codevault 1.6.0 → 1.6.1
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/chunking/file-grouper.d.ts +39 -0
- package/dist/chunking/file-grouper.d.ts.map +1 -0
- package/dist/chunking/file-grouper.js +164 -0
- package/dist/chunking/file-grouper.js.map +1 -0
- package/dist/chunking/semantic-chunker.d.ts +37 -0
- package/dist/chunking/semantic-chunker.d.ts.map +1 -0
- package/dist/chunking/semantic-chunker.js +157 -0
- package/dist/chunking/semantic-chunker.js.map +1 -0
- package/dist/chunking/token-counter.d.ts +28 -0
- package/dist/chunking/token-counter.d.ts.map +1 -0
- package/dist/chunking/token-counter.js +207 -0
- package/dist/chunking/token-counter.js.map +1 -0
- package/dist/cli/commands/ask-cmd.d.ts +3 -0
- package/dist/cli/commands/ask-cmd.d.ts.map +1 -0
- package/dist/cli/commands/ask-cmd.js +130 -0
- package/dist/cli/commands/ask-cmd.js.map +1 -0
- package/dist/cli/commands/chat-cmd.d.ts +3 -0
- package/dist/cli/commands/chat-cmd.d.ts.map +1 -0
- package/dist/cli/commands/chat-cmd.js +194 -0
- package/dist/cli/commands/chat-cmd.js.map +1 -0
- package/dist/cli/commands/config-cmd.d.ts +3 -0
- package/dist/cli/commands/config-cmd.d.ts.map +1 -0
- package/dist/cli/commands/config-cmd.js +245 -0
- package/dist/cli/commands/config-cmd.js.map +1 -0
- package/dist/cli/commands/context.d.ts +3 -0
- package/dist/cli/commands/context.d.ts.map +1 -0
- package/dist/cli/commands/context.js +98 -0
- package/dist/cli/commands/context.js.map +1 -0
- package/dist/cli/commands/interactive-config.d.ts +2 -0
- package/dist/cli/commands/interactive-config.d.ts.map +1 -0
- package/dist/cli/commands/interactive-config.js +274 -0
- package/dist/cli/commands/interactive-config.js.map +1 -0
- package/dist/cli.d.ts +3 -0
- package/dist/cli.d.ts.map +1 -0
- package/dist/cli.js +403 -0
- package/dist/cli.js.map +1 -0
- package/dist/codemap/io.d.ts +5 -0
- package/dist/codemap/io.d.ts.map +1 -0
- package/dist/codemap/io.js +30 -0
- package/dist/codemap/io.js.map +1 -0
- package/dist/config/apply-env.d.ts +15 -0
- package/dist/config/apply-env.d.ts.map +1 -0
- package/dist/config/apply-env.js +91 -0
- package/dist/config/apply-env.js.map +1 -0
- package/dist/config/constants.d.ts +326 -0
- package/dist/config/constants.d.ts.map +1 -0
- package/dist/config/constants.js +214 -0
- package/dist/config/constants.js.map +1 -0
- package/dist/config/loader.d.ts +57 -0
- package/dist/config/loader.d.ts.map +1 -0
- package/dist/config/loader.js +287 -0
- package/dist/config/loader.js.map +1 -0
- package/dist/config/types.d.ts +46 -0
- package/dist/config/types.d.ts.map +1 -0
- package/dist/config/types.js +2 -0
- package/dist/config/types.js.map +1 -0
- package/dist/context/packs.d.ts +33 -0
- package/dist/context/packs.d.ts.map +1 -0
- package/dist/context/packs.js +180 -0
- package/dist/context/packs.js.map +1 -0
- package/dist/core/batch-indexer.d.ts +56 -0
- package/dist/core/batch-indexer.d.ts.map +1 -0
- package/dist/core/batch-indexer.js +189 -0
- package/dist/core/batch-indexer.js.map +1 -0
- package/dist/core/indexer.d.ts +3 -0
- package/dist/core/indexer.d.ts.map +1 -0
- package/dist/core/indexer.js +600 -0
- package/dist/core/indexer.js.map +1 -0
- package/dist/core/metadata.d.ts +19 -0
- package/dist/core/metadata.d.ts.map +1 -0
- package/dist/core/metadata.js +161 -0
- package/dist/core/metadata.js.map +1 -0
- package/dist/core/search.d.ts +7 -0
- package/dist/core/search.d.ts.map +1 -0
- package/dist/core/search.js +545 -0
- package/dist/core/search.js.map +1 -0
- package/dist/core/symbol-extractor.d.ts +3 -0
- package/dist/core/symbol-extractor.d.ts.map +1 -0
- package/dist/core/symbol-extractor.js +78 -0
- package/dist/core/symbol-extractor.js.map +1 -0
- package/dist/core/types.d.ts +104 -0
- package/dist/core/types.d.ts.map +1 -0
- package/dist/core/types.js +2 -0
- package/dist/core/types.js.map +1 -0
- package/dist/database/db.d.ts +102 -0
- package/dist/database/db.d.ts.map +1 -0
- package/dist/database/db.js +366 -0
- package/dist/database/db.js.map +1 -0
- package/dist/indexer/merkle.d.ts +13 -0
- package/dist/indexer/merkle.d.ts.map +1 -0
- package/dist/indexer/merkle.js +86 -0
- package/dist/indexer/merkle.js.map +1 -0
- package/dist/indexer/update.d.ts +19 -0
- package/dist/indexer/update.d.ts.map +1 -0
- package/dist/indexer/update.js +40 -0
- package/dist/indexer/update.js.map +1 -0
- package/dist/indexer/watch.d.ts +21 -0
- package/dist/indexer/watch.d.ts.map +1 -0
- package/dist/indexer/watch.js +222 -0
- package/dist/indexer/watch.js.map +1 -0
- package/dist/languages/rules.d.ts +11 -0
- package/dist/languages/rules.d.ts.map +1 -0
- package/dist/languages/rules.js +371 -0
- package/dist/languages/rules.js.map +1 -0
- package/dist/languages/tree-sitter-loader.d.ts +27 -0
- package/dist/languages/tree-sitter-loader.d.ts.map +1 -0
- package/dist/languages/tree-sitter-loader.js +76 -0
- package/dist/languages/tree-sitter-loader.js.map +1 -0
- package/dist/mcp/schemas.d.ts +166 -0
- package/dist/mcp/schemas.d.ts.map +1 -0
- package/dist/mcp/schemas.js +159 -0
- package/dist/mcp/schemas.js.map +1 -0
- package/dist/mcp/tools/ask-codebase.d.ts +85 -0
- package/dist/mcp/tools/ask-codebase.d.ts.map +1 -0
- package/dist/mcp/tools/ask-codebase.js +125 -0
- package/dist/mcp/tools/ask-codebase.js.map +1 -0
- package/dist/mcp/tools/use-context-pack.d.ts +57 -0
- package/dist/mcp/tools/use-context-pack.d.ts.map +1 -0
- package/dist/mcp/tools/use-context-pack.js +91 -0
- package/dist/mcp/tools/use-context-pack.js.map +1 -0
- package/dist/mcp-server.d.ts +3 -0
- package/dist/mcp-server.d.ts.map +1 -0
- package/dist/mcp-server.js +511 -0
- package/dist/mcp-server.js.map +1 -0
- package/dist/providers/base.d.ts +39 -0
- package/dist/providers/base.d.ts.map +1 -0
- package/dist/providers/base.js +198 -0
- package/dist/providers/base.js.map +1 -0
- package/dist/providers/chat-llm.d.ts +30 -0
- package/dist/providers/chat-llm.d.ts.map +1 -0
- package/dist/providers/chat-llm.js +85 -0
- package/dist/providers/chat-llm.js.map +1 -0
- package/dist/providers/index.d.ts +5 -0
- package/dist/providers/index.d.ts.map +1 -0
- package/dist/providers/index.js +12 -0
- package/dist/providers/index.js.map +1 -0
- package/dist/providers/openai.d.ts +14 -0
- package/dist/providers/openai.d.ts.map +1 -0
- package/dist/providers/openai.js +122 -0
- package/dist/providers/openai.js.map +1 -0
- package/dist/providers/token-counter.d.ts +2 -0
- package/dist/providers/token-counter.d.ts.map +1 -0
- package/dist/providers/token-counter.js +18 -0
- package/dist/providers/token-counter.js.map +1 -0
- package/dist/ranking/api-reranker.d.ts +18 -0
