@thanh01.pmt/research-kit 1.0.0 → 1.1.1

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 (78) hide show
  1. package/LICENSE +21 -0
  2. package/README.md +213 -0
  3. package/dist/academic/index.cjs +332 -127
  4. package/dist/academic/index.cjs.map +1 -1
  5. package/dist/academic/index.d.cts +3 -3
  6. package/dist/academic/index.d.ts +3 -3
  7. package/dist/academic/index.js +2 -2
  8. package/dist/{catalogBuilder-8A-Dg_A-.d.ts → catalogBuilder-CmmQzUnf.d.ts} +11 -3
  9. package/dist/{catalogBuilder-feU79wQQ.d.cts → catalogBuilder-DgXYxL2F.d.cts} +11 -3
  10. package/dist/chunk-2JEVBCIF.js +3 -0
  11. package/dist/{chunk-KGCDYFFD.js.map → chunk-2JEVBCIF.js.map} +1 -1
  12. package/dist/chunk-2XIEWCUE.js +244 -0
  13. package/dist/chunk-2XIEWCUE.js.map +1 -0
  14. package/dist/chunk-AUGVVZTL.js +514 -0
  15. package/dist/chunk-AUGVVZTL.js.map +1 -0
  16. package/dist/chunk-DOXSPRSP.js +559 -0
  17. package/dist/chunk-DOXSPRSP.js.map +1 -0
  18. package/dist/chunk-GKF3NDEX.js +162 -0
  19. package/dist/chunk-GKF3NDEX.js.map +1 -0
  20. package/dist/chunk-J3VCGH7Z.js +3 -0
  21. package/dist/chunk-J3VCGH7Z.js.map +1 -0
  22. package/dist/{chunk-67V2GJGK.js → chunk-J7XKDMGL.js} +7 -5
  23. package/dist/chunk-J7XKDMGL.js.map +1 -0
  24. package/dist/chunk-KAFQVI2R.js +333 -0
  25. package/dist/chunk-KAFQVI2R.js.map +1 -0
  26. package/dist/chunk-XIF5YNI7.js +799 -0
  27. package/dist/chunk-XIF5YNI7.js.map +1 -0
  28. package/dist/config/index.cjs +5 -3
  29. package/dist/config/index.cjs.map +1 -1
  30. package/dist/config/index.d.cts +1 -1
  31. package/dist/config/index.d.ts +1 -1
  32. package/dist/config/index.js +1 -1
  33. package/dist/evidence/index.cjs +337 -21
  34. package/dist/evidence/index.cjs.map +1 -1
  35. package/dist/evidence/index.d.cts +44 -7
  36. package/dist/evidence/index.d.ts +44 -7
  37. package/dist/evidence/index.js +3 -2
  38. package/dist/harvester/index.cjs +805 -64
  39. package/dist/harvester/index.cjs.map +1 -1
  40. package/dist/harvester/index.d.cts +79 -8
  41. package/dist/harvester/index.d.ts +79 -8
  42. package/dist/harvester/index.js +3 -2
  43. package/dist/{index-BU5n1ip_.d.ts → index-t8pahlJ4.d.cts} +104 -15
  44. package/dist/{index-BU5n1ip_.d.cts → index-t8pahlJ4.d.ts} +104 -15
  45. package/dist/index.cjs +2001 -252
  46. package/dist/index.cjs.map +1 -1
  47. package/dist/index.d.cts +156 -8
  48. package/dist/index.d.ts +156 -8
  49. package/dist/index.js +237 -8
  50. package/dist/index.js.map +1 -1
  51. package/dist/literatureReview-BTPFe7eb.d.cts +64 -0
  52. package/dist/literatureReview-hGbfrBmw.d.ts +64 -0
  53. package/dist/pipeline/index.cjs +1570 -241
  54. package/dist/pipeline/index.cjs.map +1 -1
