLDOWN_MS = 30_000;
12
+ const CLAUDE_DEFAULT_CONTEXT_LIMIT = 200_000;
13
+ const CLAUDE_MODEL_PATTERN = /claude-(opus|sonnet|haiku)/i;
14
+ function createCompactionPrompt(projectMemories) {
15
+ const memoriesSection = projectMemories.length > 0
16
+ ? `
17
+ ## Project Knowledge (from Memory System)
18
+ The following project-specific knowledge should be preserved and referenced in the summary:
19
+ ${projectMemories.map(m => `- ${m}`).join('\n')}
20
+ `
21
+ : '';
22
+ return `[COMPACTION CONTEXT INJECTION]
23
+
24
+ When summarizing this session, you MUST include the following sections in your summary:
25
+
26
+ ## 1. User Requests (As-Is)
27
+ - List all original user requests exactly as they were stated
28
+ - Preserve the user's exact wording and intent
29
+
30
+ ## 2. Final Goal
31
+ - What the user ultimately wanted to achieve
32
+ - The end result or deliverable expected
33
+
34
+ ## 3. Work Completed
35
+ - What has been done so far
36
+ - Files created/modified
37
+ - Features implemented
38
+ - Problems solved
39
+
40
+ ## 4. Remaining Tasks
41
+ - What still needs to be done
42
+ - Pending items from the original request
43
+ - Follow-up tasks identified during the work
44
+
45
+ ## 5. MUST NOT Do (Critical Constraints)
46
+ - Things that were explicitly forbidden
47
+ - Approaches that failed and should not be retried
48
+ - User's explicit restrictions or preferences
49
+ - Anti-patterns identified during the session
50
+ ${memoriesSection}
51
+ This context is critical for maintaining continuity after compaction.
52
+ `;
53
+ }
54
+ function isSupportedModel(modelID) {
55
+ return CLAUDE_MODEL_PATTERN.test(modelID);
56
+ }
57
+ function getMessageDir(sessionID) {
58
+ if (!existsSync(MESSAGE_STORAGE))
59
+ return null;
60
+ const directPath = join(MESSAGE_STORAGE, sessionID);
61
+ if (existsSync(directPath))
62
+ return directPath;
63
+ for (const dir of readdirSync(MESSAGE_STORAGE)) {
64
+ const sessionPath = join(MESSAGE_STORAGE, dir, sessionID);
65
+ if (existsSync(sessionPath))
66
+ return sessionPath;
67
+ }
68
+ return null;
69
+ }
70
+ function getOrCreateMessageDir(sessionID) {
71
+ if (!existsSync(MESSAGE_STORAGE)) {
72
+ mkdirSync(MESSAGE_STORAGE, { recursive: true });
73
+ }
74
+ const directPath = join(MESSAGE_STORAGE, sessionID);
75
+ if (existsSync(directPath))
76
+ return directPath;
77
+ for (const dir of readdirSync(MESSAGE_STORAGE)) {
78
+ const sessionPath = join(MESSAGE_STORAGE, dir, sessionID);
79
+ if (existsSync(sessionPath))
80
+ return sessionPath;
81
+ }
82
+ mkdirSync(directPath, { recursive: true });
83
+ return directPath;
84
+ }
85
+ function findNearestMessageWithFields(messageDir) {
86
+ try {
87
+ const files = readdirSync(messageDir)
88
+ .filter((f) => f.endsWith(".json"))
89
+ .sort()
90
+ .reverse();
91
+ for (const file of files) {
92
+ try {
93
+ const content = readFileSync(join(messageDir, file), "utf-8");
94
+ const msg = JSON.parse(content);
95
+ if (msg.agent && msg.model?.providerID && msg.model?.modelID) {
96
+ return msg;
97
+ }
98
+ }
99
+ catch {
100
+ continue;
101
+ }
102
+ }
103
+ }
104
+ catch {
105
+ return null;
106
+ }
107
+ return null;
108
+ }
109
+ function generateMessageId() {
110
+ const timestamp = Date.now().toString(16);
111
+ const random = Math.random().toString(36).substring(2, 14);
112
+ return `msg_${timestamp}${random}`;
113
+ }
114
+ function generatePartId() {
115