- package/dist/ranking/api-reranker.d.ts.map +1 -0
- package/dist/ranking/api-reranker.js +134 -0
- package/dist/ranking/api-reranker.js.map +1 -0
- package/dist/ranking/symbol-boost.d.ts +15 -0
- package/dist/ranking/symbol-boost.d.ts.map +1 -0
- package/dist/ranking/symbol-boost.js +175 -0
- package/dist/ranking/symbol-boost.js.map +1 -0
- package/dist/search/bm25.d.ts +17 -0
- package/dist/search/bm25.d.ts.map +1 -0
- package/dist/search/bm25.js +56 -0
- package/dist/search/bm25.js.map +1 -0
- package/dist/search/hybrid.d.ts +21 -0
- package/dist/search/hybrid.d.ts.map +1 -0
- package/dist/search/hybrid.js +50 -0
- package/dist/search/hybrid.js.map +1 -0
- package/dist/search/scope.d.ts +5 -0
- package/dist/search/scope.d.ts.map +1 -0
- package/dist/search/scope.js +107 -0
- package/dist/search/scope.js.map +1 -0
- package/dist/storage/encrypted-chunks.d.ts +40 -0
- package/dist/storage/encrypted-chunks.d.ts.map +1 -0
- package/dist/storage/encrypted-chunks.js +238 -0
- package/dist/storage/encrypted-chunks.js.map +1 -0
- package/dist/symbols/extract.d.ts +15 -0
- package/dist/symbols/extract.d.ts.map +1 -0
- package/dist/symbols/extract.js +208 -0
- package/dist/symbols/extract.js.map +1 -0
- package/dist/symbols/graph.d.ts +3 -0
- package/dist/symbols/graph.d.ts.map +1 -0
- package/dist/symbols/graph.js +89 -0
- package/dist/symbols/graph.js.map +1 -0
- package/dist/synthesis/conversational-synthesizer.d.ts +61 -0
- package/dist/synthesis/conversational-synthesizer.d.ts.map +1 -0
- package/dist/synthesis/conversational-synthesizer.js +286 -0
- package/dist/synthesis/conversational-synthesizer.js.map +1 -0
- package/dist/synthesis/markdown-formatter.d.ts +13 -0
- package/dist/synthesis/markdown-formatter.d.ts.map +1 -0
- package/dist/synthesis/markdown-formatter.js +104 -0
- package/dist/synthesis/markdown-formatter.js.map +1 -0
- package/dist/synthesis/prompt-builder.d.ts +21 -0
- package/dist/synthesis/prompt-builder.d.ts.map +1 -0
- package/dist/synthesis/prompt-builder.js +129 -0
- package/dist/synthesis/prompt-builder.js.map +1 -0
- package/dist/synthesis/synthesizer.d.ts +30 -0
- package/dist/synthesis/synthesizer.d.ts.map +1 -0
- package/dist/synthesis/synthesizer.js +210 -0
- package/dist/synthesis/synthesizer.js.map +1 -0
- package/dist/types/ast.d.ts +3 -0
- package/dist/types/ast.d.ts.map +1 -0
- package/dist/types/ast.js +2 -0
- package/dist/types/ast.js.map +1 -0
- package/dist/types/codemap.d.ts +58 -0
- package/dist/types/codemap.d.ts.map +1 -0
- package/dist/types/codemap.js +224 -0
- package/dist/types/codemap.js.map +1 -0
- package/dist/types/context-pack.d.ts +47 -0
- package/dist/types/context-pack.d.ts.map +1 -0
- package/dist/types/context-pack.js +44 -0
- package/dist/types/context-pack.js.map +1 -0
- package/dist/types/search.d.ts +15 -0
- package/dist/types/search.d.ts.map +1 -0
- package/dist/types/search.js +11 -0
- package/dist/types/search.js.map +1 -0
- package/dist/utils/cli-ui.d.ts +44 -0
- package/dist/utils/cli-ui.d.ts.map +1 -0
- package/dist/utils/cli-ui.js +139 -0
- package/dist/utils/cli-ui.js.map +1 -0
- package/dist/utils/indexer-with-progress.d.ts +10 -0
- package/dist/utils/indexer-with-progress.d.ts.map +1 -0
- package/dist/utils/indexer-with-progress.js +43 -0
- package/dist/utils/indexer-with-progress.js.map +1 -0
- package/dist/utils/logger.d.ts +55 -0
- package/dist/utils/logger.d.ts.map +1 -0
- package/dist/utils/logger.js +121 -0
- package/dist/utils/logger.js.map +1 -0
- package/dist/utils/mutex.d.ts +63 -0
- package/dist/utils/mutex.d.ts.map +1 -0
- package/dist/utils/mutex.js +123 -0
- package/dist/utils/mutex.js.map +1 -0
- package/dist/utils/path-helpers.d.ts +27 -0
- package/dist/utils/path-helpers.d.ts.map +1 -0
- package/dist/utils/path-helpers.js +55 -0
- package/dist/utils/path-helpers.js.map +1 -0
- package/dist/utils/rate-limiter.d.ts +34 -0
- package/dist/utils/rate-limiter.d.ts.map +1 -0
- package/dist/utils/rate-limiter.js +178 -0
- package/dist/utils/rate-limiter.js.map +1 -0
- package/dist/utils/scan-patterns.d.ts +5 -0
- package/dist/utils/scan-patterns.d.ts.map +1 -0
- package/dist/utils/scan-patterns.js +29 -0
- package/dist/utils/scan-patterns.js.map +1 -0
- package/package.json +1 -1
|
@@ -0,0 +1,39 @@
|
|
|
1
|
+
export interface ModelProfile {
|
|
2
|
+
maxTokens: number;
|
|
3
|
+
optimalTokens: number;
|
|
4
|
+
minChunkTokens: number;
|
|
5
|
+
maxChunkTokens: number;
|
|
6
|
+
overlapTokens: number;
|
|
7
|
+
optimalChars: number;
|
|
8
|
+
minChunkChars: number;
|
|
9
|
+
maxChunkChars: number;
|
|
10
|
+
overlapChars: number;
|
|
11
|
+
dimensions: number;
|
|
12
|
+
useTokens: boolean;
|
|
13
|
+
tokenizerType: string;
|
|
14
|
+
encoding?: string;
|
|
15
|
+
tokenCounter?: (text: string) => number | Promise<number>;
|
|
16
|
+
}
|
|
17
|
+
export declare abstract class EmbeddingProvider {
|
|
18
|
+
abstract generateEmbedding(text: string): Promise<number[]>;
|
|
19
|
+
abstract getDimensions(): number;
|
|
20
|
+
abstract getName(): string;
|
|
21
|
+
abstract getModelName?(): string;
|
|
22
|
+
abstract init?(): Promise<void>;
|
|
23
|
+
generateEmbeddings(texts: string[]): Promise<number[][]>;
|
|
24
|
+
rateLimiter?: any;
|
|
25
|
+
}
|
|
26
|
+
export declare const BATCH_SIZE = 50;
|
|
27
|
+
export declare const MAX_BATCH_TOKENS = 100000;
|
|
28
|
+
export declare const MAX_ITEM_TOKENS = 8191;
|
|
29
|
+
export declare function estimateTokens(text: string): number;
|
|
30
|
+
export declare const MODEL_PROFILES: Record<string, Omit<ModelProfile, 'tokenCounter'>>;
|
|
31
|
+
export declare function getModelProfile(providerName: string, modelName: string | null): Promise<ModelProfile>;
|
|
32
|
+
export declare function getSizeLimits(profile: ModelProfile): {
|
|
33
|
+
optimal: number;
|
|
34
|
+
min: number;
|
|
35
|
+
max: number;
|
|
36
|
+
overlap: number;
|
|
37
|
+
unit: string;
|
|
38
|
+
};
|
|
39
|
+