  55. package/dist/pipeline/index.d.cts +49 -9
  56. package/dist/pipeline/index.d.ts +49 -9
  57. package/dist/pipeline/index.js +5 -4
  58. package/dist/webScraper-BhiQ8s9x.d.ts +68 -0
  59. package/dist/webScraper-D9Ldus9p.d.cts +68 -0
  60. package/package.json +8 -9
  61. package/dist/chunk-47L7ZHMY.js +0 -316
  62. package/dist/chunk-47L7ZHMY.js.map +0 -1
  63. package/dist/chunk-63X2HPCI.js +0 -47
  64. package/dist/chunk-63X2HPCI.js.map +0 -1
  65. package/dist/chunk-67V2GJGK.js.map +0 -1
  66. package/dist/chunk-6Q2XCUG7.js +0 -40
  67. package/dist/chunk-6Q2XCUG7.js.map +0 -1
  68. package/dist/chunk-KGCDYFFD.js +0 -3
  69. package/dist/chunk-MHAYD555.js +0 -239
  70. package/dist/chunk-MHAYD555.js.map +0 -1
  71. package/dist/chunk-MPLDSHFM.js +0 -396
  72. package/dist/chunk-MPLDSHFM.js.map +0 -1
  73. package/dist/chunk-QZNLB5X4.js +0 -65
  74. package/dist/chunk-QZNLB5X4.js.map +0 -1
  75. package/dist/literatureReview-C4AgVljZ.d.ts +0 -43
  76. package/dist/literatureReview-Dlods4Aw.d.cts +0 -43
  77. package/dist/webScraper-pKzdXhfI.d.ts +0 -46
  78. package/dist/webScraper-yBcYOHoL.d.cts +0 -46
@@ -1,7 +1,7 @@
1
- import { T as TechnicalDomain, S as ScrapedDocument } from '../webScraper-yBcYOHoL.cjs';
2
- import { e as AcademicPaper, d as ResearchConfig, a as CustomSearchAdapter, b as CustomCrawlerAdapter, c as CustomAcademicSearchAdapter } from '../index-BU5n1ip_.cjs';
3
- import { C as CS2023KnowledgeArea, L as LiteratureReviewReport } from '../literatureReview-Dlods4Aw.cjs';
4
- import { R as ResourceCatalogBundle } from '../catalogBuilder-feU79wQQ.cjs';
1
+ import { T as TechnicalDomain, S as ScrapedDocument } from '../webScraper-D9Ldus9p.cjs';
2
+ import { C as CourseResearchContext, L as LlmTextFn, A as AcademicPaper, E as ExtractedConceptNode, f as SynthesizedResearchInsights, R as ResearchApiKeys, h as ResearchConfig, a as CustomSearchAdapter, b as CustomCrawlerAdapter, c as CustomAcademicSearchAdapter, d as CustomSynthesizerAdapter } from '../index-t8pahlJ4.cjs';
3
+ import { C as CS2023KnowledgeArea, a as CSTAK12Strand, L as LiteratureReviewReport } from '../literatureReview-BTPFe7eb.cjs';
4
+ import { R as ResourceCatalogBundle } from '../catalogBuilder-DgXYxL2F.cjs';
5
5
  import { a as ProvenanceSidecar } from '../provenanceTracker-ct4s5lJV.cjs';
6
6
  import 'zod';
7
7
 
@@ -15,19 +15,30 @@ interface ResearchPlan {
15
15
  planMd: string;
16
16
  }
17
17
  /**
18
- * Creates and formats MASTER_RESEARCH_PLAN.md with Scale Decision (English Canonical)
18
+ * Creates and formats MASTER_RESEARCH_PLAN.md with Context-Aware Multi-Angle Queries
19
+ * (Technical Ground Truth + Pedagogical Exemplars + Learner Misconceptions/Pitfalls).
