scrapex 1.0.0-alpha.1 → 1.0.0-beta.1

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "scrapex",
-  "version": "1.0.0-alpha.1",
+  "version": "1.0.0-beta.1",
   "description": "Modern web scraper with LLM-enhanced extraction, extensible pipeline, and pluggable parsers",
   "type": "module",
   "exports": {
@@ -18,6 +18,11 @@
       "types": "./dist/llm/index.d.mts",
       "import": "./dist/llm/index.mjs",
       "require": "./dist/llm/index.cjs"
+    },
+    "./embeddings": {
+      "types": "./dist/embeddings/index.d.mts",
+      "import": "./dist/embeddings/index.mjs",
+      "require": "./dist/embeddings/index.cjs"
     }
   },
   "main": "./dist/index.cjs",
@@ -53,7 +58,9 @@
     "extraction",
     "readability",
     "markdown",
-    "parser"
+    "parser",
+    "embeddings",
+    "vector-search"
   ],
   "author": "Rakesh Paul <https://binaryroute.com/authors/rk-paul/>",
   "license": "MIT",
@@ -63,35 +70,27 @@
   "dependencies": {
     "@mozilla/readability": "^0.6.0",
     "cheerio": "^1.1.2",
-    "jsdom": "^27.2.0",
+    "jsdom": "^27.4.0",
     "mdast-util-from-markdown": "^2.0.2",
     "mdast-util-to-string": "^4.0.0",
     "turndown": "^7.2.2",
     "unist-util-visit": "^5.0.0",
-    "zod": "^4.1.13"
+    "zod": "^4.3.4"
   },
   "devDependencies": {
-    "@biomejs/biome": "^2.3.8",
+    "@biomejs/biome": "^2.3.10",
     "@types/jsdom": "^27.0.0",
     "@types/mdast": "^4.0.4",
     "@types/node": "^22.10.0",
     "@types/turndown": "^5.0.6",
-    "tsdown": "^0.17.0",
+    "tsdown": "^0.18.4",
     "typescript": "^5.9.3",
-    "vitest": "^4.0.15"
+    "vitest": "^4.0.16"
   },
   "peerDependencies": {
-    "@anthropic-ai/sdk": ">=0.30.0",
-    "openai": ">=4.0.0",
-    "puppeteer": ">=23.0.0"
+    "puppeteer": "^24.34.0"
   },
   "peerDependenciesMeta": {
-    "@anthropic-ai/sdk": {
-      "optional": true
-    },
-    "openai": {
-      "optional": true
-    },
     "puppeteer": {
       "optional": true
     }

package/dist/enhancer-Q6CSc1gA.mjs

@@ -1,220 +0,0 @@
-import { z } from "zod";
-
-//#region src/core/errors.ts
-/**
-* Custom error class for scraping failures with structured error codes
-*/
-var ScrapeError = class ScrapeError extends Error {
-	code;
-	statusCode;
-	constructor(message, code, statusCode, cause) {
-		super(message, { cause });
-		this.name = "ScrapeError";
-		this.code = code;
-		this.statusCode = statusCode;
-		if (Error.captureStackTrace) Error.captureStackTrace(this, ScrapeError);
-	}
-	/**
-	* Create a ScrapeError from an unknown error
-	*/
-	static from(error, code = "FETCH_FAILED") {
-		if (error instanceof ScrapeError) return error;
-		if (error instanceof Error) return new ScrapeError(error.message, code, void 0, error);
-		return new ScrapeError(String(error), code);
-	}
-	/**
-	* Check if error is retryable (network issues, timeouts)
-	*/
-	isRetryable() {
-		return this.code === "FETCH_FAILED" || this.code === "TIMEOUT";
-	}
-	/**
-	* Convert to a plain object for serialization
-	*/
-	toJSON() {
-		return {
-			name: this.name,
-			message: this.message,
-			code: this.code,
-			statusCode: this.statusCode,
-			stack: this.stack
-		};
-	}
-};
-
-//#endregion
-//#region src/llm/types.ts
-/**
-* Zod schemas for LLM outputs
-*/
-const SummarySchema = z.object({ summary: z.string().describe("A concise 2-3 sentence summary of the content") });
-const TagsSchema = z.object({ tags: z.array(z.string()).describe("5-10 relevant tags/keywords") });
-const EntitiesSchema = z.object({
-	people: z.array(z.string()).describe("People mentioned"),
-	organizations: z.array(z.string()).describe("Organizations/companies"),
-	technologies: z.array(z.string()).describe("Technologies/tools/frameworks"),
-	locations: z.array(z.string()).describe("Locations/places"),
-	concepts: z.array(z.string()).describe("Key concepts/topics")
-});
-const ClassifySchema = z.object({
-	contentType: z.enum([
-		"article",
-		"repo",
-		"docs",
-		"package",
-		"video",
-		"tool",
-		"product",
-		"unknown"
-	]).describe("The type of content"),
-	confidence: z.number().min(0).max(1).describe("Confidence score 0-1")
-});
-
-//#endregion
-//#region src/llm/enhancer.ts
-/**
-* Enhance scraped data with LLM-powered features
-*/
-async function enhance(data, provider, types) {
-	const results = {};
-	const content = data.excerpt || data.textContent.slice(0, 1e4);
-	const context = `Title: ${data.title}\nURL: ${data.url}\n\nContent:\n${content}`;
-	const promises = [];
-	if (types.includes("summarize")) promises.push(summarize(context, provider).then((summary) => {
-		results.summary = summary;
-	}));
-	if (types.includes("tags")) promises.push(extractTags(context, provider).then((tags) => {
-		results.suggestedTags = tags;
-	}));
-	if (types.includes("entities")) promises.push(extractEntities(context, provider).then((entities) => {
-		results.entities = entities;
-	}));
-	if (types.includes("classify")) promises.push(classify(context, provider).then((classification) => {
-		if (classification.confidence > .7) results.contentType = classification.contentType;
-	}));
-	await Promise.all(promises);
-	return results;
-}
-/**
-* Ask a custom question about the scraped content
-* Results are stored in the `custom` field of ScrapedData
-*/
-async function ask(data, provider, prompt, options) {
-	const key = options?.key || "response";
-	const content = data.excerpt || data.textContent.slice(0, 1e4);
-	const processedPrompt = applyPlaceholders(prompt, data, content);
-	if (options?.schema) {
-		const result = await extract(data, provider, options.schema, processedPrompt);
-		return { custom: { [key]: result } };
-	}
-	const fullPrompt = prompt.includes("{{content}}") ? processedPrompt : `${processedPrompt}\n\nTitle: ${data.title}\nURL: ${data.url}\n\nContent:\n${content}`;
-	const response = await provider.complete(fullPrompt);
-	return { custom: { [key]: response } };
-}
-/**
-* Apply placeholder replacements to a prompt template
-*/
-function applyPlaceholders(prompt, data, content) {
-	const domain = (() => {
-		try {
-			return new URL(data.url).hostname;
-		} catch {
-			return "";
-		}
-	})();
-	return prompt.replace(/\{\{title\}\}/g, data.title).replace(/\{\{url\}\}/g, data.url).replace(/\{\{content\}\}/g, content).replace(/\{\{description\}\}/g, data.description || "").replace(/\{\{excerpt\}\}/g, data.excerpt || "").replace(/\{\{domain\}\}/g, domain);
-}
-/**
-* Extract structured data using LLM and a custom schema
-*/
-async function extract(data, provider, schema, promptTemplate) {
-	const zodShape = {};
-	for (const [key, type] of Object.entries(schema)) {
-		const isOptional = type.endsWith("?");
-		const baseType = isOptional ? type.slice(0, -1) : type;
-		let zodType;
-		switch (baseType) {
-			case "string":
-				zodType = z.string();
-				break;
-			case "number":
-				zodType = z.number();
-				break;
-			case "boolean":
-				zodType = z.boolean();
-				break;
-			case "string[]":
-				zodType = z.array(z.string());
-				break;
-			case "number[]":
-				zodType = z.array(z.number());
-				break;
-			default: zodType = z.string();
-		}
-		zodShape[key] = isOptional ? zodType.optional() : zodType;
-	}
-	const zodSchema = z.object(zodShape);
-	const content = data.textContent.slice(0, 4e3);
-	let prompt;
-	if (promptTemplate) {
-		prompt = applyPlaceholders(promptTemplate, data, content);
-		if (!promptTemplate.includes("{{content}}")) prompt += `\n\nContext:\n${content}`;
-	} else prompt = `Extract the following information from this content:
-
-Title: ${data.title}
-URL: ${data.url}
-
-Content:
-${content}
-
-Extract these fields:
-${Object.entries(schema).map(([key, type]) => `- ${key} (${type})`).join("\n")}`;
-	return provider.completeJSON(prompt, zodSchema);
-}
-/**
-* Generate a summary of the content