+ const timestamp = Date.now().toString(16);
116
+ const random = Math.random().toString(36).substring(2, 10);
117
+ return `prt_${timestamp}${random}`;
118
+ }
119
+ function injectHookMessage(sessionID, hookContent, originalMessage) {
120
+ if (!hookContent || hookContent.trim().length === 0) {
121
+ log("[compaction] attempted to inject empty content, skipping");
122
+ return false;
123
+ }
124
+ const messageDir = getOrCreateMessageDir(sessionID);
125
+ const fallback = findNearestMessageWithFields(messageDir);
126
+ const now = Date.now();
127
+ const messageID = generateMessageId();
128
+ const partID = generatePartId();
129
+ const resolvedAgent = originalMessage.agent ?? fallback?.agent ?? "general";
130
+ const resolvedModel = originalMessage.model?.providerID && originalMessage.model?.modelID
131
+ ? { providerID: originalMessage.model.providerID, modelID: originalMessage.model.modelID }
132
+ : fallback?.model?.providerID && fallback?.model?.modelID
133
+ ? { providerID: fallback.model.providerID, modelID: fallback.model.modelID }
134
+ : undefined;
135
+ const messageMeta = {
136
+ id: messageID,
137
+ sessionID,
138
+ role: "user",
139
+ time: { created: now },
140
+ agent: resolvedAgent,
141
+ model: resolvedModel,
142
+ path: originalMessage.path?.cwd
143
+ ? { cwd: originalMessage.path.cwd, root: originalMessage.path.root ?? "/" }
144
+ : undefined,
145
+ };
146
+ const textPart = {
147
+ id: partID,
148
+ type: "text",
149
+ text: hookContent,
150
+ synthetic: true,
151
+ time: { start: now, end: now },
152
+ messageID,
153
+ sessionID,
154
+ };
155
+ try {
156
+ writeFileSync(join(messageDir, `${messageID}.json`), JSON.stringify(messageMeta, null, 2));
157
+ const partDir = join(PART_STORAGE, messageID);
158
+ if (!existsSync(partDir)) {
159
+ mkdirSync(partDir, { recursive: true });
160
+ }
161
+ writeFileSync(join(partDir, `${partID}.json`), JSON.stringify(textPart, null, 2));
162
+ log("[compaction] hook message injected", { sessionID, messageID });
163
+ return true;
164
+ }
165
+ catch (err) {
166
+ log("[compaction] failed to inject hook message", { error: String(err) });
167
+ return false;
168
+ }
169
+ }
170
+ export function createCompactionHook(ctx, tags, options) {
171
+ const state = {
172
+ lastCompactionTime: new Map(),
173
+ compactionInProgress: new Set(),
174
+ summarizedSessions: new Set(),
175
+ };
176
+ const threshold = options?.threshold ?? DEFAULT_THRESHOLD;
177
+ const getModelLimit = options?.getModelLimit;
178
+ async function fetchProjectMemoriesForCompaction() {
179
+ try {
180
+ const result = await memoryClient.listMemories(tags.project, CONFIG.maxProjectMemories);
181
+ const memories = result.memories || [];
182
+ return memories.map((m) => m.summary || m.content || "").filter(Boolean);
183
+ }
184
+ catch (err) {
185
+ log("[compaction] failed to fetch project memories", { error: String(err) });
186
+ return [];
187
+ }
188
+ }
189
+ async function injectCompactionContext(summarizeCtx) {
190
+ log("[compaction] injecting context", { sessionID: summarizeCtx.sessionID });
191
+ const projectMemories = await fetchProjectMemoriesForCompaction();
192
+ const prompt = createCompactionPrompt(projectMemories);
193
+ const success = injectHookMessage(summarizeCtx.sessionID, prompt, {
194
+ agent: summarizeCtx.agent,
195
+ model: { providerID: summarizeCtx.providerID, modelID: summarizeCtx.modelID },
196
+ path: { cwd: summarizeCtx.directory },