19
20
  */
20
21
  declare class CampaignPlanner {
21
22
  createPlan(topic: string, domain: TechnicalDomain, options?: {
22
23
  maxRounds?: number;
23
24
  isDeep?: boolean;
25
+ context?: CourseResearchContext;
24
26
  }): ResearchPlan;
27
+ /**
28
+ * Optional LLM-augmented query planner when the host app provides an `llm` callback.
29
+ */
30
+ createPlanWithLlm(topic: string, domain: TechnicalDomain, options?: {
31
+ maxRounds?: number;
32
+ isDeep?: boolean;
33
+ context?: CourseResearchContext;
34
+ llm?: LlmTextFn;
35
+ }): Promise<ResearchPlan>;
25
36
  }
26
37
 
27
38
  interface ReferenceRegistryItem {
28
39
  id: string;
29
40
  title: string;
30
- sourceType: 'official_docs' | 'academic_paper' | 'community_repo' | 'industry_standard';
41
+ sourceType: 'official_docs' | 'academic_paper' | 'video_transcript' | 'community_repo' | 'industry_standard';
31
42
  url: string;
32
43
  relevanceScore: number;
33
44
  extractedKeyTakeaways: string[];
@@ -37,17 +48,23 @@ interface ReferencePackOutput {
37
48
  referencePackMd: string;
38
49
  }
39
50
  /**
40
- * Compiles the canonical REFERENCE_PACK.md & REFERENCE_REGISTRY.json (English Canonical)
51
+ * Compiles the canonical REFERENCE_PACK.md & REFERENCE_REGISTRY.json with real extracted takeaways,
52
+ * accurate code fence languages, synthesized misconceptions/exemplars, and bilingual concept anchors.
41
53
  */
42
54
  declare class ReferencePackBuilder {
43
55
  buildReferencePack(topic: string, domain: TechnicalDomain, options?: {
44
56
  cs2023Area?: CS2023KnowledgeArea;
57
+ k12Strand?: CSTAK12Strand;
58
+ context?: CourseResearchContext;
45
59
  scrapedDocs?: {
46
60
  doc: ScrapedDocument;
47
61
  score: number;
48
62
  summary: string;
63
+ keyTakeaways?: string[];
49
64
  }[];
50
65
  academicPapers?: AcademicPaper[];
66
+ concepts?: ExtractedConceptNode[];
67
+ synthesizedInsights?: SynthesizedResearchInsights;
51
68
  }): ReferencePackOutput;
52
69
  }
53
70
 
@@ -55,13 +72,21 @@ interface HarvestPipelineOptions {
55
72
  domain?: string;
56
73
  maxRounds?: number;
57
74
  isDeep?: boolean;
75
+ crawl4aiHost?: string;
58
76
  tavilyKey?: string;
59
77
  exaKey?: string;
78
+ jinaKey?: string;
79
+ serperKey?: string;
60
80
  semanticScholarKey?: string;
81
+ apiKeys?: ResearchApiKeys;
82
+ allowEnvFallback?: boolean;
83
+ context?: CourseResearchContext;
84
+ llm?: LlmTextFn;
61
85
  config?: Partial<ResearchConfig>;
62
86
  customSearchAdapter?: CustomSearchAdapter;
63
87
  customCrawlerAdapter?: CustomCrawlerAdapter;
64
88
  customAcademicSearchAdapter?: CustomAcademicSearchAdapter;
89
+ customSynthesizerAdapter?: CustomSynthesizerAdapter;
65
90
  onProgress?: (stage: string, message: string) => void;
66
91
  }
67
92
  interface HarvestPipelineResult {
@@ -72,10 +97,14 @@ interface HarvestPipelineResult {
72
97
  literatureReview: LiteratureReviewReport;
73
98
  referencePack: ReferencePackOutput;
74
99
  provenance: ProvenanceSidecar;
100
+ concepts: ExtractedConceptNode[];
101
+ conceptDagTsv: string;
102
+ synthesizedInsights: SynthesizedResearchInsights;
75
103
  totalDocumentsScraped: number;
76
104
  }
105
+ declare function normalizeUrlForDedup(rawUrl: string): string;
77
106
  /**
78