-*/
-async function summarize(context, provider) {
-	const prompt = `Summarize the following content in 2-3 concise sentences:
-
-${context}`;
-	return (await provider.completeJSON(prompt, SummarySchema)).summary;
-}
-/**
-* Extract relevant tags/keywords
-*/
-async function extractTags(context, provider) {
-	const prompt = `Extract 5-10 relevant tags or keywords from the following content. Focus on technologies, concepts, and topics mentioned:
-
-${context}`;
-	return (await provider.completeJSON(prompt, TagsSchema)).tags;
-}
-/**
-* Extract named entities from content
-*/
-async function extractEntities(context, provider) {
-	const prompt = `Extract named entities from the following content. Identify people, organizations, technologies, locations, and key concepts:
-
-${context}`;
-	return provider.completeJSON(prompt, EntitiesSchema);
-}
-/**
-* Classify content type using LLM
-*/
-async function classify(context, provider) {
-	const prompt = `Classify the following content into one of these categories:
-- article: Blog post, news article, essay
-- repo: Code repository, open source project
-- docs: Documentation, API reference, guides
-- package: npm/pip package page
-- video: Video content, YouTube
-- tool: Software tool, web application
-- product: Commercial product, e-commerce
-
-${context}`;
-	return provider.completeJSON(prompt, ClassifySchema);
-}
-
-//#endregion
-export { EntitiesSchema as a, ScrapeError as c, ClassifySchema as i, enhance as n, SummarySchema as o, extract as r, TagsSchema as s, ask as t };
-//# sourceMappingURL=enhancer-Q6CSc1gA.mjs.map

package/dist/enhancer-Q6CSc1gA.mjs.map

@@ -1 +0,0 @@
-{"version":3,"file":"enhancer-Q6CSc1gA.mjs","names":["results: Partial<ScrapedData>","promises: Promise<void>[]","zodShape: Record<string, z.ZodTypeAny>","zodType: z.ZodTypeAny","prompt: string"],"sources":["../src/core/errors.ts","../src/llm/types.ts","../src/llm/enhancer.ts"],"sourcesContent":["/**\n * Error codes for scraping failures\n */\nexport type ScrapeErrorCode =\n  | 'FETCH_FAILED'\n  | 'TIMEOUT'\n  | 'INVALID_URL'\n  | 'BLOCKED'\n  | 'NOT_FOUND'\n  | 'ROBOTS_BLOCKED'\n  | 'PARSE_ERROR'\n  | 'LLM_ERROR'\n  | 'VALIDATION_ERROR';\n\n/**\n * Custom error class for scraping failures with structured error codes\n */\nexport class ScrapeError extends Error {\n  public readonly code: ScrapeErrorCode;\n  public readonly statusCode?: number;\n\n  constructor(message: string, code: ScrapeErrorCode, statusCode?: number, cause?: Error) {\n    super(message, { cause });\n    this.name = 'ScrapeError';\n    this.code = code;\n    this.statusCode = statusCode;\n\n    // Maintains proper stack trace in V8 environments\n    if (Error.captureStackTrace) {\n      Error.captureStackTrace(this, ScrapeError);\n    }\n  }\n\n  /**\n   * Create a ScrapeError from an unknown error\n   */\n  static from(error: unknown, code: ScrapeErrorCode = 'FETCH_FAILED'): ScrapeError {\n    if (error instanceof ScrapeError) {\n      return error;\n    }\n\n    if (error instanceof Error) {\n      return new ScrapeError(error.message, code, undefined, error);\n    }\n\n    return new ScrapeError(String(error), code);\n  }\n\n  /**\n   * Check if error is retryable (network issues, timeouts)\n   */\n  isRetryable(): boolean {\n    return this.code === 'FETCH_FAILED' || this.code === 'TIMEOUT';\n  }\n\n  /**\n   * Convert to a plain object for serialization\n   */\n  toJSON(): Record<string, unknown> {\n    return {\n      name: this.name,\n      message: this.message,\n      code: this.code,\n      statusCode: this.statusCode,\n      stack: this.stack,\n    };\n  }\n}\n","import { z } from 'zod';\n\n/**\n * LLM completion options\n */\nexport interface CompletionOptions {\n  maxTokens?: number;\n  temperature?: number;\n  systemPrompt?: string;\n}\n\n/**\n * LLM Provider interface - implemented by all providers\n */\nexport interface LLMProvider {\n  readonly name: string;\n\n  /**\n   * Generate a text completion\n   */\n  complete(prompt: string, options?: CompletionOptions): Promise<string>;\n\n  /**\n   * Generate a structured JSON completion with Zod validation\n   */\n  completeJSON<T>(prompt: string, schema: z.ZodType<T>, options?: CompletionOptions): Promise<T>;\n}\n\n/**\n * Provider configuration for Anthropic\n */\nexport interface AnthropicConfig {\n  apiKey?: string; // Falls back to ANTHROPIC_API_KEY env var\n  model?: string; // Default: claude-3-haiku-20240307\n  baseUrl?: string;\n}\n\n/**\n * Provider configuration for OpenAI-compatible APIs\n * Works with: OpenAI, Ollama, LM Studio, LocalAI, vLLM, etc.\n */\nexport interface OpenAICompatibleConfig {\n  apiKey?: string; // Falls back to OPENAI_API_KEY env var\n  model?: string; // Default: gpt-4o-mini\n  baseUrl?: string; // Default: https://api.openai.com/v1\n}\n\n/**\n * Enhancement result types\n */\nexport interface SummaryResult {\n  summary: string;\n}\n\nexport interface TagsResult {\n  tags: string[];\n}\n\nexport interface EntitiesResult {\n  people: string[];\n  organizations: string[];\n  technologies: string[];\n  locations: string[];\n  concepts: string[];\n}\n\nexport interface ClassifyResult {\n  contentType: string;\n  confidence: number;\n}\n\n/**\n * Zod schemas for LLM outputs\n */\nexport const SummarySchema = z.object({\n  summary: z.string().describe('A concise 2-3 sentence summary of the content'),\n});\n\nexport const TagsSchema = z.object({\n  tags: z.array(z.string()).describe('5-10 relevant tags/keywords'),\n});\n\nexport const EntitiesSchema = z.object({\n  people: z.array(z.string()).describe('People mentioned'),\n  organizations: z.array(z.string()).describe('Organizations/companies'),\n  technologies: z.array(z.string()).describe('Technologies/tools/frameworks'),\n  locations: z.array(z.string()).describe('Locations/places'),\n  concepts: z.array(z.string()).describe('Key concepts/topics'),\n});\n\nexport const ClassifySchema = z.object({\n  contentType: z\n    .enum(['article', 'repo', 'docs', 'package', 'video', 'tool', 'product', 'unknown'])\n    .describe('The type of content'),\n  confidence: z.number().min(0).max(1).describe('Confidence score 0-1'),\n});\n","import { z } from 'zod';\nimport type {\n  EnhancementType,\n  ExtractedEntities,\n  ExtractionSchema,\n  ScrapedData,\n} from '@/core/types.js';\nimport type { LLMProvider } from './types.js';\nimport { ClassifySchema, EntitiesSchema, SummarySchema, TagsSchema } from './types.js';\n\n/**\n * Enhance scraped data with LLM-powered features\n */\nexport async function enhance(\n  data: ScrapedData,\n  provider: LLMProvider,\n  types: EnhancementType[]\n): Promise<Partial<ScrapedData>> {\n  const results: Partial<ScrapedData> = {};\n\n  // Prepare content for LLM (use excerpt/textContent to save tokens)\n  const content = data.excerpt || data.textContent.slice(0, 10000);\n  const context = `Title: ${data.title}\\nURL: ${data.url}\\n\\nContent:\\n${content}`;\n\n  // Run enhancements in parallel\n  const promises: Promise<void>[] = [];\n\n  if (types.includes('summarize')) {\n    promises.push(\n      summarize(context, provider).then((summary) => {\n        results.summary = summary;\n      })\n    );\n  }\n\n  if (types.includes('tags')) {\n    promises.push(\n      extractTags(context, provider).then((tags) => {\n        results.suggestedTags = tags;\n      })\n    );\n  }\n\n  if (types.includes('entities')) {\n    promises.push(\n      extractEntities(context, provider).then((entities) => {\n        results.entities = entities;\n      })\n    );\n  }\n\n  if (types.includes('classify')) {\n    promises.push(\n      classify(context, provider).then((classification) => {\n        if (classification.confidence > 0.7) {\n          results.contentType = classification.contentType