@vivantel/rag-core 1.1.2 → 2.0.0
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.
- package/dist/bin/rag-update.d.ts +3 -0
- package/dist/bin/rag-update.d.ts.map +1 -0
- package/dist/bin/rag-update.js +116 -0
- package/dist/bin/rag-update.js.map +1 -0
- package/dist/cli/init.d.ts +2 -0
- package/dist/cli/init.d.ts.map +1 -0
- package/dist/cli/init.js +262 -0
- package/dist/cli/init.js.map +1 -0
- package/dist/cli/validate.d.ts +2 -0
- package/dist/cli/validate.d.ts.map +1 -0
- package/dist/cli/validate.js +54 -0
- package/dist/cli/validate.js.map +1 -0
- package/dist/config-loader.d.ts.map +1 -1
- package/dist/config-loader.js +73 -7
- package/dist/config-loader.js.map +1 -1
- package/dist/core/chunk-processor.d.ts +1 -1
- package/dist/core/chunk-processor.d.ts.map +1 -1
- package/dist/core/chunk-processor.js +50 -21
- package/dist/core/chunk-processor.js.map +1 -1
- package/dist/core/embedder.d.ts +5 -1
- package/dist/core/embedder.d.ts.map +1 -1
- package/dist/core/embedder.js +40 -29
- package/dist/core/embedder.js.map +1 -1
- package/dist/core/errors.d.ts +16 -0
- package/dist/core/errors.d.ts.map +1 -0
- package/dist/core/errors.js +17 -0
- package/dist/core/errors.js.map +1 -0
- package/dist/core/git-tracker.d.ts.map +1 -1
- package/dist/core/git-tracker.js +9 -59
- package/dist/core/git-tracker.js.map +1 -1
- package/dist/core/orchestrator.d.ts +8 -0
- package/dist/core/orchestrator.d.ts.map +1 -1
- package/dist/core/orchestrator.js +153 -37
- package/dist/core/orchestrator.js.map +1 -1
- package/dist/core/plugin-discovery.d.ts +19 -0
- package/dist/core/plugin-discovery.d.ts.map +1 -0
- package/dist/core/plugin-discovery.js +47 -0
- package/dist/core/plugin-discovery.js.map +1 -0
- package/dist/core/telemetry.d.ts +61 -0
- package/dist/core/telemetry.d.ts.map +1 -0
- package/dist/core/telemetry.js +50 -0
- package/dist/core/telemetry.js.map +1 -0
- package/dist/core/uploader.d.ts +5 -1
- package/dist/core/uploader.d.ts.map +1 -1
- package/dist/core/uploader.js +23 -7
- package/dist/core/uploader.js.map +1 -1
- package/dist/core/utils.d.ts +7 -0
- package/dist/core/utils.d.ts.map +1 -1
- package/dist/core/utils.js +35 -0
- package/dist/core/utils.js.map +1 -1
- package/dist/index.d.ts +14 -1
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +14 -2
- package/dist/index.js.map +1 -1
- package/dist/interfaces/embedder.d.ts +2 -0
- package/dist/interfaces/embedder.d.ts.map +1 -1
- package/dist/interfaces/vector-store.d.ts +2 -0
- package/dist/interfaces/vector-store.d.ts.map +1 -1
- package/dist/strategies/chunk/token.js +1 -1
- package/dist/strategies/chunk/token.js.map +1 -1
- package/package.json +12 -2
- package/.github/config/release-please.json +0 -38
- package/.github/dependabot.yaml +0 -28
- package/.github/workflows/ci.yaml +0 -119
- package/.github/workflows/publish.yaml +0 -155
- package/.github/workflows/release.yaml +0 -54
- package/.release-please-manifest.json +0 -3
- package/.versionrc.json +0 -19
- package/CHANGELOG.md +0 -51
- package/bin/rag-update.ts +0 -49
- package/eslint.config.js +0 -25
- package/src/config-loader.ts +0 -21
- package/src/core/chunk-processor.test.ts +0 -36
- package/src/core/chunk-processor.ts +0 -92
- package/src/core/embedder.ts +0 -189
- package/src/core/git-tracker.test.ts +0 -64
- package/src/core/git-tracker.ts +0 -202
- package/src/core/orchestrator.test.ts +0 -53
- package/src/core/orchestrator.ts +0 -97
- package/src/core/uploader.ts +0 -123
- package/src/core/utils.ts +0 -27
- package/src/helpers/create-chunker.test.ts +0 -31
- package/src/helpers/create-chunker.ts +0 -40
- package/src/index.test.ts +0 -33
- package/src/index.ts +0 -30
- package/src/interfaces/chunker.ts +0 -59
- package/src/interfaces/embedder.ts +0 -36
- package/src/interfaces/index.test.ts +0 -9
- package/src/interfaces/index.ts +0 -3
- package/src/interfaces/vector-store.ts +0 -71
- package/src/strategies/chunk/index.ts +0 -4
- package/src/strategies/chunk/markdown-headers.test.ts +0 -37
- package/src/strategies/chunk/markdown-headers.ts +0 -106
- package/src/strategies/chunk/semantic.test.ts +0 -21
- package/src/strategies/chunk/semantic.ts +0 -80
- package/src/strategies/chunk/token.test.ts +0 -41
- package/src/strategies/chunk/token.ts +0 -72
- package/src/strategies/chunk/whole-file.test.ts +0 -24
- package/src/strategies/chunk/whole-file.ts +0 -35
- package/tsconfig.json +0 -21
- package/typedoc.json +0 -11
- package/vitest.config.ts +0 -19
|
@@ -2,6 +2,48 @@ import { GitTracker } from "./git-tracker.js";
|
|
|
2
2
|
import { ChunkProcessor } from "./chunk-processor.js";
|
|
3
3
|
import { EmbedderProcessor } from "./embedder.js";
|
|
4
4
|
import { Uploader } from "./uploader.js";
|
|
5
|
+
import { TelemetryCollector } from "./telemetry.js";
|
|
6
|
+
import { readFile } from "fs/promises";
|
|
7
|
+
async function loadPreviousState(chunksFile) {
|
|
8
|
+
try {
|
|
9
|
+
const content = await readFile(chunksFile, "utf-8");
|
|
10
|
+
const parsed = JSON.parse(content);
|
|
11
|
+
if (!Array.isArray(parsed))
|
|
12
|
+
return new Map();
|
|
13
|
+
const state = new Map();
|
|
14
|
+
for (const chunk of parsed) {
|
|
15
|
+
if (chunk.sourceFile && chunk.commitHash) {
|
|
16
|
+
state.set(chunk.sourceFile, chunk.commitHash);
|
|
17
|
+
}
|
|
18
|
+
}
|
|
19
|
+
return state;
|
|
20
|
+
}
|
|
21
|
+
catch {
|
|
22
|
+
return new Map();
|
|
23
|
+
}
|
|
24
|
+
}
|
|
25
|
+
async function loadExistingChunks(chunksFile) {
|
|
26
|
+
try {
|
|
27
|
+
const content = await readFile(chunksFile, "utf-8");
|
|
28
|
+
const parsed = JSON.parse(content);
|
|
29
|
+
return Array.isArray(parsed) ? parsed : [];
|
|
30
|
+
}
|
|
31
|
+
catch {
|
|
32
|
+
return [];
|
|
33
|
+
}
|
|
34
|
+
}
|
|
35
|
+
async function notifyWebhook(url, status, payload) {
|
|
36
|
+
try {
|
|
37
|
+
await fetch(url, {
|
|
38
|
+
method: "POST",
|
|
39
|
+
headers: { "Content-Type": "application/json" },
|
|
40
|
+
body: JSON.stringify({ status, ...payload }),
|
|
41
|
+
});
|
|
42
|
+
}
|
|
43
|
+
catch {
|
|
44
|
+
// Notification failures are non-fatal
|
|
45
|
+
}
|
|
46
|
+
}
|
|
5
47
|
export class Orchestrator {
|
|
6
48
|
config;
|
|
7
49
|
chunksFile;
|
|
@@ -13,45 +55,119 @@ export class Orchestrator {
|
|
|
13
55
|
config.options?.embeddingsFile || "./docs/rag/embeddings.json";
|
|
14
56
|
}
|
|
15
57
|
async run() {
|
|
16
|
-
|
|
17
|
-
|
|
18
|
-
|
|
19
|
-
|
|
20
|
-
|
|
21
|
-
|
|
22
|
-
|
|
23
|
-
|
|
24
|
-
|
|
25
|
-
|
|
26
|
-
|
|
58
|
+
const opts = this.config.options ?? {};
|
|
59
|
+
const telemetry = opts.telemetry ? new TelemetryCollector() : null;
|
|
60
|
+
telemetry?.start();
|
|
61
|
+
let uploadStats = { uploaded: 0, deleted: 0 };
|
|
62
|
+
try {
|
|
63
|
+
console.log("🚀 Starting RAG pipeline...\n");
|
|
64
|
+
// Step 1: Scan for changes
|
|
65
|
+
