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.
Files changed (237) hide show
  1. package/dist/chunking/file-grouper.d.ts +39 -0
  2. package/dist/chunking/file-grouper.d.ts.map +1 -0
  3. package/dist/chunking/file-grouper.js +164 -0
  4. package/dist/chunking/file-grouper.js.map +1 -0
  5. package/dist/chunking/semantic-chunker.d.ts +37 -0
  6. package/dist/chunking/semantic-chunker.d.ts.map +1 -0
  7. package/dist/chunking/semantic-chunker.js +157 -0
  8. package/dist/chunking/semantic-chunker.js.map +1 -0
  9. package/dist/chunking/token-counter.d.ts +28 -0
  10. package/dist/chunking/token-counter.d.ts.map +1 -0
  11. package/dist/chunking/token-counter.js +207 -0
  12. package/dist/chunking/token-counter.js.map +1 -0
  13. package/dist/cli/commands/ask-cmd.d.ts +3 -0
  14. package/dist/cli/commands/ask-cmd.d.ts.map +1 -0
  15. package/dist/cli/commands/ask-cmd.js +130 -0
  16. package/dist/cli/commands/ask-cmd.js.map +1 -0
  17. package/dist/cli/commands/chat-cmd.d.ts +3 -0
  18. package/dist/cli/commands/chat-cmd.d.ts.map +1 -0
  19. package/dist/cli/commands/chat-cmd.js +194 -0
  20. package/dist/cli/commands/chat-cmd.js.map +1 -0
  21. package/dist/cli/commands/config-cmd.d.ts +3 -0
  22. package/dist/cli/commands/config-cmd.d.ts.map +1 -0
  23. package/dist/cli/commands/config-cmd.js +245 -0
  24. package/dist/cli/commands/config-cmd.js.map +1 -0
  25. package/dist/cli/commands/context.d.ts +3 -0
  26. package/dist/cli/commands/context.d.ts.map +1 -0
  27. package/dist/cli/commands/context.js +98 -0
  28. package/dist/cli/commands/context.js.map +1 -0
  29. package/dist/cli/commands/interactive-config.d.ts +2 -0
  30. package/dist/cli/commands/interactive-config.d.ts.map +1 -0
  31. package/dist/cli/commands/interactive-config.js +274 -0
  32. package/dist/cli/commands/interactive-config.js.map +1 -0
  33. package/dist/cli.d.ts +3 -0
  34. package/dist/cli.d.ts.map +1 -0
  35. package/dist/cli.js +403 -0
  36. package/dist/cli.js.map +1 -0
  37. package/dist/codemap/io.d.ts +5 -0
  38. package/dist/codemap/io.d.ts.map +1 -0
  39. package/dist/codemap/io.js +30 -0
  40. package/dist/codemap/io.js.map +1 -0
  41. package/dist/config/apply-env.d.ts +15 -0
  42. package/dist/config/apply-env.d.ts.map +1 -0
  43. package/dist/config/apply-env.js +91 -0
  44. package/dist/config/apply-env.js.map +1 -0
  45. package/dist/config/constants.d.ts +326 -0
  46. package/dist/config/constants.d.ts.map +1 -0
  47. package/dist/config/constants.js +214 -0
  48. package/dist/config/constants.js.map +1 -0
  49. package/dist/config/loader.d.ts +57 -0
  50. package/dist/config/loader.d.ts.map +1 -0
  51. package/dist/config/loader.js +287 -0
  52. package/dist/config/loader.js.map +1 -0
  53. package/dist/config/types.d.ts +46 -0
  54. package/dist/config/types.d.ts.map +1 -0
  55. package/dist/config/types.js +2 -0
  56. package/dist/config/types.js.map +1 -0
  57. package/dist/context/packs.d.ts +33 -0
  58. package/dist/context/packs.d.ts.map +1 -0
  59. package/dist/context/packs.js +180 -0
  60. package/dist/context/packs.js.map +1 -0
  61. package/dist/core/batch-indexer.d.ts +56 -0
  62. package/dist/core/batch-indexer.d.ts.map +1 -0
  63. package/dist/core/batch-indexer.js +189 -0
  64. package/dist/core/batch-indexer.js.map +1 -0
  65. package/dist/core/indexer.d.ts +3 -0
  66. package/dist/core/indexer.d.ts.map +1 -0
  67. package/dist/core/indexer.js +600 -0
  68. package/dist/core/indexer.js.map +1 -0
  69. package/dist/core/metadata.d.ts +19 -0
  70. package/dist/core/metadata.d.ts.map +1 -0
  71. package/dist/core/metadata.js +161 -0
  72. package/dist/core/metadata.js.map +1 -0
  73. package/dist/core/search.d.ts +7 -0
  74. package/dist/core/search.d.ts.map +1 -0
  75. package/dist/core/search.js +545 -0
  76. package/dist/core/search.js.map +1 -0
  77. package/dist/core/symbol-extractor.d.ts +3 -0
  78. package/dist/core/symbol-extractor.d.ts.map +1 -0
  79. package/dist/core/symbol-extractor.js +78 -0
  80. package/dist/core/symbol-extractor.js.map +1 -0
  81. package/dist/core/types.d.ts +104 -0
  82. package/dist/core/types.d.ts.map +1 -0
  83. package/dist/core/types.js +2 -0
  84. package/dist/core/types.js.map +1 -0
  85. package/dist/database/db.d.ts +102 -0
  86. package/dist/database/db.d.ts.map +1 -0
  87. package/dist/database/db.js +366 -0
  88. package/dist/database/db.js.map +1 -0
  89. package/dist/indexer/merkle.d.ts +13 -0
  90. package/dist/indexer/merkle.d.ts.map +1 -0
  91. package/dist/indexer/merkle.js +86 -0
  92. package/dist/indexer/merkle.js.map +1 -0
  93. package/dist/indexer/update.d.ts +19 -0
  94. package/dist/indexer/update.d.ts.map +1 -0
  95. package/dist/indexer/update.js +40 -0
  96. package/dist/indexer/update.js.map +1 -0
  97. package/dist/indexer/watch.d.ts +21 -0
  98. package/dist/indexer/watch.d.ts.map +1 -0
  99. package/dist/indexer/watch.js +222 -0
  100. package/dist/indexer/watch.js.map +1 -0
  101. package/dist/languages/rules.d.ts +11 -0
  102. package/dist/languages/rules.d.ts.map +1 -0
  103. package/dist/languages/rules.js +371 -0
  104. package/dist/languages/rules.js.map +1 -0
  105. package/dist/languages/tree-sitter-loader.d.ts +27 -0
  106. package/dist/languages/tree-sitter-loader.d.ts.map +1 -0
  107. package/dist/languages/tree-sitter-loader.js +76 -0
  108. package/dist/languages/tree-sitter-loader.js.map +1 -0
  109. package/dist/mcp/schemas.d.ts +166 -0
  110. package/dist/mcp/schemas.d.ts.map +1 -0
  111. package/dist/mcp/schemas.js +159 -0
  112. package/dist/mcp/schemas.js.map +1 -0
  113. package/dist/mcp/tools/ask-codebase.d.ts +85 -0
  114. package/dist/mcp/tools/ask-codebase.d.ts.map +1 -0
  115. package/dist/mcp/tools/ask-codebase.js +125 -0
  116. package/dist/mcp/tools/ask-codebase.js.map +1 -0
  117. package/dist/mcp/tools/use-context-pack.d.ts +57 -0
  118. package/dist/mcp/tools/use-context-pack.d.ts.map +1 -0
  119. package/dist/mcp/tools/use-context-pack.js +91 -0
  120. package/dist/mcp/tools/use-context-pack.js.map +1 -0
  121. package/dist/mcp-server.d.ts +3 -0
  122. package/dist/mcp-server.d.ts.map +1 -0
  123. package/dist/mcp-server.js +511 -0
  124. package/dist/mcp-server.js.map +1 -0
  125. package/dist/providers/base.d.ts +39 -0
  126. package/dist/providers/base.d.ts.map +1 -0
  127. package/dist/providers/base.js +198 -0