- * Autonomous Multi-Round Deep Research & Academic Harvester Pipeline with Pluggable Tool Adapters
107
+ * Autonomous Multi-Round Deep Research & Academic Harvester Pipeline with Host API Key & LLM Injection
79
108
  */
80
109
  declare class HarvestPipeline {
81
110
  private searchEngine;
@@ -88,15 +117,26 @@ declare class HarvestPipeline {
88
117
  private litReviewSynthesizer;
89
118
  private planner;
90
119
  private refPackBuilder;
120
+ private termExtractor;
121
+ private customSynthesizer?;
122
+ private defaultLlm?;
91
123
  constructor(options?: {
92
124
  tavilyKey?: string;
93
125
  exaKey?: string;
126
+ jinaKey?: string;
127
+ serperKey?: string;
128
+ semanticScholarKey?: string;
129
+ apiKeys?: ResearchApiKeys;
130
+ allowEnvFallback?: boolean;
94
131
  crawl4aiHost?: string;
132
+ llm?: LlmTextFn;
95
133
  customSearchAdapter?: CustomSearchAdapter;
96
134
  customCrawlerAdapter?: CustomCrawlerAdapter;
97
135
  customAcademicSearchAdapter?: CustomAcademicSearchAdapter;
136
+ customSynthesizerAdapter?: CustomSynthesizerAdapter;
98
137
  });
99
138
  run(topic: string, options?: HarvestPipelineOptions): Promise<HarvestPipelineResult>;
139
+ private synthesizeInsights;
100
140
  }
101
141
 
102
- export { CampaignPlanner, HarvestPipeline, type HarvestPipelineOptions, type HarvestPipelineResult, ReferencePackBuilder, type ReferencePackOutput, type ReferenceRegistryItem, type ResearchPlan };
142
+ export { CampaignPlanner, HarvestPipeline, type HarvestPipelineOptions, type HarvestPipelineResult, ReferencePackBuilder, type ReferencePackOutput, type ReferenceRegistryItem, type ResearchPlan, normalizeUrlForDedup };
@@ -1,7 +1,7 @@
1
- import { T as TechnicalDomain, S as ScrapedDocument } from '../webScraper-pKzdXhfI.js';
2
- import { e as AcademicPaper, d as ResearchConfig, a as CustomSearchAdapter, b as CustomCrawlerAdapter, c as CustomAcademicSearchAdapter } from '../index-BU5n1ip_.js';
3
- import { C as CS2023KnowledgeArea, L as LiteratureReviewReport } from '../literatureReview-C4AgVljZ.js';
4
- import { R as ResourceCatalogBundle } from '../catalogBuilder-8A-Dg_A-.js';
1
+ import { T as TechnicalDomain, S as ScrapedDocument } from '../webScraper-BhiQ8s9x.js';
2
+ import { C as CourseResearchContext, L as LlmTextFn, A as AcademicPaper, E as ExtractedConceptNode, f as SynthesizedResearchInsights, R as ResearchApiKeys, h as ResearchConfig, a as CustomSearchAdapter, b as CustomCrawlerAdapter, c as CustomAcademicSearchAdapter, d as CustomSynthesizerAdapter } from '../index-t8pahlJ4.js';
3
+ import { C as CS2023KnowledgeArea, a as CSTAK12Strand, L as LiteratureReviewReport } from '../literatureReview-hGbfrBmw.js';
4
+ import { R as ResourceCatalogBundle } from '../catalogBuilder-CmmQzUnf.js';
5
5
  import { a as ProvenanceSidecar } from '../provenanceTracker-ct4s5lJV.js';
6
6
  import 'zod';
7
7
 
@@ -15,19 +15,30 @@ interface ResearchPlan {
15
15
  planMd: string;
16
16
  }
17
17
  /**
18
- * Creates and formats MASTER_RESEARCH_PLAN.md with Scale Decision (English Canonical)
18
+ * Creates and formats MASTER_RESEARCH_PLAN.md with Context-Aware Multi-Angle Queries
19
+ * (Technical Ground Truth + Pedagogical Exemplars + Learner Misconceptions/Pitfalls).