console.log("📂 Step 1: Scanning for changes...");
|
|
66
|
+
const t1 = Date.now();
|
|
67
|
+
const gitTracker = new GitTracker(this.config.chunkers);
|
|
68
|
+
const currentState = await gitTracker.getCurrentState();
|
|
69
|
+
const previousState = opts.force
|
|
70
|
+
? new Map()
|
|
71
|
+
: await loadPreviousState(this.chunksFile);
|
|
72
|
+
const { toProcess, toDelete } = await gitTracker.getChangedFiles(previousState);
|
|
73
|
+
telemetry?.recordGitTracking({
|
|
74
|
+
durationMs: Date.now() - t1,
|
|
75
|
+
filesScanned: currentState.size,
|
|
76
|
+
toProcess: toProcess.length,
|
|
77
|
+
toDelete: toDelete.length,
|
|
78
|
+
});
|
|
79
|
+
if (toProcess.length === 0 && toDelete.length === 0 && !opts.force) {
|
|
80
|
+
console.log("\n✨ No changes detected.");
|
|
81
|
+
return;
|
|
82
|
+
}
|
|
83
|
+
console.log(`\n📊 Changes: ${toProcess.length} to process, ${toDelete.length} to delete\n`);
|
|
84
|
+
// Step 2: Generate chunks (with resume)
|
|
85
|
+
console.log("🔪 Step 2: Generating chunks...");
|
|
86
|
+
const t2 = Date.now();
|
|
87
|
+
const chunkProcessor = new ChunkProcessor(this.config.chunkers);
|
|
88
|
+
const fileState = new Map();
|
|
89
|
+
for (const file of toProcess) {
|
|
90
|
+
const info = currentState.get(file);
|
|
91
|
+
if (info)
|
|
92
|
+
fileState.set(file, info);
|
|
93
|
+
}
|
|
94
|
+
const existingChunks = opts.force
|
|
95
|
+
? []
|
|
96
|
+
: await loadExistingChunks(this.chunksFile);
|
|
97
|
+
const chunks = await chunkProcessor.processFiles(toProcess, fileState, existingChunks);
|
|
98
|
+
await chunkProcessor.saveChunksLocal(chunks, this.chunksFile);
|
|
99
|
+
telemetry?.recordChunking({
|
|
100
|
+
durationMs: Date.now() - t2,
|
|
101
|
+
filesProcessed: toProcess.length,
|
|
102
|
+
chunksGenerated: chunks.length,
|
|
103
|
+
errors: 0,
|
|
104
|
+
});
|
|
105
|
+
if (chunks.length === 0) {
|
|
106
|
+
console.log("\n⚠️ No chunks generated. Exiting.");
|
|
107
|
+
return;
|
|
108
|
+
}
|
|
109
|
+
// Step 3: Generate embeddings
|
|
110
|
+
console.log("\n🔢 Step 3: Generating embeddings...");
|
|
111
|
+
const t3 = Date.now();
|
|
112
|
+
const embedder = new EmbedderProcessor(this.config.embedder, {
|
|
113
|
+
rateLimitMs: opts.rateLimitMs,
|
|
114
|
+
batchSize: opts.batchSize,
|
|
115
|
+
retry: opts.retry,
|
|
116
|
+
concurrency: opts.concurrency,
|
|
117
|
+
});
|
|
118
|
+
const newEmbeddings = await embedder.run(this.chunksFile, opts.force || false);
|
|
119
|
+
telemetry?.recordEmbedding({
|
|
120
|
+
durationMs: Date.now() - t3,
|
|
121
|
+
chunksEmbedded: newEmbeddings.length,
|
|
122
|
+
chunksSkipped: chunks.length - newEmbeddings.length,
|
|
123
|
+
});
|
|
124
|
+
// Step 4: Upload
|
|
125
|
+
if (opts.dryRun) {
|
|
126
|
+
const uploader = new Uploader(this.config.vectorStore);
|
|
127
|
+
const { toUpload: dryUpload, toDelete: dryDelete } = await uploader.getItemsToUpload(this.embeddingsFile, opts.force || false);
|
|
128
|
+
console.log("\n📤 Step 4: Upload (dry-run — no changes written)");
|
|
129
|
+
console.log(` Would upload: ${dryUpload.length} document(s)`);
|
|
130
|
+
console.log(` Would delete: ${dryDelete.length} source file(s)`);
|
|
131
|
+
}
|
|
132
|
+
else if (!opts.skipUpload) {
|
|
133
|
+
console.log("\n📤 Step 4: Uploading to vector store...");
|
|
134
|
+
const t4 = Date.now();
|
|
135
|
+
const uploader = new Uploader(this.config.vectorStore, {
|
|
136
|
+
retry: opts.retry,
|
|
137
|
+
});
|
|
138
|
+
uploadStats = await uploader.sync(this.embeddingsFile, opts.force || false);
|
|
139
|
+
telemetry?.recordUpload({
|
|
140
|
+
durationMs: Date.now() - t4,
|
|
141
|
+
uploaded: uploadStats.uploaded,
|
|
142
|
+
deleted: uploadStats.deleted,
|
|
143
|
+
});
|
|
144
|
+
}
|
|
145
|
+
console.log("\n✨ RAG pipeline complete!");
|
|
146
|
+
if (telemetry) {
|
|
147
|
+
telemetry.finish();
|
|
148
|
+
telemetry.printSummary();
|
|
149
|
+
const telemetryFile = this.chunksFile.replace("chunks.json", "telemetry.json");
|
|
150
|
+
await telemetry.save(telemetryFile);
|
|
151
|
+
}
|
|
152
|
+
if (opts.notifications?.webhookUrl) {
|
|
153
|
+
await notifyWebhook(opts.notifications.webhookUrl, "success", {
|
|
154
|
+
durationMs: telemetry?.getData().durationMs,
|
|
155
|
+
stages: telemetry?.getData().stages,
|
|
156
|
+
uploaded: uploadStats.uploaded,
|
|
157
|
+
deleted: uploadStats.deleted,
|
|
158
|
+
});
|
|
159
|
+
}
|
|
27
160
|
}
|
|
28
|
-
|
|
29
|
-
|
|
30
|
-
|
|
31
|
-
|
|
32
|
-
|
|
33
|
-
|
|
34
|
-
|
|
35
|
-
|
|
36
|
-
|
|
37
|
-
const chunks = await chunkProcessor.processFiles(toProcess, fileState);
|
|
38
|
-
await chunkProcessor.saveChunksLocal(chunks, this.chunksFile);
|
|
39
|
-
if (chunks.length === 0) {
|
|
40
|
-
console.log("\n⚠️ No chunks generated. Exiting.");
|
|
41
|
-
return;
|
|
42
|
-
}
|
|
43
|
-
console.log("\n🔢 Step 3: Generating embeddings...");
|
|
44
|
-
const embedder = new EmbedderProcessor(this.config.embedder, {
|
|
45
|
-
rateLimitMs: this.config.options?.rateLimitMs,
|
|
46
|
-
batchSize: this.config.options?.batchSize,
|
|
47
|
-
});
|
|
48
|
-
await embedder.run(this.chunksFile, this.config.options?.force || false);
|
|
49
|
-
if (!this.config.options?.skipUpload) {
|
|
50
|
-
console.log("\n📤 Step 4: Uploading to vector store...");
|
|
51
|
-
const uploader = new Uploader(this.config.vectorStore);
|
|
52
|
-
await uploader.sync(this.embeddingsFile, this.config.options?.force || false);
|
|
161
|
+
catch (err) {
|
|
162
|
+
telemetry?.finish();
|
|
163
|
+
if (opts.notifications?.webhookUrl) {
|
|
164
|
+
await notifyWebhook(opts.notifications.webhookUrl, "error", {
|
|
165
|
+
error: err instanceof Error ? err.message : String(err),
|
|
166
|
+
durationMs: telemetry?.getData().durationMs,
|
|
167
|
+
});
|
|
168
|
+
}
|
|
169
|
+
throw err;
|
|
53
170
|
}
|
|
54
|
-
console.log("\n✨ RAG pipeline complete!");
|
|
55
171
|
}
|
|
56
172
|
}
|
|
57
173
|
//# sourceMappingURL=orchestrator.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"orchestrator.js","sourceRoot":"","sources":["../../src/core/orchestrator.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC9C,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,OAAO,EAAE,iBAAiB,EAAE,MAAM,eAAe,CAAC;AAClD,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAC;
|
|
1
|
+