  128. package/dist/providers/base.js.map +1 -0
  129. package/dist/providers/chat-llm.d.ts +30 -0
  130. package/dist/providers/chat-llm.d.ts.map +1 -0
  131. package/dist/providers/chat-llm.js +85 -0
  132. package/dist/providers/chat-llm.js.map +1 -0
  133. package/dist/providers/index.d.ts +5 -0
  134. package/dist/providers/index.d.ts.map +1 -0
  135. package/dist/providers/index.js +12 -0
  136. package/dist/providers/index.js.map +1 -0
  137. package/dist/providers/openai.d.ts +14 -0
  138. package/dist/providers/openai.d.ts.map +1 -0
  139. package/dist/providers/openai.js +122 -0
  140. package/dist/providers/openai.js.map +1 -0
  141. package/dist/providers/token-counter.d.ts +2 -0
  142. package/dist/providers/token-counter.d.ts.map +1 -0
  143. package/dist/providers/token-counter.js +18 -0
  144. package/dist/providers/token-counter.js.map +1 -0
  145. package/dist/ranking/api-reranker.d.ts +18 -0
  146. package/dist/ranking/api-reranker.d.ts.map +1 -0
  147. package/dist/ranking/api-reranker.js +134 -0
  148. package/dist/ranking/api-reranker.js.map +1 -0
  149. package/dist/ranking/symbol-boost.d.ts +15 -0
  150. package/dist/ranking/symbol-boost.d.ts.map +1 -0
  151. package/dist/ranking/symbol-boost.js +175 -0
  152. package/dist/ranking/symbol-boost.js.map +1 -0
  153. package/dist/search/bm25.d.ts +17 -0
  154. package/dist/search/bm25.d.ts.map +1 -0
  155. package/dist/search/bm25.js +56 -0
  156. package/dist/search/bm25.js.map +1 -0
  157. package/dist/search/hybrid.d.ts +21 -0
  158. package/dist/search/hybrid.d.ts.map +1 -0
  159. package/dist/search/hybrid.js +50 -0
  160. package/dist/search/hybrid.js.map +1 -0
  161. package/dist/search/scope.d.ts +5 -0
  162. package/dist/search/scope.d.ts.map +1 -0
  163. package/dist/search/scope.js +107 -0
  164. package/dist/search/scope.js.map +1 -0
  165. package/dist/storage/encrypted-chunks.d.ts +40 -0
  166. package/dist/storage/encrypted-chunks.d.ts.map +1 -0
  167. package/dist/storage/encrypted-chunks.js +238 -0
  168. package/dist/storage/encrypted-chunks.js.map +1 -0
  169. package/dist/symbols/extract.d.ts +15 -0
  170. package/dist/symbols/extract.d.ts.map +1 -0
  171. package/dist/symbols/extract.js +208 -0
  172. package/dist/symbols/extract.js.map +1 -0
  173. package/dist/symbols/graph.d.ts +3 -0
  174. package/dist/symbols/graph.d.ts.map +1 -0
  175. package/dist/symbols/graph.js +89 -0
  176. package/dist/symbols/graph.js.map +1 -0
  177. package/dist/synthesis/conversational-synthesizer.d.ts +61 -0
  178. package/dist/synthesis/conversational-synthesizer.d.ts.map +1 -0
  179. package/dist/synthesis/conversational-synthesizer.js +286 -0
  180. package/dist/synthesis/conversational-synthesizer.js.map +1 -0
  181. package/dist/synthesis/markdown-formatter.d.ts +13 -0
  182. package/dist/synthesis/markdown-formatter.d.ts.map +1 -0
  183. package/dist/synthesis/markdown-formatter.js +104 -0
  184. package/dist/synthesis/markdown-formatter.js.map +1 -0
  185. package/dist/synthesis/prompt-builder.d.ts +21 -0
  186. package/dist/synthesis/prompt-builder.d.ts.map +1 -0
  187. package/dist/synthesis/prompt-builder.js +129 -0
  188. package/dist/synthesis/prompt-builder.js.map +1 -0
  189. package/dist/synthesis/synthesizer.d.ts +30 -0
  190. package/dist/synthesis/synthesizer.d.ts.map +1 -0
  191. package/dist/synthesis/synthesizer.js +210 -0
  192. package/dist/synthesis/synthesizer.js.map +1 -0
  193. package/dist/types/ast.d.ts +3 -0
  194. package/dist/types/ast.d.ts.map +1 -0
  195. package/dist/types/ast.js +2 -0
  196. package/dist/types/ast.js.map +1 -0
  197. package/dist/types/codemap.d.ts +58 -0
  198. package/dist/types/codemap.d.ts.map +1 -0
  199. package/dist/types/codemap.js +224 -0
  200. package/dist/types/codemap.js.map +1 -0
  201. package/dist/types/context-pack.d.ts +47 -0
  202. package/dist/types/context-pack.d.ts.map +1 -0
  203. package/dist/types/context-pack.js +44 -0
  204. package/dist/types/context-pack.js.map +1 -0
  205. package/dist/types/search.d.ts +15 -0
  206. package/dist/types/search.d.ts.map +1 -0
  207. package/dist/types/search.js +11 -0
  208. package/dist/types/search.js.map +1 -0
  209. package/dist/utils/cli-ui.d.ts +44 -0
  210. package/dist/utils/cli-ui.d.ts.map +1 -0
  211. package/dist/utils/cli-ui.js +139 -0
  212. package/dist/utils/cli-ui.js.map +1 -0
  213. package/dist/utils/indexer-with-progress.d.ts +10 -0
  214. package/dist/utils/indexer-with-progress.d.ts.map +1 -0
  215. package/dist/utils/indexer-with-progress.js +43 -0
  216. package/dist/utils/indexer-with-progress.js.map +1 -0
  217. package/dist/utils/logger.d.ts +55 -0
  218. package/dist/utils/logger.d.ts.map +1 -0
  219. package/dist/utils/logger.js +121 -0
  220. package/dist/utils/logger.js.map +1 -0
  221. package/dist/utils/mutex.d.ts +63 -0
  222. package/dist/utils/mutex.d.ts.map +1 -0
  223. package/dist/utils/mutex.js +123 -0
  224. package/dist/utils/mutex.js.map +1 -0
  225. package/dist/utils/path-helpers.d.ts +27 -0
  226. package/dist/utils/path-helpers.d.ts.map +1 -0
  227. package/dist/utils/path-helpers.js +55 -0
  228. package/dist/utils/path-helpers.js.map +1 -0
  229. package/dist/utils/rate-limiter.d.ts +34 -0
  230. package/dist/utils/rate-limiter.d.ts.map +1 -0
  231. package/dist/utils/rate-limiter.js +178 -0
  232. package/dist/utils/rate-limiter.js.map +1 -0
  233. package/dist/utils/scan-patterns.d.ts +5 -0
  234. package/dist/utils/scan-patterns.d.ts.map +1 -0
  235. package/dist/utils/scan-patterns.js +29 -0
  236. package/dist/utils/scan-patterns.js.map +1 -0
  237. package/package.json +1 -1
@@ -0,0 +1,39 @@
1
+ import type { ModelProfile } from '../providers/base.js';
2
+ import type { TreeSitterNode } from '../types/ast.js';
3
+ interface LanguageRule {
4
+ subdivisionTypes?: Record<string, string[]>;
5
+ [key: string]: any;
6
+ }
7
+ interface NodeAnalysis {
8
+ node: TreeSitterNode;
9
+ size: number;
10
+ code: string;
11
+ }
12
+ interface SemanticGroup {
13
+ type: string;
14
+ containerType?: string;
15
+ nodes: TreeSitterNode[];
16
+ analyses: NodeAnalysis[];
17
+ parentNode: TreeSitterNode | null;
18
+ }
19
+ export interface NodeGroup {
20
+ nodes: TreeSitterNode[];
21
+ totalSize: number;
22
+ groupInfo: SemanticGroup[];
23
+ }
24