//# sourceMappingURL=base.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"base.d.ts","sourceRoot":"","sources":["../../src/providers/base.ts"],"names":[],"mappings":"AAAA,MAAM,WAAW,YAAY;IAC3B,SAAS,EAAE,MAAM,CAAC;IAClB,aAAa,EAAE,MAAM,CAAC;IACtB,cAAc,EAAE,MAAM,CAAC;IACvB,cAAc,EAAE,MAAM,CAAC;IACvB,aAAa,EAAE,MAAM,CAAC;IACtB,YAAY,EAAE,MAAM,CAAC;IACrB,aAAa,EAAE,MAAM,CAAC;IACtB,aAAa,EAAE,MAAM,CAAC;IACtB,YAAY,EAAE,MAAM,CAAC;IACrB,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,OAAO,CAAC;IACnB,aAAa,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,YAAY,CAAC,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;CAC3D;AAED,8BAAsB,iBAAiB;IACrC,QAAQ,CAAC,iBAAiB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC;IAC3D,QAAQ,CAAC,aAAa,IAAI,MAAM;IAChC,QAAQ,CAAC,OAAO,IAAI,MAAM;IAC1B,QAAQ,CAAC,YAAY,CAAC,IAAI,MAAM;IAChC,QAAQ,CAAC,IAAI,CAAC,IAAI,OAAO,CAAC,IAAI,CAAC;IAGzB,kBAAkB,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC;IAU9D,WAAW,CAAC,EAAE,GAAG,CAAC;CACnB;AAGD,eAAO,MAAM,UAAU,KAAK,CAAC;AAC7B,eAAO,MAAM,gBAAgB,SAAS,CAAC;AACvC,eAAO,MAAM,eAAe,OAAO,CAAC;AAGpC,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAEnD;AAED,eAAO,MAAM,cAAc,EAAE,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,YAAY,EAAE,cAAc,CAAC,CAyG7E,CAAC;AAEF,wBAAsB,eAAe,CAAC,YAAY,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,GAAG,IAAI,GAAG,OAAO,CAAC,YAAY,CAAC,CA2D3G;AAED,wBAAgB,aAAa,CAAC,OAAO,EAAE,YAAY,GAAG;IACpD,OAAO,EAAE,MAAM,CAAC;IAChB,GAAG,EAAE,MAAM,CAAC;IACZ,GAAG,EAAE,MAAM,CAAC;IACZ,OAAO,EAAE,MAAM,CAAC;IAChB,IAAI,EAAE,MAAM,CAAC;CACd,CAiBA"}
|
|
@@ -0,0 +1,198 @@
|
|
|
1
|
+
export class EmbeddingProvider {
|
|
2
|
+
// Batch processing support (optional - providers can override)
|
|
3
|
+
async generateEmbeddings(texts) {
|
|
4
|
+
// Default implementation: process one by one (backward compatible)
|
|
5
|
+
const embeddings = [];
|
|
6
|
+
for (const text of texts) {
|
|
7
|
+
const embedding = await this.generateEmbedding(text);
|
|
8
|
+
embeddings.push(embedding);
|
|
9
|
+
}
|
|
10
|
+
return embeddings;
|
|
11
|
+
}
|
|
12
|
+
rateLimiter;
|
|
13
|
+
}
|
|
14
|
+
// Batching constants
|
|
15
|
+
export const BATCH_SIZE = 50; // Number of chunks per batch
|
|
16
|
+
export const MAX_BATCH_TOKENS = 100000; // OpenAI limit
|
|
17
|
+
export const MAX_ITEM_TOKENS = 8191; // Per-item limit
|
|
18
|
+
// Estimate tokens from text (4 chars ≈ 1 token)
|
|
19
|
+
export function estimateTokens(text) {
|
|
20
|
+
return Math.ceil(text.length / 4);
|
|
21
|
+
}
|
|
22
|
+
export const MODEL_PROFILES = {
|
|
23
|
+
'text-embedding-3-large': {
|
|
24
|
+
maxTokens: 8191,
|
|
25
|
+
optimalTokens: 4000,
|
|
26
|
+
minChunkTokens: 400,
|
|
27
|
+
maxChunkTokens: 6000,
|
|
28
|
+
overlapTokens: 100,
|
|
29
|
+
optimalChars: 16000,
|
|
30
|
+
minChunkChars: 1600,
|
|
31
|
+
maxChunkChars: 24000,
|
|
32
|
+
overlapChars: 400,
|
|
33
|
+
dimensions: 3072,
|
|
34
|
+
useTokens: true,
|
|
35
|
+
tokenizerType: 'tiktoken',
|
|
36
|
+
encoding: 'cl100k_base'
|
|
37
|
+
},
|
|
38
|
+
'text-embedding-3-small': {
|
|
39
|
+
maxTokens: 8191,
|
|
40
|
+
optimalTokens: 4000,
|
|
41
|
+
minChunkTokens: 400,
|
|
42
|
+
maxChunkTokens: 6000,
|
|
43
|
+
overlapTokens: 100,
|
|
44
|
+
optimalChars: 16000,
|
|
45
|
+
minChunkChars: 1600,
|
|
46
|
+
maxChunkChars: 24000,
|
|
47
|
+
overlapChars: 400,
|
|
48
|
+
dimensions: 1536,
|
|
49
|
+
useTokens: true,
|
|
50
|
+
tokenizerType: 'tiktoken',
|
|
51
|
+
encoding: 'cl100k_base'
|
|
52
|
+
},
|
|
53
|
+
'text-embedding-ada-002': {
|
|
54
|
+
maxTokens: 8191,
|
|
55
|
+
optimalTokens: 4000,
|
|
56
|
+
minChunkTokens: 400,
|
|
57
|
+
maxChunkTokens: 6000,
|
|
58
|
+
overlapTokens: 100,
|
|
59
|
+
optimalChars: 16000,
|
|
60
|
+
minChunkChars: 1600,
|
|
61
|
+
maxChunkChars: 24000,
|
|
62
|
+
overlapChars: 400,
|
|
63
|
+
dimensions: 1536,
|
|
64
|
+
useTokens: true,
|
|
65
|
+
tokenizerType: 'tiktoken',
|
|
66
|
+
encoding: 'cl100k_base'
|
|
67
|
+
},
|
|
68
|
+
'nomic-embed-text': {
|
|
69
|
+
maxTokens: 8192,
|
|
70
|
+
optimalTokens: 4000,
|
|
71
|
+
minChunkTokens: 400,
|
|
72
|
+
maxChunkTokens: 6000,
|
|
73
|
+
overlapTokens: 100,
|
|
74
|
+
optimalChars: 16000,
|
|
75
|
+
minChunkChars: 1600,
|
|
76
|
+
maxChunkChars: 24000,
|
|
77
|
+
overlapChars: 400,
|
|
78
|
+
dimensions: 768,
|
|
79
|
+
useTokens: true,
|
|
80
|
+
tokenizerType: 'tiktoken',
|
|
81
|
+
encoding: 'cl100k_base'
|
|
82
|
+
},
|
|
83
|
+
'Qwen/Qwen3-Embedding-8B': {
|
|
84
|
+
maxTokens: 32000,
|
|
85
|
+
optimalTokens: 16000,
|
|
86
|
+
minChunkTokens: 1000,
|
|
87
|
+
maxChunkTokens: 28000,
|
|
88
|
+
overlapTokens: 500,
|
|
89
|
+
optimalChars: 64000,
|
|
90
|
+
minChunkChars: 4000,
|
|
91
|
+
maxChunkChars: 112000,
|
|
92
|
+
overlapChars: 2000,
|
|
93
|
+
dimensions: 4096,
|
|
94
|
+
useTokens: true,
|
|
95
|
+
tokenizerType: 'tiktoken',
|
|
96
|
+
encoding: 'cl100k_base'
|
|
97
|
+
},
|
|
98
|
+
'qwen3-embedding:0.6b': {
|
|
99
|
+
maxTokens: 32000,
|
|
100
|
+
optimalTokens: 26000,
|
|
101
|
+
minChunkTokens: 1000,
|
|
102
|
+
maxChunkTokens: 30000,
|
|
103
|
+
overlapTokens: 500,
|
|
104
|
+
optimalChars: 128000,
|
|
105
|
+
minChunkChars: 4000,
|
|
106
|
+
maxChunkChars: 120000,
|
|
107
|
+
overlapChars: 2000,
|
|
108
|
+
dimensions: 1024,
|
|
109
|
+
useTokens: true,
|
|
110
|
+
tokenizerType: 'tiktoken',
|
|
111
|
+
encoding: 'cl100k_base'
|
|
112
|
+
},
|
|
113
|
+
'default': {
|
|
114
|
+
maxTokens: 512,
|
|
115
|
+
optimalTokens: 400,
|
|
116
|
+
minChunkTokens: 50,
|
|
117
|
+
maxChunkTokens: 480,
|
|
118
|
+
overlapTokens: 30,
|
|
119
|
+
optimalChars: 1600,
|
|
120
|
+
minChunkChars: 200,
|
|
121
|
+
maxChunkChars: 1920,
|
|
122
|
+
overlapChars: 120,
|
|
123
|
+
dimensions: 384,
|
|
124
|
+
useTokens: false,
|
|
125
|
+
tokenizerType: 'estimate'
|
|
126
|
+
}
|
|
127
|
+
};
|
|
128
|
+
export async function getModelProfile(providerName, modelName) {
|
|
129
|
+
let baseProfile = modelName ? MODEL_PROFILES[modelName] : undefined;
|
|
130
|
+
if (!baseProfile) {
|
|
131
|
+
const providerDefaults = {
|
|
132
|
+
'OpenAI': MODEL_PROFILES['text-embedding-3-large']
|
|
133
|
+
};
|
|
134
|
+
baseProfile = providerDefaults[providerName] || MODEL_PROFILES['default'];
|
|
135
|
+
}
|
|
136
|
+
const profile = { ...baseProfile };
|
|
137
|
+
if (process.env.CODEVAULT_EMBEDDING_MAX_TOKENS || process.env.CODEVAULT_MAX_TOKENS) {
|
|
138
|
+
const maxTokens = parseInt(process.env.CODEVAULT_EMBEDDING_MAX_TOKENS || process.env.CODEVAULT_MAX_TOKENS || '0', 10);
|
|
139
|
+
if (!isNaN(maxTokens) && maxTokens > 0) {
|
|
140
|
+
const originalMaxTokens = profile.maxTokens;
|
|
141
|
+
const scalingRatio = maxTokens / originalMaxTokens;
|
|
142
|
+
profile.maxTokens = maxTokens;
|
|
143
|
+
profile.optimalTokens = Math.floor(maxTokens * 0.82);
|
|
144
|
+
profile.minChunkTokens = Math.max(Math.floor(profile.minChunkTokens * scalingRatio), 50);
|
|
145
|
+
profile.maxChunkTokens = Math.floor(maxTokens * 0.95);
|
|
146
|
+
profile.overlapTokens = Math.floor(profile.overlapTokens * scalingRatio);
|
|
147
|
+
if (!process.env.CODEVAULT_QUIET) {
|
|
148
|
+
console.log(`Using custom max tokens: ${maxTokens}`);
|
|
149
|
+
console.log(`Auto-scaled optimal tokens: ${profile.optimalTokens} (82% of max)`);
|
|
150
|
+
console.log(`Auto-scaled min tokens: ${profile.minChunkTokens}`);
|
|
151
|
+
console.log(`Auto-scaled max chunk tokens: ${profile.maxChunkTokens} (95% of max)`);
|
|
152
|
+
}
|
|
153