19
20
  */
20
21
  declare class CampaignPlanner {
21
22
  createPlan(topic: string, domain: TechnicalDomain, options?: {
22
23
  maxRounds?: number;
23
24
  isDeep?: boolean;
25
+ context?: CourseResearchContext;
24
26
  }): ResearchPlan;
27
+ /**
28
+ * Optional LLM-augmented query planner when the host app provides an `llm` callback.
29
+ */
30
+ createPlanWithLlm(topic: string, domain: TechnicalDomain, options?: {
31
+ maxRounds?: number;
32
+ isDeep?: boolean;
33
+ context?: CourseResearchContext;
34
+ llm?: LlmTextFn;
35
+ }): Promise<ResearchPlan>;
25
36
  }
26
37
 
27
38
  interface ReferenceRegistryItem {
28
39
  id: string;
29
40
  title: string;
30
- sourceType: 'official_docs' | 'academic_paper' | 'community_repo' | 'industry_standard';
41
+ sourceType: 'official_docs' | 'academic_paper' | 'video_transcript' | 'community_repo' | 'industry_standard';
31
42
  url: string;
32
43
  relevanceScore: number;
33
44
  extractedKeyTakeaways: string[];
@@ -37,17 +48,23 @@ interface ReferencePackOutput {
37
48
  referencePackMd: string;
38
49
  }
39
50
  /**
40
- * Compiles the canonical REFERENCE_PACK.md & REFERENCE_REGISTRY.json (English Canonical)
51
+ * Compiles the canonical REFERENCE_PACK.md & REFERENCE_REGISTRY.json with real extracted takeaways,
52
+ * accurate code fence languages, synthesized misconceptions/exemplars, and bilingual concept anchors.
41
53
  */
42
54
  declare class ReferencePackBuilder {
43
55
  buildReferencePack(topic: string, domain: TechnicalDomain, options?: {
44
56
  cs2023Area?: CS2023KnowledgeArea;
57
+ k12Strand?: CSTAK12Strand;
58
+ context?: CourseResearchContext;
45
59
  scrapedDocs?: {
46
60
  doc: ScrapedDocument;
47
61
  score: number;
48
62
  summary: string;
63
+ keyTakeaways?: string[];
49
64
  }[];
50
65
  academicPapers?: AcademicPaper[];
66
+ concepts?: ExtractedConceptNode[];
67
+ synthesizedInsights?: SynthesizedResearchInsights;
51
68
  }): ReferencePackOutput;
52
69
  }
53
70
 
@@ -55,13 +72,21 @@ interface HarvestPipelineOptions {
55
72
  domain?: string;
56
73
  maxRounds?: number;
57
74
  isDeep?: boolean;
75
+ crawl4aiHost?: string;
58
76
  tavilyKey?: string;
59
77
  exaKey?: string;
78
+ jinaKey?: string;
79
+ serperKey?: string;
60
80
  semanticScholarKey?: string;
81
+ apiKeys?: ResearchApiKeys;
82
+ allowEnvFallback?: boolean;
83
+ context?: CourseResearchContext;
84
+ llm?: LlmTextFn;
61
85
  config?: Partial<ResearchConfig>;
62
86
  customSearchAdapter?: CustomSearchAdapter;
63
87
  customCrawlerAdapter?: CustomCrawlerAdapter;
64
88
  customAcademicSearchAdapter?: CustomAcademicSearchAdapter;
89
+ customSynthesizerAdapter?: CustomSynthesizerAdapter;
65
90
  onProgress?: (stage: string, message: string) => void;
66
91
  }
67
92
  interface HarvestPipelineResult {
@@ -72,10 +97,14 @@ interface HarvestPipelineResult {
72
97
  literatureReview: LiteratureReviewReport;
73
98
  referencePack: ReferencePackOutput;
74
99
  provenance: ProvenanceSidecar;