+ export declare function groupNodesForChunking(nodes: TreeSitterNode[], source: string, profile: ModelProfile, rule: LanguageRule): Promise<NodeGroup[]>;
25
+ export interface CombinedChunk {
26
+ code: string;
27
+ node: TreeSitterNode & {
28
+ type: string;
29
+ };
30
+ metadata: {
31
+ isGroup: boolean;
32
+ nodeCount: number;
33
+ totalSize: number;
34
+ groupTypes: string[];
35
+ };
36
+ }
37
+ export declare function createCombinedChunk(nodeGroup: NodeGroup, source: string): CombinedChunk | null;
38
+ export {};
39
+ //# sourceMappingURL=file-grouper.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"file-grouper.d.ts","sourceRoot":"","sources":["../../src/chunking/file-grouper.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAEzD,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AAEtD,UAAU,YAAY;IACpB,gBAAgB,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,EAAE,CAAC,CAAC;IAC5C,CAAC,GAAG,EAAE,MAAM,GAAG,GAAG,CAAC;CACpB;AAED,UAAU,YAAY;IACpB,IAAI,EAAE,cAAc,CAAC;IACrB,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;CACd;AAED,UAAU,aAAa;IACrB,IAAI,EAAE,MAAM,CAAC;IACb,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,KAAK,EAAE,cAAc,EAAE,CAAC;IACxB,QAAQ,EAAE,YAAY,EAAE,CAAC;IACzB,UAAU,EAAE,cAAc,GAAG,IAAI,CAAC;CACnC;AAED,MAAM,WAAW,SAAS;IACxB,KAAK,EAAE,cAAc,EAAE,CAAC;IACxB,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,aAAa,EAAE,CAAC;CAC5B;AAwJD,wBAAsB,qBAAqB,CACzC,KAAK,EAAE,cAAc,EAAE,EACvB,MAAM,EAAE,MAAM,EACd,OAAO,EAAE,YAAY,EACrB,IAAI,EAAE,YAAY,GACjB,OAAO,CAAC,SAAS,EAAE,CAAC,CAkBtB;AAED,MAAM,WAAW,aAAa;IAC5B,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,cAAc,GAAG;QAAE,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC;IACxC,QAAQ,EAAE;QACR,OAAO,EAAE,OAAO,CAAC;QACjB,SAAS,EAAE,MAAM,CAAC;QAClB,SAAS,EAAE,MAAM,CAAC;QAClB,UAAU,EAAE,MAAM,EAAE,CAAC;KACtB,CAAC;CACH;AAED,wBAAgB,mBAAmB,CAAC,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,MAAM,GAAG,aAAa,GAAG,IAAI,CA4B9F"}
@@ -0,0 +1,164 @@
1
+ import { batchAnalyzeCodeSize } from './token-counter.js';
2
+ import { getSizeLimits } from '../providers/base.js';
3
+ async function batchAnalyzeNodesInternal(nodes, source, profile) {
4
+ const codes = nodes.map(node => source.slice(node.startIndex, node.endIndex));
5
+ const limits = getSizeLimits(profile);
6
+ if (profile.useTokens && profile.tokenCounter) {
7
+ const analyses = await batchAnalyzeCodeSize(codes, limits, profile.tokenCounter, false);
8
+ return nodes.map((node, i) => ({
9
+ node,
10
+ size: analyses[i].size,
11
+ code: codes[i]
12
+ }));
13
+ }
14
+ return nodes.map((node, i) => ({
15
+ node,
16
+ size: codes[i].length,
17
+ code: codes[i]
18
+ }));
19
+ }
20
+ function isContainerNode(node, rule) {
21
+ const containerTypes = [
22
+ 'class_declaration',
23
+ 'class_definition',
24
+ 'interface_declaration',
25
+ 'module_declaration',
26
+ 'namespace_declaration',
27
+ 'trait_declaration',
28
+ 'enum_declaration'
29
+ ];
30
+ return containerTypes.includes(node.type);
31
+ }
32
+ function identifySemanticGroups(nodes, source, nodeAnalyses, rule) {
33
+ const groups = [];
34
+ let currentGroup = {
35
+ type: 'file_section',
36
+ nodes: [],
37
+ analyses: [],
38
+ parentNode: null
39
+ };
40
+ for (let i = 0; i < nodes.length; i++) {
41
+ const node = nodes[i];
42
+ const analysis = nodeAnalyses[i];
43
+ if (isContainerNode(node, rule)) {
44
+ if (currentGroup.nodes.length > 0) {
45
+ groups.push(currentGroup);
46
+ }
47
+ currentGroup = {
48
+ type: 'container',
49
+ containerType: node.type,
50
+ nodes: [node],
51
+ analyses: [analysis],
52
+ parentNode: node
53
+ };
54
+ groups.push(currentGroup);
55
+ currentGroup = {
56
+ type: 'file_section',
57
+ nodes: [],
58
+ analyses: [],
59
+ parentNode: null
60
+ };
61
+ }
62
+ else {
63
+ currentGroup.nodes.push(node);
64
+ currentGroup.analyses.push(analysis);
65
+ }
66
+ }
67
+ if (currentGroup.nodes.length > 0) {
68
+ groups.push(currentGroup);
69
+ }
70
+ return groups;
71
+ }
72
+ async function combineGroupsToOptimalSize(semanticGroups, source, profile, limits) {
73
+ const optimalGroups = [];
74
+ let currentCombinedGroup = {
75
+ nodes: [],
76
+ totalSize: 0,
77
+ groupInfo: []
78
+ };
79
+ for (const group of semanticGroups) {
80
+ const groupTotalSize = group.analyses.reduce((sum, a) => sum + a.size, 0);
81
+ if (groupTotalSize > limits.optimal) {
82
+ if (currentCombinedGroup.nodes.length > 0) {
83
+ optimalGroups.push(currentCombinedGroup);
84
+ }
85
+ optimalGroups.push({
86
+ nodes: group.nodes,
87
+ totalSize: groupTotalSize,
88
+ groupInfo: [group]
89
+ });
90
+ currentCombinedGroup = {
91
+ nodes: [],
92
+ totalSize: 0,
93
+ groupInfo: []
94
+ };
95
+ continue;
96
+ }
97
+ if (currentCombinedGroup.totalSize + groupTotalSize > limits.max) {
98
+ if (currentCombinedGroup.nodes.length > 0) {
99
+ optimalGroups.push(currentCombinedGroup);
100
+ }
101
+ currentCombinedGroup = {
102
+ nodes: group.nodes,
103
+ totalSize: groupTotalSize,
104
+ groupInfo: [group]
105
+ };
106
+ continue;
107
+ }
108
+ currentCombinedGroup.nodes.push(...group.nodes);
109
+ currentCombinedGroup.totalSize += groupTotalSize;
110
+ currentCombinedGroup.groupInfo.push(group);
111
+ if (currentCombinedGroup.totalSize >= limits.optimal * 0.9) {
112
+ optimalGroups.push(currentCombinedGroup);
113
+ currentCombinedGroup = {
114
+ nodes: [],
115
+ totalSize: 0,
116
+ groupInfo: []
117
+ };
118
+ }
119
+ }
120
+ if (currentCombinedGroup.nodes.length > 0) {
121
+ optimalGroups.push(currentCombinedGroup);
122
+ }
123
+ return optimalGroups;
124
+ }
125
+ export async function groupNodesForChunking(nodes, source, profile, rule) {
126
+ if (!nodes || nodes.length === 0)
127
+ return [];
128
+ const limits = getSizeLimits(profile);
129
+ if (nodes.length <= 10) {
130
+ return nodes.map(node => ({
131
+ nodes: [node],
132
+ totalSize: 0,
133
+ groupInfo: []
134
+ }));
135
+ }
136
+ const nodeAnalyses = await batchAnalyzeNodesInternal(nodes, source, profile);
137
+ const semanticGroups = identifySemanticGroups(nodes, source, nodeAnalyses, rule);
138
+ const optimalGroups = await combineGroupsToOptimalSize(semanticGroups, source, profile, limits);
139
+ return optimalGroups;
140
+ }
141
+ export function createCombinedChunk(nodeGroup, source) {
142
+ if (!nodeGroup.nodes || nodeGroup.nodes.length === 0) {
143
+ return null;
144
+ }
145
+ const codes = nodeGroup.nodes.map(node => source.slice(node.startIndex, node.endIndex));
146
