@opendataloader/pdf 1.4.3 → 1.5.1

package/README.md

@@ -136,12 +136,11 @@ opendataloader_pdf.convert(
     output_dir="output/",
     format="json,markdown,pdf",
 
-    # Reading order
-    reading_order="xycut",           # XY-Cut++ for multi-column
+    # Image output mode: "off", "embedded" (Base64), or "external" (default)
+    image_output="embedded",
 
-    # Images
-    embed_images=True,               # Base64 in output
-    image_format="png",
+    # Image format: "png" or "jpeg"
+    image_format="jpeg",
 
     # Tagged PDF
     use_struct_tree=True,            # Use native PDF structure
@@ -223,12 +222,12 @@ We continuously benchmark against real-world documents.
 
 ### Quick Comparison
 
-| Engine             | Accuracy |      | Speed (s/page) |      | Reading Order |      | Table    |      | Heading  |      |
-|--------------------|----------|------|----------------|------|---------------|------|----------|------|----------|------|
-| **opendataloader** | 0.82     | #2   | **0.05**       | #1   | **0.91**      | #1   | 0.49     | #2   | 0.65     | #2   |
-| docling            | **0.88** | #1   | 0.73           | #4   | 0.90          | #2   | **0.89** | #1   | **0.80** | #1   |
-| pymupdf4llm        | 0.73     | #3   | 0.09           | #2   | 0.89          | #3   | 0.40     | #3   | 0.41     | #3   |
-| markitdown         | 0.58     | #4   | **0.04**       | #1   | 0.88          | #4   | 0.00     | #4   | 0.00     | #4   |
+| Engine             | Speed (s/page) | Reading Order | Table    | Heading  |
+|--------------------|----------------|---------------|----------|----------|
+| **opendataloader** | **0.05**       | **0.91**      | 0.49     | 0.65     |
+| docling            | 0.73           | 0.90          | **0.89** | **0.80** |
+| pymupdf4llm        | 0.09           | 0.89          | 0.40     | 0.41     |
+| markitdown         | **0.04**       | 0.88          | 0.00     | 0.00     |
 
 > Scores are normalized to [0, 1]. Higher is better for accuracy metrics; lower is better for speed. **Bold** indicates best performance.
 
@@ -236,7 +235,7 @@ We continuously benchmark against real-world documents.
 
 | Use Case                 | Recommended Engine | Why                                                    |
 |--------------------------|--------------------|--------------------------------------------------------|
-| Best overall balance     | **opendataloader** | Fast (0.05s/page) with high reading order accuracy     |
+| Best overall balance     | **opendataloader** | Fast with high reading order accuracy     |
 | Maximum accuracy         | docling            | Highest scores for tables and headings, but 16x slower |
 | Speed-critical pipelines | markitdown         | Fastest, but no table/heading extraction               |
 | PyMuPDF ecosystem        | pymupdf4llm        | Good balance if already using PyMuPDF                  |

package/dist/cli.cjs

@@ -74,8 +74,8 @@ function buildConvertOptions(cliOptions) {
   if (cliOptions.htmlPageSeparator) {
     convertOptions.htmlPageSeparator = cliOptions.htmlPageSeparator;
   }
-  if (cliOptions.embedImages) {
-    convertOptions.embedImages = true;
+  if (cliOptions.imageOutput) {
+    convertOptions.imageOutput = cliOptions.imageOutput;
   }
   if (cliOptions.imageFormat) {
     convertOptions.imageFormat = cliOptions.imageFormat;
@@ -135,8 +135,8 @@ function buildArgs(options) {
   if (options.htmlPageSeparator) {
     args.push("--html-page-separator", options.htmlPageSeparator);
   }
-  if (options.embedImages) {
-    args.push("--embed-images");
+  if (options.imageOutput) {
+    args.push("--image-output", options.imageOutput);
   }
   if (options.imageFormat) {
     args.push("--image-format", options.imageFormat);
@@ -229,12 +229,12 @@ function registerCliOptions(program) {
   program.option("--keep-line-breaks", "Preserve original line breaks in extracted text");
   program.option("--replace-invalid-chars <value>", "Replacement character for invalid/unrecognized characters. Default: space");
   program.option("--use-struct-tree", "Use PDF structure tree (tagged PDF) for reading order and semantic structure");
-  program.option("--table-method <value>", "Table detection method. Values: cluster");
-  program.option("--reading-order <value>", "Reading order algorithm. Values: none, xycut. Default: none");
+  program.option("--table-method <value>", "Table detection method. Values: default (border-based), cluster (border + cluster). Default: default");
+  program.option("--reading-order <value>", "Reading order algorithm. Values: off, xycut. Default: xycut");
   program.option("--markdown-page-separator <value>", "Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none");
   program.option("--text-page-separator <value>", "Separator between pages in text output. Use %page-number% for page numbers. Default: none");
   program.option("--html-page-separator <value>", "Separator between pages in HTML output. Use %page-number% for page numbers. Default: none");
-  program.option("--embed-images", "Embed images as Base64 data URIs instead of file path references");
+  program.option("--image-output <value>", "Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external");
   program.option("--image-format <value>", "Output format for extracted images. Values: png, jpeg. Default: png");
 }
 