{"version":3,"file":"orchestrator.js","sourceRoot":"","sources":["../../src/core/orchestrator.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC9C,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,OAAO,EAAE,iBAAiB,EAAE,MAAM,eAAe,CAAC;AAClD,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAC;AACzC,OAAO,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAC;AAOpD,OAAO,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAsBvC,KAAK,UAAU,iBAAiB,CAC9B,UAAkB;IAElB,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,UAAU,EAAE,OAAO,CAAC,CAAC;QACpD,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;QACnC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC;YAAE,OAAO,IAAI,GAAG,EAAE,CAAC;QAC7C,MAAM,KAAK,GAAG,IAAI,GAAG,EAAkB,CAAC;QACxC,KAAK,MAAM,KAAK,IAAI,MAAiB,EAAE,CAAC;YACtC,IAAI,KAAK,CAAC,UAAU,IAAI,KAAK,CAAC,UAAU,EAAE,CAAC;gBACzC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,UAAU,EAAE,KAAK,CAAC,UAAU,CAAC,CAAC;YAChD,CAAC;QACH,CAAC;QACD,OAAO,KAAK,CAAC;IACf,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,GAAG,EAAE,CAAC;IACnB,CAAC;AACH,CAAC;AAED,KAAK,UAAU,kBAAkB,CAAC,UAAkB;IAClD,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,UAAU,EAAE,OAAO,CAAC,CAAC;QACpD,MAAM,MAAM,GAAY,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;QAC5C,OAAO,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAE,MAAkB,CAAC,CAAC,CAAC,EAAE,CAAC;IAC1D,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED,KAAK,UAAU,aAAa,CAC1B,GAAW,EACX,MAA2B,EAC3B,OAAgC;IAEhC,IAAI,CAAC;QACH,MAAM,KAAK,CAAC,GAAG,EAAE;YACf,MAAM,EAAE,MAAM;YACd,OAAO,EAAE,EAAE,cAAc,EAAE,kBAAkB,EAAE;YAC/C,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,EAAE,MAAM,EAAE,GAAG,OAAO,EAAE,CAAC;SAC7C,CAAC,CAAC;IACL,CAAC;IAAC,MAAM,CAAC;QACP,sCAAsC;IACxC,CAAC;AACH,CAAC;AAED,MAAM,OAAO,YAAY;IACf,MAAM,CAAoB;IAC1B,UAAU,CAAS;IACnB,cAAc,CAAS;IAE/B,YAAY,MAAyB;QACnC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;QACrB,IAAI,CAAC,UAAU,GAAG,MAAM,CAAC,OAAO,EAAE,UAAU,IAAI,wBAAwB,CAAC;QACzE,IAAI,CAAC,cAAc;YACjB,MAAM,CAAC,OAAO,EAAE,cAAc,IAAI,4BAA4B,CAAC;IACnE,CAAC;IAED,KAAK,CAAC,GAAG;QACP,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,IAAI,EAAE,CAAC;QACvC,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,kBAAkB,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;QACnE,SAAS,EAAE,KAAK,EAAE,CAAC;QAEnB,IAAI,WAAW,GAAG,EAAE,QAAQ,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC;QAE9C,IAAI,CAAC;YACH,OAAO,CAAC,GAAG,CAAC,+BAA+B,CAAC,CAAC;YAE7C,2BAA2B;YAC3B,OAAO,CAAC,GAAG,CAAC,oCAAoC,CAAC,CAAC;YAClD,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;YACtB,MAAM,UAAU,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;YACxD,MAAM,YAAY,GAAG,MAAM,UAAU,CAAC,eAAe,EAAE,CAAC;YAExD,MAAM,aAAa,GAAG,IAAI,CAAC,KAAK;gBAC9B,CAAC,CAAC,IAAI,GAAG,EAAkB;gBAC3B,CAAC,CAAC,MAAM,iBAAiB,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;YAE7C,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,GAC3B,MAAM,UAAU,CAAC,eAAe,CAAC,aAAa,CAAC,CAAC;YAElD,SAAS,EAAE,iBAAiB,CAAC;gBAC3B,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE;gBAC3B,YAAY,EAAE,YAAY,CAAC,IAAI;gBAC/B,SAAS,EAAE,SAAS,CAAC,MAAM;gBAC3B,QAAQ,EAAE,QAAQ,CAAC,MAAM;aAC1B,CAAC,CAAC;YAEH,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC;gBACnE,OAAO,CAAC,GAAG,CAAC,0BAA0B,CAAC,CAAC;gBACxC,OAAO;YACT,CAAC;YAED,OAAO,CAAC,GAAG,CACT,iBAAiB,SAAS,CAAC,MAAM,gBAAgB,QAAQ,CAAC,MAAM,cAAc,CAC/E,CAAC;YAEF,wCAAwC;YACxC,OAAO,CAAC,GAAG,CAAC,iCAAiC,CAAC,CAAC;YAC/C,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;YACtB,MAAM,cAAc,GAAG,IAAI,cAAc,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;YAEhE,MAAM,SAAS,GAAG,IAAI,GAAG,EAGtB,CAAC;YACJ,KAAK,MAAM,IAAI,IAAI,SAAS,EAAE,CAAC;gBAC7B,MAAM,IAAI,GAAG,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;gBACpC,IAAI,IAAI;oBAAE,SAAS,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;YACtC,CAAC;YAED,MAAM,cAAc,GAAG,IAAI,CAAC,KAAK;gBAC/B,CAAC,CAAC,EAAE;gBACJ,CAAC,CAAC,MAAM,kBAAkB,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;YAE9C,MAAM,MAAM,GAAG,MAAM,cAAc,CAAC,YAAY,CAC9C,SAAS,EACT,SAAS,EACT,cAAc,CACf,CAAC;YACF,MAAM,cAAc,CAAC,eAAe,CAAC,MAAM,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC;YAE9D,SAAS,EAAE,cAAc,CAAC;gBACxB,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE;gBAC3B,cAAc,EAAE,SAAS,CAAC,MAAM;gBAChC,eAAe,EAAE,MAAM,CAAC,MAAM;gBAC9B,MAAM,EAAE,CAAC;aACV,CAAC,CAAC;YAEH,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACxB,OAAO,CAAC,GAAG,CAAC,oCAAoC,CAAC,CAAC;gBAClD,OAAO;YACT,CAAC;YAED,8BAA8B;YAC9B,OAAO,CAAC,GAAG,CAAC,uCAAuC,CAAC,CAAC;YACrD,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;YACtB,MAAM,QAAQ,GAAG,IAAI,iBAAiB,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE;gBAC3D,WAAW,EAAE,IAAI,CAAC,WAAW;gBAC7B,SAAS,EAAE,IAAI,CAAC,SAAS;gBACzB,KAAK,EAAE,IAAI,CAAC,KAAK;gBACjB,WAAW,EAAE,IAAI,CAAC,WAAW;aAC9B,CAAC,CAAC;YAEH,MAAM,aAAa,GAAG,MAAM,QAAQ,CAAC,GAAG,CACtC,IAAI,CAAC,UAAU,EACf,IAAI,CAAC,KAAK,IAAI,KAAK,CACpB,CAAC;YAEF,SAAS,EAAE,eAAe,CAAC;gBACzB,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE;gBAC3B,cAAc,EAAE,aAAa,CAAC,MAAM;gBACpC,aAAa,EAAE,MAAM,CAAC,MAAM,GAAG,aAAa,CAAC,MAAM;aACpD,CAAC,CAAC;YAEH,iBAAiB;YACjB,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;gBAChB,MAAM,QAAQ,GAAG,IAAI,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC;gBACvD,MAAM,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE,GAChD,MAAM,QAAQ,CAAC,gBAAgB,CAC7B,IAAI,CAAC,cAAc,EACnB,IAAI,CAAC,KAAK,IAAI,KAAK,CACpB,CAAC;gBACJ,OAAO,CAAC,GAAG,CAAC,oDAAoD,CAAC,CAAC;gBAClE,OAAO,CAAC,GAAG,CAAC,oBAAoB,SAAS,CAAC,MAAM,cAAc,CAAC,CAAC;gBAChE,OAAO,CAAC,GAAG,CAAC,oBAAoB,SAAS,CAAC,MAAM,iBAAiB,CAAC,CAAC;YACrE,CAAC;iBAAM,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,CAAC;gBAC5B,OAAO,CAAC,GAAG,CAAC,2CAA2C,CAAC,CAAC;gBACzD,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;gBACtB,MAAM,QAAQ,GAAG,IAAI,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE;oBACrD,KAAK,EAAE,IAAI,CAAC,KAAK;iBAClB,CAAC,CAAC;gBACH,WAAW,GAAG,MAAM,QAAQ,CAAC,IAAI,CAC/B,IAAI,CAAC,cAAc,EACnB,IAAI,CAAC,KAAK,IAAI,KAAK,CACpB,CAAC;gBAEF,SAAS,EAAE,YAAY,CAAC;oBACtB,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE;oBAC3B,QAAQ,EAAE,WAAW,CAAC,QAAQ;oBAC9B,OAAO,EAAE,WAAW,CAAC,OAAO;iBAC7B,CAAC,CAAC;YACL,CAAC;YAED,OAAO,CAAC,GAAG,CAAC,4BAA4B,CAAC,CAAC;YAE1C,IAAI,SAAS,EAAE,CAAC;gBACd,SAAS,CAAC,MAAM,EAAE,CAAC;gBACnB,SAAS,CAAC,YAAY,EAAE,CAAC;gBACzB,MAAM,aAAa,GAAG,IAAI,CAAC,UAAU,CAAC,OAAO,CAC3C,aAAa,EACb,gBAAgB,CACjB,CAAC;gBACF,MAAM,SAAS,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;YACtC,CAAC;YAED,IAAI,IAAI,CAAC,aAAa,EAAE,UAAU,EAAE,CAAC;gBACnC,MAAM,aAAa,CAAC,IAAI,CAAC,aAAa,CAAC,UAAU,EAAE,SAAS,EAAE;oBAC5D,UAAU,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC,UAAU;oBAC3C,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC,MAAM;oBACnC,QAAQ,EAAE,WAAW,CAAC,QAAQ;oBAC9B,OAAO,EAAE,WAAW,CAAC,OAAO;iBAC7B,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,SAAS,EAAE,MAAM,EAAE,CAAC;YAEpB,IAAI,IAAI,CAAC,aAAa,EAAE,UAAU,EAAE,CAAC;gBACnC,MAAM,aAAa,CAAC,IAAI,CAAC,aAAa,CAAC,UAAU,EAAE,OAAO,EAAE;oBAC1D,KAAK,EAAE,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC;oBACvD,UAAU,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC,UAAU;iBAC5C,CAAC,CAAC;YACL,CAAC;YAED,MAAM,GAAG,CAAC;QACZ,CAAC;IACH,CAAC;CACF"}
|
|
@@ -0,0 +1,19 @@
|
|
|
1
|
+
export interface RagPlugin {
|
|
2
|
+
name: string;
|
|
3
|
+
type: "embedder" | "store" | "chunker";
|
|
4
|
+
factory: () => unknown;
|
|
5
|
+
}
|
|
6
|
+
/**
|
|
7
|
+
* Discover RAG plugins from installed npm packages.