100
+ concepts: ExtractedConceptNode[];
101
+ conceptDagTsv: string;
102
+ synthesizedInsights: SynthesizedResearchInsights;
75
103
  totalDocumentsScraped: number;
76
104
  }
105
+ declare function normalizeUrlForDedup(rawUrl: string): string;
77
106
  /**
78
- * Autonomous Multi-Round Deep Research & Academic Harvester Pipeline with Pluggable Tool Adapters
107
+ * Autonomous Multi-Round Deep Research & Academic Harvester Pipeline with Host API Key & LLM Injection
79
108
  */
80
109
  declare class HarvestPipeline {
81
110
  private searchEngine;
@@ -88,15 +117,26 @@ declare class HarvestPipeline {
88
117
  private litReviewSynthesizer;
89
118
  private planner;
90
119
  private refPackBuilder;
120
+ private termExtractor;
121
+ private customSynthesizer?;
122
+ private defaultLlm?;
91
123
  constructor(options?: {
92
124
  tavilyKey?: string;
93
125
  exaKey?: string;
126
+ jinaKey?: string;
127
+ serperKey?: string;
128
+ semanticScholarKey?: string;
129
+ apiKeys?: ResearchApiKeys;
130
+ allowEnvFallback?: boolean;
94
131
  crawl4aiHost?: string;
132
+ llm?: LlmTextFn;
95
133
  customSearchAdapter?: CustomSearchAdapter;
96
134
  customCrawlerAdapter?: CustomCrawlerAdapter;
97
135
  customAcademicSearchAdapter?: CustomAcademicSearchAdapter;
136
+ customSynthesizerAdapter?: CustomSynthesizerAdapter;
98
137
  });
99
138
  run(topic: string, options?: HarvestPipelineOptions): Promise<HarvestPipelineResult>;
139
+ private synthesizeInsights;
100
140
  }
101
141
 
102
- export { CampaignPlanner, HarvestPipeline, type HarvestPipelineOptions, type HarvestPipelineResult, ReferencePackBuilder, type ReferencePackOutput, type ReferenceRegistryItem, type ResearchPlan };
142
+ export { CampaignPlanner, HarvestPipeline, type HarvestPipelineOptions, type HarvestPipelineResult, ReferencePackBuilder, type ReferencePackOutput, type ReferenceRegistryItem, type ResearchPlan, normalizeUrlForDedup };
@@ -1,6 +1,7 @@
1
- export { CampaignPlanner, HarvestPipeline, ReferencePackBuilder } from '../chunk-MHAYD555.js';
2
- import '../chunk-47L7ZHMY.js';
3
- import '../chunk-MPLDSHFM.js';
4
- import '../chunk-QZNLB5X4.js';
1
+ export { CampaignPlanner, HarvestPipeline, ReferencePackBuilder, normalizeUrlForDedup } from '../chunk-AUGVVZTL.js';
2
+ import '../chunk-DOXSPRSP.js';
3
+ import '../chunk-XIF5YNI7.js';
4
+ import '../chunk-GKF3NDEX.js';
5
+ import '../chunk-KAFQVI2R.js';
5
6
  //# sourceMappingURL=index.js.map
6
7
  //# sourceMappingURL=index.js.map
@@ -0,0 +1,68 @@
1
+ import { C as CourseResearchContext, b as CustomCrawlerAdapter } from './index-t8pahlJ4.js';
2
+
3
+ type TechnicalDomain = 'hardware' | 'software' | 'platform' | 'academic';
4
+ interface DomainClassificationResult {
5
+ domain: TechnicalDomain;
6
+ confidence: number;
7
+ rationale: string;
8
+ recommendedExtractors: string[];
9
+ primaryCodeLanguage: string;
10
+ }
11
+ /**
12
+ * Infers the primary code fence language from topic and optional course context.