package/dist/cli.cjs.map

@@ -1 +1 @@
-{"version":3,"sources":["../src/cli.ts","../src/index.ts","../src/convert-options.generated.ts","../src/cli-options.generated.ts"],"sourcesContent":["#!/usr/bin/env node\nimport { Command, CommanderError } from 'commander';\nimport { convert } from './index.js';\nimport { CliOptions, buildConvertOptions } from './convert-options.generated.js';\nimport { registerCliOptions } from './cli-options.generated.js';\n\nfunction createProgram(): Command {\n  const program = new Command();\n\n  program\n    .name('opendataloader-pdf')\n    .usage('[options] <input...>')\n    .description('Convert PDFs using the OpenDataLoader CLI.')\n    .showHelpAfterError(\"Use '--help' to see available options.\")\n    .showSuggestionAfterError(false)\n    .argument('<input...>', 'Input files or directories to convert');\n\n  // Register CLI options from auto-generated file\n  registerCliOptions(program);\n\n  program.configureOutput({\n    writeErr: (str) => {\n      console.error(str.trimEnd());\n    },\n    outputError: (str, write) => {\n      write(str);\n    },\n  });\n\n  return program;\n}\n\nasync function main(): Promise<number> {\n  const program = createProgram();\n\n  program.exitOverride();\n\n  try {\n    program.parse(process.argv);\n  } catch (err) {\n    if (err instanceof CommanderError) {\n      if (err.code === 'commander.helpDisplayed') {\n        return 0;\n      }\n      return err.exitCode ?? 1;\n    }\n\n    const message = err instanceof Error ? err.message : String(err);\n    console.error(message);\n    console.error(\"Use '--help' to see available options.\");\n    return 1;\n  }\n\n  const cliOptions = program.opts<CliOptions>();\n  const inputPaths = program.args;\n  const convertOptions = buildConvertOptions(cliOptions);\n\n  try {\n    const output = await convert(inputPaths, convertOptions);\n    if (output && !convertOptions.quiet) {\n      process.stdout.write(output);\n      if (!output.endsWith('\\n')) {\n        process.stdout.write('\\n');\n      }\n    }\n    return 0;\n  } catch (err) {\n    const message = err instanceof Error ? err.message : String(err);\n    console.error(message);\n    return 1;\n  }\n}\n\nmain().then((code) => {\n  if (code !== 0) {\n    process.exit(code);\n  }\n});\n","import { spawn } from 'child_process';\nimport * as path from 'path';\nimport * as fs from 'fs';\nimport { fileURLToPath } from 'url';\n\n// Re-export types and utilities from auto-generated file\nexport type { ConvertOptions } from './convert-options.generated.js';\nexport { buildArgs } from './convert-options.generated.js';\nimport type { ConvertOptions } from './convert-options.generated.js';\nimport { buildArgs } from './convert-options.generated.js';\n\nconst __filename = fileURLToPath(import.meta.url);\nconst __dirname = path.dirname(__filename);\n\nconst JAR_NAME = 'opendataloader-pdf-cli.jar';\n\ninterface JarExecutionOptions {\n  streamOutput?: boolean;\n}\n\nfunction executeJar(args: string[], executionOptions: JarExecutionOptions = {}): Promise<string> {\n  const { streamOutput = false } = executionOptions;\n\n  return new Promise((resolve, reject) => {\n    const jarPath = path.join(__dirname, '..', 'lib', JAR_NAME);\n\n    if (!fs.existsSync(jarPath)) {\n      return reject(\n        new Error(`JAR file not found at ${jarPath}. Please run the build script first.`),\n      );\n    }\n\n    const command = 'java';\n    const commandArgs = ['-jar', jarPath, ...args];\n\n    const javaProcess = spawn(command, commandArgs);\n\n    let stdout = '';\n    let stderr = '';\n\n    javaProcess.stdout.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stdout.write(chunk);\n      }\n      stdout += chunk;\n    });\n\n    javaProcess.stderr.