197
+ });
198
+ if (success) {
199
+ log("[compaction] context injected with project memories", {
200
+ sessionID: summarizeCtx.sessionID,
201
+ memoriesCount: projectMemories.length
202
+ });
203
+ }
204
+ }
205
+ async function saveSummaryAsMemory(sessionID, summaryContent) {
206
+ if (!summaryContent || summaryContent.length < 100) {
207
+ log("[compaction] summary too short to save", { sessionID, length: summaryContent.length });
208
+ return;
209
+ }
210
+ try {
211
+ const result = await memoryClient.addMemory(`[Session Summary]\n${summaryContent}`, tags.project, { type: "conversation" });
212
+ if (result.success) {
213
+ log("[compaction] summary saved as memory", { sessionID, memoryId: result.id });
214
+ }
215
+ else {
216
+ log("[compaction] failed to save summary", { error: result.error });
217
+ }
218
+ }
219
+ catch (err) {
220
+ log("[compaction] failed to save summary", { error: String(err) });
221
+ }
222
+ }
223
+ async function checkAndTriggerCompaction(sessionID, lastAssistant) {
224
+ if (state.compactionInProgress.has(sessionID))
225
+ return;
226
+ const lastCompaction = state.lastCompactionTime.get(sessionID) ?? 0;
227
+ if (Date.now() - lastCompaction < COMPACTION_COOLDOWN_MS)
228
+ return;
229
+ if (lastAssistant.summary === true)
230
+ return;
231
+ const tokens = lastAssistant.tokens;
232
+ if (!tokens)
233
+ return;
234
+ let modelID = lastAssistant.modelID ?? "";
235
+ let providerID = lastAssistant.providerID ?? "";
236
+ let agent;
237
+ // Fallback: find model/agent from stored messages if not available
238
+ const messageDir = getMessageDir(sessionID);
239
+ const storedMessage = messageDir ? findNearestMessageWithFields(messageDir) : null;
240
+ if (!providerID || !modelID) {
241
+ if (storedMessage?.model?.providerID)
242
+ providerID = storedMessage.model.providerID;
243
+ if (storedMessage?.model?.modelID)
244
+ modelID = storedMessage.model.modelID;
245
+ }
246
+ agent = storedMessage?.agent;
247
+ if (!isSupportedModel(modelID)) {
248
+ log("[compaction] skipping unsupported model", { modelID });
249
+ return;
250
+ }
251
+ const configLimit = getModelLimit?.(providerID, modelID);
252
+ const contextLimit = configLimit ?? CLAUDE_DEFAULT_CONTEXT_LIMIT;
253
+ const totalUsed = tokens.input + tokens.cache.read + tokens.output;
254
+ if (totalUsed < MIN_TOKENS_FOR_COMPACTION)
255
+ return;
256
+ const usageRatio = totalUsed / contextLimit;
257
+ log("[compaction] checking", {
258
+ sessionID,
259
+ totalUsed,
260
+ contextLimit,
261
+ usageRatio: usageRatio.toFixed(2),
262
+ threshold,
263
+ });
264
+ if (usageRatio < threshold)
265
+ return;
266
+ state.compactionInProgress.add(sessionID);
267
+ state.lastCompactionTime.set(sessionID, Date.now());
268
+ if (!providerID || !modelID) {
269
+ state.compactionInProgress.delete(sessionID);
270
+ return;
271
+ }
272
+ await ctx.client.tui.showToast({
273
+ body: {
274
+ title: "Preemptive Compaction",
275
+ message: `Context at ${(usageRatio * 100).toFixed(0)}% - compacting with memory context...`,
276
+ variant: "warning",
277
+ duration: 3000,
278
+ },
279
+ }).catch(() => { });
280
+ log("[compaction] triggering compaction", { sessionID, usageRatio });
281
+ try {
282
+ await injectCompactionContext({
283
+ sessionID,
284
+ providerID,
285
+ modelID,
286
+ usageRatio,
287
+ directory: ctx.directory,
288
+ agent,
289
+ });
290
+ state.summarizedSessions.add(sessionID);
291