+ const combinedCode = codes.join('\n\n');
147
+ const firstNode = nodeGroup.nodes[0];
148
+ const lastNode = nodeGroup.nodes[nodeGroup.nodes.length - 1];
149
+ return {
150
+ code: combinedCode,
151
+ node: {
152
+ ...firstNode,
153
+ type: `${firstNode.type}_group_${nodeGroup.nodes.length}`,
154
+ endIndex: lastNode.endIndex
155
+ },
156
+ metadata: {
157
+ isGroup: true,
158
+ nodeCount: nodeGroup.nodes.length,
159
+ totalSize: nodeGroup.totalSize,
160
+ groupTypes: nodeGroup.groupInfo?.map(g => g.type) || ['combined']
161
+ }
162
+ };
163
+ }
164
+ //# sourceMappingURL=file-grouper.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"file-grouper.js","sourceRoot":"","sources":["../../src/chunking/file-grouper.ts"],"names":[],"mappings":"AAAA,OAAO,EAAmB,oBAAoB,EAAyB,MAAM,oBAAoB,CAAC;AAElG,OAAO,EAAE,aAAa,EAAE,MAAM,sBAAsB,CAAC;AA4BrD,KAAK,UAAU,yBAAyB,CAAC,KAAuB,EAAE,MAAc,EAAE,OAAqB;IACrG,MAAM,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC;IAC9E,MAAM,MAAM,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC;IAEtC,IAAI,OAAO,CAAC,SAAS,IAAI,OAAO,CAAC,YAAY,EAAE,CAAC;QAC9C,MAAM,QAAQ,GAAG,MAAM,oBAAoB,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,YAAY,EAAE,KAAK,CAAC,CAAC;QACxF,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;YAC7B,IAAI;YACJ,IAAI,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC,IAAI;YACtB,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC;SACf,CAAC,CAAC,CAAC;IACN,CAAC;IAED,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;QAC7B,IAAI;QACJ,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM;QACrB,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC;KACf,CAAC,CAAC,CAAC;AACN,CAAC;AAED,SAAS,eAAe,CAAC,IAAoB,EAAE,IAAkB;IAC/D,MAAM,cAAc,GAAG;QACrB,mBAAmB;QACnB,kBAAkB;QAClB,uBAAuB;QACvB,oBAAoB;QACpB,uBAAuB;QACvB,mBAAmB;QACnB,kBAAkB;KACnB,CAAC;IAEF,OAAO,cAAc,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC5C,CAAC;AAED,SAAS,sBAAsB,CAAC,KAAuB,EAAE,MAAc,EAAE,YAA4B,EAAE,IAAkB;IACvH,MAAM,MAAM,GAAoB,EAAE,CAAC;IACnC,IAAI,YAAY,GAAkB;QAChC,IAAI,EAAE,cAAc;QACpB,KAAK,EAAE,EAAE;QACT,QAAQ,EAAE,EAAE;QACZ,UAAU,EAAE,IAAI;KACjB,CAAC;IAEF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACtB,MAAM,QAAQ,GAAG,YAAY,CAAC,CAAC,CAAC,CAAC;QAEjC,IAAI,eAAe,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,CAAC;YAChC,IAAI,YAAY,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAClC,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;YAC5B,CAAC;YAED,YAAY,GAAG;gBACb,IAAI,EAAE,WAAW;gBACjB,aAAa,EAAE,IAAI,CAAC,IAAI;gBACxB,KAAK,EAAE,CAAC,IAAI,CAAC;gBACb,QAAQ,EAAE,CAAC,QAAQ,CAAC;gBACpB,UAAU,EAAE,IAAI;aACjB,CAAC;YAEF,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;YAE1B,YAAY,GAAG;gBACb,IAAI,EAAE,cAAc;gBACpB,KAAK,EAAE,EAAE;gBACT,QAAQ,EAAE,EAAE;gBACZ,UAAU,EAAE,IAAI;aACjB,CAAC;QACJ,CAAC;aAAM,CAAC;YACN,YAAY,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC9B,YAAY,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QACvC,CAAC;IACH,CAAC;IAED,IAAI,YAAY,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClC,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;IAC5B,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,KAAK,UAAU,0BAA0B,CACvC,cAA+B,EAC/B,MAAc,EACd,OAAqB,EACrB,MAAwC;IAExC,MAAM,aAAa,GAAgB,EAAE,CAAC;IACtC,IAAI,oBAAoB,GAAc;QACpC,KAAK,EAAE,EAAE;QACT,SAAS,EAAE,CAAC;QACZ,SAAS,EAAE,EAAE;KACd,CAAC;IAEF,KAAK,MAAM,KAAK,IAAI,cAAc,EAAE,CAAC;QACnC,MAAM,cAAc,GAAG,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;QAE1E,IAAI,cAAc,GAAG,MAAM,CAAC,OAAO,EAAE,CAAC;YACpC,IAAI,oBAAoB,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC1C,aAAa,CAAC,IAAI,CAAC,oBAAoB,CAAC,CAAC;YAC3C,CAAC;YAED,aAAa,CAAC,IAAI,CAAC;gBACjB,KAAK,EAAE,KAAK,CAAC,KAAK;gBAClB,SAAS,EAAE,cAAc;gBACzB,SAAS,EAAE,CAAC,KAAK,CAAC;aACnB,CAAC,CAAC;YAEH,oBAAoB,GAAG;gBACrB,KAAK,EAAE,EAAE;gBACT,SAAS,EAAE,CAAC;gBACZ,SAAS,EAAE,EAAE;aACd,CAAC;YACF,SAAS;QACX,CAAC;QAED,IAAI,oBAAoB,CAAC,SAAS,GAAG,cAAc,GAAG,MAAM,CAAC,GAAG,EAAE,CAAC;YACjE,IAAI,oBAAoB,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC1C,aAAa,CAAC,IAAI,CAAC,oBAAoB,CAAC,CAAC;YAC3C,CAAC;YAED,oBAAoB,GAAG;gBACrB,KAAK,EAAE,KAAK,CAAC,KAAK;gBAClB,SAAS,EAAE,cAAc;gBACzB,SAAS,EAAE,CAAC,KAAK,CAAC;aACnB,CAAC;YACF,SAAS;QACX,CAAC;QAED,oBAAoB,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC;QAChD,oBAAoB,CAAC,SAAS,IAAI,cAAc,CAAC;QACjD,oBAAoB,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAE3C,IAAI,oBAAoB,CAAC,SAAS,IAAI,MAAM,CAAC,OAAO,GAAG,GAAG,EAAE,CAAC;YAC3D,aAAa,CAAC,IAAI,CAAC,oBAAoB,CAAC,CAAC;YACzC,oBAAoB,GAAG;gBACrB,KAAK,EAAE,EAAE;gBACT,SAAS,EAAE,CAAC;gBACZ,SAAS,EAAE,EAAE;aACd,CAAC;QACJ,CAAC;IACH,CAAC;IAED,IAAI,oBAAoB,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC1C,aAAa,CAAC,IAAI,CAAC,oBAAoB,CAAC,CAAC;IAC3C,CAAC;IAED,OAAO,aAAa,CAAC;AACvB,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,qBAAqB,CACzC,KAAuB,EACvB,MAAc,EACd,OAAqB,EACrB,IAAkB;IAElB,IAAI,CAAC,KAAK,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAE5C,MAAM,MAAM,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC;IAEtC,IAAI,KAAK,CAAC,MAAM,IAAI,EAAE,EAAE,CAAC;QACvB,OAAO,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACxB,KAAK,EAAE,CAAC,IAAI,CAAC;YACb,SAAS,EAAE,CAAC;YACZ,SAAS,EAAE,EAAE;SACd,CAAC,CAAC,CAAC;IACN,CAAC;IAED,MAAM,YAAY,GAAG,MAAM,yBAAyB,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC;IAC7E,MAAM,cAAc,GAAG,sBAAsB,CAAC,KAAK,EAAE,MAAM,EAAE,YAAY,EAAE,IAAI,CAAC,CAAC;IACjF,MAAM,aAAa,GAAG,MAAM,0BAA0B,CAAC,cAAc,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;IAEhG,OAAO,aAAa,CAAC;AACvB,CAAC;AAaD,MAAM,UAAU,mBAAmB,CAAC,SAAoB,EAAE,MAAc;IACtE,IAAI,CAAC,SAAS,CAAC,KAAK,IAAI,SAAS,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACrD,OAAO,IAAI,CAAC;IACd,CAAC;IAED,MAAM,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CACvC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,QAAQ,CAAC,CAC7C,CAAC;IAEF,MAAM,YAAY,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAExC,MAAM,SAAS,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACrC,MAAM,QAAQ,GAAG,SAAS,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAE7D,OAAO;QACL,IAAI,EAAE,YAAY;QAClB,IAAI,EAAE;YACJ,GAAG,SAAS;YACZ,IAAI,EAAE,GAAG,SAAS,CAAC,IAAI,UAAU,SAAS,CAAC,KAAK,CAAC,MAAM,EAAE;YACzD,QAAQ,EAAE,QAAQ,CAAC,QAAQ;SAC5B;QACD,QAAQ,EAAE;YACR,OAAO,EAAE,IAAI;YACb,SAAS,EAAE,SAAS,CAAC,KAAK,CAAC,MAAM;YACjC,SAAS,EAAE,SAAS,CAAC,SAAS;YAC9B,UAAU,EAAE,SAAS,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC;SAClE;KACF,CAAC;AACJ,CAAC"}
@@ -0,0 +1,37 @@