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stderr.write(chunk);\n      }\n      stderr += chunk;\n    });\n\n    javaProcess.on('close', (code) => {\n      if (code === 0) {\n        resolve(stdout);\n      } else {\n        const errorOutput = stderr || stdout;\n        const error = new Error(\n          `The opendataloader-pdf CLI exited with code ${code}.\\n\\n${errorOutput}`,\n        );\n        reject(error);\n      }\n    });\n\n    javaProcess.on('error', (err: Error) => {\n      if (err.message.includes('ENOENT')) {\n        reject(\n          new Error(\n            \"'java' command not found. Please ensure Java is installed and in your system's PATH.\",\n          ),\n        );\n      } else {\n        reject(err);\n      }\n    });\n  });\n}\n\nexport function convert(\n  inputPaths: string | string[],\n  options: ConvertOptions = {},\n): Promise<string> {\n  const inputList = Array.isArray(inputPaths) ? inputPaths : [inputPaths];\n  if (inputList.length === 0) {\n    return Promise.reject(new Error('At least one input path must be provided.'));\n  }\n\n  for (const input of inputList) {\n    if (!fs.existsSync(input)) {\n      return Promise.reject(new Error(`Input file or folder not found: ${input}`));\n    }\n  }\n\n  const args: string[] = [...inputList, ...buildArgs(options)];\n\n  return executeJar(args, {\n    streamOutput: !options.quiet,\n  });\n}\n\n/**\n * @deprecated Use `convert()` and `ConvertOptions` instead. This function will be removed in a future version.\n */\nexport interface RunOptions {\n  outputFolder?: string;\n  password?: string;\n  replaceInvalidChars?: string;\n  generateMarkdown?: boolean;\n  generateHtml?: boolean;\n  generateAnnotatedPdf?: boolean;\n  keepLineBreaks?: boolean;\n  contentSafetyOff?: string;\n  htmlInMarkdown?: boolean;\n  addImageToMarkdown?: boolean;\n  noJson?: boolean;\n  debug?: boolean;\n  useStructTree?: boolean;\n}\n\n/**\n * @deprecated Use `convert()` instead. This function will be removed in a future version.\n */\nexport function run(inputPath: string, options: RunOptions = {}): Promise<string> {\n  console.warn(\n    'Warning: run() is deprecated and will be removed in a future version. Use convert() instead.',\n  );\n\n  // Build format array based on legacy boolean options\n  const formats: string[] = [];\n  if (!options.noJson) {\n    formats.push('json');\n  }\n  if (options.generateMarkdown) {\n    if (options.addImageToMarkdown) {\n      formats.push('markdown-with-images');\n    } else if (options.htmlInMarkdown) {\n      formats.push('markdown-with-html');\n    } else {\n      formats.push('markdown');\n    }\n  }\n  if (options.generateHtml) {\n    formats.push('html');\n  }\n  if (options.generateAnnotatedPdf) {\n    formats.push('pdf');\n  }\n\n  return convert(inputPath, {\n    outputDir: options.outputFolder,\n    password: options.password,\n    replaceInvalidChars: options.replaceInvalidChars,\n    keepLineBreaks: options.keepLineBreaks,\n    contentSafetyOff: options.contentSafetyOff,\n    useStructTree: options.useStructTree,\n    format: formats.length > 0 ? formats : undefined,\n    quiet: !options.debug,\n  });\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\n/**\n * Options for the convert function.\n */\nexport interface ConvertOptions {\n  /** Directory where output files are written. Default: input file directory */\n  outputDir?: string;\n  /** Password for encrypted PDF files */\n  password?: string;\n  /** Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json */\n  format?: string | string[];\n  /** Suppress console logging output */\n  quiet?: boolean;\n  /** Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg */\n  contentSafetyOff?: string | string[];\n  /** Preserve original line breaks in extracted text */\n  keepLineBreaks?: boolean;\n  /** Replacement character for invalid/unrecognized characters. Default: space */\n  replaceInvalidChars?: string;\n  /** Use PDF structure tree (tagged PDF) for reading order and semantic structure */\n  useStructTree?: boolean;\n  /** Table detection method. Values: cluster */\n  