+ await ctx.client.session.summarize({
292
+ path: { id: sessionID },
293
+ body: { providerID, modelID },
294
+ query: { directory: ctx.directory },
295
+ });
296
+ await ctx.client.tui.showToast({
297
+ body: {
298
+ title: "Compaction Complete",
299
+ message: "Session compacted with memory context. Resuming...",
300
+ variant: "success",
301
+ duration: 2000,
302
+ },
303
+ }).catch(() => { });
304
+ state.compactionInProgress.delete(sessionID);
305
+ setTimeout(async () => {
306
+ try {
307
+ const messageDir = getMessageDir(sessionID);
308
+ const storedMessage = messageDir ? findNearestMessageWithFields(messageDir) : null;
309
+ await ctx.client.session.promptAsync({
310
+ path: { id: sessionID },
311
+ body: {
312
+ agent: storedMessage?.agent,
313
+ parts: [{ type: "text", text: "Continue" }],
314
+ },
315
+ query: { directory: ctx.directory },
316
+ });
317
+ }
318
+ catch { }
319
+ }, 500);
320
+ }
321
+ catch (err) {
322
+ log("[compaction] compaction failed", { sessionID, error: String(err) });
323
+ state.compactionInProgress.delete(sessionID);
324
+ }
325
+ }
326
+ async function handleSummaryMessage(sessionID, _messageInfo) {
327
+ log("[compaction] handleSummaryMessage called", { sessionID, inSet: state.summarizedSessions.has(sessionID) });
328
+ if (!state.summarizedSessions.has(sessionID))
329
+ return;
330
+ state.summarizedSessions.delete(sessionID);
331
+ log("[compaction] capturing summary for memory", { sessionID });
332
+ try {
333
+ const resp = await ctx.client.session.messages({
334
+ path: { id: sessionID },
335
+ query: { directory: ctx.directory },
336
+ });
337
+ const messages = (resp.data ?? resp);
338
+ const summaryMessage = messages.find(m => m.info.role === "assistant" &&
339
+ m.info.summary === true);
340
+ log("[compaction] looking for summary message", {
341
+ sessionID,
342
+ found: !!summaryMessage,
343
+ hasParts: !!summaryMessage?.parts
344
+ });
345
+ if (summaryMessage?.parts) {
346
+ const textParts = summaryMessage.parts.filter(p => p.type === "text" && p.text);
347
+ const summaryContent = textParts.map(p => p.text).join("\n");
348
+ log("[compaction] summary content", {
349
+ sessionID,
350
+ textPartsCount: textParts.length,
351
+ contentLength: summaryContent.length
352
+ });
353
+ if (summaryContent) {
354
+ await saveSummaryAsMemory(sessionID, summaryContent);
355
+ }
356
+ }
357
+ }
358
+ catch (err) {
359
+ log("[compaction] failed to capture summary", { error: String(err) });
360
+ }
361
+ }
362
+ return {
363
+ compactionTracker: state.lastCompactionTime,
364
+ async event({ event }) {
365
+ const props = event.properties;
366
+ if (event.type === "session.deleted") {
367
+ const sessionInfo = props?.info;
368
+ if (sessionInfo?.id) {
369
+ state.lastCompactionTime.delete(sessionInfo.id);
370
+ state.compactionInProgress.delete(sessionInfo.id);
371
+ state.summarizedSessions.delete(sessionInfo.id);
372
+ }
373
+ return;
374
+ }
375
+ if (event.type === "message.updated") {
376
+ const info = props?.info;
377
+ if (!info)
378
+ return;
379
+ const sessionID = info.sessionID;
380
+ if (!sessionID)
381
+ return;
382
+ if (info.role === "assistant" && info.summary === true && info.finish) {
383
+ await handleSummaryMessage(sessionID, info);
384
+ return;
385
+ }
386
+ if (info.role !== "assistant" || !info.finish)
387
+ return;
388
+ await checkAndTriggerCompaction(sessionID, info);
389
+ return;
390
+ }
391