|
+
}
|
|
154
|
+
}
|
|
155
|
+
if (process.env.CODEVAULT_EMBEDDING_DIMENSIONS || process.env.CODEVAULT_DIMENSIONS) {
|
|
156
|
+
const dimensions = parseInt(process.env.CODEVAULT_EMBEDDING_DIMENSIONS || process.env.CODEVAULT_DIMENSIONS || '0', 10);
|
|
157
|
+
if (!isNaN(dimensions) && dimensions > 0) {
|
|
158
|
+
profile.dimensions = dimensions;
|
|
159
|
+
if (!process.env.CODEVAULT_QUIET) {
|
|
160
|
+
console.log(`Using custom dimensions: ${dimensions}`);
|
|
161
|
+
}
|
|
162
|
+
}
|
|
163
|
+
}
|
|
164
|
+
if (profile.useTokens) {
|
|
165
|
+
const { getTokenCounter } = await import('./token-counter.js');
|
|
166
|
+
const counter = await getTokenCounter(modelName || providerName);
|
|
167
|
+
if (!counter) {
|
|
168
|
+
if (!process.env.CODEVAULT_QUIET) {
|
|
169
|
+
console.warn(`Token counter unavailable for ${modelName}, using character estimation`);
|
|
170
|
+
}
|
|
171
|
+
profile.useTokens = false;
|
|
172
|
+
profile.tokenCounter = undefined;
|
|
173
|
+
}
|
|
174
|
+
else {
|
|
175
|
+
profile.tokenCounter = counter;
|
|
176
|
+
}
|
|
177
|
+
}
|
|
178
|
+
return profile;
|
|
179
|
+
}
|
|
180
|
+
export function getSizeLimits(profile) {
|
|
181
|
+
if (profile.useTokens && profile.tokenCounter) {
|
|
182
|
+
return {
|
|
183
|
+
optimal: profile.optimalTokens,
|
|
184
|
+
min: profile.minChunkTokens,
|
|
185
|
+
max: profile.maxChunkTokens,
|
|
186
|
+
overlap: profile.overlapTokens,
|
|
187
|
+
unit: 'tokens'
|
|
188
|
+
};
|
|
189
|
+
}
|
|
190
|
+
return {
|
|
191
|
+
optimal: profile.optimalChars,
|
|
192
|
+
min: profile.minChunkChars,
|
|
193
|
+
max: profile.maxChunkChars,
|
|
194
|
+
overlap: profile.overlapChars,
|
|
195
|
+
unit: 'characters'
|
|
196
|
+
};
|
|
197
|
+
}
|
|
198
|
+
//# sourceMappingURL=base.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"base.js","sourceRoot":"","sources":["../../src/providers/base.ts"],"names":[],"mappings":"AAiBA,MAAM,OAAgB,iBAAiB;IAOrC,+DAA+D;IAC/D,KAAK,CAAC,kBAAkB,CAAC,KAAe;QACtC,mEAAmE;QACnE,MAAM,UAAU,GAAe,EAAE,CAAC;QAClC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC;YACrD,UAAU,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAC7B,CAAC;QACD,OAAO,UAAU,CAAC;IACpB,CAAC;IAED,WAAW,CAAO;CACnB;AAED,qBAAqB;AACrB,MAAM,CAAC,MAAM,UAAU,GAAG,EAAE,CAAC,CAAC,6BAA6B;AAC3D,MAAM,CAAC,MAAM,gBAAgB,GAAG,MAAM,CAAC,CAAC,eAAe;AACvD,MAAM,CAAC,MAAM,eAAe,GAAG,IAAI,CAAC,CAAC,iBAAiB;AAEtD,gDAAgD;AAChD,MAAM,UAAU,cAAc,CAAC,IAAY;IACzC,OAAO,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;AACpC,CAAC;AAED,MAAM,CAAC,MAAM,cAAc,GAAuD;IAChF,wBAAwB,EAAE;QACxB,SAAS,EAAE,IAAI;QACf,aAAa,EAAE,IAAI;QACnB,cAAc,EAAE,GAAG;QACnB,cAAc,EAAE,IAAI;QACpB,aAAa,EAAE,GAAG;QAClB,YAAY,EAAE,KAAK;QACnB,aAAa,EAAE,IAAI;QACnB,aAAa,EAAE,KAAK;QACpB,YAAY,EAAE,GAAG;QACjB,UAAU,EAAE,IAAI;QAChB,SAAS,EAAE,IAAI;QACf,aAAa,EAAE,UAAU;QACzB,QAAQ,EAAE,aAAa;KACxB;IACD,wBAAwB,EAAE;QACxB,SAAS,EAAE,IAAI;QACf,aAAa,EAAE,IAAI;QACnB,cAAc,EAAE,GAAG;QACnB,cAAc,EAAE,IAAI;QACpB,aAAa,EAAE,GAAG;QAClB,YAAY,EAAE,KAAK;QACnB,aAAa,EAAE,IAAI;QACnB,aAAa,EAAE,KAAK;QACpB,YAAY,EAAE,GAAG;QACjB,UAAU,EAAE,IAAI;QAChB,SAAS,EAAE,IAAI;QACf,aAAa,EAAE,UAAU;QACzB,QAAQ,EAAE,aAAa;KACxB;IACD,wBAAwB,EAAE;QACxB,SAAS,EAAE,IAAI;QACf,aAAa,EAAE,IAAI;QACnB,cAAc,EAAE,GAAG;QACnB,cAAc,EAAE,IAAI;QACpB,aAAa,EAAE,GAAG;QAClB,YAAY,EAAE,KAAK;QACnB,aAAa,EAAE,IAAI;QACnB,aAAa,EAAE,KAAK;QACpB,YAAY,EAAE,GAAG;QACjB,UAAU,EAAE,IAAI;QAChB,SAAS,EAAE,IAAI;QACf,aAAa,EAAE,UAAU;QACzB,QAAQ,EAAE,aAAa;KACxB;IACD,kBAAkB,EAAE;QAClB,SAAS,EAAE,IAAI;QACf,aAAa,EAAE,IAAI;QACnB,cAAc,EAAE,GAAG;QACnB,cAAc,EAAE,IAAI;QACpB,aAAa,EAAE,GAAG;QAClB,YAAY,EAAE,KAAK;QACnB,aAAa,EAAE,IAAI;QACnB,aAAa,EAAE,KAAK;QACpB,YAAY,EAAE,GAAG;QACjB,UAAU,EAAE,GAAG;QACf,SAAS,EAAE,IAAI;QACf,aAAa,EAAE,UAAU;QACzB,QAAQ,EAAE,aAAa;KACxB;IACD,yBAAyB,EAAE;QACzB,SAAS,EAAE,KAAK;QAChB,aAAa,EAAE,KAAK;QACpB,cAAc,EAAE,IAAI;QACpB,cAAc,EAAE,KAAK;QACrB,aAAa,EAAE,GAAG;QAClB,YAAY,EAAE,KAAK;QACnB,aAAa,EAAE,IAAI;QACnB,aAAa,EAAE,MAAM;QACrB,YAAY,EAAE,IAAI;QAClB,UAAU,EAAE,IAAI;QAChB,SAAS,EAAE,IAAI;QACf,aAAa,EAAE,UAAU;QACzB,QAAQ,EAAE,aAAa;KACxB;IACD,sBAAsB,EAAE;QACtB,SAAS,EAAE,KAAK;QAChB,aAAa,EAAE,KAAK;QACpB,cAAc,EAAE,IAAI;QACpB,cAAc,EAAE,KAAK;QACrB,aAAa,EAAE,GAAG;QAClB,YAAY,EAAE,MAAM;QACpB,aAAa,EAAE,IAAI;QACnB,aAAa,EAAE,MAAM;QACrB,YAAY,EAAE,IAAI;QAClB,UAAU,EAAE,IAAI;QAChB,SAAS,EAAE,IAAI;QACf,aAAa,EAAE,UAAU;QACzB,QAAQ,EAAE,aAAa;KACxB;IACD,SAAS,EAAE;QACT,SAAS,EAAE,GAAG;QACd,aAAa,EAAE,GAAG;QAClB,cAAc,EAAE,EAAE;QAClB,cAAc,EAAE,GAAG;QACnB,aAAa,EAAE,EAAE;QACjB,YAAY,EAAE,IAAI;QAClB,aAAa,EAAE,GAAG;QAClB,aAAa,EAAE,IAAI;QACnB,YAAY,EAAE,GAAG;QACjB,UAAU,EAAE,GAAG;QACf,SAAS,EAAE,KAAK;QAChB,aAAa,EAAE,UAAU;KAC1B;CACF,CAAC;AAEF,MAAM,CAAC,KAAK,UAAU,eAAe,CAAC,YAAoB,EAAE,SAAwB;IAClF,IAAI,WAAW,GAAG,SAAS,CAAC,CAAC,CAAC,cAAc,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IAEpE,IAAI,CAAC,WAAW,EAAE,CAAC;QACjB,MAAM,gBAAgB,GAAuD;YAC3E,QAAQ,EAAE,cAAc,CAAC,wBAAwB,CAAC;SACnD,CAAC;QACF,WAAW,GAAG,gBAAgB,CAAC,YAAY,CAAC,IAAI,cAAc,CAAC,SAAS,CAAC,CAAC;IAC5E,CAAC;IAED,MAAM,OAAO,GAAiB,EAAE,GAAG,WAAW,EAAkB,CAAC;IAEjE,IAAI,OAAO,CAAC,GAAG,CAAC,8BAA8B,IAAI,OAAO,CAAC,GAAG,CAAC,oBAAoB,EAAE,CAAC;QACnF,MAAM,SAAS,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,8BAA8B,IAAI,OAAO,CAAC,GAAG,CAAC,oBAAoB,IAAI,GAAG,EAAE,EAAE,CAAC,CAAC;QACtH,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;YACvC,MAAM,iBAAiB,GAAG,OAAO,CAAC,SAAS,CAAC;YAC5C,MAAM,YAAY,GAAG,SAAS,GAAG,iBAAiB,CAAC;YAEnD,OAAO,CAAC,SAAS,GAAG,SAAS,CAAC;YAC9B,OAAO,CAAC,aAAa,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,GAAG,IAAI,CAAC,CAAC;YACrD,OAAO,CAAC,cAAc,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,cAAc,GAAG,YAAY,CAAC,EAAE,EAAE,CAAC,CAAC;YACzF,OAAO,CAAC,cAAc,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,GAAG,IAAI,CAAC,CAAC;YACtD,OAAO,CAAC,aAAa,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,aAAa,GAAG,YAAY,CAAC,CAAC;YAEzE,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,eAAe,EAAE,CAAC;gBACjC,OAAO,CAAC,GAAG,CAAC,4BAA4B,SAAS,EAAE,CAAC,CAAC;gBACrD,OAAO,CAAC,GAAG,CAAC,+BAA+B,OAAO,CAAC,aAAa,eAAe,CAAC,CAAC;gBACjF,OAAO,CAAC,GAAG,CAAC,2BAA2B,OAAO,CAAC,cAAc,EAAE,CAAC,CAAC;gBACjE,OAAO,CAAC,GAAG,CAAC,iCAAiC,OAAO,CAAC,cAAc,eAAe,CAAC,CAAC;YACtF,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,OAAO,CAAC,GAAG,CAAC,8BAA8B,IAAI,OAAO,CAAC,GAAG,CAAC,oBAAoB,EAAE,CAAC;QACnF,MAAM,UAAU,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,8BAA8B,IAAI,OAAO,CAAC,GAAG,CAAC,oBAAoB,IAAI,GAAG,EAAE,EAAE,CAAC,CAAC;QACvH,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,IAAI,UAAU,GAAG,CAAC,EAAE,CAAC;YACzC,OAAO,CAAC,UAAU,GAAG,UAAU,CAAC;YAChC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,eAAe,EAAE,CAAC;gBACjC,OAAO,CAAC,GAAG,CAAC,4BAA4B,UAAU,EAAE,CAAC,CAAC;YACxD,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