|
|
8
|
+
*
|
|
9
|
+
* A plugin package must export either:
|
|
10
|
+
* `ragPlugin: RagPlugin` — single plugin
|
|
11
|
+
* `ragPlugins: RagPlugin[]` — multiple plugins
|
|
12
|
+
*
|
|
13
|
+
* Naming convention (not enforced):
|
|
14
|
+
* rag-embedder-* EmbeddingProvider implementations
|
|
15
|
+
* rag-store-* VectorStore implementations
|
|
16
|
+
* rag-chunker-* FileChunker / ChunkStrategy implementations
|
|
17
|
+
*/
|
|
18
|
+
export declare function discoverPlugins(packageNames: string[]): Promise<RagPlugin[]>;
|
|
19
|
+
//# sourceMappingURL=plugin-discovery.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"plugin-discovery.d.ts","sourceRoot":"","sources":["../../src/core/plugin-discovery.ts"],"names":[],"mappings":"AAAA,MAAM,WAAW,SAAS;IACxB,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,UAAU,GAAG,OAAO,GAAG,SAAS,CAAC;IACvC,OAAO,EAAE,MAAM,OAAO,CAAC;CACxB;AAED;;;;;;;;;;;GAWG;AACH,wBAAsB,eAAe,CACnC,YAAY,EAAE,MAAM,EAAE,GACrB,OAAO,CAAC,SAAS,EAAE,CAAC,CA0BtB"}
|
|
@@ -0,0 +1,47 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Discover RAG plugins from installed npm packages.
|
|
3
|
+
*
|
|
4
|
+
* A plugin package must export either:
|
|
5
|
+
* `ragPlugin: RagPlugin` — single plugin
|
|
6
|
+
* `ragPlugins: RagPlugin[]` — multiple plugins
|
|
7
|
+
*
|
|
8
|
+
* Naming convention (not enforced):
|
|
9
|
+
* rag-embedder-* EmbeddingProvider implementations
|
|
10
|
+
* rag-store-* VectorStore implementations
|
|
11
|
+
* rag-chunker-* FileChunker / ChunkStrategy implementations
|
|
12
|
+
*/
|
|
13
|
+
export async function discoverPlugins(packageNames) {
|
|
14
|
+
const discovered = [];
|
|
15
|
+
for (const pkg of packageNames) {
|
|
16
|
+
let mod;
|
|
17
|
+
try {
|
|
18
|
+
mod = (await import(pkg));
|
|
19
|
+
}
|
|
20
|
+
catch {
|
|
21
|
+
console.warn(` ⚠️ Could not load plugin package: ${pkg}`);
|
|
22
|
+
continue;
|
|
23
|
+
}
|
|
24
|
+
if (isRagPlugin(mod.ragPlugin)) {
|
|
25
|
+
discovered.push(mod.ragPlugin);
|
|
26
|
+
}
|
|
27
|
+
else if (Array.isArray(mod.ragPlugins)) {
|
|
28
|
+
for (const p of mod.ragPlugins) {
|
|
29
|
+
if (isRagPlugin(p))
|
|
30
|
+
discovered.push(p);
|
|
31
|
+
}
|
|
32
|
+
}
|
|
33
|
+
else {
|
|
34
|
+
console.warn(` ⚠️ Package "${pkg}" has no ragPlugin or ragPlugins export — skipping`);
|
|
35
|
+
}
|
|
36
|
+
}
|
|
37
|
+
return discovered;
|
|
38
|
+
}
|
|
39
|
+
function isRagPlugin(value) {
|
|
40
|
+
if (!value || typeof value !== "object")
|
|
41
|
+
return false;
|
|
42
|
+
const p = value;
|
|
43
|
+
return (typeof p.name === "string" &&
|
|
44
|
+
(p.type === "embedder" || p.type === "store" || p.type === "chunker") &&
|
|
45
|
+
typeof p.factory === "function");
|
|
46
|
+
}
|
|
47
|
+
//# sourceMappingURL=plugin-discovery.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"plugin-discovery.js","sourceRoot":"","sources":["../../src/core/plugin-discovery.ts"],"names":[],"mappings":"AAMA;;;;;;;;;;;GAWG;AACH,MAAM,CAAC,KAAK,UAAU,eAAe,CACnC,YAAsB;IAEtB,MAAM,UAAU,GAAgB,EAAE,CAAC;IAEnC,KAAK,MAAM,GAAG,IAAI,YAAY,EAAE,CAAC;QAC/B,IAAI,GAA4B,CAAC;QACjC,IAAI,CAAC;YACH,GAAG,GAAG,CAAC,MAAM,MAAM,CAAC,GAAG,CAAC,CAA4B,CAAC;QACvD,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,CAAC,IAAI,CAAC,uCAAuC,GAAG,EAAE,CAAC,CAAC;YAC3D,SAAS;QACX,CAAC;QAED,IAAI,WAAW,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,CAAC;YAC/B,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QACjC,CAAC;aAAM,IAAI,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,UAAU,CAAC,EAAE,CAAC;YACzC,KAAK,MAAM,CAAC,IAAI,GAAG,CAAC,UAAU,EAAE,CAAC;gBAC/B,IAAI,WAAW,CAAC,CAAC,CAAC;oBAAE,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACzC,CAAC;QACH,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,IAAI,CACV,iBAAiB,GAAG,oDAAoD,CACzE,CAAC;QACJ,CAAC;IACH,CAAC;IAED,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,SAAS,WAAW,CAAC,KAAc;IACjC,IAAI,CAAC,KAAK,IAAI,OAAO,KAAK,KAAK,QAAQ;QAAE,OAAO,KAAK,CAAC;IACtD,MAAM,CAAC,GAAG,KAAgC,CAAC;IAC3C,OAAO,CACL,OAAO,CAAC,CAAC,IAAI,KAAK,QAAQ;QAC1B,CAAC,CAAC,CAAC,IAAI,KAAK,UAAU,IAAI,CAAC,CAAC,IAAI,KAAK,OAAO,IAAI,CAAC,CAAC,IAAI,KAAK,SAAS,CAAC;QACrE,OAAO,CAAC,CAAC,OAAO,KAAK,UAAU,CAChC,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,61 @@
|
|
|
1
|
+
interface StageStats {
|
|
2
|
+
durationMs: number;
|
|
3
|
+
[key: string]: unknown;
|
|
4
|
+
}
|
|
5
|
+
interface TelemetryData {
|
|
6
|
+
runAt: string;
|
|
7
|
+
durationMs: number;
|
|
8
|
+
stages: {
|
|
9
|
+
gitTracking?: StageStats & {
|
|
10
|
+
filesScanned: number;
|
|
11
|
+
toProcess: number;
|
|
12
|
+
toDelete: number;
|
|
13
|
+
};
|
|
14
|
+
chunking?: StageStats & {
|
|
15
|
+
filesProcessed: number;
|
|
16
|
+
chunksGenerated: number;
|
|
17
|
+
errors: number;
|
|
18
|
+
};
|
|
19
|
+
embedding?: StageStats & {
|
|
20
|
+
chunksEmbedded: number;
|
|
21
|
+
chunksSkipped: number;
|
|
22
|
+
};
|
|
23
|
+
upload?: StageStats & {
|
|
24
|
+
uploaded: number;
|
|
25
|
+
deleted: number;
|
|
26
|
+
};
|
|
27
|
+
};
|
|
28
|
+
}
|
|
29
|
+
export declare class TelemetryCollector {
|
|
30
|
+
private startedAt;
|
|
31
|
+
private data;
|
|
32
|
+
start(): void;
|
|
33
|
+
recordGitTracking(stats: {
|
|
34
|
+
filesScanned: number;
|
|
35
|
+
toProcess: number;
|
|
36
|
+
toDelete: number;
|
|
37
|
+
durationMs: number;
|
|
38
|
+
}): void;
|
|
39
|
+
recordChunking(stats: {
|
|
40
|
+
filesProcessed: number;
|
|
41
|
+
chunksGenerated: number;
|
|
42
|
+
errors: number;
|
|
43
|
+
durationMs: number;
|
|
44
|
+
}): void;
|
|
45
|
+
recordEmbedding(stats: {
|
|
46
|
+
chunksEmbedded: number;
|
|
47
|
+
chunksSkipped: number;
|
|
48
|
+
durationMs: number;
|
|
49
|
+
}): void;
|
|
50
|
+
recordUpload(stats: {
|
|
51
|
+
uploaded: number;
|
|
52
|
+
deleted: number;
|
|
53
|
+
durationMs: number;
|
|
54
|
+
}): void;
|
|
55
|
+
finish(): void;
|
|
56
|
+