+ if (event.type === "session.idle") {
392
+ const sessionID = props?.sessionID;
393
+ if (!sessionID)
394
+ return;
395
+ try {
396
+ const resp = await ctx.client.session.messages({
397
+ path: { id: sessionID },
398
+ query: { directory: ctx.directory },
399
+ });
400
+ const messages = (resp.data ?? resp);
401
+ const assistants = messages
402
+ .filter((m) => m.info.role === "assistant")
403
+ .map((m) => m.info);
404
+ if (assistants.length === 0)
405
+ return;
406
+ const lastAssistant = assistants[assistants.length - 1];
407
+ if (!lastAssistant.providerID || !lastAssistant.modelID) {
408
+ const messageDir = getMessageDir(sessionID);
409
+ const storedMessage = messageDir ? findNearestMessageWithFields(messageDir) : null;
410
+ if (storedMessage?.model?.providerID && storedMessage?.model?.modelID) {
411
+ lastAssistant.providerID = storedMessage.model.providerID;
412
+ lastAssistant.modelID = storedMessage.model.modelID;
413
+ }
414
+ }
415
+ await checkAndTriggerCompaction(sessionID, lastAssistant);
416
+ }
417
+ catch { }
418
+ }
419
+ },
420
+ };
421
+ }
@@ -0,0 +1,17 @@
1
+ interface MemoryResultMinimal {
2
+ similarity: number;
3
+ memory?: string;
4
+ chunk?: string;
5
+ }
6
+ interface MemoriesResponseMinimal {
7
+ results?: MemoryResultMinimal[];
8
+ }
9
+ interface ProfileResponse {
10
+ profile?: {
11
+ static: string[];
12
+ dynamic: string[];
13
+ };
14
+ }
15
+ export declare function formatContextForPrompt(profile: ProfileResponse | null, userMemories: MemoriesResponseMinimal, projectMemories: MemoriesResponseMinimal): string;
16
+ export {};
17
+ //# sourceMappingURL=context.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"context.d.ts","sourceRoot":"","sources":["../../src/services/context.ts"],"names":[],"mappings":"AAEA,UAAU,mBAAmB;IAC3B,UAAU,EAAE,MAAM,CAAC;IACnB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED,UAAU,uBAAuB;IAC/B,OAAO,CAAC,EAAE,mBAAmB,EAAE,CAAC;CACjC;AAED,UAAU,eAAe;IACvB,OAAO,CAAC,EAAE;QACR,MAAM,EAAE,MAAM,EAAE,CAAC;QACjB,OAAO,EAAE,MAAM,EAAE,CAAC;KACnB,CAAC;CACH;AAED,wBAAgB,sBAAsB,CACpC,OAAO,EAAE,eAAe,GAAG,IAAI,EAC/B,YAAY,EAAE,uBAAuB,EACrC,eAAe,EAAE,uBAAuB,GACvC,MAAM,CA8CR"}
@@ -0,0 +1,41 @@
1
+ import { CONFIG } from "../config.js";
2
+ export function formatContextForPrompt(profile, userMemories, projectMemories) {
3
+ const parts = ["[MEMORY]"];
4
+ if (CONFIG.injectProfile && profile?.profile) {
5
+ const { static: staticFacts, dynamic: dynamicFacts } = profile.profile;
6
+ if (staticFacts.length > 0) {
7
+ parts.push("\nUser Profile:");
8
+ staticFacts.slice(0, CONFIG.maxProfileItems).forEach((fact) => {
9
+ parts.push(`- ${fact}`);
10
+ });
11
+ }
12
+ if (dynamicFacts.length > 0) {
13
+ parts.push("\nRecent Context:");
14
+ dynamicFacts.slice(0, CONFIG.maxProfileItems).forEach((fact) => {
15
+ parts.push(`- ${fact}`);
16
+ });
17
+ }
18
+ }
19
+ const projectResults = projectMemories.results || [];
20
+ if (projectResults.length > 0) {
21
+ parts.push("\nProject Knowledge:");
22
+ projectResults.forEach((mem) => {
23
+ const similarity = Math.round(mem.similarity * 100);
24
+ const content = mem.memory || mem.chunk || "";
25
+ parts.push(`- [${similarity}%] ${content}`);
26
+ });
27
+ }
28
+ const userResults = userMemories.results || [];
29
+ if (userResults.length > 0) {
30
+ parts.push("\nRelevant Memories:");
31
+ userResults.forEach((mem) => {
32
+ const similarity = Math.round(mem.similarity * 100);
33