1
+ import type { ModelProfile } from '../providers/base.js';
2
+ import type { TreeSitterNode } from '../types/ast.js';
3
+ interface LanguageRule {
4
+ subdivisionTypes?: Record<string, string[]>;
5
+ [key: string]: any;
6
+ }
7
+ export declare function findSemanticSubdivisions(node: TreeSitterNode, rule: LanguageRule): TreeSitterNode[];
8
+ export declare function findLastCompleteBoundary(code: string, maxSize: number): number;
9
+ export declare function extractSignature(node: TreeSitterNode, source: string): string;
10
+ export declare function extractLinesBeforeNode(node: TreeSitterNode, source: string, numLines: number): string;
11
+ export declare function extractParentContext(node: TreeSitterNode, source: string): {
12
+ signature: string;
13
+ startLine: number;
14
+ endLine: number;
15
+ };
16
+ export declare function getLineNumber(byteOffset: number, source: string): number;
17
+ export interface NodeAnalysis {
18
+ isSingleChunk: boolean;
19
+ needsSubdivision: boolean;
20
+ subdivisionCandidates: TreeSitterNode[];
21
+ size: number;
22
+ unit: string;
23
+ method: string;
24
+ estimatedSubchunks: number;
25
+ }
26
+ export declare function analyzeNodeForChunking(node: TreeSitterNode, source: string, rule: LanguageRule, profile: ModelProfile): Promise<NodeAnalysis>;
27
+ export declare function batchAnalyzeNodes(nodes: TreeSitterNode[], source: string, rule: LanguageRule, profile: ModelProfile, isSubdivision?: boolean): Promise<Array<NodeAnalysis & {
28
+ node: TreeSitterNode;
29
+ }>>;
30
+ export interface StatementChunk {
31
+ code: string;
32
+ size: number;
33
+ unit: string;
34
+ }
35
+ export declare function yieldStatementChunks(node: TreeSitterNode, source: string, maxSize: number, overlapSize: number, profile: ModelProfile): Promise<StatementChunk[]>;
36
+ export {};
37
+ //# sourceMappingURL=semantic-chunker.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"semantic-chunker.d.ts","sourceRoot":"","sources":["../../src/chunking/semantic-chunker.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAEzD,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AAEtD,UAAU,YAAY;IACpB,gBAAgB,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,EAAE,CAAC,CAAC;IAC5C,CAAC,GAAG,EAAE,MAAM,GAAG,GAAG,CAAC;CACpB;AAED,wBAAgB,wBAAwB,CAAC,IAAI,EAAE,cAAc,EAAE,IAAI,EAAE,YAAY,GAAG,cAAc,EAAE,CAwBnG;AAED,wBAAgB,wBAAwB,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,GAAG,MAAM,CAgB9E;AAED,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,cAAc,EAAE,MAAM,EAAE,MAAM,GAAG,MAAM,CAO7E;AAED,wBAAgB,sBAAsB,CAAC,IAAI,EAAE,cAAc,EAAE,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,MAAM,CAIrG;AAED,wBAAgB,oBAAoB,CAAC,IAAI,EAAE,cAAc,EAAE,MAAM,EAAE,MAAM,GAAG;IAC1E,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;CACjB,CAMA;AAED,wBAAgB,aAAa,CAAC,UAAU,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,MAAM,CAGxE;AAED,MAAM,WAAW,YAAY;IAC3B,aAAa,EAAE,OAAO,CAAC;IACvB,gBAAgB,EAAE,OAAO,CAAC;IAC1B,qBAAqB,EAAE,cAAc,EAAE,CAAC;IACxC,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,EAAE,MAAM,CAAC;IACf,kBAAkB,EAAE,MAAM,CAAC;CAC5B;AAED,wBAAsB,sBAAsB,CAC1C,IAAI,EAAE,cAAc,EACpB,MAAM,EAAE,MAAM,EACd,IAAI,EAAE,YAAY,EAClB,OAAO,EAAE,YAAY,GACpB,OAAO,CAAC,YAAY,CAAC,CA2BvB;AAED,wBAAsB,iBAAiB,CACrC,KAAK,EAAE,cAAc,EAAE,EACvB,MAAM,EAAE,MAAM,EACd,IAAI,EAAE,YAAY,EAClB,OAAO,EAAE,YAAY,EACrB,aAAa,UAAQ,GACpB,OAAO,CAAC,KAAK,CAAC,YAAY,GAAG;IAAE,IAAI,EAAE,cAAc,CAAA;CAAE,CAAC,CAAC,CAgCzD;AAED,MAAM,WAAW,cAAc;IAC7B,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;CACd;AAED,wBAAsB,oBAAoB,CACxC,IAAI,EAAE,cAAc,EACpB,MAAM,EAAE,MAAM,EACd,OAAO,EAAE,MAAM,EACf,WAAW,EAAE,MAAM,EACnB,OAAO,EAAE,YAAY,GACpB,OAAO,CAAC,cAAc,EAAE,CAAC,CA0C3B"}
@@ -0,0 +1,157 @@
1
+ import { analyzeCodeSize, batchAnalyzeCodeSize } from './token-counter.js';
2
+ import { getSizeLimits } from '../providers/base.js';
3
+ export function findSemanticSubdivisions(node, rule) {
4
+ if (!node || !rule)
5
+ return [];
6
+ const subdivisionTypes = rule.subdivisionTypes?.[node.type] || [];
7
+ if (subdivisionTypes.length === 0)
8
+ return [];
9
+ const candidates = [];
10
+ function walk(n, depth = 0) {
11
+ if (depth > 0 && subdivisionTypes.includes(n.type)) {
12
+ candidates.push(n);
13
+ return;
14
+ }
15
+ for (let i = 0; i < n.childCount; i++) {
16
+ const child = n.child(i);
17
+ if (child) {
18
+ walk(child, depth + 1);
19
+ }
20
+ }
21
+ }
22
+ walk(node);
23
+ return candidates;
24
+ }
25
+ export function findLastCompleteBoundary(code, maxSize) {
26
+ const boundaries = [
27
+ { pattern: /\n\s*}\s*$/gm, priority: 1 },
28
+ { pattern: /;\s*$/gm, priority: 2 },
29
+ { pattern: /\n\s*$/gm, priority: 3 }
30
+ ];
31
+ for (const boundary of boundaries) {
32
+ const matches = [...code.substring(0, maxSize).matchAll(boundary.pattern)];
33
+ if (matches.length > 0) {
34
+ const lastMatch = matches[matches.length - 1];
35
+ return lastMatch.index + lastMatch[0].length;
36
+ }
37
+ }
38
+ return maxSize;
39
+ }
40
+ export function extractSignature(node, source) {
41
+ const code = source.slice(node.startIndex, node.endIndex);
42
+ const firstBrace = code.indexOf('{');
43
+ if (firstBrace !== -1) {
44
+ return code.substring(0, firstBrace).trim() + ' {';
45
+ }
46
+ return code.split('\n')[0];
47
+ }
48
+ export function extractLinesBeforeNode(node, source, numLines) {
49
+ const beforeCode = source.substring(0, node.startIndex);
50
+ const lines = beforeCode.split('\n');
51
+ return lines.slice(-numLines).join('\n');
52
+ }
53
+ export function extractParentContext(node, source) {
54
+ return {
55
+ signature: extractSignature(node, source),
56
+ startLine: getLineNumber(node.startIndex, source),
57
+ endLine: getLineNumber(node.endIndex, source)
58
+ };
59
+ }
60
+ export function getLineNumber(byteOffset, source) {
61
+ const before = source.substring(0, byteOffset);
62
+ return before.split('\n').length;
63
+ }
64
+ export async function analyzeNodeForChunking(node, source, rule, profile) {
65
+ const code = source.slice(node.startIndex, node.endIndex);
66
+ const limits = getSizeLimits(profile);
67
+ let actualSize;
68
+ let method;
69
+ if (profile.useTokens && profile.tokenCounter) {
70
+ const analysis = await analyzeCodeSize(code, limits, profile.tokenCounter);
71
+ actualSize = analysis.size;
72
+ method = analysis.method;
73
+ }
74
+ else {
75
+ actualSize = code.length;
76