getData(): TelemetryData;
|
|
57
|
+
printSummary(): void;
|
|
58
|
+
save(outputPath: string): Promise<void>;
|
|
59
|
+
}
|
|
60
|
+
export {};
|
|
61
|
+
//# sourceMappingURL=telemetry.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"telemetry.d.ts","sourceRoot":"","sources":["../../src/core/telemetry.ts"],"names":[],"mappings":"AAGA,UAAU,UAAU;IAClB,UAAU,EAAE,MAAM,CAAC;IACnB,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC;CACxB;AAED,UAAU,aAAa;IACrB,KAAK,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,MAAM,CAAC;IACnB,MAAM,EAAE;QACN,WAAW,CAAC,EAAE,UAAU,GAAG;YACzB,YAAY,EAAE,MAAM,CAAC;YACrB,SAAS,EAAE,MAAM,CAAC;YAClB,QAAQ,EAAE,MAAM,CAAC;SAClB,CAAC;QACF,QAAQ,CAAC,EAAE,UAAU,GAAG;YACtB,cAAc,EAAE,MAAM,CAAC;YACvB,eAAe,EAAE,MAAM,CAAC;YACxB,MAAM,EAAE,MAAM,CAAC;SAChB,CAAC;QACF,SAAS,CAAC,EAAE,UAAU,GAAG;YACvB,cAAc,EAAE,MAAM,CAAC;YACvB,aAAa,EAAE,MAAM,CAAC;SACvB,CAAC;QACF,MAAM,CAAC,EAAE,UAAU,GAAG;YACpB,QAAQ,EAAE,MAAM,CAAC;YACjB,OAAO,EAAE,MAAM,CAAC;SACjB,CAAC;KACH,CAAC;CACH;AAED,qBAAa,kBAAkB;IAC7B,OAAO,CAAC,SAAS,CAAK;IACtB,OAAO,CAAC,IAAI,CAIV;IAEF,KAAK,IAAI,IAAI;IAKb,iBAAiB,CAAC,KAAK,EAAE;QACvB,YAAY,EAAE,MAAM,CAAC;QACrB,SAAS,EAAE,MAAM,CAAC;QAClB,QAAQ,EAAE,MAAM,CAAC;QACjB,UAAU,EAAE,MAAM,CAAC;KACpB,GAAG,IAAI;IAIR,cAAc,CAAC,KAAK,EAAE;QACpB,cAAc,EAAE,MAAM,CAAC;QACvB,eAAe,EAAE,MAAM,CAAC;QACxB,MAAM,EAAE,MAAM,CAAC;QACf,UAAU,EAAE,MAAM,CAAC;KACpB,GAAG,IAAI;IAIR,eAAe,CAAC,KAAK,EAAE;QACrB,cAAc,EAAE,MAAM,CAAC;QACvB,aAAa,EAAE,MAAM,CAAC;QACtB,UAAU,EAAE,MAAM,CAAC;KACpB,GAAG,IAAI;IAIR,YAAY,CAAC,KAAK,EAAE;QAClB,QAAQ,EAAE,MAAM,CAAC;QACjB,OAAO,EAAE,MAAM,CAAC;QAChB,UAAU,EAAE,MAAM,CAAC;KACpB,GAAG,IAAI;IAIR,MAAM,IAAI,IAAI;IAId,OAAO,IAAI,aAAa;IAIxB,YAAY,IAAI,IAAI;IAYd,IAAI,CAAC,UAAU,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;CAK9C"}
|
|
@@ -0,0 +1,50 @@
|
|
|
1
|
+
import { writeFile, mkdir } from "fs/promises";
|
|
2
|
+
import { dirname } from "path";
|
|
3
|
+
export class TelemetryCollector {
|
|
4
|
+
startedAt = 0;
|
|
5
|
+
data = {
|
|
6
|
+
runAt: new Date().toISOString(),
|
|
7
|
+
durationMs: 0,
|
|
8
|
+
stages: {},
|
|
9
|
+
};
|
|
10
|
+
start() {
|
|
11
|
+
this.startedAt = Date.now();
|
|
12
|
+
this.data.runAt = new Date().toISOString();
|
|
13
|
+
}
|
|
14
|
+
recordGitTracking(stats) {
|
|
15
|
+
this.data.stages.gitTracking = stats;
|
|
16
|
+
}
|
|
17
|
+
recordChunking(stats) {
|
|
18
|
+
this.data.stages.chunking = stats;
|
|
19
|
+
}
|
|
20
|
+
recordEmbedding(stats) {
|
|
21
|
+
this.data.stages.embedding = stats;
|
|
22
|
+
}
|
|
23
|
+
recordUpload(stats) {
|
|
24
|
+
this.data.stages.upload = stats;
|
|
25
|
+
}
|
|
26
|
+
finish() {
|
|
27
|
+
this.data.durationMs = Date.now() - this.startedAt;
|
|
28
|
+
}
|
|
29
|
+
getData() {
|
|
30
|
+
return this.data;
|
|
31
|
+
}
|
|
32
|
+
printSummary() {
|
|
33
|
+
console.log(`\n📊 Telemetry summary (${this.data.durationMs}ms total):`);
|
|
34
|
+
for (const [stage, stats] of Object.entries(this.data.stages)) {
|
|
35
|
+
if (!stats)
|
|
36
|
+
continue;
|
|
37
|
+
const { durationMs, ...rest } = stats;
|
|
38
|
+
const details = Object.entries(rest)
|
|
39
|
+
.map(([k, v]) => `${k}=${v}`)
|
|
40
|
+
.join(", ");
|
|
41
|
+
console.log(` ${stage}: ${details} (${durationMs}ms)`);
|
|
42
|
+
}
|
|
43
|
+
}
|
|
44
|
+
async save(outputPath) {
|
|
45
|
+
await mkdir(dirname(outputPath), { recursive: true });
|
|
46
|
+
await writeFile(outputPath, JSON.stringify(this.data, null, 2));
|
|
47
|
+
console.log(`\n📈 Telemetry saved to ${outputPath}`);
|
|
48
|
+
}
|
|
49
|
+
}
|
|
50
|
+
//# sourceMappingURL=telemetry.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"telemetry.js","sourceRoot":"","sources":["../../src/core/telemetry.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAC/C,OAAO,EAAE,OAAO,EAAE,MAAM,MAAM,CAAC;AAgC/B,MAAM,OAAO,kBAAkB;IACrB,SAAS,GAAG,CAAC,CAAC;IACd,IAAI,GAAkB;QAC5B,KAAK,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QAC/B,UAAU,EAAE,CAAC;QACb,MAAM,EAAE,EAAE;KACX,CAAC;IAEF,KAAK;QACH,IAAI,CAAC,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAC5B,IAAI,CAAC,IAAI,CAAC,KAAK,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IAC7C,CAAC;IAED,iBAAiB,CAAC,KAKjB;QACC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,WAAW,GAAG,KAAK,CAAC;IACvC,CAAC;IAED,cAAc,CAAC,KAKd;QACC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,GAAG,KAAK,CAAC;IACpC,CAAC;IAED,eAAe,CAAC,KAIf;QACC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,GAAG,KAAK,CAAC;IACrC,CAAC;IAED,YAAY,CAAC,KAIZ;QACC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,GAAG,KAAK,CAAC;IAClC,CAAC;IAED,MAAM;QACJ,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,GAAG,EAAE,GAAG,IAAI,CAAC,SAAS,CAAC;IACrD,CAAC;IAED,OAAO;QACL,OAAO,IAAI,CAAC,IAAI,CAAC;IACnB,CAAC;IAED,YAAY;QACV,OAAO,CAAC,GAAG,CAAC,2BAA2B,IAAI,CAAC,IAAI,CAAC,UAAU,YAAY,CAAC,CAAC;QACzE,KAAK,MAAM,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;YAC9D,IAAI,CAAC,KAAK;gBAAE,SAAS;YACrB,MAAM,EAAE,UAAU,EAAE,GAAG,IAAI,EAAE,GAAG,KAAK,CAAC;YACtC,MAAM,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC;iBACjC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;iBAC5B,IAAI,CAAC,IAAI,CAAC,CAAC;YACd,OAAO,CAAC,GAAG,CAAC,MAAM,KAAK,KAAK,OAAO,KAAK,UAAU,KAAK,CAAC,CAAC;QAC3D,CAAC;IACH,CAAC;IAED,KAAK,CAAC,IAAI,CAAC,UAAkB;QAC3B,MAAM,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;QACtD,MAAM,SAAS,CAAC,UAAU,EAAE,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;QAChE,OAAO,CAAC,GAAG,CAAC,2BAA2B,UAAU,EAAE,CAAC,CAAC;IACvD,CAAC;CACF"}
|
package/dist/core/uploader.d.ts
CHANGED
|
@@ -1,7 +1,11 @@
|
|
|
1
1
|
import { VectorStore, EmbeddedChunk } from "../interfaces/index.js";
|
|
2
|
+
import { RetryOptions } from "./utils.js";
|
|
2
3
|
export declare class Uploader {
|
|
3
4
|
private vectorStore;
|
|
4
|
-
|
|
5
|
+
private retryOptions;
|
|
6
|
+
constructor(vectorStore: VectorStore, options?: {
|
|
7