+ const content = mem.memory || mem.chunk || "";
34
+ parts.push(`- [${similarity}%] ${content}`);
35
+ });
36
+ }
37
+ if (parts.length === 1) {
38
+ return "";
39
+ }
40
+ return parts.join("\n");
41
+ }
@@ -0,0 +1,30 @@
1
+ interface DuplicateGroup {
2
+ representative: {
3
+ id: string;
4
+ content: string;
5
+ containerTag: string;
6
+ createdAt: number;
7
+ };
8
+ duplicates: Array<{
9
+ id: string;
10
+ content: string;
11
+ similarity: number;
12
+ }>;
13
+ }
14
+ interface DeduplicationResult {
15
+ exactDuplicatesDeleted: number;
16
+ nearDuplicateGroups: DuplicateGroup[];
17
+ }
18
+ export declare class DeduplicationService {
19
+ private isRunning;
20
+ detectAndRemoveDuplicates(): Promise<DeduplicationResult>;
21
+ private cosineSimilarity;
22
+ getStatus(): {
23
+ enabled: boolean;
24
+ threshold: number;
25
+ isRunning: boolean;
26
+ };
27
+ }
28
+ export declare const deduplicationService: DeduplicationService;
29
+ export {};
30
+ //# sourceMappingURL=deduplication-service.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"deduplication-service.d.ts","sourceRoot":"","sources":["../../src/services/deduplication-service.ts"],"names":[],"mappings":"AAOA,UAAU,cAAc;IACtB,cAAc,EAAE;QACd,EAAE,EAAE,MAAM,CAAC;QACX,OAAO,EAAE,MAAM,CAAC;QAChB,YAAY,EAAE,MAAM,CAAC;QACrB,SAAS,EAAE,MAAM,CAAC;KACnB,CAAC;IACF,UAAU,EAAE,KAAK,CAAC;QAChB,EAAE,EAAE,MAAM,CAAC;QACX,OAAO,EAAE,MAAM,CAAC;QAChB,UAAU,EAAE,MAAM,CAAC;KACpB,CAAC,CAAC;CACJ;AAED,UAAU,mBAAmB;IAC3B,sBAAsB,EAAE,MAAM,CAAC;IAC/B,mBAAmB,EAAE,cAAc,EAAE,CAAC;CACvC;AAED,qBAAa,oBAAoB;IAC/B,OAAO,CAAC,SAAS,CAAkB;IAE7B,yBAAyB,IAAI,OAAO,CAAC,mBAAmB,CAAC;IAgH/D,OAAO,CAAC,gBAAgB;IAoBxB,SAAS;;;;;CAOV;AAED,eAAO,MAAM,oBAAoB,sBAA6B,CAAC"}
@@ -0,0 +1,131 @@
1
+ import { embeddingService } from "./embedding.js";
2
+ import { shardManager } from "./sqlite/shard-manager.js";
3
+ import { vectorSearch } from "./sqlite/vector-search.js";
4
+ import { connectionManager } from "./sqlite/connection-manager.js";
5
+ import { CONFIG } from "../config.js";
6
+ import { log } from "./logger.js";
7
+ export class DeduplicationService {
8
+ isRunning = false;
9
+ async detectAndRemoveDuplicates() {
10
+ if (this.isRunning) {
11
+ throw new Error("Deduplication already running");
12
+ }
13
+ if (!CONFIG.deduplicationEnabled) {
14
+ throw new Error("Deduplication is disabled in config");
15
+ }
16
+ this.isRunning = true;
17
+ try {
18
+ log("Deduplication: starting", {
19
+ threshold: CONFIG.deduplicationSimilarityThreshold,
20
+ });
21
+ const userShards = shardManager.getAllShards("user", "");
22
+ const projectShards = shardManager.getAllShards("project", "");
23
+ const allShards = [...userShards, ...projectShards];
24
+ let exactDeleted = 0;
25
+ const nearDuplicateGroups = [];
26
+ for (const shard of allShards) {
27
+ const db = connectionManager.getConnection(shard.dbPath);
28
+ const memories = vectorSearch.getAllMemories(db);
29
+ const contentMap = new Map();
30
+ for (const memory of memories) {
31
+ const key = `${memory.container_tag}:${memory.content}`;
32
+ if (!contentMap.has(key)) {
33
+ contentMap.set(key, []);
34
+ }
35
+ contentMap.get(key).push(memory);
36
+ }
37
+ for (const [key, duplicates] of contentMap) {
38
+ if (duplicates.length > 1) {
39
+ duplicates.sort((a, b) => Number(b.created_at) - Number(a.created_at));
40
+ const keep = duplicates[0];
41
+ const toDelete = duplicates.slice(1);
42
+ for (const dup of toDelete) {
43