as ScrapedData['contentType'];\n        }\n      })\n    );\n  }\n\n  await Promise.all(promises);\n\n  return results;\n}\n\n/**\n * Options for the ask() function\n */\nexport interface AskOptions {\n  /** Key to store the result under in custom field */\n  key?: string;\n  /** Schema for structured response */\n  schema?: ExtractionSchema;\n}\n\n/**\n * Ask a custom question about the scraped content\n * Results are stored in the `custom` field of ScrapedData\n */\nexport async function ask(\n  data: ScrapedData,\n  provider: LLMProvider,\n  prompt: string,\n  options?: AskOptions\n): Promise<Partial<ScrapedData>> {\n  const key = options?.key || 'response';\n  const content = data.excerpt || data.textContent.slice(0, 10000);\n\n  // Apply placeholder replacements\n  const processedPrompt = applyPlaceholders(prompt, data, content);\n\n  if (options?.schema) {\n    // Use structured extraction\n    const result = await extract(data, provider, options.schema, processedPrompt);\n    return { custom: { [key]: result } };\n  }\n\n  // Simple string response\n  const fullPrompt = prompt.includes('{{content}}')\n    ? processedPrompt\n    : `${processedPrompt}\\n\\nTitle: ${data.title}\\nURL: ${data.url}\\n\\nContent:\\n${content}`;\n\n  const response = await provider.complete(fullPrompt);\n  return { custom: { [key]: response } };\n}\n\n/**\n * Apply placeholder replacements to a prompt template\n */\nfunction applyPlaceholders(prompt: string, data: ScrapedData, content: string): string {\n  const domain = (() => {\n    try {\n      return new URL(data.url).hostname;\n    } catch {\n      return '';\n    }\n  })();\n\n  return prompt\n    .replace(/\\{\\{title\\}\\}/g, data.title)\n    .replace(/\\{\\{url\\}\\}/g, data.url)\n    .replace(/\\{\\{content\\}\\}/g, content)\n    .replace(/\\{\\{description\\}\\}/g, data.description || '')\n    .replace(/\\{\\{excerpt\\}\\}/g, data.excerpt || '')\n    .replace(/\\{\\{domain\\}\\}/g, domain);\n}\n\n/**\n * Extract structured data using LLM and a custom schema\n */\nexport async function extract<T>(\n  data: ScrapedData,\n  provider: LLMProvider,\n  schema: ExtractionSchema,\n  promptTemplate?: string\n): Promise<T> {\n  // Convert simple schema to Zod schema\n  const zodShape: Record<string, z.ZodTypeAny> = {};\n\n  for (const [key, type] of Object.entries(schema)) {\n    const isOptional = type.endsWith('?');\n    const baseType = isOptional ? type.slice(0, -1) : type;\n\n    let zodType: z.ZodTypeAny;\n    switch (baseType) {\n      case 'string':\n        zodType = z.string();\n        break;\n      case 'number':\n        zodType = z.number();\n        break;\n      case 'boolean':\n        zodType = z.boolean();\n        break;\n      case 'string[]':\n        zodType = z.array(z.string());\n        break;\n      case 'number[]':\n        zodType = z.array(z.number());\n        break;\n      default:\n        zodType = z.string();\n    }\n\n    zodShape[key] = isOptional ? zodType.optional() : zodType;\n  }\n\n  const zodSchema = z.object(zodShape) as unknown as z.ZodType<T>;\n\n  const content = data.textContent.slice(0, 4000);\n\n  let prompt: string;\n\n  if (promptTemplate) {\n    // Apply all placeholder replacements\n    prompt = applyPlaceholders(promptTemplate, data, content);\n\n    // If content wasn't included via placeholder, append it\n    if (!promptTemplate.includes('{{content}}')) {\n      prompt += `\\n\\nContext:\\n${content}`;\n    }\n  } else {\n    prompt = `Extract the following information from this content:\n\nTitle: ${data.title}\nURL: ${data.url}\n\nContent:\n${content}\n\nExtract these fields:\n${Object.entries(schema)\n  .map(([key, type]) => `- ${key} (${type})`)\n  .join('\\n')}`;\n  }\n\n  return provider.completeJSON<T>(prompt, zodSchema as z.ZodType<T>);\n}\n\n/**\n * Generate a summary of the content\n */\nasync function summarize(context: string, provider: LLMProvider): Promise<string> {\n  const prompt = `Summarize the following content in 2-3 concise sentences:\n\n${context}`;\n\n  const result = await provider.completeJSON(prompt, SummarySchema);\n  return result.summary;\n}\n\n/**\n * Extract relevant tags/keywords\n */\nasync function extractTags(context: string, provider: LLMProvider): Promise<string[]> {\n  const prompt = `Extract 5-10 relevant tags or keywords from the following content. Focus on technologies, concepts, and topics mentioned:\n\n${context}`;\n\n  const result = await provider.completeJSON(prompt, TagsSchema);\n  return result.tags;\n}\n\n/**\n * Extract named entities from content\n */\nasync function extractEntities(context: string, provider: LLMProvider): Promise<ExtractedEntities> {\n  const prompt = `Extract named entities from the following content. Identify people, organizations, technologies, locations, and key concepts:\n\n${context}`;\n\n  return provider.completeJSON(prompt, EntitiesSchema);\n}\n\n/**\n * Classify content type using LLM\n */\nasync function classify(\n  context: string,\n  provider: LLMProvider\n): Promise<{ contentType: string; confidence: number }> {\n  const prompt = `Classify the following content into one of these categories:\n- article: Blog post, news article, essay\n- repo: Code repository, open source project\n- docs: Documentation, API reference, guides\n- package: npm/pip package page\n- video: Video content, YouTube\n- tool: Software tool, web application\n- product: Commercial product, e-commerce\n\n${context}`;\n\n  return provider.completeJSON(prompt, ClassifySchema);\n}\n"],"mappings":";;;;;;AAiBA,IAAa,cAAb,MAAa,oBAAoB,MAAM;CACrC,AAAgB;CAChB,AAAgB;CAEhB,YAAY,SAAiB,MAAuB,YAAqB,OAAe;AACtF,QAAM,SAAS,EAAE,OAAO,CAAC;AACzB,OAAK,OAAO;AACZ,OAAK,OAAO;AACZ,OAAK,aAAa;AAGlB,MAAI,MAAM,kBACR,OAAM,kBAAkB,MAAM,YAAY;;;;;CAO9C,OAAO,KAAK,OAAgB,OAAwB,gBAA6B;AAC/E,MAAI,iBAAiB,YACnB,QAAO;AAGT,MAAI,iBAAiB,MACnB,QAAO,IAAI,YAAY,MAAM,SAAS,MAAM,QAAW,MAAM;AAG/D,SAAO,IAAI,YAAY,OAAO,MAAM,EAAE,KAAK;;;;;CAM7C,cAAuB;AACrB,SAAO,KAAK,SAAS,kBAAkB,KAAK,SAAS;;;;;CAMvD,SAAkC;AAChC,SAAO;GACL,MAAM,KAAK;GACX,SAAS,KAAK;GACd,MAAM,KAAK;GACX,YAAY,KAAK;GACjB,OAAO,KAAK;GACb;;;;;;;;;ACSL,MAAa,gBAAgB,EAAE,OAAO,EACpC,SAAS,EAAE,QAAQ,CAAC,SAAS,gDAAgD,EAC9E,CAAC;AAEF,MAAa,aAAa,EAAE,OAAO,EACjC,MAAM,EAAE,MAAM,EAAE,QAAQ,CAAC,CAAC,SAAS,8BAA8B,EAClE,CAAC;AAEF,MAAa,iBAAiB,EAAE,OAAO;CACrC,QAAQ,EAAE,MAAM,EAAE,QAAQ,CAAC,CAAC,SAAS,mBAAmB;CACxD,eAAe,EAAE,MAAM,EAAE,QAAQ,CAAC,CAAC,SAAS,0BAA0B;CACtE,cAAc,EAAE,MAAM,EAAE,QAAQ,CAAC,CAAC,SAAS,gCAAgC;CAC3E,WAAW,EAAE,MAAM,EAAE,QAAQ,CAAC,CAAC,SAAS,mBAAmB;CAC3D,UAAU,EAAE,MAAM,EAAE,QAAQ,CAAC,CAAC,SAAS,sBAAsB;CAC9D,CAAC;AAEF,MAAa,iBAAiB,EAAE,OAAO;CACrC,aAAa,EACV,KAAK;EAAC;EAAW;EAAQ;EAAQ;EAAW;EAAS;EAAQ;EAAW;EAAU,CAAC,CACnF,SAAS,sBAAsB;CAClC,YAAY,EAAE,QAAQ,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC,SAAS,uBAAuB;CACtE,CAAC;;;;;;;AClFF,eAAsB,QACpB,MACA,UACA,OAC+B;CAC/B,MAAMA,UAAgC,EAAE;CAGxC,MAAM,UAAU,KAAK,WAAW,KAAK,YAAY,MAAM,GAAG,IAAM;CAChE,MAAM,UAAU,UAAU,KAAK,MAAM,SAAS,KAAK,IAAI,gBAAgB;CAGvE,MAAMC,WAA4B,EAAE;AAEpC,KAAI,MAAM,SAAS,YAAY,CAC7B,UAAS,KACP,UAAU,SAAS,SAAS,CAAC,MAAM,YAAY;AAC7C,UAAQ,UAAU;GAClB,CACH;AAGH,KAAI,MAAM,SAAS,OAAO,CACxB,UAAS,KACP,YAAY,SAAS,SAAS,CAAC,MAAM,SAAS;AAC5C,UAAQ,gBAAgB;GACxB,CACH;AAGH,KAAI,MAAM,SAAS,WAAW,CAC5B,UAAS,KACP,gBAAgB,SAAS,SAAS,CAAC,MAAM,aAAa;AACpD,UAAQ,WAAW;GACnB,CACH;AAGH,KAAI,MAAM,SAAS,WAAW,CAC5B,UAAS,KACP,SAAS,SAAS,SAAS,CAAC,MAAM,mBAAmB;AACnD,MAAI,eAAe,aAAa,GAC9B,SAAQ,cAAc,eAAe;GAEvC,CACH;AAGH,OAAM,QAAQ,IAAI,SAAS;AAE3B,QAAO;;;;;;AAiBT,eAAsB,IACpB,MACA,UACA,QACA,SAC+B;CAC/B,MAAM,MAAM,SAAS,OAAO;CAC5B,MAAM,UAAU,KAAK,WAAW,KAAK,YAAY,MAAM,GAAG,IAAM;CAGhE,MAAM,kBAAkB,kBAAkB,QAAQ,MAAM,QAAQ;AAEhE,KAAI,SAAS,QAAQ;EAEnB,MAAM,SAAS,MAAM,QAAQ,MAAM,UAAU,QAAQ,QAAQ,gBAAgB;AAC7E,SAAO,EAAE,QAAQ,GAAG,MAAM,QAAQ,EAAE;;CAItC,MAAM,aAAa,OAAO,SAAS,cAAc,GAC7C,kBACA,GAAG,gBAAgB,aAAa,KAAK,MAAM,SAAS,KAAK,IAAI,gBAAgB;CAEjF,MAAM,WAAW,MAAM,SAAS,SAAS,WAAW;AACpD,QAAO,EAAE,QAAQ,GAAG,MAAM,UAAU,EAAE;;;;;AAMxC,SAAS,kBAAkB,QAAgB,MAAmB,SAAyB;CACrF,MAAM,gBAAgB;AACpB,MAAI;AACF,UAAO,IAAI,IAAI,KAAK,IAAI,CAAC;UACnB;AACN,UAAO;;KAEP;AAEJ,QAAO,OACJ,QAAQ,kBAAkB,KAAK,MAAM,CACrC,QAAQ,gBAAgB,KAAK,IAAI,CACjC,QAAQ,oBAAoB,QAAQ,CACpC,QAAQ,wBAAwB,KAAK,eAAe,GAAG,CACvD,QAAQ,oBAAoB,KAAK,WAAW,GAAG,CAC/C,QAAQ,mBAAmB,OAAO;;;;;AAMvC,eAAsB,QACpB,MACA,UACA,QACA,gBACY;CAEZ,MAAMC,WAAyC,EAAE;AAEjD,MAAK,MAAM,CAAC,KAAK,SAAS,OAAO,QAAQ,OAAO,EAAE;EAChD,MAAM,aAAa,KAAK,SAAS,IAAI;EACrC,MAAM,WAAW,aAAa,KAAK,MAAM,GAAG,GAAG,GAAG;EAElD,IAAIC;AACJ,UAAQ,UAAR;GACE,KAAK;AACH,cAAU,EAAE,QAAQ;AACpB;GACF,KAAK;AACH,cAAU,EAAE,QAAQ;AACpB;GACF,KAAK;AACH,cAAU,EAAE,SAAS;AACrB;GACF,KAAK;AACH,cAAU,EAAE,MAAM,EAAE,QAAQ,CAAC;AAC7B;GACF,KAAK;AACH,cAAU,EAAE,MAAM,EAAE,QAAQ,CAAC;AAC7B;GACF,QACE,WAAU,EAAE,QAAQ;;AAGxB,WAAS,OAAO,aAAa,QAAQ,UAAU,GAAG;;CAGpD,MAAM,YAAY,EAAE,OAAO,SAAS;CAEpC,MAAM,UAAU,KAAK,YAAY,MAAM,GAAG,IAAK;CAE/C,IAAIC;AAEJ,KAAI,gBAAgB;AAElB,WAAS,kBAAkB,gBAAgB,MAAM,QAAQ;AAGzD,MAAI,CAAC,eAAe,SAAS,cAAc,CACzC,WAAU,iBAAiB;OAG7B,UAAS;;SAEJ,KAAK,MAAM;OACb,KAAK,IAAI;;;EAGd,QAAQ;;;EAGR,OAAO,QAAQ,OAAO,CACrB,KAAK,CAAC,KAAK,UAAU,KAAK,IAAI,IAAI,KAAK,GAAG,CAC1C,KAAK,KAAK;AAGX,QAAO,SAAS,aAAgB,QAAQ,UAA0B;;;;;AAMpE,eAAe,UAAU,SAAiB,UAAwC;CAChF,MAAM,SAAS;;EAEf;AAGA,SADe,MAAM,SAAS,aAAa,QAAQ,cAAc,EACnD;;;;;AAMhB,eAAe,YAAY,SAAiB,UAA0C;CACpF,MAAM,SAAS;;EAEf;AAGA,SADe,MAAM,SAAS,aAAa,QAAQ,WAAW,EAChD;;;;;AAMhB,eAAe,gBAAgB,SAAiB,UAAmD;CACjG,MAAM,SAAS;;EAEf;AAEA,QAAO,SAAS,aAAa,QAAQ,eAAe;;;;;AAMtD,eAAe,SACb,SACA,UACsD;CACtD,MAAM,SAAS;;;;;;;;;EASf;AAEA,QAAO,SAAS,aAAa,QAAQ,eAAe"}

package/dist/enhancer-oM4BhYYS.cjs

@@ -1,268 +0,0 @@
-const require_index = require('./index.cjs');
-let zod = require("zod");
-
-//#region src/core/errors.ts
-/**
-* Custom error class for scraping failures with structured error codes
-*/
-var ScrapeError = class ScrapeError extends Error {
-	code;
-	statusCode;
-	constructor(message, code, statusCode, cause) {
-		super(message, { cause });
-		this.name = "ScrapeError";
-		this.code = code;
-		this.statusCode = statusCode;
-		if (Error.captureStackTrace) Error.captureStackTrace(this, ScrapeError);
-	}
-	/**
-	* Create a ScrapeError from an unknown error
-	*/
-	static from(error, code = "FETCH_FAILED") {
-		if (error instanceof ScrapeError) return error;
-		if (error instanceof Error) return new ScrapeError(error.message, code, void 0, error);
-		return new ScrapeError(String(error), code);
-	}
-	/**
-	* Check if error is retryable (network issues, timeouts)
-	*/
-	isRetryable() {
-		return this.code === "FETCH_FAILED" || this.code === "TIMEOUT";
-	}
-	/**
-	* Convert to a plain object for serialization
-	*/
-	toJSON() {
-		return {
-			name: this.name,
-			message: this.message,
-			code: this.code,
-			statusCode: this.statusCode,
-			stack: this.stack
-		};
-	}
-};
-
-//#endregion
-//#region src/llm/types.ts
-/**
-* Zod schemas for LLM outputs
-*/
-const SummarySchema = zod.z.object({ summary: zod.z.string().describe("A concise 2-3 sentence summary of the content") });
-const TagsSchema = zod.z.object({ tags: zod.z.array(zod.z.string()).describe("5-10 relevant tags/keywords") });
-const EntitiesSchema = zod.z.object({
-	people: zod.z.array(zod.z.string()).describe("People mentioned"),
-	organizations: zod.z.array(zod.z.string()).describe("Organizations/companies"),
-	technologies: zod.z.array(zod.z.string()).describe("Technologies/tools/frameworks"),
-	locations: zod.z.array(zod.z.string()).describe("Locations/places"),
-	concepts: zod.z.array(zod.z.string()).describe("Key concepts/topics")
-});
-const ClassifySchema = zod.z.object({
-	contentType: zod.z.enum([
-		"article",
-		"repo",
-		"docs",
-		"package",
-		"video",
-		"tool",
-		"product",
-		"unknown"
-	]).describe("The type of content"),
-	confidence: zod.z.number().min(0).max(1).describe("Confidence score 0-1")
-});
-
-//#endregion
-//#region src/llm/enhancer.ts
-/**
-* Enhance scraped data with LLM-powered features
-*/
-async function enhance(data, provider, types) {
-	const results = {};
-	const content = data.excerpt || data.textContent.slice(0, 1e4);
-	const context = `Title: ${data.title}\nURL: ${data.url}\n\nContent:\n${content}`;
-	const promises = [];
-	if (types.includes("summarize")) promises.push(summarize(context, provider).then((summary) => {
-		results.summary = summary;
-	}));
-	if (types.includes("tags")) promises.push(extractTags(context, provider).then((tags) => {
-		results.suggestedTags = tags;
-	}));
-	if (types.includes("entities")) promises.push(extractEntities(context, provider).then((entities) => {
-		results.entities = entities;
-	}));
-	if (types.includes("classify")) promises.push(classify(context, provider).then((classification) => {
-		if (classification.confidence > .7) results.contentType = classification.contentType;
-	}));
-	await Promise.all(promises);
-	return results;
-}
-/**
-* Ask a custom question about the scraped content
-* Results are stored in the `custom` field of ScrapedData
-*/
-async function ask(data, provider, prompt, options) {
-	const key = options?.key || "response";
-	const content = data.excerpt || data.textContent.slice(0, 1e4);
-	const processedPrompt = applyPlaceholders(prompt, data, content);
-	if (options?.schema) {
-		const result = await extract(data, provider, options.schema, processedPrompt);
-		return { custom: { [key]: result } };
-	}
-	const fullPrompt = prompt.includes("{{content}}") ? processedPrompt : `${processedPrompt}\n\nTitle: ${data.title}\nURL: ${data.url}\n\nContent:\n${content}`;
-	const response = await provider.complete(fullPrompt);
-	return { custom: { [key]: response } };
-}
-/**
-* Apply placeholder replacements to a prompt template
-*/
-function applyPlaceholders(prompt, data, content) {
-	const domain = (() => {
-		try {
-			return new URL(data.url).hostname;
-		} catch {
-			return "";
-		}
-	})();
-	return prompt.replace(/\{\{title\}\}/g, data.title).replace(/\{\{url\}\}/g, data.url).replace(/\{\{content\}\}/g, content).replace(/\{\{description\}\}/g, data.description || "").replace(/\{\{excerpt\}\}/g, data.excerpt || "").replace(/\{\{domain\}\}/g, domain);
-}
-/**
-* Extract structured data using LLM and a custom schema
-*/
-async function extract(data, provider, schema, promptTemplate) {
-	const zodShape = {};
-	for (const [key, type] of Object.entries(schema)) {
-		const isOptional = type.endsWith("?");
-		const baseType = isOptional ? type.slice(0, -1) : type;
-		let zodType;
-		switch (baseType) {
-			case "string":
-				zodType = zod.z.string();
-				break;
-			case "number":
-				zodType = zod.z.number();
-				break;
-			case "boolean":
-				zodType = zod.z.boolean();
-				break;
-			case "string[]":
-				zodType = zod.z.array(zod.z.string());
-				break;
-			case "number[]":
-				zodType = zod.z.array(zod.z.number());
-				break;
-			default: zodType = zod.z.string();
-		}
-		zodShape[key] = isOptional ? zodType.optional() : zodType;
-	}
-	const zodSchema = zod.z.object(zodShape);
-	const content = data.textContent.slice(0, 4e3);
-	let prompt;
-	if (promptTemplate) {
-		prompt = applyPlaceholders(promptTemplate, data, content);
-		if (!promptTemplate.includes("{{content}}")) prompt += `\n\nContext:\n${content}`;
-	} else prompt = `Extract the following information from this content:
-
-Title: ${data.title}
-URL: ${data.url}
-
-Content:
-${content}
-
-Extract these fields:
-${Object.entries(schema).map(([key, type]) => `- ${key} (${type})`).join("\n")}`;
-	return provider.completeJSON(prompt, zodSchema);
-}
-/**
-* Generate a summary of the content
-*/
-async function summarize(context, provider) {
-	const prompt = `Summarize the following content in 2-3 concise sentences:
-
-${context}`;