|
+
retry?: RetryOptions;
|
|
8
|
+
});
|
|
5
9
|
private chunkToDocument;
|
|
6
10
|
getItemsToUpload(embeddingsFile: string, force?: boolean): Promise<{
|
|
7
11
|
toUpload: EmbeddedChunk[];
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"uploader.d.ts","sourceRoot":"","sources":["../../src/core/uploader.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,WAAW,EAEX,aAAa,EACd,MAAM,wBAAwB,CAAC;
|
|
1
|
+
{"version":3,"file":"uploader.d.ts","sourceRoot":"","sources":["../../src/core/uploader.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,WAAW,EAEX,aAAa,EACd,MAAM,wBAAwB,CAAC;AAIhC,OAAO,EAAa,YAAY,EAAE,MAAM,YAAY,CAAC;AASrD,qBAAa,QAAQ;IACnB,OAAO,CAAC,WAAW,CAAc;IACjC,OAAO,CAAC,YAAY,CAAe;gBAGjC,WAAW,EAAE,WAAW,EACxB,OAAO,GAAE;QAAE,KAAK,CAAC,EAAE,YAAY,CAAA;KAAO;IAMxC,OAAO,CAAC,eAAe;IAejB,gBAAgB,CACpB,cAAc,EAAE,MAAM,EACtB,KAAK,GAAE,OAAe,GACrB,OAAO,CAAC;QACT,QAAQ,EAAE,aAAa,EAAE,CAAC;QAC1B,QAAQ,EAAE,MAAM,EAAE,CAAC;KACpB,CAAC;IAkDI,IAAI,CACR,cAAc,EAAE,MAAM,EACtB,KAAK,GAAE,OAAe,GACrB,OAAO,CAAC;QACT,QAAQ,EAAE,MAAM,CAAC;QACjB,OAAO,EAAE,MAAM,CAAC;KACjB,CAAC;CAgDH"}
|
package/dist/core/uploader.js
CHANGED
|
@@ -1,8 +1,17 @@
|
|
|
1
1
|
import { readFile } from "fs/promises";
|
|
2
|
+
import { createHash } from "crypto";
|
|
3
|
+
import { UploadError } from "./errors.js";
|
|
4
|
+
import { withRetry } from "./utils.js";
|
|
5
|
+
function contentHash(chunk) {
|
|
6
|
+
return (chunk.contentHash ??
|
|
7
|
+
createHash("sha256").update(chunk.content).digest("hex").slice(0, 16));
|
|
8
|
+
}
|
|
2
9
|
export class Uploader {
|
|
3
10
|
vectorStore;
|
|
4
|
-
|
|
11
|
+
retryOptions;
|
|
12
|
+
constructor(vectorStore, options = {}) {
|
|
5
13
|
this.vectorStore = vectorStore;
|
|
14
|
+
this.retryOptions = options.retry ?? {};
|
|
6
15
|
}
|
|
7
16
|
chunkToDocument(chunk, collection) {
|
|
8
17
|
return {
|
|
@@ -11,7 +20,7 @@ export class Uploader {
|
|
|
11
20
|
embedding: chunk.embedding,
|
|
12
21
|
sourceFile: chunk.sourceFile,
|
|
13
22
|
commitHash: chunk.commitHash,
|
|
14
|
-
contentHash: chunk
|
|
23
|
+
contentHash: contentHash(chunk),
|
|
15
24
|
collection,
|
|
16
25
|
};
|
|
17
26
|
}
|
|
@@ -19,10 +28,17 @@ export class Uploader {
|
|
|
19
28
|
let embeddings;
|
|
20
29
|
try {
|
|
21
30
|
const content = await readFile(embeddingsFile, "utf-8");
|
|
22
|
-
|
|
31
|
+
const parsed = JSON.parse(content);
|
|
32
|
+
if (!Array.isArray(parsed)) {
|
|
33
|
+
throw new Error("embeddings file does not contain a JSON array");
|
|
34
|
+
}
|
|
35
|
+
embeddings = parsed;
|
|
23
36
|
}
|
|
24
|
-
catch {
|
|
25
|
-
throw new
|
|
37
|
+
catch (err) {
|
|
38
|
+
throw new UploadError(`Failed to load embeddings from ${embeddingsFile}: ${err instanceof Error ? err.message : String(err)}`, {
|
|
39
|
+
suggestion: "Run the pipeline without --skip-upload to regenerate embeddings first.",
|
|
40
|
+
cause: err,
|
|
41
|
+
});
|
|
26
42
|
}
|
|
27
43
|
console.log(`📖 Loaded ${embeddings.length} embeddings from ${embeddingsFile}`);
|
|
28
44
|
if (force) {
|
|
@@ -58,7 +74,7 @@ export class Uploader {
|
|
|
58
74
|
return { uploaded: 0, deleted: 0 };
|
|
59
75
|
}
|
|
60
76
|
if (toDelete.length > 0) {
|
|
61
|
-
await this.vectorStore.deleteBySourceFile(toDelete);
|
|
77
|
+
await withRetry(() => this.vectorStore.deleteBySourceFile(toDelete), this.retryOptions);
|
|
62
78
|
console.log(` 🗑️ Deleted ${toDelete.length} obsolete documents`);
|
|
63
79
|
}
|
|
64
80
|
if (toUpload.length > 0) {
|
|
@@ -66,7 +82,7 @@ export class Uploader {
|
|
|
66
82
|
const batchSize = 50;
|
|
67
83
|
for (let i = 0; i < documents.length; i += batchSize) {
|
|
68
84
|
const batch = documents.slice(i, i + batchSize);
|
|
69
|
-
await this.vectorStore.upsert(batch);
|
|
85
|
+
await withRetry(() => this.vectorStore.upsert(batch), this.retryOptions);
|
|
70
86
|
console.log(` ✅ Uploaded batch ${Math.floor(i / batchSize) + 1}/${Math.ceil(documents.length / batchSize)}`);
|
|
71
87
|
}
|
|
72
88
|
}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"uploader.js","sourceRoot":"","sources":["../../src/core/uploader.ts"],"names":[],"mappings":"AAKA,OAAO,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;
|
|
1
|
+
{"version":3,"file":"uploader.js","sourceRoot":"","sources":["../../src/core/uploader.ts"],"names":[],"mappings":"AAKA,OAAO,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AACvC,OAAO,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAC;AACpC,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AAC1C,OAAO,EAAE,SAAS,EAAgB,MAAM,YAAY,CAAC;AAErD,SAAS,WAAW,CAAC,KAAoB;IACvC,OAAO,CACL,KAAK,CAAC,WAAW;QACjB,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CACtE,CAAC;AACJ,CAAC;AAED,MAAM,OAAO,QAAQ;IACX,WAAW,CAAc;IACzB,YAAY,CAAe;IAEnC,YACE,WAAwB,EACxB,UAAoC,EAAE;QAEtC,IAAI,CAAC,WAAW,GAAG,WAAW,CAAC;QAC/B,IAAI,CAAC,YAAY,GAAG,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC;IAC1C,CAAC;IAEO,eAAe,CACrB,KAAoB,EACpB,UAAmB;QAEnB,OAAO;YACL,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,QAAQ,EAAE,KAAK,CAAC,QAAQ;YACxB,SAAS,EAAE,KAAK,CAAC,SAAS;YAC1B,UAAU,EAAE,KAAK,CAAC,UAAU;YAC5B,UAAU,EAAE,KAAK,CAAC,UAAU;YAC5B,WAAW,EAAE,WAAW,CAAC,KAAK,CAAC;YAC/B,UAAU;SACX,CAAC;IACJ,CAAC;IAED,KAAK,CAAC,gBAAgB,CACpB,cAAsB,EACtB,QAAiB,KAAK;QAKtB,IAAI,UAA2B,CAAC;QAChC,IAAI,CAAC;YACH,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,cAAc,EAAE,OAAO,CAAC,CAAC;YACxD,MAAM,MAAM,GAAY,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;YAC5C,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC3B,MAAM,IAAI,KAAK,CAAC,+CAA+C,CAAC,CAAC;YACnE,CAAC;YACD,UAAU,GAAG,MAAyB,CAAC;QACzC,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,IAAI,WAAW,CACnB,kCAAkC,cAAc,KAAK,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,EACvG;gBACE,UAAU,EACR,wEAAwE;gBAC1E,KAAK,EAAE,GAAG;aACX,CACF,CAAC;QACJ,CAAC;QAED,OAAO,CAAC,GAAG,CACT,aAAa,UAAU,CAAC,MAAM,oBAAoB,cAAc,EAAE,CACnE,CAAC;QAEF,IAAI,KAAK,EAAE,CAAC;YACV,MAAM,cAAc,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;YACzE,OAAO,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ,EAAE,cAAc,EAAE,CAAC;QAC5D,CAAC;QAED,MAAM,aAAa,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,eAAe,EAAE,CAAC;QAC/D,MAAM,YAAY,GAAoB,EAAE,CAAC;QACzC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;QAEtC,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;YAC7B,MAAM,YAAY,GAAG,aAAa,CAAC,GAAG,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;YAEvD,IAAI,CAAC,YAAY,EAAE,CAAC;gBAClB,YAAY,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACzB,CAAC;iBAAM,IAAI,YAAY,KAAK,GAAG,CAAC,UAAU,EAAE,CAAC;gBAC3C,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;gBAChC,YAAY,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACzB,CAAC;QACH,CAAC;QAED,OAAO;YACL,QAAQ,EAAE,YAAY;YACtB,QAAQ,EAAE,CAAC,GAAG,WAAW,CAAC;SAC3B,CAAC;IACJ,CAAC;IAED,KAAK,CAAC,IAAI,CACR,cAAsB,EACtB,QAAiB,KAAK;QAKtB,OAAO,CAAC,GAAG,CAAC,mCAAmC,CAAC,CAAC;QAEjD,MAAM,IAAI,CAAC,WAAW,CAAC,UAAU,EAAE,CAAC;QAEpC,MAAM,EAAE,QAAQ,EAAE,QAAQ,EAAE,GAAG,MAAM,IAAI,CAAC,gBAAgB,CACxD,cAAc,EACd,KAAK,CACN,CAAC;QAEF,OAAO,CAAC,GAAG,CAAC,wBAAwB,QAAQ,CAAC,MAAM,YAAY,CAAC,CAAC;QACjE,OAAO,CAAC,GAAG,CAAC,sBAAsB,QAAQ,CAAC,MAAM,QAAQ,CAAC,CAAC;QAE3D,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACnD,OAAO,CAAC,GAAG,CAAC,0BAA0B,CAAC,CAAC;YACxC,OAAO,EAAE,QAAQ,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC;QACrC,CAAC;QAED,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACxB,MAAM,SAAS,CACb,GAAG,EAAE,CAAC,IAAI,CAAC,WAAW,CAAC,kBAAkB,CAAC,QAAQ,CAAC,EACnD,IAAI,CAAC,YAAY,CAClB,CAAC;YACF,OAAO,CAAC,GAAG,CAAC,iBAAiB,QAAQ,CAAC,MAAM,qBAAqB,CAAC,CAAC;QACrE,CAAC;QAED,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACxB,MAAM,SAAS,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC,CAAC;YAE/D,MAAM,SAAS,GAAG,EAAE,CAAC;YACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;gBACrD,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,CAAC;gBAChD,MAAM,SAAS,CACb,GAAG,EAAE,CAAC,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,KAAK,CAAC,EACpC,IAAI,CAAC,YAAY,CAClB,CAAC;gBACF,OAAO,CAAC,GAAG,CACT,sBAAsB,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,SAAS,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,GAAG,SAAS,CAAC,EAAE,CACjG,CAAC;YACJ,CAAC;QACH,CAAC;QAED,OAAO,CAAC,GAAG,CAAC,sBAAsB,CAAC,CAAC;QACpC,OAAO,CAAC,GAAG,CAAC,gBAAgB,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC;QAC/C,OAAO,CAAC,GAAG,CAAC,eAAe,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC;QAE9C,OAAO,EAAE,QAAQ,EAAE,QAAQ,CAAC,MAAM,EAAE,OAAO,EAAE,QAAQ,CAAC,MAAM,EAAE,CAAC;IACjE,CAAC;CACF"}
|
package/dist/core/utils.d.ts
CHANGED
|
@@ -1,5 +1,12 @@
|
|
|
1
1
|
export declare function computeContentHash(content: string): string;
|
|
2
2
|
export declare function sleep(ms: number): Promise<void>;
|
|
3
|
+
export interface RetryOptions {
|
|
4
|
+
maxRetries?: number;
|
|
5
|
+
retryDelayMs?: number;
|
|
6
|
+
retryBackoffFactor?: number;
|
|
7
|
+
}
|
|
8
|
+
export declare function withRetry<T>(fn: () => Promise<T>, opts?: RetryOptions): Promise<T>;
|
|
9
|
+
export declare function withConcurrency<T>(tasks: Array<() => Promise<T>>, limit: number): Promise<T[]>;
|
|
3
10
|
export declare function batchArray<T>(array: T[], batchSize: number): T[][];
|
|
4
11
|
export declare function extractFileName(filePath: string): string;
|
|
5
12
|
export declare function extractDirectory(filePath: string): string;
|
package/dist/core/utils.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"utils.d.ts","sourceRoot":"","sources":["../../src/core/utils.ts"],"names":[],"mappings":"AAEA,wBAAgB,kBAAkB,CAAC,OAAO,EAAE,MAAM,GAAG,MAAM,CAE1D;AAED,wBAAgB,KAAK,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAE/C;AAED,wBAAgB,UAAU,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,EAAE,SAAS,EAAE,MAAM,GAAG,CAAC,EAAE,EAAE,CAMlE;AAED,wBAAgB,eAAe,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,CAExD;AAED,wBAAgB,gBAAgB,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,CAIzD"}
|
|
1
|
+
{"version":3,"file":"utils.d.ts","sourceRoot":"","sources":["../../src/core/utils.ts"],"names":[],"mappings":"AAEA,wBAAgB,kBAAkB,CAAC,OAAO,EAAE,MAAM,GAAG,MAAM,CAE1D;AAED,wBAAgB,KAAK,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAE/C;AAED,MAAM,WAAW,YAAY;IAC3B,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,kBAAkB,CAAC,EAAE,MAAM,CAAC;CAC7B;AAED,wBAAsB,SAAS,CAAC,CAAC,EAC/B,EAAE,EAAE,MAAM,OAAO,CAAC,CAAC,CAAC,EACpB,IAAI,GAAE,YAAiB,GACtB,OAAO,CAAC,CAAC,CAAC,CAqBZ;AAED,wBAAsB,eAAe,CAAC,CAAC,EACrC,KAAK,EAAE,KAAK,CAAC,MAAM,OAAO,CAAC,CAAC,CAAC,CAAC,EAC9B,KAAK,EAAE,MAAM,GACZ,OAAO,CAAC,CAAC,EAAE,CAAC,CAed;AAED,wBAAgB,UAAU,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,EAAE,SAAS,EAAE,MAAM,GAAG,CAAC,EAAE,EAAE,CAMlE;AAED,wBAAgB,eAAe,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,CAExD;AAED,wBAAgB,gBAAgB,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,CAIzD"}
|
package/dist/core/utils.js
CHANGED
|
@@ -5,6 +5,41 @@ export function computeContentHash(content) {
|
|
|
5
5
|
export function sleep(ms) {
|
|
6
6
|
return new Promise((resolve) => setTimeout(resolve, ms));
|
|
7
7
|
}
|
|
8
|
+
export async function withRetry(fn, opts = {}) {
|
|
9
|
+
const maxRetries = opts.maxRetries ?? 3;
|
|
10
|
+
const delayMs = opts.retryDelayMs ?? 1000;
|
|
11
|
+
const factor = opts.retryBackoffFactor ?? 2;
|
|
12
|
+
let lastErr;
|
|
13
|
+
for (let attempt = 0; attempt <= maxRetries; attempt++) {
|
|
14
|
+
try {
|
|
15
|
+
return await fn();
|
|
16
|
+
}
|
|
17
|
+
catch (err) {
|
|
18
|
+
lastErr = err;
|
|
19
|
+
if (attempt < maxRetries) {
|
|
20
|
+
const wait = Math.round(delayMs * Math.pow(factor, attempt));
|
|
21
|
+
console.warn(` ⚠️ Attempt ${attempt + 1}/${maxRetries + 1} failed, retrying in ${wait}ms...`);
|