13
+ * Never silently defaults to 'swift' unless Swift is actually present.
14
+ */
15
+ declare function inferPrimaryCodeLanguage(topic: string, context?: CourseResearchContext, domain?: TechnicalDomain): string;
16
+ /**
17
+ * Domain Classifier supporting English & Vietnamese CS/STEM/EdTech topics and CourseResearchContext
18
+ */
19
+ declare function classifyTechnicalDomain(topic: string, explicitDomain?: string, context?: CourseResearchContext): DomainClassificationResult;
20
+
21
+ interface ExtractedCodeBlock {
22
+ language: string;
23
+ code: string;
24
+ }
25
+ interface ScrapedDocument {
26
+ url: string;
27
+ title: string;
28
+ markdownContent: string;
29
+ byteSize: number;
30
+ extractedCodeSnippets: string[];
31
+ extractedCodeBlocks?: ExtractedCodeBlock[];
32
+ extractedTables?: string[];
33
+ extractedHeadings?: string[];
34
+ keyExcerpts?: string[];
35
+ tablesCount: number;
36
+ fetchProvider: 'jina' | 'crawl4ai' | 'youtube' | 'direct' | 'custom';
37
+ }
38
+ interface WebScraperOptions {
39
+ crawl4aiHost?: string;
40
+ jinaKey?: string;
41
+ allowEnvFallback?: boolean;
42
+ customAdapter?: CustomCrawlerAdapter;
43
+ }
44
+ /**
45
+ * Extracts top substantive paragraphs from Markdown (stripping navigation, links-only lines, and cookie banners).
46
+ */
47
+ declare function extractSubstantiveExcerpts(markdown: string, maxExcerpts?: number): string[];
48
+ /**
49
+ * High-fidelity web content & YouTube transcript extractor with Custom Crawler & Jina Auth support
50
+ */
51
+ declare class WebScraper {
52
+ private crawl4aiHost?;
53
+ private jinaKey?;
54
+ private allowEnvFallback;
55
+ private customAdapter?;
56
+ private youtubeService;
57
+ constructor(options?: WebScraperOptions);
58
+ setCustomAdapter(adapter: CustomCrawlerAdapter): void;
59
+ setJinaKey(key?: string): void;
60
+ scrapeUrl(targetUrl: string, options?: {
61
+ customAdapter?: CustomCrawlerAdapter;
62
+ jinaKey?: string;
63
+ preferredLanguages?: string[];
64
+ }): Promise<ScrapedDocument>;
65
+ private packageDocument;
66
+ }
67
+
68
+ export { type DomainClassificationResult as D, type ExtractedCodeBlock as E, type ScrapedDocument as S, type TechnicalDomain as T, type WebScraperOptions as W, WebScraper as a, classifyTechnicalDomain as c, extractSubstantiveExcerpts as e, inferPrimaryCodeLanguage as i };
@@ -0,0 +1,68 @@
1
+ import { C as CourseResearchContext, b as CustomCrawlerAdapter } from './index-t8pahlJ4.cjs';
2
+
3
+ type TechnicalDomain = 'hardware' | 'software' | 'platform' | 'academic';
4
+ interface DomainClassificationResult {
5
+ domain: TechnicalDomain;
6
+ confidence: number;
7
+ rationale: string;
8
+ recommendedExtractors: string[];
9
+ primaryCodeLanguage: string;
10
+ }
11
+ /**
12
+ * Infers the primary code fence language from topic and optional course context.
13
+ * Never silently defaults to 'swift' unless Swift is actually present.