-	return (await provider.completeJSON(prompt, SummarySchema)).summary;
-}
-/**
-* Extract relevant tags/keywords
-*/
-async function extractTags(context, provider) {
-	const prompt = `Extract 5-10 relevant tags or keywords from the following content. Focus on technologies, concepts, and topics mentioned:
-
-${context}`;
-	return (await provider.completeJSON(prompt, TagsSchema)).tags;
-}
-/**
-* Extract named entities from content
-*/
-async function extractEntities(context, provider) {
-	const prompt = `Extract named entities from the following content. Identify people, organizations, technologies, locations, and key concepts:
-
-${context}`;
-	return provider.completeJSON(prompt, EntitiesSchema);
-}
-/**
-* Classify content type using LLM
-*/
-async function classify(context, provider) {
-	const prompt = `Classify the following content into one of these categories:
-- article: Blog post, news article, essay
-- repo: Code repository, open source project
-- docs: Documentation, API reference, guides
-- package: npm/pip package page
-- video: Video content, YouTube
-- tool: Software tool, web application
-- product: Commercial product, e-commerce
-
-${context}`;
-	return provider.completeJSON(prompt, ClassifySchema);
-}
-
-//#endregion
-Object.defineProperty(exports, 'ClassifySchema', {
-  enumerable: true,
-  get: function () {
-    return ClassifySchema;
-  }
-});
-Object.defineProperty(exports, 'EntitiesSchema', {
-  enumerable: true,
-  get: function () {
-    return EntitiesSchema;
-  }
-});
-Object.defineProperty(exports, 'ScrapeError', {
-  enumerable: true,
-  get: function () {
-    return ScrapeError;
-  }
-});
-Object.defineProperty(exports, 'SummarySchema', {
-  enumerable: true,
-  get: function () {
-    return SummarySchema;
-  }
-});
-Object.defineProperty(exports, 'TagsSchema', {
-  enumerable: true,
-  get: function () {
-    return TagsSchema;
-  }
-});
-Object.defineProperty(exports, 'ask', {
-  enumerable: true,
-  get: function () {
-    return ask;
-  }
-});
-Object.defineProperty(exports, 'enhance', {
-  enumerable: true,
-  get: function () {
-    return enhance;
-  }
-});
-Object.defineProperty(exports, 'extract', {
-  enumerable: true,
-  get: function () {
-    return extract;
-  }
-});
-//# sourceMappingURL=enhancer-oM4BhYYS.cjs.map

package/dist/enhancer-oM4BhYYS.cjs.map

@@ -1 +0,0 @@
-{"version":3,"file":"enhancer-oM4BhYYS.cjs","names":["z","results: Partial<ScrapedData>","promises: Promise<void>[]","zodShape: Record<string, z.ZodTypeAny>","zodType: z.ZodTypeAny","z","prompt: string"],"sources":["../src/core/errors.ts","../src/llm/types.ts","../src/llm/enhancer.ts"],"sourcesContent":["/**\n * Error codes for scraping failures\n */\nexport type ScrapeErrorCode =\n  | 'FETCH_FAILED'\n  | 'TIMEOUT'\n  | 'INVALID_URL'\n  | 'BLOCKED'\n  | 'NOT_FOUND'\n  | 'ROBOTS_BLOCKED'\n  | 'PARSE_ERROR'\n  | 'LLM_ERROR'\n  | 'VALIDATION_ERROR';\n\n/**\n * Custom error class for scraping failures with structured error codes\n */\nexport class ScrapeError extends Error {\n  public readonly code: ScrapeErrorCode;\n  public readonly statusCode?: number;\n\n  constructor(message: string, code: ScrapeErrorCode, statusCode?: number, cause?: Error) {\n    super(message, { cause });\n    this.name = 'ScrapeError';\n    this.code = code;\n    this.statusCode = statusCode;\n\n    // Maintains proper stack trace in V8 environments\n    if (Error.captureStackTrace) {\n      Error.captureStackTrace(this, ScrapeError);\n    }\n  }\n\n  /**\n   * Create a ScrapeError from an unknown error\n   */\n  static from(error: unknown, code: ScrapeErrorCode = 'FETCH_FAILED'): ScrapeError {\n    if (error instanceof ScrapeError) {\n      return error;\n    }\n\n    if (error instanceof Error) {\n      return new ScrapeError(error.message, code, undefined, error);\n    }\n\n    return new ScrapeError(String(error), code);\n  }\n\n  /**\n   * Check if error is retryable (network issues, timeouts)\n   */\n  isRetryable(): boolean {\n    return this.code === 'FETCH_FAILED' || this.code === 'TIMEOUT';\n  }\n\n  /**\n   * Convert to a plain object for serialization\n   */\n  toJSON(): Record<string, unknown> {\n    return {\n      name: this.name,\n      message: this.message,\n      code: this.code,\n      statusCode: this.statusCode,\n      stack: this.stack,\n    };\n  }\n}\n","import { z } from 'zod';\n\n/**\n * LLM completion options\n */\nexport interface CompletionOptions {\n  maxTokens?: number;\n  temperature?: number;\n  systemPrompt?: string;\n}\n\n/**\n * LLM Provider interface - implemented by all providers\n */\nexport interface LLMProvider {\n  readonly name: string;\n\n  /**\n   * Generate a text completion\n   */\n  complete(prompt: string, options?: CompletionOptions): Promise<string>;\n\n  /**\n   * Generate a structured JSON completion with Zod validation\n   */\n  completeJSON<T>(prompt: string, schema: z.ZodType<T>, options?: CompletionOptions): Promise<T>;\n}\n\n/**\n * Provider configuration for Anthropic\n */\nexport interface AnthropicConfig {\n  apiKey?: string; // Falls back to ANTHROPIC_API_KEY env var\n  model?: string; // Default: claude-3-haiku-20240307\n  baseUrl?: string;\n}\n\n/**\n * Provider configuration for OpenAI-compatible APIs\n * Works with: OpenAI, Ollama, LM Studio, LocalAI, vLLM, etc.\n */\nexport interface OpenAICompatibleConfig {\n  apiKey?: string; // Falls back to OPENAI_API_KEY env var\n  model?: string; // Default: gpt-4o-mini\n  baseUrl?: string; // Default: https://api.openai.com/v1\n}\n\n/**\n * Enhancement result types\n */\nexport interface SummaryResult {\n  summary: string;\n}\n\nexport interface TagsResult {\n  tags: string[];\n}\n\nexport interface EntitiesResult {\n  people: string[];\n  organizations: string[];\n  technologies: string[];\n  locations: string[];\n  concepts: string[];\n}\n\nexport interface ClassifyResult {\n  contentType: string;\n  confidence: number;\n}\n\n/**\n * Zod schemas for LLM outputs\n */\nexport const SummarySchema = z.object({\n  summary: z.string().describe('A concise 2-3 sentence summary of the content'),\n});\n\nexport const TagsSchema = z.object({\n  tags: z.array(z.string()).describe('5-10 relevant tags/keywords'),\n});\n\nexport const EntitiesSchema = z.object({\n  people: z.array(z.string()).describe('People mentioned'),\n  organizations: z.array(z.string()).describe('Organizations/companies'),\n  technologies: z.array(z.string()).describe('Technologies/tools/frameworks'),\n  locations: z.array(z.string()).describe('Locations/places'),\n  concepts: z.array(z.string()).describe('Key concepts/topics'),\n});\n\nexport const ClassifySchema = z.object({\n  contentType: z\n    .enum(['article', 'repo', 'docs', 'package', 'video', 'tool', 'product', 'unknown'])\n    .describe('The type of content'),\n  confidence: z.number().min(0).max(1).describe('Confidence score 0-1'),\n});\n","import { z } from 'zod';\nimport type {\n  EnhancementType,\n  ExtractedEntities,\n  ExtractionSchema,\n  ScrapedData,\n} from '@/core/types.js';\nimport type { LLMProvider } from './types.js';\nimport { ClassifySchema, EntitiesSchema, SummarySchema, TagsSchema } from './types.js';\n\n/**\n * Enhance scraped data with LLM-powered features\n */\nexport async function enhance(\n  data: ScrapedData,\n  provider: LLMProvider,\n  types: EnhancementType[]\n): Promise<Partial<ScrapedData>> {\n  const results: Partial<ScrapedData> = {};\n\n  // Prepare content for LLM (use excerpt/textContent to save tokens)\n  const content = data.excerpt || data.textContent.slice(0, 10000);\n  const context = `Title: ${data.title}\\nURL: ${data.url}\\n\\nContent:\\n${content}`;\n\n  // Run enhancements in parallel\n  const promises: Promise<void>[] = [];\n\n  if (types.includes('summarize')) {\n    promises.push(\n      