|
22
|
+
await sleep(wait);
|
|
23
|
+
}
|
|
24
|
+
}
|
|
25
|
+
}
|
|
26
|
+
throw lastErr;
|
|
27
|
+
}
|
|
28
|
+
export async function withConcurrency(tasks, limit) {
|
|
29
|
+
if (tasks.length === 0)
|
|
30
|
+
return [];
|
|
31
|
+
const results = new Array(tasks.length);
|
|
32
|
+
let index = 0;
|
|
33
|
+
async function worker() {
|
|
34
|
+
while (index < tasks.length) {
|
|
35
|
+
const i = index++;
|
|
36
|
+
results[i] = await tasks[i]();
|
|
37
|
+
}
|
|
38
|
+
}
|
|
39
|
+
const workerCount = Math.min(limit, tasks.length);
|
|
40
|
+
await Promise.all(Array.from({ length: workerCount }, worker));
|
|
41
|
+
return results;
|
|
42
|
+
}
|
|
8
43
|
export function batchArray(array, batchSize) {
|
|
9
44
|
const batches = [];
|
|
10
45
|
for (let i = 0; i < array.length; i += batchSize) {
|
package/dist/core/utils.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../src/core/utils.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAC;AAEpC,MAAM,UAAU,kBAAkB,CAAC,OAAe;IAChD,OAAO,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;AACzE,CAAC;AAED,MAAM,UAAU,KAAK,CAAC,EAAU;IAC9B,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC;AAC3D,CAAC;AAED,MAAM,UAAU,UAAU,CAAI,KAAU,EAAE,SAAiB;IACzD,MAAM,OAAO,GAAU,EAAE,CAAC;IAC1B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;QACjD,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,CAAC,CAAC;IAC9C,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,MAAM,UAAU,eAAe,CAAC,QAAgB;IAC9C,OAAO,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,QAAQ,CAAC;AAC/C,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,QAAgB;IAC/C,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAClC,KAAK,CAAC,GAAG,EAAE,CAAC;IACZ,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AACzB,CAAC"}
|
|
1
|
+
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../src/core/utils.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAC;AAEpC,MAAM,UAAU,kBAAkB,CAAC,OAAe;IAChD,OAAO,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;AACzE,CAAC;AAED,MAAM,UAAU,KAAK,CAAC,EAAU;IAC9B,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC;AAC3D,CAAC;AAQD,MAAM,CAAC,KAAK,UAAU,SAAS,CAC7B,EAAoB,EACpB,OAAqB,EAAE;IAEvB,MAAM,UAAU,GAAG,IAAI,CAAC,UAAU,IAAI,CAAC,CAAC;IACxC,MAAM,OAAO,GAAG,IAAI,CAAC,YAAY,IAAI,IAAI,CAAC;IAC1C,MAAM,MAAM,GAAG,IAAI,CAAC,kBAAkB,IAAI,CAAC,CAAC;IAE5C,IAAI,OAAgB,CAAC;IACrB,KAAK,IAAI,OAAO,GAAG,CAAC,EAAE,OAAO,IAAI,UAAU,EAAE,OAAO,EAAE,EAAE,CAAC;QACvD,IAAI,CAAC;YACH,OAAO,MAAM,EAAE,EAAE,CAAC;QACpB,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,OAAO,GAAG,GAAG,CAAC;YACd,IAAI,OAAO,GAAG,UAAU,EAAE,CAAC;gBACzB,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;gBAC7D,OAAO,CAAC,IAAI,CACV,gBAAgB,OAAO,GAAG,CAAC,IAAI,UAAU,GAAG,CAAC,wBAAwB,IAAI,OAAO,CACjF,CAAC;gBACF,MAAM,KAAK,CAAC,IAAI,CAAC,CAAC;YACpB,CAAC;QACH,CAAC;IACH,CAAC;IACD,MAAM,OAAO,CAAC;AAChB,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,eAAe,CACnC,KAA8B,EAC9B,KAAa;IAEb,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAClC,MAAM,OAAO,GAAG,IAAI,KAAK,CAAI,KAAK,CAAC,MAAM,CAAC,CAAC;IAC3C,IAAI,KAAK,GAAG,CAAC,CAAC;IAEd,KAAK,UAAU,MAAM;QACnB,OAAO,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;YAC5B,MAAM,CAAC,GAAG,KAAK,EAAE,CAAC;YAClB,OAAO,CAAC,CAAC,CAAC,GAAG,MAAM,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;QAChC,CAAC;IACH,CAAC;IAED,MAAM,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;IAClD,MAAM,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,WAAW,EAAE,EAAE,MAAM,CAAC,CAAC,CAAC;IAC/D,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,MAAM,UAAU,UAAU,CAAI,KAAU,EAAE,SAAiB;IACzD,MAAM,OAAO,GAAU,EAAE,CAAC;IAC1B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;QACjD,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,CAAC,CAAC;IAC9C,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,MAAM,UAAU,eAAe,CAAC,QAAgB;IAC9C,OAAO,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,QAAQ,CAAC;AAC/C,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,QAAgB;IAC/C,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAClC,KAAK,CAAC,GAAG,EAAE,CAAC;IACZ,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AACzB,CAAC"}
|
package/dist/index.d.ts
CHANGED
|
@@ -1,11 +1,24 @@
|
|
|
1
1
|
export * from "./interfaces/index.js";
|
|
2
|
+
export { RagError, ConfigError, ChunkError, EmbedError, UploadError, } from "./core/errors.js";
|
|
2
3
|
export { GitTracker } from "./core/git-tracker.js";
|
|
3
4
|
export { ChunkProcessor } from "./core/chunk-processor.js";
|
|
4
5
|
export { EmbedderProcessor } from "./core/embedder.js";
|
|
5
6
|
export { Uploader } from "./core/uploader.js";
|
|
6
7
|
export { Orchestrator, RAGPipelineConfig } from "./core/orchestrator.js";
|
|
7
|
-
export { computeContentHash, sleep, batchArray, extractFileName, extractDirectory, } from "./core/utils.js";
|
|
8
|
+
export { computeContentHash, sleep, batchArray, extractFileName, extractDirectory, withRetry, withConcurrency, } from "./core/utils.js";
|
|
9
|
+
export type { RetryOptions } from "./core/utils.js";
|
|
10
|
+
export { discoverPlugins, RagPlugin } from "./core/plugin-discovery.js";
|
|
11
|
+
/**
|
|
12
|
+
* Chunking strategies.
|
|
13
|
+
*
|
|
14
|
+
* @deprecated Import from `@vivantel/rag-strategies` instead:
|
|
15
|
+
* ```ts
|
|
16
|
+
* import { tokenStrategy } from '@vivantel/rag-strategies';
|
|
17
|
+
* ```
|
|
18
|
+
* Direct exports from `@vivantel/rag-core` will be removed in v3.0.
|
|
19
|
+
*/
|
|
8
20
|
export { tokenStrategy, markdownHeadersStrategy, semanticStrategy, wholeFileStrategy, } from "./strategies/chunk/index.js";
|
|
21
|
+
export type { TokenStrategyOptions } from "./strategies/chunk/token.js";
|
|
9
22
|
export { createChunker } from "./helpers/create-chunker.js";
|
|
10
23
|
export { loadConfig } from "./config-loader.js";
|
|
11
24
|
//# sourceMappingURL=index.d.ts.map
|