import Extract2MDConverter from 'extract2md';

// Scenario 1: Quick conversion only
const markdown1 = await Extract2MDConverter.quickConvertOnly(pdfFile);

// Scenario 2: High accuracy OCR conversion only  
const markdown2 = await Extract2MDConverter.highAccuracyConvertOnly(pdfFile);

// Scenario 3: Quick conversion + LLM enhancement
const markdown3 = await Extract2MDConverter.quickConvertWithLLM(pdfFile);

// Scenario 4: High accuracy conversion + LLM enhancement
const markdown4 = await Extract2MDConverter.highAccuracyConvertWithLLM(pdfFile);

// Scenario 5: Combined extraction + LLM enhancement (most comprehensive)
const markdown5 = await Extract2MDConverter.combinedConvertWithLLM(pdfFile);

const config = {
  // PDF.js Worker
  pdfJsWorkerSrc: "../pdf.worker.min.mjs",
  
  // Tesseract OCR Settings
  tesseract: {
    workerPath: "./tesseract-worker.min.js",
    corePath: "./tesseract-core.wasm.js", 
    langPath: "./lang-data/",
    language: "eng",
    options: {}
  },
  
  // LLM Configuration
  webllm: {
    model: "Qwen3-0.6B-q4f16_1-MLC",
    // Optional: Custom model
    customModel: {
      model: "https://huggingface.co/mlc-ai/your-model/resolve/main/",
      model_id: "YourModel-ID",
      model_lib: "https://example.com/your-model.wasm",
      required_features: ["shader-f16"],
      overrides: { conv_template: "qwen" }
    },
    options: {
      temperature: 0.7,
      maxTokens: 4096
    }
  },
  
  // System Prompt Customizations
  systemPrompts: {
    singleExtraction: "Focus on preserving code examples exactly.",
    combinedExtraction: "Pay attention to tables and diagrams from OCR."
  },
  
  // Processing Options
  processing: {
    splitPascalCase: false,
    pdfRenderScale: 2.5,
    postProcessRules: [
      { find: /\bAPI\b/g, replace: "API" }
    ]
  },
  
  // Progress Tracking
  progressCallback: (progress) => {
    console.log(`${progress.stage}: ${progress.message}`);
    if (progress.currentPage) {
      console.log(`Page ${progress.currentPage}/${progress.totalPages}`);
    }
  }
};

// Use configuration
const markdown = await Extract2MDConverter.combinedConvertWithLLM(pdfFile, config);

import { 
  WebLLMEngine, 
  OutputParser, 
  SystemPrompts,
  ConfigValidator 
} from 'extract2md';

// Validate configuration
const validatedConfig = ConfigValidator.validate(userConfig);

// Initialize WebLLM engine
const engine = new WebLLMEngine(validatedConfig);
await engine.initialize();

// Generate text
const result = await engine.generate("Your prompt here");

// Parse output
const parser = new OutputParser();
const cleanMarkdown = parser.parse(result);

// For scenarios 3 & 4 (single extraction)
const singlePrompt = SystemPrompts.getSingleExtractionPrompt(
  "Additional instruction: Preserve all technical terms."
);

// For scenario 5 (combined extraction) 
const combinedPrompt = SystemPrompts.getCombinedExtractionPrompt(
  "Focus on creating comprehensive documentation."
);

import { ConfigValidator } from 'extract2md';

// Load from JSON string
const config = ConfigValidator.fromJSON(configJsonString);

// Use with any scenario
const result = await Extract2MDConverter.quickConvertWithLLM(pdfFile, config);

const config = {
  progressCallback: (progress) => {
    switch (progress.stage) {
      case 'scenario_5_start':
        console.log('Starting combined conversion...');
        break;
      case 'webllm_load_progress':
        console.log(`Loading model: ${progress.progress}%`);
        break;
      case 'ocr_page_process':
        console.log(`OCR: ${progress.currentPage}/${progress.totalPages}`);
        break;
      case 'webllm_generate_start':
        console.log('AI enhancement in progress...');
        break;
      case 'scenario_5_complete':
        console.log('Conversion completed!');
        break;
      default:
        console.log(`${progress.stage}: ${progress.message}`);
    }
    
    if (progress.error) {
      console.error('Error:', progress.error);
    }
  }
};

try {
  const result = await Extract2MDConverter.combinedConvertWithLLM(pdfFile, config);
  console.log('Success:', result);
} catch (error) {
  console.error('Conversion failed:', error.message);
}

import { LegacyExtract2MDConverter } from 'extract2md';

// Old way
const converter = new LegacyExtract2MDConverter(options);
const quick = await converter.quickConvert(pdfFile);
const ocr = await converter.highAccuracyConvert(pdfFile);
const enhanced = await converter.llmRewrite(text);

// New way (recommended)
const quick = await Extract2MDConverter.quickConvertOnly(pdfFile, config);
const ocr = await Extract2MDConverter.highAccuracyConvertOnly(pdfFile, config);
const enhanced = await Extract2MDConverter.quickConvertWithLLM(pdfFile, config);

// Default worker paths (adjust for your deployment)
const config = {
    pdfJsWorkerSrc: "/pdf.worker.min.mjs",
    tesseract: {
        workerPath: "/tesseract-worker.min.js",
        corePath: "/tesseract-core.wasm.js"
    }
};