;QACtB,MAAM,EAAE,eAAe,EAAE,GAAG,MAAM,MAAM,CAAC,oBAAoB,CAAC,CAAC;QAC/D,MAAM,OAAO,GAAG,MAAM,eAAe,CAAC,SAAS,IAAI,YAAY,CAAC,CAAC;QAEjE,IAAI,CAAC,OAAO,EAAE,CAAC;YACb,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,eAAe,EAAE,CAAC;gBACjC,OAAO,CAAC,IAAI,CAAC,iCAAiC,SAAS,8BAA8B,CAAC,CAAC;YACzF,CAAC;YACD,OAAO,CAAC,SAAS,GAAG,KAAK,CAAC;YAC1B,OAAO,CAAC,YAAY,GAAG,SAAS,CAAC;QACnC,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,YAAY,GAAG,OAAO,CAAC;QACjC,CAAC;IACH,CAAC;IAED,OAAO,OAAuB,CAAC;AACjC,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,OAAqB;IAOjD,IAAI,OAAO,CAAC,SAAS,IAAI,OAAO,CAAC,YAAY,EAAE,CAAC;QAC9C,OAAO;YACL,OAAO,EAAE,OAAO,CAAC,aAAa;YAC9B,GAAG,EAAE,OAAO,CAAC,cAAc;YAC3B,GAAG,EAAE,OAAO,CAAC,cAAc;YAC3B,OAAO,EAAE,OAAO,CAAC,aAAa;YAC9B,IAAI,EAAE,QAAQ;SACf,CAAC;IACJ,CAAC;IACD,OAAO;QACL,OAAO,EAAE,OAAO,CAAC,YAAY;QAC7B,GAAG,EAAE,OAAO,CAAC,aAAa;QAC1B,GAAG,EAAE,OAAO,CAAC,aAAa;QAC1B,OAAO,EAAE,OAAO,CAAC,YAAY;QAC7B,IAAI,EAAE,YAAY;KACnB,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,30 @@
|
|
|
1
|
+
export interface ChatMessage {
|
|
2
|
+
role: 'system' | 'user' | 'assistant';
|
|
3
|
+
content: string;
|
|
4
|
+
}
|
|
5
|
+
export interface ChatCompletionOptions {
|
|
6
|
+
temperature?: number;
|
|
7
|
+
maxTokens?: number;
|
|
8
|
+
stream?: boolean;
|
|
9
|
+
}
|
|
10
|
+
export declare abstract class ChatLLMProvider {
|
|
11
|
+
abstract generateCompletion(messages: ChatMessage[], options?: ChatCompletionOptions): Promise<string>;
|
|
12
|
+
abstract generateStreamingCompletion(messages: ChatMessage[], options?: ChatCompletionOptions): AsyncGenerator<string>;
|
|
13
|
+
abstract getName(): string;
|
|
14
|
+
abstract getModelName?(): string;
|
|
15
|
+
abstract init?(): Promise<void>;
|
|
16
|
+
rateLimiter?: any;
|
|
17
|
+
}
|
|
18
|
+
export declare class OpenAIChatProvider extends ChatLLMProvider {
|
|
19
|
+
private openai;
|
|
20
|
+
private model;
|
|
21
|
+
rateLimiter: any;
|
|
22
|
+
constructor();
|
|
23
|
+
init(): Promise<void>;
|
|
24
|
+
generateCompletion(messages: ChatMessage[], options?: ChatCompletionOptions): Promise<string>;
|
|
25
|
+
generateStreamingCompletion(messages: ChatMessage[], options?: ChatCompletionOptions): AsyncGenerator<string>;
|
|
26
|
+
getName(): string;
|
|
27
|
+
getModelName(): string;
|
|
28
|
+
}
|
|
29
|
+
export declare function createChatLLMProvider(providerName?: string): ChatLLMProvider;
|
|
30
|
+
//# sourceMappingURL=chat-llm.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"chat-llm.d.ts","sourceRoot":"","sources":["../../src/providers/chat-llm.ts"],"names":[],"mappings":"AAGA,MAAM,WAAW,WAAW;IAC1B,IAAI,EAAE,QAAQ,GAAG,MAAM,GAAG,WAAW,CAAC;IACtC,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,MAAM,WAAW,qBAAqB;IACpC,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,MAAM,CAAC,EAAE,OAAO,CAAC;CAClB;AAED,8BAAsB,eAAe;IACnC,QAAQ,CAAC,kBAAkB,CAAC,QAAQ,EAAE,WAAW,EAAE,EAAE,OAAO,CAAC,EAAE,qBAAqB,GAAG,OAAO,CAAC,MAAM,CAAC;IACtG,QAAQ,CAAC,2BAA2B,CAAC,QAAQ,EAAE,WAAW,EAAE,EAAE,OAAO,CAAC,EAAE,qBAAqB,GAAG,cAAc,CAAC,MAAM,CAAC;IACtH,QAAQ,CAAC,OAAO,IAAI,MAAM;IAC1B,QAAQ,CAAC,YAAY,CAAC,IAAI,MAAM;IAChC,QAAQ,CAAC,IAAI,CAAC,IAAI,OAAO,CAAC,IAAI,CAAC;IAE/B,WAAW,CAAC,EAAE,GAAG,CAAC;CACnB;AAED,qBAAa,kBAAmB,SAAQ,eAAe;IACrD,OAAO,CAAC,MAAM,CAAuB;IACrC,OAAO,CAAC,KAAK,CAAS;IACtB,WAAW,EAAE,GAAG,CAAC;;IAWX,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC;IAgBrB,kBAAkB,CAAC,QAAQ,EAAE,WAAW,EAAE,EAAE,OAAO,GAAE,qBAA0B,GAAG,OAAO,CAAC,MAAM,CAAC;IAqBhG,2BAA2B,CAAC,QAAQ,EAAE,WAAW,EAAE,EAAE,OAAO,GAAE,qBAA0B,GAAG,cAAc,CAAC,MAAM,CAAC;IA4BxH,OAAO,IAAI,MAAM;IAIjB,YAAY,IAAI,MAAM;CAGvB;AAED,wBAAgB,qBAAqB,CAAC,YAAY,SAAS,GAAG,eAAe,CAO5E"}
|
|
@@ -0,0 +1,85 @@
|
|
|
1
|
+
import { OpenAI } from 'openai';
|
|
2
|
+
import { createRateLimiter } from '../utils/rate-limiter.js';
|
|
3
|
+
export class ChatLLMProvider {
|
|
4
|
+
rateLimiter;
|
|
5
|
+
}
|
|
6
|
+
export class OpenAIChatProvider extends ChatLLMProvider {
|
|
7
|
+
openai = null;
|
|
8
|
+
model;
|
|
9
|
+
rateLimiter;
|
|
10
|
+
constructor() {
|
|
11
|
+
super();
|
|
12
|
+
this.model = process.env.CODEVAULT_CHAT_MODEL
|
|
13
|
+
|| process.env.CODEVAULT_OPENAI_CHAT_MODEL // Backward compatibility
|
|
14
|
+
|| 'gpt-4o';
|
|
15
|
+
// Use 'OpenAI' to match rate limiter defaults (rpm: 50)
|
|
16
|
+
this.rateLimiter = createRateLimiter('OpenAI');
|
|
17
|
+
}
|
|
18
|
+
async init() {
|
|
19
|
+
if (!this.openai) {
|
|
20
|
+
const config = {};
|
|
21
|
+
if (process.env.CODEVAULT_CHAT_API_KEY || process.env.OPENAI_API_KEY) {
|
|
22
|
+
config.apiKey = process.env.CODEVAULT_CHAT_API_KEY || process.env.OPENAI_API_KEY;
|
|
23
|
+
}
|
|
24
|
+
if (process.env.CODEVAULT_CHAT_BASE_URL || process.env.OPENAI_BASE_URL) {
|
|
25
|
+
config.baseURL = process.env.CODEVAULT_CHAT_BASE_URL || process.env.OPENAI_BASE_URL;
|
|
26
|
+
}
|
|
27
|
+
this.openai = new OpenAI(config);
|
|
28
|
+
}
|
|
29
|
+
}
|
|
30
|
+
async generateCompletion(messages, options = {}) {
|
|
31
|
+
await this.init();
|
|
32
|
+
const temperature = options.temperature ?? parseFloat(process.env.CODEVAULT_CHAT_TEMPERATURE || '0.7');
|
|
33
|
+
const maxTokens = options.maxTokens ?? parseInt(process.env.CODEVAULT_CHAT_MAX_TOKENS || '4096', 10);
|
|
34
|
+
return await this.rateLimiter.execute(async () => {
|
|
35
|
+
const completion = await this.openai.chat.completions.create({
|
|
36
|
+
model: this.model,
|
|
37
|
+
messages: messages.map(msg => ({
|
|
38
|
+
role: msg.role,
|
|
39
|
+
content: msg.content
|
|
40
|
+
})),
|
|
41
|
+
temperature,
|
|
42
|
+
max_tokens: maxTokens
|
|
43
|
+
});
|
|
44
|
+
return completion.choices[0]?.message?.content || '';
|
|
45
|
+
});
|
|
46
|
+
}
|
|
47
|
+
async *generateStreamingCompletion(messages, options = {}) {
|
|
48
|
+
await this.init();
|
|
49
|
+
const temperature = options.temperature ?? parseFloat(process.env.CODEVAULT_CHAT_TEMPERATURE || '0.7');
|
|
50
|
+
const maxTokens = options.maxTokens ?? parseInt(process.env.CODEVAULT_CHAT_MAX_TOKENS || '4096', 10);
|
|
51
|
+
// Apply rate limiting to streaming requests to prevent overwhelming the provider
|
|
52
|
+
await this.rateLimiter.execute(async () => Promise.resolve(), 0, 0);
|
|
53
|
+
const stream = await this.openai.chat.completions.create({
|
|
54
|
+
model: this.model,
|
|
55
|
+
messages: messages.map(msg => ({
|
|
56
|
+
role: msg.role,
|
|
57
|
+
content: msg.content
|
|
58
|
+
})),
|
|
59
|
+
temperature,
|
|
60
|
+
max_tokens: maxTokens,
|
|
61
|
+
stream: true
|
|
62
|
+
});
|
|
63
|
+
for await (const chunk of stream) {
|
|
64
|
+
const content = chunk.choices[0]?.delta?.content;
|
|
65
|
+
if (content) {
|
|
66
|
+
yield content;
|
|
67
|
+
}