14
+ */
15
+ declare function inferPrimaryCodeLanguage(topic: string, context?: CourseResearchContext, domain?: TechnicalDomain): string;
16
+ /**
17
+ * Domain Classifier supporting English & Vietnamese CS/STEM/EdTech topics and CourseResearchContext
18
+ */
19
+ declare function classifyTechnicalDomain(topic: string, explicitDomain?: string, context?: CourseResearchContext): DomainClassificationResult;
20
+
21
+ interface ExtractedCodeBlock {
22
+ language: string;
23
+ code: string;
24
+ }
25
+ interface ScrapedDocument {
26
+ url: string;
27
+ title: string;
28
+ markdownContent: string;
29
+ byteSize: number;
30
+ extractedCodeSnippets: string[];
31
+ extractedCodeBlocks?: ExtractedCodeBlock[];
32
+ extractedTables?: string[];
33
+ extractedHeadings?: string[];
34
+ keyExcerpts?: string[];
35
+ tablesCount: number;
36
+ fetchProvider: 'jina' | 'crawl4ai' | 'youtube' | 'direct' | 'custom';
37
+ }
38
+ interface WebScraperOptions {
39
+ crawl4aiHost?: string;
40
+ jinaKey?: string;
41
+ allowEnvFallback?: boolean;
42
+ customAdapter?: CustomCrawlerAdapter;
43
+ }
44
+ /**
45
+ * Extracts top substantive paragraphs from Markdown (stripping navigation, links-only lines, and cookie banners).
46
+ */
47
+ declare function extractSubstantiveExcerpts(markdown: string, maxExcerpts?: number): string[];
48
+ /**
49
+ * High-fidelity web content & YouTube transcript extractor with Custom Crawler & Jina Auth support
50
+ */
51
+ declare class WebScraper {
52
+ private crawl4aiHost?;
53
+ private jinaKey?;
54
+ private allowEnvFallback;
55
+ private customAdapter?;
56
+ private youtubeService;
57
+ constructor(options?: WebScraperOptions);
58
+ setCustomAdapter(adapter: CustomCrawlerAdapter): void;
59
+ setJinaKey(key?: string): void;
60
+ scrapeUrl(targetUrl: string, options?: {
61
+ customAdapter?: CustomCrawlerAdapter;
62
+ jinaKey?: string;
63
+ preferredLanguages?: string[];
64
+ }): Promise<ScrapedDocument>;
65
+ private packageDocument;
66
+ }
67
+
68
+ export { type DomainClassificationResult as D, type ExtractedCodeBlock as E, type ScrapedDocument as S, type TechnicalDomain as T, type WebScraperOptions as W, WebScraper as a, classifyTechnicalDomain as c, extractSubstantiveExcerpts as e, inferPrimaryCodeLanguage as i };
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@thanh01.pmt/research-kit",
3
- "version": "1.0.0",
3
+ "version": "1.1.1",
4
4
  "description": "An autonomous Deep Research, Academic Harvest & Evidence Verification Engine for Curriculum OS & AI Agents.",
5
5
  "author": "Thanh Pham <https://github.com/thanh01pmt>",
6
6
  "license": "MIT",
@@ -46,13 +46,6 @@
46
46
  "publishConfig": {
47
47
  "access": "public"
48
48
  },
49
- "scripts": {
50
- "clean": "rm -rf dist",
51
- "build": "npm run clean && tsup",
52
- "test": "vitest run",
53
- "typecheck": "tsc --noEmit",
54
- "prepublishOnly": "npm run build && npm test"
55
- },
56
49
  "dependencies": {
57
50
  "zod": "^3.23.8"
58
51
  },
@@ -60,5 +53,11 @@
60
53
  "tsup": "^8.3.5",
61
54
  "typescript": "^5.6.3",
62
55
  "vitest": "^2.1.8"
56
+ },
57
+ "scripts": {
58
+ "clean": "rm -rf dist",
59
+ "build": "npm run clean && tsup",
60
+ "test": "vitest run",
61
+ "typecheck": "tsc --noEmit"
63
62
  }
64
- }
63
+ }