+ method = 'chars';
77
+ }
78
+ const subdivisionThreshold = limits.max;
79
+ return {
80
+ isSingleChunk: actualSize <= subdivisionThreshold,
81
+ needsSubdivision: actualSize > subdivisionThreshold,
82
+ subdivisionCandidates: findSemanticSubdivisions(node, rule),
83
+ size: actualSize,
84
+ unit: limits.unit,
85
+ method,
86
+ estimatedSubchunks: Math.ceil(actualSize / limits.optimal)
87
+ };
88
+ }
89
+ export async function batchAnalyzeNodes(nodes, source, rule, profile, isSubdivision = false) {
90
+ const codes = nodes.map(node => source.slice(node.startIndex, node.endIndex));
91
+ const limits = getSizeLimits(profile);
92
+ let analyses;
93
+ if (profile.useTokens && profile.tokenCounter) {
94
+ analyses = await batchAnalyzeCodeSize(codes, limits, profile.tokenCounter, isSubdivision);
95
+ }
96
+ else {
97
+ analyses = codes.map(code => ({
98
+ size: code.length,
99
+ decision: code.length < limits.min ? 'too_small'
100
+ : code.length > limits.max ? 'too_large'
101
+ : code.length <= limits.optimal ? 'optimal'
102
+ : 'needs_tokenization',
103
+ method: 'chars'
104
+ }));
105
+ }
106
+ return nodes.map((node, i) => {
107
+ const analysis = analyses[i];
108
+ const subdivisionThreshold = limits.max;
109
+ return {
110
+ node,
111
+ isSingleChunk: analysis.size <= subdivisionThreshold,
112
+ needsSubdivision: analysis.size > subdivisionThreshold,
113
+ subdivisionCandidates: findSemanticSubdivisions(node, rule),
114
+ size: analysis.size,
115
+ unit: limits.unit,
116
+ method: analysis.method,
117
+ estimatedSubchunks: Math.ceil(analysis.size / limits.optimal)
118
+ };
119
+ });
120
+ }
121
+ export async function yieldStatementChunks(node, source, maxSize, overlapSize, profile) {
122
+ const code = source.slice(node.startIndex, node.endIndex);
123
+ const lines = code.split('\n');
124
+ const chunks = [];
125
+ let currentChunk = [];
126
+ let currentSize = 0;
127
+ for (const line of lines) {
128
+ const lineSize = profile.useTokens && profile.tokenCounter
129
+ ? await profile.tokenCounter(line)
130
+ : line.length;
131
+ if (currentSize + lineSize > maxSize && currentChunk.length > 0) {
132
+ chunks.push({
133
+ code: currentChunk.join('\n'),
134
+ size: currentSize,
135
+ unit: profile.useTokens ? 'tokens' : 'characters'
136
+ });
137
+ // Use the provided overlapSize parameter instead of hardcoded 20%
138
+ const overlapRatio = Math.min(1, Math.max(0, overlapSize / maxSize));
139
+ const overlapLines = Math.max(1, Math.floor(currentChunk.length * overlapRatio));
140
+ currentChunk = currentChunk.slice(-overlapLines);
141
+ currentSize = profile.useTokens && profile.tokenCounter
142
+ ? await profile.tokenCounter(currentChunk.join('\n'))
143
+ : currentChunk.join('\n').length;
144
+ }
145
+ currentChunk.push(line);
146
+ currentSize += lineSize;
147
+ }
148
+ if (currentChunk.length > 0) {
149
+ chunks.push({
150
+ code: currentChunk.join('\n'),
151
+ size: currentSize,
152
+ unit: profile.useTokens ? 'tokens' : 'characters'
153
+ });
154
+ }
155
+ return chunks;
156
+ }
157
+ //# sourceMappingURL=semantic-chunker.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"semantic-chunker.js","sourceRoot":"","sources":["../../src/chunking/semantic-chunker.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,eAAe,EAAE,oBAAoB,EAAyB,MAAM,oBAAoB,CAAC;AAElG,OAAO,EAAE,aAAa,EAAE,MAAM,sBAAsB,CAAC;AAQrD,MAAM,UAAU,wBAAwB,CAAC,IAAoB,EAAE,IAAkB;IAC/E,IAAI,CAAC,IAAI,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAC;IAE9B,MAAM,gBAAgB,GAAG,IAAI,CAAC,gBAAgB,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;IAClE,IAAI,gBAAgB,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAE7C,MAAM,UAAU,GAAqB,EAAE,CAAC;IAExC,SAAS,IAAI,CAAC,CAAiB,EAAE,KAAK,GAAG,CAAC;QACxC,IAAI,KAAK,GAAG,CAAC,IAAI,gBAAgB,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC;YACnD,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACnB,OAAO;QACT,CAAC;QAED,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,UAAU,EAAE,CAAC,EAAE,EAAE,CAAC;YACtC,MAAM,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;YACzB,IAAI,KAAK,EAAE,CAAC;gBACV,IAAI,CAAC,KAAK,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;YACzB,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,CAAC,IAAI,CAAC,CAAC;IACX,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,MAAM,UAAU,wBAAwB,CAAC,IAAY,EAAE,OAAe;IACpE,MAAM,UAAU,GAAG;QACjB,EAAE,OAAO,EAAE,cAAc,EAAE,QAAQ,EAAE,CAAC,EAAE;QACxC,EAAE,OAAO,EAAE,SAAS,EAAE,QAAQ,EAAE,CAAC,EAAE;QACnC,EAAE,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAE,CAAC,EAAE;KACrC,CAAC;IAEF,KAAK,MAAM,QAAQ,IAAI,UAAU,EAAE,CAAC;QAClC,MAAM,OAAO,GAAG,CAAC,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,QAAQ,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC;QAC3E,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACvB,MAAM,SAAS,GAAG,OAAO,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;YAC9C,OAAO,SAAS,CAAC,KAAM,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;QAChD,CAAC;IACH,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,IAAoB,EAAE,MAAc;IACnE,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,QAAQ,CAAC,CAAC;IAC1D,MAAM,UAAU,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;IACrC,IAAI,UAAU,KAAK,CAAC,CAAC,EAAE,CAAC;QACtB,OAAO,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC,IAAI,EAAE,GAAG,IAAI,CAAC;IACrD,CAAC;IACD,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;AAC7B,CAAC;AAED,MAAM,UAAU,sBAAsB,CAAC,IAAoB,EAAE,MAAc,EAAE,QAAgB;IAC3F,MAAM,UAAU,GAAG,MAAM,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC;IACxD,MAAM,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACrC,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC3C,CAAC;AAED,MAAM,UAAU,oBAAoB,CAAC,IAAoB,EAAE,MAAc;IAKvE,OAAO;QACL,SAAS,EAAE,gBAAgB,CAAC,IAAI,EAAE,MAAM,CAAC;QACzC,SAAS,EAAE,aAAa,CAAC,IAAI,CAAC,UAAU,EAAE,MAAM,CAAC;QACjD,OAAO,EAAE,aAAa,CAAC,IAAI,CAAC,QAAQ,EAAE,MAAM,CAAC;KAC9C,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,UAAkB,EAAE,MAAc;IAC9D,MAAM,MAAM,GAAG,MAAM,CAAC,SAAS,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC;IAC/C,OAAO,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC;AACnC,CAAC;AAYD,MAAM,CAAC,KAAK,UAAU,sBAAsB,CAC1C,IAAoB,EACpB,MAAc,EACd,IAAkB,EAClB,OAAqB;IAErB,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,QAAQ,CAAC,CAAC;IAC1D,MAAM,MAAM,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC;IAEtC,IAAI,UAAkB,CAAC;IACvB,IAAI,MAAc,CAAC;IAEnB,IAAI,OAAO,CAAC,SAAS,IAAI,OAAO,CAAC,YAAY,EAAE,CAAC;QAC9C,MAAM,QAAQ,GAAG,MAAM,eAAe,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,CAAC,YAAY,CAAC,CAAC;QAC3E,UAAU,GAAG,QAAQ,CAAC,IAAI,CAAC;QAC3B,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC;IAC3B,CAAC;SAAM,CAAC;QACN,UAAU,GAAG,IAAI,CAAC,MAAM,CAAC;QACzB,MAAM,GAAG,OAAO,CAAC;IACnB,CAAC;IAED,MAAM,oBAAoB,GAAG,MAAM,CAAC,GAAG,CAAC;IAExC,OAAO;QACL,aAAa,EAAE,UAAU,IAAI,oBAAoB;QACjD,gBAAgB,EAAE,UAAU,GAAG,oBAAoB;QACnD,qBAAqB,EAAE,wBAAwB,CAAC,IAAI,EAAE,IAAI,CAAC;QAC3D,IAAI,EAAE,UAAU;QAChB,IAAI,EAAE,MAAM,CAAC,IAAI;QACjB,MAAM;QACN,kBAAkB,EAAE,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,MAAM,CAAC,OAAO,CAAC;KAC3D,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,iBAAiB,CACrC,KAAuB,EACvB,MAAc,EACd,IAAkB,EAClB,OAAqB,EACrB,aAAa,GAAG,KAAK;IAErB,MAAM,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC;IAC9E,MAAM,MAAM,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC;IAEtC,IAAI,QAA4B,CAAC;IACjC,IAAI,OAAO,CAAC,SAAS,IAAI,OAAO,CAAC,YAAY,EAAE,CAAC;QAC9C,QAAQ,GAAG,MAAM,oBAAoB,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,YAAY,EAAE,aAAa,CAAC,CAAC;IAC5F,CAAC;SAAM,CAAC;QACN,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YAC5B,IAAI,EAAE,IAAI,CAAC,MAAM;YACjB,QAAQ,EAAE,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,WAAoB;gBACnD,CAAC,CAAC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,WAAoB;oBACjD,CAAC,CAAC,IAAI,CAAC,MAAM,IAAI,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,SAAkB;wBACpD,CAAC,CAAC,oBAA6B;YACrC,MAAM,EAAE,OAAO;SAChB,CAAC,CAAC,CAAC;IACN,CAAC;IAED,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,EAAE;QAC3B,MAAM,QAAQ,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;QAC7B,MAAM,oBAAoB,GAAG,MAAM,CAAC,GAAG,CAAC;QACxC,OAAO;YACL,IAAI;YACJ,aAAa,EAAE,QAAQ,CAAC,IAAI,IAAI,oBAAoB;YACpD,gBAAgB,EAAE,QAAQ,CAAC,IAAI,GAAG,oBAAoB;YACtD,qBAAqB,EAAE,wBAAwB,CAAC,IAAI,EAAE,IAAI,CAAC;YAC3D,IAAI,EAAE,QAAQ,CAAC,IAAI;YACnB,IAAI,EAAE,MAAM,CAAC,IAAI;YACjB,MAAM,EAAE,QAAQ,CAAC,MAAM;YACvB,kBAAkB,EAAE,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC;SAC9D,CAAC;IACJ,CAAC,CAAC,CAAC;AACL,CAAC;AAQD,MAAM,CAAC,KAAK,UAAU,oBAAoB,CACxC,IAAoB,EACpB,MAAc,EACd,OAAe,EACf,WAAmB,EACnB,OAAqB;IAErB,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,QAAQ,CAAC,CAAC;IAC1D,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAE/B,MAAM,MAAM,GAAqB,EAAE,CAAC;IACpC,IAAI,YAAY,GAAa,EAAE,CAAC;IAChC,IAAI,WAAW,GAAG,CAAC,CAAC;IAEpB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,QAAQ,GAAG,OAAO,CAAC,SAAS,IAAI,OAAO,CAAC,YAAY;YACxD,CAAC,CAAC,MAAM,OAAO,CAAC,YAAY,CAAC,IAAI,CAAC;YAClC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC;QAEhB,IAAI,WAAW,GAAG,QAAQ,GAAG,OAAO,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAChE,MAAM,CAAC,IAAI,CAAC;gBACV,IAAI,EAAE,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC;gBAC7B,IAAI,EAAE,WAAW;gBACjB,IAAI,EAAE,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,YAAY;aAClD,CAAC,CAAC;YAEH,kEAAkE;YAClE,MAAM,YAAY,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,WAAW,GAAG,OAAO,CAAC,CAAC,CAAC;YACrE,MAAM,YAAY,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,MAAM,GAAG,YAAY,CAAC,CAAC,CAAC;YACjF,YAAY,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC,YAAY,CAAC,CAAC;YACjD,WAAW,GAAG,OAAO,CAAC,SAAS,IAAI,OAAO,CAAC,YAAY;gBACrD,CAAC,CAAC,MAAM,OAAO,CAAC,YAAY,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACrD,CAAC,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC;QACrC,CAAC;QAED,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACxB,WAAW,IAAI,QAAQ,CAAC;IAC1B,CAAC;IAED,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,MAAM,CAAC,IAAI,CAAC;YACV,IAAI,EAAE,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC;YAC7B,IAAI,EAAE,WAAW;YACjB,IAAI,EAAE,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,YAAY;SAClD,CAAC,CAAC;IACL,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}
@@ -0,0 +1,28 @@
1
+ interface TokenCountStats {
2
+ totalRequests: number;
3
+ cacheHits: number;
4
+ charFilterSkips: number;
5
+ actualTokenizations: number;
6
+ batchTokenizations: number;
7
+ cacheHitRate: string;
8
+ charFilterRate: string;
9
+ tokenizationRate: string;
10
+ }
11
+ interface SizeLimits {
12
+ min: number;
13
+ max: number;
14
+ optimal: number;
15
+ }
16
+ type Decision = 'too_small' | 'too_large' | 'needs_tokenization' | 'optimal';
17
+ export interface CodeSizeAnalysis {
18
+ size: number;
19
+ decision: Decision;
20
+ method: string;
21
+ }
22
+ export declare function analyzeCodeSize(code: string, limits: SizeLimits, tokenCounter: (text: string) => number | Promise<number>, allowEstimateForSkip?: boolean): Promise<CodeSizeAnalysis>;
23
+ export declare function batchAnalyzeCodeSize(codeSnippets: string[], limits: SizeLimits, tokenCounter: (text: string) => number | Promise<number>, allowEstimateForSkip?: boolean): Promise<CodeSizeAnalysis[]>;
24
+ export declare function getTokenCountStats(): TokenCountStats;
25
+ export declare function resetTokenCountStats(): void;
26
+ export declare function clearTokenCache(): void;
27
+ export {};
28
+ //# sourceMappingURL=token-counter.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"token-counter.d.ts","sourceRoot":"","sources":["../../src/chunking/token-counter.ts"],"names":[],"mappings":"AAqCA,UAAU,eAAe;IACvB,aAAa,EAAE,MAAM,CAAC;IACtB,SAAS,EAAE,MAAM,CAAC;IAClB,eAAe,EAAE,MAAM,CAAC;IACxB,mBAAmB,EAAE,MAAM,CAAC;IAC5B,kBAAkB,EAAE,MAAM,CAAC;IAC3B,YAAY,EAAE,MAAM,CAAC;IACrB,cAAc,EAAE,MAAM,CAAC;IACvB,gBAAgB,EAAE,MAAM,CAAC;CAC1B;AAeD,UAAU,UAAU;IAClB,GAAG,EAAE,MAAM,CAAC;IACZ,GAAG,EAAE,MAAM,CAAC;IACZ,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,KAAK,QAAQ,GAAG,WAAW,GAAG,WAAW,GAAG,oBAAoB,GAAG,SAAS,CAAC;AAwF7E,MAAM,WAAW,gBAAgB;IAC/B,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,EAAE,QAAQ,CAAC;IACnB,MAAM,EAAE,MAAM,CAAC;CAChB;AAED,wBAAsB,eAAe,CACnC,IAAI,EAAE,MAAM,EACZ,MAAM,EAAE,UAAU,EAClB,YAAY,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,EACxD,oBAAoB,UAAQ,GAC3B,OAAO,CAAC,gBAAgB,CAAC,CAgC3B;AAED,wBAAsB,oBAAoB,CACxC,YAAY,EAAE,MAAM,EAAE,EACtB,MAAM,EAAE,UAAU,EAClB,YAAY,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,EACxD,oBAAoB,UAAQ,GAC3B,OAAO,CAAC,gBAAgB,EAAE,CAAC,CAiD7B;AAED,wBAAgB,kBAAkB,IAAI,eAAe,CAapD;AAED,wBAAgB,oBAAoB,IAAI,IAAI,CAM3C;AAED,wBAAgB,eAAe,IAAI,IAAI,CAEtC"}
@@ -0,0 +1,207 @@
1
+ class LRUCache {
2
+ maxSize;
3