|
|
68
|
+
}
|
|
69
|
+
}
|
|
70
|
+
getName() {
|
|
71
|
+
return 'OpenAI-Chat';
|
|
72
|
+
}
|
|
73
|
+
getModelName() {
|
|
74
|
+
return this.model;
|
|
75
|
+
}
|
|
76
|
+
}
|
|
77
|
+
export function createChatLLMProvider(providerName = 'auto') {
|
|
78
|
+
switch (providerName.toLowerCase()) {
|
|
79
|
+
case 'openai':
|
|
80
|
+
case 'auto':
|
|
81
|
+
default:
|
|
82
|
+
return new OpenAIChatProvider();
|
|
83
|
+
}
|
|
84
|
+
}
|
|
85
|
+
//# sourceMappingURL=chat-llm.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"chat-llm.js","sourceRoot":"","sources":["../../src/providers/chat-llm.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAChC,OAAO,EAAE,iBAAiB,EAAE,MAAM,0BAA0B,CAAC;AAa7D,MAAM,OAAgB,eAAe;IAOnC,WAAW,CAAO;CACnB;AAED,MAAM,OAAO,kBAAmB,SAAQ,eAAe;IAC7C,MAAM,GAAkB,IAAI,CAAC;IAC7B,KAAK,CAAS;IACtB,WAAW,CAAM;IAEjB;QACE,KAAK,EAAE,CAAC;QACR,IAAI,CAAC,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,oBAAoB;eAC7B,OAAO,CAAC,GAAG,CAAC,2BAA2B,CAAC,yBAAyB;eACjE,QAAQ,CAAC;QACzB,wDAAwD;QACxD,IAAI,CAAC,WAAW,GAAG,iBAAiB,CAAC,QAAQ,CAAC,CAAC;IACjD,CAAC;IAED,KAAK,CAAC,IAAI;QACR,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YACjB,MAAM,MAAM,GAAQ,EAAE,CAAC;YAEvB,IAAI,OAAO,CAAC,GAAG,CAAC,sBAAsB,IAAI,OAAO,CAAC,GAAG,CAAC,cAAc,EAAE,CAAC;gBACrE,MAAM,CAAC,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC,sBAAsB,IAAI,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC;YACnF,CAAC;YAED,IAAI,OAAO,CAAC,GAAG,CAAC,uBAAuB,IAAI,OAAO,CAAC,GAAG,CAAC,eAAe,EAAE,CAAC;gBACvE,MAAM,CAAC,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,uBAAuB,IAAI,OAAO,CAAC,GAAG,CAAC,eAAe,CAAC;YACtF,CAAC;YAED,IAAI,CAAC,MAAM,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,CAAC;QACnC,CAAC;IACH,CAAC;IAED,KAAK,CAAC,kBAAkB,CAAC,QAAuB,EAAE,UAAiC,EAAE;QACnF,MAAM,IAAI,CAAC,IAAI,EAAE,CAAC;QAElB,MAAM,WAAW,GAAG,OAAO,CAAC,WAAW,IAAI,UAAU,CAAC,OAAO,CAAC,GAAG,CAAC,0BAA0B,IAAI,KAAK,CAAC,CAAC;QACvG,MAAM,SAAS,GAAG,OAAO,CAAC,SAAS,IAAI,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,yBAAyB,IAAI,MAAM,EAAE,EAAE,CAAC,CAAC;QAErG,OAAO,MAAM,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,KAAK,IAAI,EAAE;YAC/C,MAAM,UAAU,GAAG,MAAM,IAAI,CAAC,MAAO,CAAC,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC;gBAC5D,KAAK,EAAE,IAAI,CAAC,KAAK;gBACjB,QAAQ,EAAE,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;oBAC7B,IAAI,EAAE,GAAG,CAAC,IAAI;oBACd,OAAO,EAAE,GAAG,CAAC,OAAO;iBACrB,CAAC,CAAC;gBACH,WAAW;gBACX,UAAU,EAAE,SAAS;aACtB,CAAC,CAAC;YAEH,OAAO,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,OAAO,IAAI,EAAE,CAAC;QACvD,CAAC,CAAC,CAAC;IACL,CAAC;IAED,KAAK,CAAC,CAAC,2BAA2B,CAAC,QAAuB,EAAE,UAAiC,EAAE;QAC7F,MAAM,IAAI,CAAC,IAAI,EAAE,CAAC;QAElB,MAAM,WAAW,GAAG,OAAO,CAAC,WAAW,IAAI,UAAU,CAAC,OAAO,CAAC,GAAG,CAAC,0BAA0B,IAAI,KAAK,CAAC,CAAC;QACvG,MAAM,SAAS,GAAG,OAAO,CAAC,SAAS,IAAI,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,yBAAyB,IAAI,MAAM,EAAE,EAAE,CAAC,CAAC;QAErG,iFAAiF;QACjF,MAAM,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;QAEpE,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,MAAO,CAAC,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC;YACxD,KAAK,EAAE,IAAI,CAAC,KAAK;YACjB,QAAQ,EAAE,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;gBAC7B,IAAI,EAAE,GAAG,CAAC,IAAI;gBACd,OAAO,EAAE,GAAG,CAAC,OAAO;aACrB,CAAC,CAAC;YACH,WAAW;YACX,UAAU,EAAE,SAAS;YACrB,MAAM,EAAE,IAAI;SACb,CAAC,CAAC;QAEH,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YACjC,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,OAAO,CAAC;YACjD,IAAI,OAAO,EAAE,CAAC;gBACZ,MAAM,OAAO,CAAC;YAChB,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO;QACL,OAAO,aAAa,CAAC;IACvB,CAAC;IAED,YAAY;QACV,OAAO,IAAI,CAAC,KAAK,CAAC;IACpB,CAAC;CACF;AAED,MAAM,UAAU,qBAAqB,CAAC,YAAY,GAAG,MAAM;IACzD,QAAQ,YAAY,CAAC,WAAW,EAAE,EAAE,CAAC;QACnC,KAAK,QAAQ,CAAC;QACd,KAAK,MAAM,CAAC;QACZ;YACE,OAAO,IAAI,kBAAkB,EAAE,CAAC;IACpC,CAAC;AACH,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/providers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,MAAM,WAAW,CAAC;AAG9C,cAAc,WAAW,CAAC;AAC1B,cAAc,aAAa,CAAC;AAE5B,wBAAgB,uBAAuB,CAAC,YAAY,SAAS,GAAG,iBAAiB,CAOhF"}
|
|
@@ -0,0 +1,12 @@
|
|
|
1
|
+
import { OpenAIProvider } from './openai.js';
|
|
2
|
+
export * from './base.js';
|
|
3
|
+
export * from './openai.js';
|
|
4
|
+
export function createEmbeddingProvider(providerName = 'auto') {
|
|
5
|
+
switch (providerName.toLowerCase()) {
|
|
6
|
+
case 'openai':
|
|
7
|
+
case 'auto':
|
|
8
|
+
default:
|
|
9
|
+
return new OpenAIProvider();
|
|
10
|
+
}
|
|
11
|
+
}
|
|
12
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/providers/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,cAAc,EAAE,MAAM,aAAa,CAAC;AAE7C,cAAc,WAAW,CAAC;AAC1B,cAAc,aAAa,CAAC;AAE5B,MAAM,UAAU,uBAAuB,CAAC,YAAY,GAAG,MAAM;IAC3D,QAAQ,YAAY,CAAC,WAAW,EAAE,EAAE,CAAC;QACnC,KAAK,QAAQ,CAAC;QACd,KAAK,MAAM,CAAC;QACZ;YACE,OAAO,IAAI,cAAc,EAAE,CAAC;IAChC,CAAC;AACH,CAAC"}
|
|
@@ -0,0 +1,14 @@
|
|
|
1
|
+
import { EmbeddingProvider } from './base.js';
|
|
2
|
+
export declare class OpenAIProvider extends EmbeddingProvider {
|
|
3
|
+
private openai;
|
|
4
|
+
private model;
|
|
5
|
+
rateLimiter: any;
|
|
6
|
+
constructor();
|
|
7
|
+
init(): Promise<void>;
|
|
8
|
+
generateEmbedding(text: string): Promise<number[]>;
|
|
9
|
+
getDimensions(): number;
|
|
10
|
+
getName(): string;
|
|
11
|
+
getModelName(): string;
|
|
12
|
+
generateEmbeddings(texts: string[]): Promise<number[][]>;
|
|
13
|
+
}
|
|
14
|
+
//# sourceMappingURL=openai.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"openai.d.ts","sourceRoot":"","sources":["../../src/providers/openai.ts"],"names":[],"mappings":"AAEA,OAAO,EACL,iBAAiB,EAMlB,MAAM,WAAW,CAAC;AAEnB,qBAAa,cAAe,SAAQ,iBAAiB;IACnD,OAAO,CAAC,MAAM,CAAuB;IACrC,OAAO,CAAC,KAAK,CAAS;IACtB,WAAW,EAAE,GAAG,CAAC;;IAWX,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC;IAgBrB,iBAAiB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC;IAgBxD,aAAa,IAAI,MAAM;IAWvB,OAAO,IAAI,MAAM;IAIjB,YAAY,IAAI,MAAM;IAKhB,kBAAkB,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC;CAwE/D"}
|
|
@@ -0,0 +1,122 @@
|
|
|
1
|
+
import { OpenAI } from 'openai';
|
|
2
|
+
import { createRateLimiter } from '../utils/rate-limiter.js';
|
|
3
|
+
import { EmbeddingProvider, getModelProfile, getSizeLimits, MAX_BATCH_TOKENS, MAX_ITEM_TOKENS, estimateTokens } from './base.js';
|
|
4
|
+
export class OpenAIProvider extends EmbeddingProvider {
|
|
5
|
+
openai = null;
|
|
6
|
+
model;
|
|
7
|
+
rateLimiter;
|
|
8
|
+
constructor() {
|
|
9
|
+
super();
|
|
10
|
+
this.model = process.env.CODEVAULT_EMBEDDING_MODEL
|
|
11
|
+
|| process.env.CODEVAULT_OPENAI_EMBEDDING_MODEL // Backward compatibility
|
|
12
|
+
|| process.env.OPENAI_MODEL // Backward compatibility