summarize(context, provider).then((summary) => {\n        results.summary = summary;\n      })\n    );\n  }\n\n  if (types.includes('tags')) {\n    promises.push(\n      extractTags(context, provider).then((tags) => {\n        results.suggestedTags = tags;\n      })\n    );\n  }\n\n  if (types.includes('entities')) {\n    promises.push(\n      extractEntities(context, provider).then((entities) => {\n        results.entities = entities;\n      })\n    );\n  }\n\n  if (types.includes('classify')) {\n    promises.push(\n      classify(context, provider).then((classification) => {\n        if (classification.confidence > 0.7) {\n          results.contentType = classification.contentType as ScrapedData['contentType'];\n        }\n      })\n    );\n  }\n\n  await Promise.all(promises);\n\n  return results;\n}\n\n/**\n * Options for the ask() function\n */\nexport interface AskOptions {\n  /** Key to store the result under in custom field */\n  key?: string;\n  /** Schema for structured response */\n  schema?: ExtractionSchema;\n}\n\n/**\n * Ask a custom question about the scraped content\n * Results are stored in the `custom` field of ScrapedData\n */\nexport async function ask(\n  data: ScrapedData,\n  provider: LLMProvider,\n  prompt: string,\n  options?: AskOptions\n): Promise<Partial<ScrapedData>> {\n  const key = options?.key || 'response';\n  const content = data.excerpt || data.textContent.slice(0, 10000);\n\n  // Apply placeholder replacements\n  const processedPrompt = applyPlaceholders(prompt, data, content);\n\n  if (options?.schema) {\n    // Use structured extraction\n    const result = await extract(data, provider, options.schema, processedPrompt);\n    return { custom: { [key]: result } };\n  }\n\n  // Simple string response\n  const fullPrompt = prompt.includes('{{content}}')\n    ? processedPrompt\n    : `${processedPrompt}\\n\\nTitle: ${data.title}\\nURL: ${data.url}\\n\\nContent:\\n${content}`;\n\n  const response = await provider.complete(fullPrompt);\n  return { custom: { [key]: response } };\n}\n\n/**\n * Apply placeholder replacements to a prompt template\n */\nfunction applyPlaceholders(prompt: string, data: ScrapedData, content: string): string {\n  const domain = (() => {\n    try {\n      return new URL(data.url).hostname;\n    } catch {\n      return '';\n    }\n  })();\n\n  return prompt\n    .replace(/\\{\\{title\\}\\}/g, data.title)\n    .replace(/\\{\\{url\\}\\}/g, data.url)\n    .replace(/\\{\\{content\\}\\}/g, content)\n    .replace(/\\{\\{description\\}\\}/g, data.description || '')\n    .replace(/\\{\\{excerpt\\}\\}/g, data.excerpt || '')\n    .replace(/\\{\\{domain\\}\\}/g, domain);\n}\n\n/**\n * Extract structured data using LLM and a custom schema\n */\nexport async function extract<T>(\n  data: ScrapedData,\n  provider: LLMProvider,\n  schema: ExtractionSchema,\n  promptTemplate?: string\n): Promise<T> {\n  // Convert simple schema to Zod schema\n  const zodShape: Record<string, z.ZodTypeAny> = {};\n\n  for (const [key, type] of Object.entries(schema)) {\n    const isOptional = type.endsWith('?');\n    const baseType = isOptional ? type.slice(0, -1) : type;\n\n    let zodType: z.ZodTypeAny;\n    switch (baseType) {\n      case 'string':\n        zodType = z.string();\n        break;\n      case 'number':\n        zodType = z.number();\n        break;\n      case 'boolean':\n        zodType = z.boolean();\n        break;\n      case 'string[]':\n        zodType = z.array(z.string());\n        break;\n      case 'number[]':\n        zodType = z.array(z.number());\n        break;\n      default:\n        zodType = z.string();\n    }\n\n    zodShape[key] = isOptional ? zodType.optional() : zodType;\n  }\n\n  const zodSchema = z.object(zodShape) as unknown as z.ZodType<T>;\n\n  const content = data.textContent.slice(0, 4000);\n\n  let prompt: string;\n\n  if (promptTemplate) {\n    // Apply all placeholder replacements\n    prompt = applyPlaceholders(promptTemplate, data, content);\n\n    // If content wasn't included via placeholder, append it\n    if (!promptTemplate.includes('{{content}}')) {\n      prompt += `\\n\\nContext:\\n${content}`;\n    }\n  } else {\n    prompt = `Extract the following information from this content:\n\nTitle: ${data.title}\nURL: ${data.url}\n\nContent:\n${content}\n\nExtract these fields:\n${Object.entries(schema)\n  .map(([key, type]) => `- ${key} (${type})`)\n  .join('\\n')}`;\n  }\n\n  return provider.completeJSON<T>(prompt, zodSchema as z.ZodType<T>);\n}\n\n/**\n * Generate a summary of the content\n */\nasync function summarize(context: string, provider: LLMProvider): Promise<string> {\n  const prompt = `Summarize the following content in 2-3 concise sentences:\n\n${context}`;\n\n  const result = await provider.completeJSON(prompt, SummarySchema);\n  return result.summary;\n}\n\n/**\n * Extract relevant tags/keywords\n */\nasync function extractTags(context: string, provider: LLMProvider): Promise<string[]> {\n  const prompt = `Extract 5-10 relevant tags or keywords from the following content. Focus on technologies, concepts, and topics mentioned:\n\n${context}`;\n\n  const result = await provider.completeJSON(prompt, TagsSchema);\n  return result.tags;\n}\n\n/**\n * Extract named entities from content\n */\nasync function extractEntities(context: string, provider: LLMProvider): Promise<ExtractedEntities> {\n  const prompt = `Extract named entities from the following content. Identify people, organizations, technologies, locations, and key concepts:\n\n${context}`;\n\n  return provider.completeJSON(prompt, EntitiesSchema);\n}\n\n/**\n * Classify content type using LLM\n */\nasync function classify(\n  context: string,\n  provider: LLMProvider\n): Promise<{ contentType: string; confidence: number }> {\n  const prompt = `Classify the following content into one of these categories:\n- article: Blog post, news article, essay\n- repo: Code repository, open source project\n- docs: Documentation, API reference, guides\n- package: npm/pip package page\n- video: Video content, YouTube\n- tool: Software tool, web application\n- product: Commercial product, e-commerce\n\n${context}`;\n\n  return provider.completeJSON(prompt, ClassifySchema);\n}\n"],"mappings":";;;;;;;AAiBA,IAAa,cAAb,MAAa,oBAAoB,MAAM;CACrC,AAAgB;CAChB,AAAgB;CAEhB,YAAY,SAAiB,MAAuB,YAAqB,OAAe;AACtF,QAAM,SAAS,EAAE,OAAO,CAAC;AACzB,OAAK,OAAO;AACZ,OAAK,OAAO;AACZ,OAAK,aAAa;AAGlB,MAAI,MAAM,kBACR,OAAM,kBAAkB,MAAM,YAAY;;;;;CAO9C,OAAO,KAAK,OAAgB,OAAwB,gBAA6B;AAC/E,MAAI,iBAAiB,YACnB,QAAO;AAGT,MAAI,iBAAiB,MACnB,QAAO,IAAI,YAAY,MAAM,SAAS,MAAM,QAAW,MAAM;AAG/D,SAAO,IAAI,YAAY,OAAO,MAAM,EAAE,KAAK;;;;;CAM7C,cAAuB;AACrB,SAAO,KAAK,SAAS,kBAAkB,KAAK,SAAS;;;;;CAMvD,SAAkC;AAChC,SAAO;GACL,MAAM,KAAK;GACX,SAAS,KAAK;GACd,MAAM,KAAK;GACX,YAAY,KAAK;GACjB,OAAO,KAAK;GACb;;;;;;;;;ACSL,MAAa,gBAAgBA,MAAE,OAAO,EACpC,SAASA,MAAE,QAAQ,CAAC,SAAS,gDAAgD,EAC9E,CAAC;AAEF,MAAa,aAAaA,MAAE,OAAO,EACjC,MAAMA,MAAE,MAAMA,MAAE,QAAQ,CAAC,CAAC,SAAS,8BAA8B,EAClE,CAAC;AAEF,MAAa,iBAAiBA,MAAE,OAAO;CACrC,QAAQA,MAAE,MAAMA,MAAE,QAAQ,CAAC,CAAC,SAAS,mBAAmB;CACxD,eAAeA,MAAE,MAAMA,MAAE,QAAQ,CAAC,CAAC,SAAS,0BAA0B;CACtE,cAAcA,MAAE,MAAMA,MAAE,QAAQ,CAAC,CAAC,SAAS,gCAAgC;CAC3E,WAAWA,MAAE,MAAMA,MAAE,QAAQ,CAAC,CAAC,SAAS,mBAAmB;CAC3D,UAAUA,MAAE,MAAMA,MAAE,QAAQ,CAAC,CAAC,SAAS,sBAAsB;CAC9D,CAAC;AAEF,MAAa,iBAAiBA,MAAE,OAAO;CACrC,aAAaA,MACV,KAAK;EAAC;EAAW;EAAQ;EAAQ;EAAW;EAAS;EAAQ;EAAW;EAAU,CAAC,CACnF,SAAS,sBAAsB;CAClC,YAAYA,MAAE,QAAQ,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC,SAAS,uBAAuB;CACtE,CAAC;;;;;;;AClFF,eAAsB,QACpB,MACA,UACA,OAC+B;CAC/B,MAAMC,UAAgC,EAAE;CAGxC,MAAM,UAAU,KAAK,WAAW,KAAK,YAAY,MAAM,GAAG,IAAM;CAChE,MAAM,UAAU,UAAU,KAAK,MAAM,SAAS,KAAK,IAAI,gBAAgB;CAGvE,MAAMC,WAA4B,EAAE;AAEpC,KAAI,MAAM,SAAS,YAAY,CAC7B,UAAS,KACP,UAAU,SAAS,SAAS,CAAC,MAAM,YAAY;AAC7C,UAAQ,UAAU;GAClB,CACH;AAGH,KAAI,MAAM,SAAS,OAAO,CACxB,UAAS,KACP,YAAY,SAAS,SAAS,CAAC,MAAM,SAAS;AAC5C,UAAQ,gBAAgB;GACxB,CACH;AAGH,KAAI,MAAM,SAAS,WAAW,CAC5B,UAAS,KACP,gBAAgB,SAAS,SAAS,CAAC,MAAM,aAAa;AACpD,UAAQ,WAAW;GACnB,CACH;AAGH,KAAI,MAAM,SAAS,WAAW,CAC5B,UAAS,KACP,SAAS,SAAS,SAAS,CAAC,MAAM,mBAAmB;AACnD,MAAI,eAAe,aAAa,GAC9B,SAAQ,cAAc,eAAe;GAEvC,CACH;AAGH,OAAM,QAAQ,IAAI,SAAS;AAE3B,QAAO;;;;;;AAiBT,eAAsB,IACpB,MACA,UACA,QACA,SAC+B;CAC/B,MAAM,MAAM,SAAS,OAAO;CAC5B,MAAM,UAAU,KAAK,WAAW,KAAK,YAAY,MAAM,GAAG,IAAM;CAGhE,MAAM,kBAAkB,kBAAkB,QAAQ,MAAM,QAAQ;AAEhE,KAAI,SAAS,QAAQ;EAEnB,MAAM,SAAS,MAAM,QAAQ,MAAM,UAAU,QAAQ,QAAQ,gBAAgB;AAC7E,SAAO,EAAE,QAAQ,GAAG,MAAM,QAAQ,EAAE;;CAItC,MAAM,aAAa,OAAO,SAAS,cAAc,GAC7C,kBACA,GAAG,gBAAgB,aAAa,KAAK,MAAM,SAAS,KAAK,IAAI,gBAAgB;CAEjF,MAAM,WAAW,MAAM,SAAS,SAAS,WAAW;AACpD,QAAO,EAAE,QAAQ,GAAG,MAAM,UAAU,EAAE;;;;;AAMxC,SAAS,kBAAkB,QAAgB,MAAmB,SAAyB;CACrF,MAAM,gBAAgB;AACpB,MAAI;AACF,UAAO,IAAI,IAAI,KAAK,IAAI,CAAC;UACnB;AACN,UAAO;;KAEP;AAEJ,QAAO,OACJ,QAAQ,kBAAkB,KAAK,MAAM,CACrC,QAAQ,gBAAgB,KAAK,IAAI,CACjC,QAAQ,oBAAoB,QAAQ,CACpC,QAAQ,wBAAwB,KAAK,eAAe,GAAG,CACvD,QAAQ,oBAAoB,KAAK,WAAW,GAAG,CAC/C,QAAQ,mBAAmB,OAAO;;;;;AAMvC,eAAsB,QACpB,MACA,UACA,QACA,gBACY;CAEZ,MAAMC,WAAyC,EAAE;AAEjD,MAAK,MAAM,CAAC,KAAK,SAAS,OAAO,QAAQ,OAAO,EAAE;EAChD,MAAM,aAAa,KAAK,SAAS,IAAI;EACrC,MAAM,WAAW,aAAa,KAAK,MAAM,GAAG,GAAG,GAAG;EAElD,IAAIC;AACJ,UAAQ,UAAR;GACE,KAAK;AACH,cAAUC,MAAE,QAAQ;AACpB;GACF,KAAK;AACH,cAAUA,MAAE,QAAQ;AACpB;GACF,KAAK;AACH,cAAUA,MAAE,SAAS;AACrB;GACF,KAAK;AACH,cAAUA,MAAE,MAAMA,MAAE,QAAQ,CAAC;AAC7B;GACF,KAAK;AACH,cAAUA,MAAE,MAAMA,MAAE,QAAQ,CAAC;AAC7B;GACF,QACE,WAAUA,MAAE,QAAQ;;AAGxB,WAAS,OAAO,aAAa,QAAQ,UAAU,GAAG;;CAGpD,MAAM,YAAYA,MAAE,OAAO,SAAS;CAEpC,MAAM,UAAU,KAAK,YAAY,MAAM,GAAG,IAAK;CAE/C,IAAIC;AAEJ,KAAI,gBAAgB;AAElB,WAAS,kBAAkB,gBAAgB,MAAM,QAAQ;AAGzD,MAAI,CAAC,eAAe,SAAS,cAAc,CACzC,WAAU,iBAAiB;OAG7B,UAAS;;SAEJ,KAAK,MAAM;OACb,KAAK,IAAI;;;EAGd,QAAQ;;;EAGR,OAAO,QAAQ,OAAO,CACrB,KAAK,CAAC,KAAK,UAAU,KAAK,IAAI,IAAI,KAAK,GAAG,CAC1C,KAAK,KAAK;AAGX,QAAO,SAAS,aAAgB,QAAQ,UAA0B;;;;;AAMpE,eAAe,UAAU,SAAiB,UAAwC;CAChF,MAAM,SAAS;;EAEf;AAGA,SADe,MAAM,SAAS,aAAa,QAAQ,cAAc,EACnD;;;;;AAMhB,eAAe,YAAY,SAAiB,UAA0C;CACpF,MAAM,SAAS;;EAEf;AAGA,SADe,MAAM,SAAS,aAAa,QAAQ,WAAW,EAChD;;;;;AAMhB,eAAe,gBAAgB,SAAiB,UAAmD;CACjG,MAAM,SAAS;;EAEf;AAEA,QAAO,SAAS,aAAa,QAAQ,eAAe;;;;;AAMtD,eAAe,SACb,SACA,UACsD;CACtD,MAAM,SAAS;;;;;;;;;EASf;AAEA,QAAO,SAAS,aAAa,QAAQ,eAAe"}

package/dist/parsers/index.cjs.map

@@ -1 +0,0 @@
-{"version":3,"file":"index.cjs","names":["sections: MarkdownSection[]","allLinks: MarkdownLink[]","codeBlocks: CodeBlock[]","frontmatter: Record<string, unknown> | undefined","currentSection: MarkdownSection | null","linkData: MarkdownLink","result: Record<string, unknown>","value: string | boolean | number","links: MarkdownLink[]"],"sources":["../../src/parsers/github.ts","../../src/parsers/markdown.ts"],"sourcesContent":["import type { GitHubMeta, MarkdownLink } from './types.js';\n\n/**\n * GitHub-specific utilities for parsing repositories.\n */\n\n/**\n * Check if a URL is a GitHub repository\n */\nexport function isGitHubRepo(url: string): boolean {\n  return /^https?:\\/\\/(www\\.)?github\\.com\\/[^/]+\\/[^/]+\\/?$/.test(url);\n}\n\n/**\n * Extract GitHub repo info from URL\n */\nexport function parseGitHubUrl(url: string): { owner: string; repo: string } | null {\n  const match = url.match(/github\\.com\\/([^/]+)\\/([^/]+)/);\n  if (!match || !match[1] || !match[2]) return null;\n  return {\n    owner: match[1],\n    repo: match[2].replace(/\\.git$/, ''),\n  };\n}\n\n/**\n * Convert a GitHub repo URL to raw content URL\n */\nexport function toRawUrl(url: string, branch = 'main', file = 'README.md'): string {\n  const info = parseGitHubUrl(url);\n  if (!info) return url;\n  return `https://raw.githubusercontent.com/${info.owner}/${info.repo}/${branch}/${file}`;\n}\n\n/**\n * Fetch GitHub API metadata for a repository\n * Note: This is a placeholder - actual implementation would need GitHub API access\n */\nexport async function fetchRepoMeta(\n  owner: string,\n  repo: string,\n  _token?: string\n): Promise<GitHubMeta> {\n  // This would make actual API calls in a full implementation\n  // For now, return basic info\n  return {\n    repoOwner: owner,\n    repoName: repo,\n  };\n}\n\n/**\n * Group links by their category/section\n */\nexport function groupByCategory(links: MarkdownLink[]): Map<string, MarkdownLink[]> {\n  const groups = new Map<string, MarkdownLink[]>();\n\n  for (const link of links) {\n    const category = link.context || 'Uncategorized';\n    const existing = groups.get(category) || [];\n    existing.push(link);\n    groups.set(category, existing);\n  }\n\n  return groups;\n}\n","import type { Code, Heading, Link, ListItem, Root } from 'mdast';\nimport { fromMarkdown } from 'mdast-util-from-markdown';\nimport { toString as mdastToString } from 'mdast-util-to-string';\nimport { visit } from 'unist-util-visit';\nimport type {\n  CodeBlock,\n  MarkdownLink,\n  MarkdownSection,\n  ParsedMarkdown,\n  ParserResult,\n  SourceParser,\n} from './types.js';\n\n/**\n * Generic Markdown parser.\n * Extracts structure, links, and code blocks from markdown content.\n *\n * @example\n * ```ts\n * const parser = new MarkdownParser();\n * const result = parser.parse(markdownContent);\n * console.log(result.data.sections);\n * console.log(result.data.links);\n * ```\n */\nexport class MarkdownParser implements SourceParser<ParsedMarkdown> {\n  readonly name = 'markdown';\n\n  canParse(content: string): boolean {\n    // Check for common markdown patterns\n    return (\n      content.includes('# ') ||\n      content.includes('## ') ||\n      content.includes('- [') ||\n      content.includes('* [') ||\n      content.includes('```')\n    );\n  }\n\n  parse(content: string): ParserResult<ParsedMarkdown> {\n    const tree = fromMarkdown(content);\n    const sections: MarkdownSection[] = [];\n    const allLinks: MarkdownLink[] = [];\n    const codeBlocks: CodeBlock[] = [];\n    let frontmatter: Record<string, unknown> | undefined;\n\n    // Extract frontmatter if present\n    if (content.startsWith('---')) {\n      const endIndex = content.indexOf('---', 3);\n      if (endIndex !== -1) {\n        const frontmatterContent = content.slice(3, endIndex).trim();\n        frontmatter = this.parseFrontmatter(frontmatterContent);\n      }\n    }\n\n    // Track current section\n    let currentSection: MarkdownSection | null = null;\n\n    // Process the AST\n    visit(tree, (node) => {\n      // Handle headings\n      if (node.type === 'heading') {\n        const heading = node as Heading;\n        const title = mdastToString(heading);\n\n        // Finalize previous section\n        if (currentSection) {\n          sections.push(currentSection);\n        }\n\n        currentSection = {\n          level: heading.depth,\n          title,\n          content: '',\n          links: [],\n        };\n      }\n\n      // Handle links\n      if (node.type === 'link') {\n        const link = node as Link;\n        const text = mdastToString(link);\n        const linkData: MarkdownLink = {\n          url: link.url,\n          text,\n          title: link.title ?? undefined,\n          context: currentSection?.title,\n        };\n\n        allLinks.push(linkData);\n        if (currentSection) {\n          currentSection.links.push(linkData);\n        }\n      }\n\n      // Handle code blocks\n      if (node.type === 'code') {\n        const code = node as Code;\n        codeBlocks.push({\n          language: code.lang ?? undefined,\n          code: code.value,\n          meta: code.meta ?? undefined,\n        });\n      }\n\n      // Accumulate content for current section\n      if (currentSection && node.type === 'paragraph') {\n        const text = mdastToString(node);\n        currentSection.content += (currentSection.content ? '\\n\\n' : '') + text;\n      }\n    });\n\n    // Finalize last section\n    if (currentSection) {\n      sections.push(currentSection);\n    }\n\n    // Extract title from first h1 or frontmatter\n    const title = (frontmatter?.title as string) ?? sections.find((s) => s.level === 1)?.title;\n\n    // Extract description from frontmatter or first paragraph before any heading\n    const description = (frontmatter?.description as string) ?? this.extractDescription(tree);\n\n    return {\n      data: {\n        title,\n        description,\n        sections,\n        links: allLinks,\n        codeBlocks,\n        frontmatter,\n      },\n    };\n  }\n\n  private parseFrontmatter(content: string): Record<string, unknown> {\n    const result: Record<string, unknown> = {};\n    const lines = content.split('\\n');\n\n    for (const line of lines) {\n      const colonIndex = line.indexOf(':');\n      if (colonIndex > 0) {\n        const key = line.slice(0, colonIndex).trim();\n        let value: string | boolean | number = line.slice(colonIndex + 1).trim();\n\n        // Parse simple types\n        if (value === 'true') value = true;\n        else if (value === 'false') value = false;\n        else if (/^-?\\d+(\\.\\d+)?$/.test(value)) value = Number(value);\n        else if (value.startsWith('\"') && value.endsWith('\"')) value = value.slice(1, -1);\n        else if (value.startsWith(\"'\") && value.endsWith(\"'\")) value = value.slice(1, -1);\n\n        result[key] = value;\n      }\n    }\n\n    return result;\n  }\n\n  private extractDescription(tree: Root): string | undefined {\n    // Find first paragraph before any heading\n    for (const node of tree.children) {\n      if (node.type === 'heading') break;\n      if (node.type === 'paragraph') {\n        return mdastToString(node);\n      }\n    }\n    return undefined;\n  }\n}\n\n/**\n * Extract links from a list-based markdown structure (like awesome lists)\n */\nexport function extractListLinks(markdown: string): MarkdownLink[] {\n  const tree = fromMarkdown(markdown);\n  const links: MarkdownLink[] = [];\n  let currentHeading = '';\n\n  visit(tree, (node) => {\n    if (node.type === 'heading') {\n      currentHeading = mdastToString(node as Heading);\n    }\n\n    if (node.type === 'listItem') {\n      const listItem = node as ListItem;\n\n      // Find links in this list item\n      visit(listItem, 'link', (linkNode: Link) => {\n        links.push({\n          url: linkNode.url,\n          text: mdastToString(linkNode),\n          title: linkNode.title ?? undefined,\n          context: currentHeading || undefined,\n        });\n      });\n    }\n  });\n\n  return links;\n}\n\n/**\n * Parse markdown into sections by heading level\n */\nexport function parseByHeadings(markdown: string, minLevel = 2): MarkdownSection[] {\n  const parser = new MarkdownParser();\n  const result = parser.parse(markdown);\n  return result.data.sections.filter((s) => s.level >= minLevel);\n}\n"],"mappings":";;;;;;;;;;;;AASA,SAAgB,aAAa,KAAsB;AACjD,QAAO,oDAAoD,KAAK,IAAI;;;;;AAMtE,SAAgB,eAAe,KAAqD;CAClF,MAAM,QAAQ,IAAI,MAAM,gCAAgC;AACxD,KAAI,CAAC,SAAS,CAAC,MAAM,MAAM,CAAC,MAAM,GAAI,QAAO;AAC7C,QAAO;EACL,OAAO,MAAM;EACb,MAAM,MAAM,GAAG,QAAQ,UAAU,GAAG;EACrC;;;;;AAMH,SAAgB,SAAS,KAAa,SAAS,QAAQ,OAAO,aAAqB;CACjF,MAAM,OAAO,eAAe,IAAI;AAChC,KAAI,CAAC,KAAM,QAAO;AAClB,QAAO,qCAAqC,KAAK,MAAM,GAAG,KAAK,KAAK,GAAG,OAAO,GAAG;;;;;;AAOnF,eAAsB,cACpB,OACA,MACA,QACqB;AAGrB,QAAO;EACL,WAAW;EACX,UAAU;EACX;;;;;AAMH,SAAgB,gBAAgB,OAAoD;CAClF,MAAM,yBAAS,IAAI,KAA6B;AAEhD,MAAK,MAAM,QAAQ,OAAO;EACxB,MAAM,WAAW,KAAK,WAAW;EACjC,MAAM,WAAW,OAAO,IAAI,SAAS,IAAI,EAAE;AAC3C,WAAS,KAAK,KAAK;AACnB,SAAO,IAAI,UAAU,SAAS;;AAGhC,QAAO;;;;;;;;;;;;;;;;;ACvCT,IAAa,iBAAb,MAAoE;CAClE,AAAS,OAAO;CAEhB,SAAS,SAA0B;AAEjC,SACE,QAAQ,SAAS,KAAK,IACtB,QAAQ,SAAS,MAAM,IACvB,QAAQ,SAAS,MAAM,IACvB,QAAQ,SAAS,MAAM,IACvB,QAAQ,SAAS,MAAM;;CAI3B,MAAM,SAA+C;EACnD,MAAM,kDAAoB,QAAQ;EAClC,MAAMA,WAA8B,EAAE;EACtC,MAAMC,WAA2B,EAAE;EACnC,MAAMC,aAA0B,EAAE;EAClC,IAAIC;AAGJ,MAAI,QAAQ,WAAW,MAAM,EAAE;GAC7B,MAAM,WAAW,QAAQ,QAAQ,OAAO,EAAE;AAC1C,OAAI,aAAa,IAAI;IACnB,MAAM,qBAAqB,QAAQ,MAAM,GAAG,SAAS,CAAC,MAAM;AAC5D,kBAAc,KAAK,iBAAiB,mBAAmB;;;EAK3D,IAAIC,iBAAyC;AAG7C,8BAAM,OAAO,SAAS;AAEpB,OAAI,KAAK,SAAS,WAAW;IAC3B,MAAM,UAAU;IAChB,MAAM,2CAAsB,QAAQ;AAGpC,QAAI,eACF,UAAS,KAAK,eAAe;AAG/B,qBAAiB;KACf,OAAO,QAAQ;KACf;KACA,SAAS;KACT,OAAO,EAAE;KACV;;AAIH,OAAI,KAAK,SAAS,QAAQ;IACxB,MAAM,OAAO;IACb,MAAM,0CAAqB,KAAK;IAChC,MAAMC,WAAyB;KAC7B,KAAK,KAAK;KACV;KACA,OAAO,KAAK,SAAS;KACrB,SAAS,gBAAgB;KAC1B;AAED,aAAS,KAAK,SAAS;AACvB,QAAI,eACF,gBAAe,MAAM,KAAK,SAAS;;AAKvC,OAAI,KAAK,SAAS,QAAQ;IACxB,MAAM,OAAO;AACb,eAAW,KAAK;KACd,UAAU,KAAK,QAAQ;KACvB,MAAM,KAAK;KACX,MAAM,KAAK,QAAQ;KACpB,CAAC;;AAIJ,OAAI,kBAAkB,KAAK,SAAS,aAAa;IAC/C,MAAM,0CAAqB,KAAK;AAChC,mBAAe,YAAY,eAAe,UAAU,SAAS,MAAM;;IAErE;AAGF,MAAI,eACF,UAAS,KAAK,eAAe;AAS/B,SAAO,EACL,MAAM;GACJ,OAPW,aAAa,SAAoB,SAAS,MAAM,MAAM,EAAE,UAAU,EAAE,EAAE;GAQjF,aALiB,aAAa,eAA0B,KAAK,mBAAmB,KAAK;GAMrF;GACA,OAAO;GACP;GACA;GACD,EACF;;CAGH,AAAQ,iBAAiB,SAA0C;EACjE,MAAMC,SAAkC,EAAE;EAC1C,MAAM,QAAQ,QAAQ,MAAM,KAAK;AAEjC,OAAK,MAAM,QAAQ,OAAO;GACxB,MAAM,aAAa,KAAK,QAAQ,IAAI;AACpC,OAAI,aAAa,GAAG;IAClB,MAAM,MAAM,KAAK,MAAM,GAAG,WAAW,CAAC,MAAM;IAC5C,IAAIC,QAAmC,KAAK,MAAM,aAAa,EAAE,CAAC,MAAM;AAGxE,QAAI,UAAU,OAAQ,SAAQ;aACrB,UAAU,QAAS,SAAQ;aAC3B,kBAAkB,KAAK,MAAM,CAAE,SAAQ,OAAO,MAAM;aACpD,MAAM,WAAW,KAAI,IAAI,MAAM,SAAS,KAAI,CAAE,SAAQ,MAAM,MAAM,GAAG,GAAG;aACxE,MAAM,WAAW,IAAI,IAAI,MAAM,SAAS,IAAI,CAAE,SAAQ,MAAM,MAAM,GAAG,GAAG;AAEjF,WAAO,OAAO;;;AAIlB,SAAO;;CAGT,AAAQ,mBAAmB,MAAgC;AAEzD,OAAK,MAAM,QAAQ,KAAK,UAAU;AAChC,OAAI,KAAK,SAAS,UAAW;AAC7B,OAAI,KAAK,SAAS,YAChB,2CAAqB,KAAK;;;;;;;AAUlC,SAAgB,iBAAiB,UAAkC;CACjE,MAAM,kDAAoB,SAAS;CACnC,MAAMC,QAAwB,EAAE;CAChC,IAAI,iBAAiB;AAErB,6BAAM,OAAO,SAAS;AACpB,MAAI,KAAK,SAAS,UAChB,qDAA+B,KAAgB;AAGjD,MAAI,KAAK,SAAS,WAIhB,6BAHiB,MAGD,SAAS,aAAmB;AAC1C,SAAM,KAAK;IACT,KAAK,SAAS;IACd,yCAAoB,SAAS;IAC7B,OAAO,SAAS,SAAS;IACzB,SAAS,kBAAkB;IAC5B,CAAC;IACF;GAEJ;AAEF,QAAO;;;;;AAMT,SAAgB,gBAAgB,UAAkB,WAAW,GAAsB;AAGjF,QAFe,IAAI,gBAAgB,CACb,MAAM,SAAS,CACvB,KAAK,SAAS,QAAQ,MAAM,EAAE,SAAS,SAAS"}

package/dist/parsers/index.d.cts.map

@@ -1 +0,0 @@
-{"version":3,"file":"index.d.cts","names":[],"sources":["../../src/parsers/types.ts","../../src/parsers/github.ts","../../src/parsers/markdown.ts"],"sourcesContent":[],"mappings":";;AAOA;;;;;AAiBA;AAQiB,UAzBA,YAyBY,CAAA,KAAA,EAAA,QAAA,OAAA,CAAA,CAAA;EAUZ,SAAA,IAAA,EAAA,MAAe;EAUf;;;EAKH,QAAA,CAAA,OAAA,EAAA,MAAA,EAAA,GAAA,CAAA,EAAA,MAAA,CAAA,EAAA,OAAA;EACE;;AAMhB;EASiB,KAAA,CAAA,OAAU,EAAA,MAAA,EAAA,GAAA,CAAA,EAAA,MAAA,CAAA,EAvDa,YAuDb,CAvD0B,KAuD1B,EAvDiC,KAuDjC,CAAA;;;;AChE3B;AAOgB,UDQC,YCRa,CAAA,KAAA,EAAA,QAAA,OAAA,CAAA,CAAA;EAYd,IAAA,EDHR,KCGgB;EAUF,IAAA,CAAA,EDZb,KCYa;AAgBtB;;;;AAA2D,UDtB1C,YAAA,CCsB0C;;;;EC7B9C,OAAA,CAAA,EAAA,MAAA;;;;;AAAsC,UFiBlC,eAAA,CEjBkC;EAqJnC,KAAA,EAAA,MAAA;EA+BA,KAAA,EAAA,MAAA;;SF/JP;;;;;UAMQ,cAAA;;;YAGL;SACH;cACK;gBACE;;;;;UAMC,SAAA;;;;;;;;UASA,UAAA;;;;;;;;AAlEjB;;;;;AAiBA;AAQiB,iBCvBD,YAAA,CDuBa,GAAA,EAAA,MAAA,CAAA,EAAA,OAAA;AAU7B;AAUA;;AAIS,iBCxCO,cAAA,CDwCP,GAAA,EAAA,MAAA,CAAA,EAAA;EACK,KAAA,EAAA,MAAA;EACE,IAAA,EAAA,MAAA;CAAM,GAAA,IAAA;AAMtB;AASA;;iBC7CgB,QAAA;;AAnBhB;AAOA;AAYA;AAUsB,iBAAA,aAAA,CAIX,KAAA,EAAR,MAAA,EAAO,IAAA,EAAA,MAAA,EAAA,MAAA,CAAA,EAAA,MAAA,CAAA,EAAP,OAAO,CAAC,UAAD,CAAA;AAYV;;;AAAwD,iBAAxC,eAAA,CAAwC,KAAA,EAAjB,YAAiB,EAAA,CAAA,EAAA,GAAA,CAAA,MAAA,EAAY,YAAZ,EAAA,CAAA;;;AD/CxD;;;;;AAiBA;AAQA;AAUA;AAUA;;;;AAMgB,cEjCH,cAAA,YAA0B,YFiCvB,CEjCoC,cFiCpC,CAAA,CAAA;EAAM,SAAA,IAAA,GAAA,UAAA;EAML,QAAA,CAAA,OAAS,EAAA,MAAA,CAAA,EAAA,OAAA;EAST,KAAA,CAAA,OAAU,EAAA,MAAA,CAAA,EElCD,YFkCC,CElCY,cFkCZ,CAAA;;;;AChE3B;AAOA;AAYA;AAUsB,iBCwIN,gBAAA,CDpIL,QAAR,EAAA,MAAO,CAAA,ECoI0C,YDpI1C,EAAA;AAYV;;;AAAwD,iBCuJxC,eAAA,CDvJwC,QAAA,EAAA,MAAA,EAAA,QAAA,CAAA,EAAA,MAAA,CAAA,ECuJS,eDvJT,EAAA"}