+ cache = new Map();
4
+ constructor(maxSize = 1000) {
5
+ this.maxSize = maxSize;
6
+ }
7
+ get(key) {
8
+ if (!this.cache.has(key))
9
+ return undefined;
10
+ const value = this.cache.get(key);
11
+ this.cache.delete(key);
12
+ this.cache.set(key, value);
13
+ return value;
14
+ }
15
+ set(key, value) {
16
+ if (this.cache.has(key)) {
17
+ this.cache.delete(key);
18
+ }
19
+ this.cache.set(key, value);
20
+ if (this.cache.size > this.maxSize) {
21
+ const firstKey = this.cache.keys().next().value;
22
+ if (firstKey !== undefined) {
23
+ this.cache.delete(firstKey);
24
+ }
25
+ }
26
+ }
27
+ clear() {
28
+ this.cache.clear();
29
+ }
30
+ }
31
+ const tokenCountCache = new LRUCache(1000);
32
+ const stats = {
33
+ totalRequests: 0,
34
+ cacheHits: 0,
35
+ charFilterSkips: 0,
36
+ actualTokenizations: 0,
37
+ batchTokenizations: 0
38
+ };
39
+ function estimateTokensFromChars(charCount) {
40
+ return Math.ceil(charCount / 4);
41
+ }
42
+ function preFilterByChars(code, limits) {
43
+ const charCount = code.length;
44
+ const estimatedTokens = estimateTokensFromChars(charCount);
45
+ const minEstimate = limits.min * 0.8;
46
+ const maxEstimate = limits.max * 1.2;
47
+ const optimalLow = limits.optimal * 0.8;
48
+ const optimalHigh = limits.optimal * 1.2;
49
+ if (estimatedTokens < minEstimate) {
50
+ return { decision: 'too_small', estimate: estimatedTokens };
51
+ }
52
+ if (estimatedTokens > maxEstimate) {
53
+ return { decision: 'too_large', estimate: estimatedTokens };
54
+ }
55
+ if (estimatedTokens >= optimalLow && estimatedTokens <= optimalHigh) {
56
+ return { decision: 'optimal', estimate: estimatedTokens };
57
+ }
58
+ return { decision: 'needs_tokenization', estimate: estimatedTokens };
59
+ }
60
+ async function countTokensWithCache(code, tokenCounter) {
61
+ stats.totalRequests++;
62
+ const cached = tokenCountCache.get(code);
63
+ if (cached !== undefined) {
64
+ stats.cacheHits++;
65
+ return cached;
66
+ }
67
+ stats.actualTokenizations++;
68
+ const result = tokenCounter(code);
69
+ const count = result instanceof Promise ? await result : result;
70
+ tokenCountCache.set(code, count);
71
+ return count;
72
+ }
73
+ async function batchCountTokens(codeSnippets, tokenCounter) {
74
+ stats.batchTokenizations++;
75
+ const results = [];
76
+ const uncached = [];
77
+ const uncachedIndices = [];
78
+ for (let i = 0; i < codeSnippets.length; i++) {
79
+ const code = codeSnippets[i];
80
+ const cached = tokenCountCache.get(code);
81
+ if (cached !== undefined) {
82
+ stats.cacheHits++;
83
+ results[i] = cached;
84
+ }
85
+ else {
86
+ uncached.push(code);
87
+ uncachedIndices.push(i);
88
+ }
89
+ }
90
+ if (uncached.length > 0) {
91
+ stats.actualTokenizations += uncached.length;
92
+ const counts = await Promise.all(uncached.map(async (code) => {
93
+ const result = tokenCounter(code);
94
+ return result instanceof Promise ? await result : result;
95
+ }));
96
+ for (let i = 0; i < counts.length; i++) {
97
+ const code = uncached[i];
98
+ const count = counts[i];
99
+ tokenCountCache.set(code, count);
100
+ results[uncachedIndices[i]] = count;
101
+ }
102
+ }
103
+ return results;
104
+ }
105
+ export async function analyzeCodeSize(code, limits, tokenCounter, allowEstimateForSkip = false) {
106
+ stats.totalRequests++;
107
+ const preFilter = preFilterByChars(code, limits);
108
+ if (allowEstimateForSkip && preFilter.decision === 'too_large') {
109
+ stats.charFilterSkips++;
110
+ return {
111
+ size: preFilter.estimate,
112
+ decision: preFilter.decision,
113
+ method: 'char_estimate'
114
+ };
115
+ }
116
+ const actualSize = await countTokensWithCache(code, tokenCounter);
117
+ let decision;
118
+ if (actualSize < limits.min) {
119
+ decision = 'too_small';
120
+ }
121
+ else if (actualSize > limits.max) {
122
+ decision = 'too_large';
123
+ }
124
+ else if (actualSize <= limits.optimal) {
125
+ decision = 'optimal';
126
+ }
127
+ else {
128
+ decision = 'needs_tokenization';
129
+ }
130
+ return {
131
+ size: actualSize,
132
+ decision,
133
+ method: 'tokenized'
134
+ };
135
+ }
136
+ export async function batchAnalyzeCodeSize(codeSnippets, limits, tokenCounter, allowEstimateForSkip = false) {
137
+ const results = [];
138
+ const needsTokenization = [];
139
+ const needsTokenizationIndices = [];
140
+ for (let i = 0; i < codeSnippets.length; i++) {
141
+ const code = codeSnippets[i];
142
+ const preFilter = preFilterByChars(code, limits);
143
+ if (allowEstimateForSkip && preFilter.decision === 'too_large') {
144
+ stats.charFilterSkips++;
145
+ results[i] = {
146
+ size: preFilter.estimate,
147
+ decision: preFilter.decision,
148
+ method: 'char_estimate'
149
+ };
150
+ }
151
+ else {
152
+ needsTokenization.push(code);
153
+ needsTokenizationIndices.push(i);
154
+ }
155
+ }
156
+ if (needsTokenization.length > 0) {
157
+ const tokenCounts = await batchCountTokens(needsTokenization, tokenCounter);
158
+ for (let i = 0; i < tokenCounts.length; i++) {
159
+ const actualSize = tokenCounts[i];
160
+ const idx = needsTokenizationIndices[i];
161
+ let decision;
162
+ if (actualSize < limits.min) {
163
+ decision = 'too_small';
164
+ }
165
+ else if (actualSize > limits.max) {
166
+ decision = 'too_large';
167
+ }
168
+ else if (actualSize <= limits.optimal) {
169
+ decision = 'optimal';
170
+ }
171
+ else {
172
+ decision = 'needs_tokenization';
173
+ }
174
+ results[idx] = {
175
+ size: actualSize,
176
+ decision,
177
+ method: 'tokenized'
178
+ };
179
+ }
180
+ }
181
+ return results;
182
+ }
183
+ export function getTokenCountStats() {
184
+ return {
185
+ ...stats,
186
+ cacheHitRate: stats.totalRequests > 0
187
+ ? (stats.cacheHits / stats.totalRequests * 100).toFixed(1) + '%'
188
+ : '0%',
189
+ charFilterRate: stats.totalRequests > 0
190
+ ? (stats.charFilterSkips / stats.totalRequests * 100).toFixed(1) + '%'
191
+ : '0%',
192
+ tokenizationRate: stats.totalRequests > 0
193
+ ? (stats.actualTokenizations / stats.totalRequests * 100).toFixed(1) + '%'
194
+ : '0%'
195
+ };
196
+ }
197
+ export function resetTokenCountStats() {
198
+ stats.totalRequests = 0;
199
+ stats.cacheHits = 0;
200
+ stats.charFilterSkips = 0;
201
+ stats.actualTokenizations = 0;
202
+ stats.batchTokenizations = 0;
203
+ }
204
+ export function clearTokenCache() {
205
+ tokenCountCache.clear();
206
+ }
207
+ //# sourceMappingURL=token-counter.js.map