|
|
13
|
+
|| 'text-embedding-3-large';
|
|
14
|
+
this.rateLimiter = createRateLimiter('OpenAI');
|
|
15
|
+
}
|
|
16
|
+
async init() {
|
|
17
|
+
if (!this.openai) {
|
|
18
|
+
const config = {};
|
|
19
|
+
if (process.env.CODEVAULT_EMBEDDING_API_KEY || process.env.OPENAI_API_KEY) {
|
|
20
|
+
config.apiKey = process.env.CODEVAULT_EMBEDDING_API_KEY || process.env.OPENAI_API_KEY;
|
|
21
|
+
}
|
|
22
|
+
if (process.env.CODEVAULT_EMBEDDING_BASE_URL || process.env.OPENAI_BASE_URL) {
|
|
23
|
+
config.baseURL = process.env.CODEVAULT_EMBEDDING_BASE_URL || process.env.OPENAI_BASE_URL;
|
|
24
|
+
}
|
|
25
|
+
this.openai = new OpenAI(config);
|
|
26
|
+
}
|
|
27
|
+
}
|
|
28
|
+
async generateEmbedding(text) {
|
|
29
|
+
await this.init();
|
|
30
|
+
const profile = await getModelProfile(this.getName(), this.model);
|
|
31
|
+
const limits = getSizeLimits(profile);
|
|
32
|
+
const maxChars = profile.maxChunkChars || 8000;
|
|
33
|
+
return await this.rateLimiter.execute(async () => {
|
|
34
|
+
const { data } = await this.openai.embeddings.create({
|
|
35
|
+
model: this.model,
|
|
36
|
+
input: text.slice(0, maxChars)
|
|
37
|
+
});
|
|
38
|
+
return data[0].embedding;
|
|
39
|
+
});
|
|
40
|
+
}
|
|
41
|
+
getDimensions() {
|
|
42
|
+
if (process.env.CODEVAULT_EMBEDDING_DIMENSIONS || process.env.CODEVAULT_DIMENSIONS) {
|
|
43
|
+
const dims = parseInt(process.env.CODEVAULT_EMBEDDING_DIMENSIONS || process.env.CODEVAULT_DIMENSIONS || '0', 10);
|
|
44
|
+
if (!isNaN(dims) && dims > 0)
|
|
45
|
+
return dims;
|
|
46
|
+
}
|
|
47
|
+
if (this.model.includes('3-small'))
|
|
48
|
+
return 1536;
|
|
49
|
+
if (this.model.includes('3-large'))
|
|
50
|
+
return 3072;
|
|
51
|
+
return 1536;
|
|
52
|
+
}
|
|
53
|
+
getName() {
|
|
54
|
+
return 'OpenAI';
|
|
55
|
+
}
|
|
56
|
+
getModelName() {
|
|
57
|
+
return this.model;
|
|
58
|
+
}
|
|
59
|
+
// Batch processing implementation for OpenAI
|
|
60
|
+
async generateEmbeddings(texts) {
|
|
61
|
+
await this.init();
|
|
62
|
+
const profile = await getModelProfile(this.getName(), this.model);
|
|
63
|
+
const maxChars = profile.maxChunkChars || 8000;
|
|
64
|
+
const allEmbeddings = [];
|
|
65
|
+
const remainingTexts = [...texts];
|
|
66
|
+
let batchCount = 0;
|
|
67
|
+
while (remainingTexts.length > 0) {
|
|
68
|
+
batchCount++;
|
|
69
|
+
const currentBatch = [];
|
|
70
|
+
let currentBatchTokens = 0;
|
|
71
|
+
const processedIndices = [];
|
|
72
|
+
// Build batch based on token limits
|
|
73
|
+
for (let i = 0; i < remainingTexts.length; i++) {
|
|
74
|
+
const text = remainingTexts[i];
|
|
75
|
+
const truncatedText = text.slice(0, maxChars);
|
|
76
|
+
const itemTokens = estimateTokens(truncatedText);
|
|
77
|
+
// Skip items that exceed single item limit
|
|
78
|
+
if (itemTokens > MAX_ITEM_TOKENS) {
|
|
79
|
+
if (!process.env.CODEVAULT_QUIET) {
|
|
80
|
+
console.warn(` ⚠️ Text at index ${i} exceeds max token limit (${itemTokens} > ${MAX_ITEM_TOKENS}), skipping`);
|
|
81
|
+
}
|
|
82
|
+
processedIndices.push(i);
|
|
83
|
+
// Add empty embedding as placeholder
|
|
84
|
+
allEmbeddings.push(new Array(this.getDimensions()).fill(0));
|
|
85
|
+
continue;
|
|
86
|
+
}
|
|
87
|
+
// Add to batch if it fits
|
|
88
|
+
if (currentBatchTokens + itemTokens <= MAX_BATCH_TOKENS) {
|
|
89
|
+
currentBatch.push(truncatedText);
|
|
90
|
+
currentBatchTokens += itemTokens;
|
|
91
|
+
processedIndices.push(i);
|
|
92
|
+
}
|
|
93
|
+
else {
|
|
94
|
+
break; // Batch is full
|
|
95
|
+
}
|
|
96
|
+
}
|
|
97
|
+
// Remove processed items from remaining texts
|
|
98
|
+
for (let i = processedIndices.length - 1; i >= 0; i--) {
|
|
99
|
+
remainingTexts.splice(processedIndices[i], 1);
|
|
100
|
+
}
|
|
101
|
+
// Process current batch if not empty
|
|
102
|
+
if (currentBatch.length > 0) {
|
|
103
|
+
if (!process.env.CODEVAULT_QUIET) {
|
|
104
|
+
console.log(` → API call ${batchCount}: ${currentBatch.length} items (${currentBatchTokens} tokens)`);
|
|
105
|
+
}
|
|
106
|
+
const batchEmbeddings = await this.rateLimiter.execute(async () => {
|
|
107
|
+
const { data } = await this.openai.embeddings.create({
|
|
108
|
+
model: this.model,
|
|
109
|
+
input: currentBatch
|
|
110
|
+
});
|
|
111
|
+
return data.map(item => item.embedding);
|
|
112
|
+
});
|
|
113
|
+
allEmbeddings.push(...batchEmbeddings);
|
|
114
|
+
}
|
|
115
|
+
}
|
|
116
|
+
if (!process.env.CODEVAULT_QUIET) {
|
|
117
|
+
console.log(` ✓ Batch complete: ${texts.length} embeddings from ${batchCount} API call(s)\n`);
|
|
118
|
+
}
|
|
119
|
+
return allEmbeddings;
|
|
120
|
+
}
|
|
121
|
+
}
|
|
122
|
+
//# sourceMappingURL=openai.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"openai.js","sourceRoot":"","sources":["../../src/providers/openai.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAChC,OAAO,EAAE,iBAAiB,EAAE,MAAM,0BAA0B,CAAC;AAC7D,OAAO,EACL,iBAAiB,EACjB,eAAe,EACf,aAAa,EACb,gBAAgB,EAChB,eAAe,EACf,cAAc,EACf,MAAM,WAAW,CAAC;AAEnB,MAAM,OAAO,cAAe,SAAQ,iBAAiB;IAC3C,MAAM,GAAkB,IAAI,CAAC;IAC7B,KAAK,CAAS;IACtB,WAAW,CAAM;IAEjB;QACE,KAAK,EAAE,CAAC;QACR,IAAI,CAAC,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,yBAAyB;eAClC,OAAO,CAAC,GAAG,CAAC,gCAAgC,CAAC,yBAAyB;eACtE,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,yBAAyB;eAClD,wBAAwB,CAAC;QACzC,IAAI,CAAC,WAAW,GAAG,iBAAiB,CAAC,QAAQ,CAAC,CAAC;IACjD,CAAC;IAED,KAAK,CAAC,IAAI;QACR,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YACjB,MAAM,MAAM,GAAQ,EAAE,CAAC;YAEvB,IAAI,OAAO,CAAC,GAAG,CAAC,2BAA2B,IAAI,OAAO,CAAC,GAAG,CAAC,cAAc,EAAE,CAAC;gBAC1E,MAAM,CAAC,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC,2BAA2B,IAAI,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC;YACxF,CAAC;YAED,IAAI,OAAO,CAAC,GAAG,CAAC,4BAA4B,IAAI,OAAO,CAAC,GAAG,CAAC,eAAe,EAAE,CAAC;gBAC5E,MAAM,CAAC,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,4BAA4B,IAAI,OAAO,CAAC,GAAG,CAAC,eAAe,CAAC;YAC3F,CAAC;YAED,IAAI,CAAC,MAAM,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,CAAC;QACnC,CAAC;IACH,CAAC;IAED,KAAK,CAAC,iBAAiB,CAAC,IAAY;QAClC,MAAM,IAAI,CAAC,IAAI,EAAE,CAAC;QAElB,MAAM,OAAO,GAAG,MAAM,eAAe,CAAC,IAAI,CAAC,OAAO,EAAE,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC;QAClE,MAAM,MAAM,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,OAAO,CAAC,aAAa,IAAI,IAAI,CAAC;QAE/C,OAAO,MAAM,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,KAAK,IAAI,EAAE;YAC/C,MAAM,EAAE,IAAI,EAAE,GAAG,MAAM,IAAI,CAAC,MAAO,CAAC,UAAU,CAAC,MAAM,CAAC;gBACpD,KAAK,EAAE,IAAI,CAAC,KAAK;gBACjB,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC;aAC/B,CAAC,CAAC;YACH,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;QAC3B,CAAC,CAAC,CAAC;IACL,CAAC;IAED,aAAa;QACX,IAAI,OAAO,CAAC,GAAG,CAAC,8BAA