+ try {
44
+ vectorSearch.deleteVector(db, dup.id);
45
+ shardManager.decrementVectorCount(shard.id);
46
+ exactDeleted++;
47
+ }
48
+ catch (error) {
49
+ log("Deduplication: delete error", {
50
+ memoryId: dup.id,
51
+ error: String(error),
52
+ });
53
+ }
54
+ }
55
+ }
56
+ }
57
+ const uniqueMemories = Array.from(contentMap.values()).map((arr) => arr[0]);
58
+ for (let i = 0; i < uniqueMemories.length; i++) {
59
+ const mem1 = uniqueMemories[i];
60
+ if (!mem1.vector)
61
+ continue;
62
+ const vector1 = new Float32Array(new Uint8Array(mem1.vector).buffer);
63
+ const similarGroup = {
64
+ representative: {
65
+ id: mem1.id,
66
+ content: mem1.content,
67
+ containerTag: mem1.container_tag,
68
+ createdAt: mem1.created_at,
69
+ },
70
+ duplicates: [],
71
+ };
72
+ for (let j = i + 1; j < uniqueMemories.length; j++) {
73
+ const mem2 = uniqueMemories[j];
74
+ if (!mem2.vector)
75
+ continue;
76
+ if (mem1.container_tag !== mem2.container_tag)
77
+ continue;
78
+ const vector2 = new Float32Array(new Uint8Array(mem2.vector).buffer);
79
+ const similarity = this.cosineSimilarity(vector1, vector2);
80
+ if (similarity >= CONFIG.deduplicationSimilarityThreshold && similarity < 1.0) {
81
+ similarGroup.duplicates.push({
82
+ id: mem2.id,
83
+ content: mem2.content,
84
+ similarity,
85
+ });
86
+ }
87
+ }
88
+ if (similarGroup.duplicates.length > 0) {
89
+ nearDuplicateGroups.push(similarGroup);
90
+ }
91
+ }
92
+ }
93
+ log("Deduplication: completed", {
94
+ exactDeleted,
95
+ nearDuplicateGroupsFound: nearDuplicateGroups.length,
96
+ });
97
+ return {
98
+ exactDuplicatesDeleted: exactDeleted,
99
+ nearDuplicateGroups,
100
+ };
101
+ }
102
+ finally {
103
+ this.isRunning = false;
104
+ }
105
+ }
106
+ cosineSimilarity(a, b) {
107
+ if (a.length !== b.length)
108
+ return 0;
109
+ let dotProduct = 0;
110
+ let normA = 0;
111
+ let normB = 0;
112
+ for (let i = 0; i < a.length; i++) {
113
+ const aVal = a[i] || 0;
114
+ const bVal = b[i] || 0;
115
+ dotProduct += aVal * bVal;
116
+ normA += aVal * aVal;
117
+ normB += bVal * bVal;
118
+ }
119
+ if (normA === 0 || normB === 0)
120
+ return 0;
121
+ return dotProduct / (Math.sqrt(normA) * Math.sqrt(normB));
122
+ }
123
+ getStatus() {
124
+ return {
125
+ enabled: CONFIG.deduplicationEnabled,
126
+ threshold: CONFIG.deduplicationSimilarityThreshold,
127
+ isRunning: this.isRunning,
128
+ };
129
+ }
130
+ }
131
+ export const deduplicationService = new DeduplicationService();
@@ -0,0 +1,10 @@
1
+ export declare class EmbeddingService {
2
+ private pipe;
3
+ private initPromise;
4
+ isWarmedUp: boolean;
5
+ warmup(progressCallback?: (progress: any) => void): Promise<void>;
6
+ embed(text: string): Promise<Float32Array>;
7
+ embedWithTimeout(text: string): Promise<Float32Array>;
8
+ }
9
+ export declare const embeddingService: EmbeddingService;
10
+ //# sourceMappingURL=embedding.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"embedding.d.ts","sourceRoot":"","sources":["../../src/services/embedding.ts"],"names":[],"mappings":"AAiBA,qBAAa,gBAAgB;IAC3B,OAAO,CAAC,IAAI,CAAa;IACzB,OAAO,CAAC,WAAW,CAA8B;IAC1C,UAAU,EAAE,OAAO,CAAS;IAE7B,MAAM,CAAC,gBAAgB,CAAC,EAAE,CAAC,QAAQ,EAAE,GAAG,KAAK,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC;IA8BjE,KAAK,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,YAAY,CAAC;IAkC1C,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,YAAY,CAAC;CAG5D;AAED,eAAO,MAAM,gBAAgB,kBAAyB,CAAC"}