8B,IAAI,OAAO,CAAC,GAAG,CAAC,oBAAoB,EAAE,CAAC;YACnF,MAAM,IAAI,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,8BAA8B,IAAI,OAAO,CAAC,GAAG,CAAC,oBAAoB,IAAI,GAAG,EAAE,EAAE,CAAC,CAAC;YACjH,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,IAAI,GAAG,CAAC;gBAAE,OAAO,IAAI,CAAC;QAC5C,CAAC;QAED,IAAI,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,SAAS,CAAC;YAAE,OAAO,IAAI,CAAC;QAChD,IAAI,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,SAAS,CAAC;YAAE,OAAO,IAAI,CAAC;QAChD,OAAO,IAAI,CAAC;IACd,CAAC;IAED,OAAO;QACL,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED,YAAY;QACV,OAAO,IAAI,CAAC,KAAK,CAAC;IACpB,CAAC;IAED,6CAA6C;IAC7C,KAAK,CAAC,kBAAkB,CAAC,KAAe;QACtC,MAAM,IAAI,CAAC,IAAI,EAAE,CAAC;QAElB,MAAM,OAAO,GAAG,MAAM,eAAe,CAAC,IAAI,CAAC,OAAO,EAAE,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC;QAClE,MAAM,QAAQ,GAAG,OAAO,CAAC,aAAa,IAAI,IAAI,CAAC;QAE/C,MAAM,aAAa,GAAe,EAAE,CAAC;QACrC,MAAM,cAAc,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC;QAClC,IAAI,UAAU,GAAG,CAAC,CAAC;QAEnB,OAAO,cAAc,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACjC,UAAU,EAAE,CAAC;YACb,MAAM,YAAY,GAAa,EAAE,CAAC;YAClC,IAAI,kBAAkB,GAAG,CAAC,CAAC;YAC3B,MAAM,gBAAgB,GAAa,EAAE,CAAC;YAEtC,oCAAoC;YACpC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,cAAc,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC/C,MAAM,IAAI,GAAG,cAAc,CAAC,CAAC,CAAC,CAAC;gBAC/B,MAAM,aAAa,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAC;gBAC9C,MAAM,UAAU,GAAG,cAAc,CAAC,aAAa,CAAC,CAAC;gBAEjD,2CAA2C;gBAC3C,IAAI,UAAU,GAAG,eAAe,EAAE,CAAC;oBACjC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,eAAe,EAAE,CAAC;wBACjC,OAAO,CAAC,IAAI,CAAC,uBAAuB,CAAC,6BAA6B,UAAU,MAAM,eAAe,aAAa,CAAC,CAAC;oBAClH,CAAC;oBACD,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;oBACzB,qCAAqC;oBACrC,aAAa,CAAC,IAAI,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,aAAa,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;oBAC5D,SAAS;gBACX,CAAC;gBAED,0BAA0B;gBAC1B,IAAI,kBAAkB,GAAG,UAAU,IAAI,gBAAgB,EAAE,CAAC;oBACxD,YAAY,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;oBACjC,kBAAkB,IAAI,UAAU,CAAC;oBACjC,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;gBAC3B,CAAC;qBAAM,CAAC;oBACN,MAAM,CAAC,gBAAgB;gBACzB,CAAC;YACH,CAAC;YAED,8CAA8C;YAC9C,KAAK,IAAI,CAAC,GAAG,gBAAgB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;gBACtD,cAAc,CAAC,MAAM,CAAC,gBAAgB,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;YAChD,CAAC;YAED,qCAAqC;YACrC,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC5B,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,eAAe,EAAE,CAAC;oBACjC,OAAO,CAAC,GAAG,CAAC,gBAAgB,UAAU,KAAK,YAAY,CAAC,MAAM,WAAW,kBAAkB,UAAU,CAAC,CAAC;gBACzG,CAAC;gBAED,MAAM,eAAe,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,KAAK,IAAI,EAAE;oBAChE,MAAM,EAAE,IAAI,EAAE,GAAG,MAAM,IAAI,CAAC,MAAO,CAAC,UAAU,CAAC,MAAM,CAAC;wBACpD,KAAK,EAAE,IAAI,CAAC,KAAK;wBACjB,KAAK,EAAE,YAAY;qBACpB,CAAC,CAAC;oBACH,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;gBAC1C,CAAC,CAAC,CAAC;gBAEH,aAAa,CAAC,IAAI,CAAC,GAAG,eAAe,CAAC,CAAC;YACzC,CAAC;QACH,CAAC;QAED,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,eAAe,EAAE,CAAC;YACjC,OAAO,CAAC,GAAG,CAAC,uBAAuB,KAAK,CAAC,MAAM,oBAAoB,UAAU,gBAAgB,CAAC,CAAC;QACjG,CAAC;QAED,OAAO,aAAa,CAAC;IACvB,CAAC;CACF"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"token-counter.d.ts","sourceRoot":"","sources":["../../src/providers/token-counter.ts"],"names":[],"mappings":"AAEA,wBAAsB,eAAe,CAAC,SAAS,EAAE,MAAM,GAAG,OAAO,CAAC,CAAC,CAAC,IAAI,EAAE,MAAM,KAAK,MAAM,CAAC,GAAG,IAAI,CAAC,CAenG"}
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
let tiktokenEncoder = null;
|
|
2
|
+
export async function getTokenCounter(modelName) {
|
|
3
|
+
if (modelName.includes('text-embedding') || modelName.includes('ada-002')) {
|
|
4
|
+
if (!tiktokenEncoder) {
|
|
5
|
+
try {
|
|
6
|
+
const tiktoken = await import('tiktoken');
|
|
7
|
+
tiktokenEncoder = tiktoken.encoding_for_model('text-embedding-3-large');
|
|
8
|
+
}
|
|
9
|
+
catch (error) {
|
|
10
|
+
console.warn('tiktoken not available, falling back to character estimation');
|
|
11
|
+
return null;
|
|
12
|
+
}
|
|
13
|
+
}
|
|
14
|
+
return (text) => tiktokenEncoder.encode(text).length;
|
|
15
|
+
}
|
|
16
|
+
return (text) => Math.ceil(text.length / 4);
|
|
17
|
+
}
|
|
18
|
+
//# sourceMappingURL=token-counter.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"token-counter.js","sourceRoot":"","sources":["../../src/providers/token-counter.ts"],"names":[],"mappings":"AAAA,IAAI,eAAe,GAAQ,IAAI,CAAC;AAEhC,MAAM,CAAC,KAAK,UAAU,eAAe,CAAC,SAAiB;IACrD,IAAI,SAAS,CAAC,QAAQ,CAAC,gBAAgB,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,CAAC;QAC1E,IAAI,CAAC,eAAe,EAAE,CAAC;YACrB,IAAI,CAAC;gBACH,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,UAAU,CAAC,CAAC;gBAC1C,eAAe,GAAG,QAAQ,CAAC,kBAAkB,CAAC,wBAAwB,CAAC,CAAC;YAC1E,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,OAAO,CAAC,IAAI,CAAC,8DAA8D,CAAC,CAAC;gBAC7E,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;QACD,OAAO,CAAC,IAAY,EAAE,EAAE,CAAC,eAAe,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC;IAC/D,CAAC;IAED,OAAO,CAAC,IAAY,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;AACtD,CAAC"}
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
export declare function isAPIRerankingConfigured(): boolean;
|
|
2
|
+
interface Candidate {
|
|
3
|
+
rerankerScore?: number;
|
|
4
|
+
rerankerRank?: number;
|
|
5
|
+
[key: string]: any;
|
|
6
|
+
}
|
|
7
|
+
interface RerankOptions {
|
|
8
|
+
max?: number;
|
|
9
|
+
maxTokens?: number;
|
|
10
|
+
getText?: (candidate: Candidate) => string;
|
|
11
|
+
getTextAsync?: (candidate: Candidate) => Promise<string>;
|
|
12
|
+
apiUrl?: string;
|
|
13
|
+
apiKey?: string;
|
|
14
|
+
model?: string;
|
|
15
|
+
}
|
|
16
|
+
export declare function rerankWithAPI(query: string, candidates: Candidate[], options?: RerankOptions): Promise<Candidate[]>;
|
|
17
|
+
export {};
|
|
18
|
+
//# sourceMappingURL=api-reranker.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"api-reranker.d.ts","sourceRoot":"","sources":["../../src/ranking/api-reranker.ts"],"names":[],"mappings":"AAoCA,wBAAgB,wBAAwB,IAAI,OAAO,CAIlD;AAuED,UAAU,SAAS;IACjB,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,CAAC,GAAG,EAAE,MAAM,GAAG,GAAG,CAAC;CACpB;AAED,UAAU,aAAa;IACrB,GAAG,CAAC,EAAE,MAAM,CAAC;IACb,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,OAAO,CAAC,EAAE,CAAC,SAAS,EAAE,SAAS,KAAK,MAAM,CAAC;IAC3C,YAAY,CAAC,EAAE,CAAC,SAAS,EAAE,SAAS,KAAK,OAAO,CAAC,MAAM,CAAC,CAAC;IACzD,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED,wBAAsB,aAAa,CAAC,KAAK,EAAE,MAAM,EAAE,UAAU,EAAE,SAAS,EAAE,EAAE,OAAO,GAAE,aAAkB,GAAG,OAAO,CAAC,SAAS,EAAE,CAAC,CAoE7H"}
|