tableMethod?: string;\n  /** Reading order algorithm. Values: none, xycut. Default: none */\n  readingOrder?: string;\n  /** Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none */\n  markdownPageSeparator?: string;\n  /** Separator between pages in text output. Use %page-number% for page numbers. Default: none */\n  textPageSeparator?: string;\n  /** Separator between pages in HTML output. Use %page-number% for page numbers. Default: none */\n  htmlPageSeparator?: string;\n  /** Embed images as Base64 data URIs instead of file path references */\n  embedImages?: boolean;\n  /** Output format for extracted images. Values: png, jpeg. Default: png */\n  imageFormat?: string;\n}\n\n/**\n * Options as parsed from CLI (all values are strings from commander).\n */\nexport interface CliOptions {\n  outputDir?: string;\n  password?: string;\n  format?: string;\n  quiet?: boolean;\n  contentSafetyOff?: string;\n  keepLineBreaks?: boolean;\n  replaceInvalidChars?: string;\n  useStructTree?: boolean;\n  tableMethod?: string;\n  readingOrder?: string;\n  markdownPageSeparator?: string;\n  textPageSeparator?: string;\n  htmlPageSeparator?: string;\n  embedImages?: boolean;\n  imageFormat?: string;\n}\n\n/**\n * Convert CLI options to ConvertOptions.\n */\nexport function buildConvertOptions(cliOptions: CliOptions): ConvertOptions {\n  const convertOptions: ConvertOptions = {};\n\n  if (cliOptions.outputDir) {\n    convertOptions.outputDir = cliOptions.outputDir;\n  }\n  if (cliOptions.password) {\n    convertOptions.password = cliOptions.password;\n  }\n  if (cliOptions.format) {\n    convertOptions.format = cliOptions.format;\n  }\n  if (cliOptions.quiet) {\n    convertOptions.quiet = true;\n  }\n  if (cliOptions.contentSafetyOff) {\n    convertOptions.contentSafetyOff = cliOptions.contentSafetyOff;\n  }\n  if (cliOptions.keepLineBreaks) {\n    convertOptions.keepLineBreaks = true;\n  }\n  if (cliOptions.replaceInvalidChars) {\n    convertOptions.replaceInvalidChars = cliOptions.replaceInvalidChars;\n  }\n  if (cliOptions.useStructTree) {\n    convertOptions.useStructTree = true;\n  }\n  if (cliOptions.tableMethod) {\n    convertOptions.tableMethod = cliOptions.tableMethod;\n  }\n  if (cliOptions.readingOrder) {\n    convertOptions.readingOrder = cliOptions.readingOrder;\n  }\n  if (cliOptions.markdownPageSeparator) {\n    convertOptions.markdownPageSeparator = cliOptions.markdownPageSeparator;\n  }\n  if (cliOptions.textPageSeparator) {\n    convertOptions.textPageSeparator = cliOptions.textPageSeparator;\n  }\n  if (cliOptions.htmlPageSeparator) {\n    convertOptions.htmlPageSeparator = cliOptions.htmlPageSeparator;\n  }\n  if (cliOptions.embedImages) {\n    convertOptions.embedImages = true;\n  }\n  if (cliOptions.imageFormat) {\n    convertOptions.imageFormat = cliOptions.imageFormat;\n  }\n\n  return convertOptions;\n}\n\n/**\n * Build CLI arguments array from ConvertOptions.\n */\nexport function buildArgs(options: ConvertOptions): string[] {\n  const args: string[] = [];\n\n  if (options.outputDir) {\n    args.push('--output-dir', options.outputDir);\n  }\n  if (options.password) {\n    args.push('--password', options.password);\n  }\n  if (options.format) {\n    if (Array.isArray(options.format)) {\n      if (options.format.length > 0) {\n        args.push('--format', options.format.join(','));\n      }\n    } else {\n      args.push('--format', options.format);\n    }\n  }\n  if (options.quiet) {\n    args.push('--quiet');\n  }\n  if (options.contentSafetyOff) {\n    if (Array.isArray(options.contentSafetyOff)) {\n      if (options.contentSafetyOff.length > 0) {\n        args.push('--content-safety-off', options.contentSafetyOff.join(','));\n      }\n    } else {\n      args.push('--content-safety-off', options.contentSafetyOff);\n    }\n  }\n  if (options.keepLineBreaks) {\n    args.push('--keep-line-breaks');\n  }\n  if (options.replaceInvalidChars) {\n    args.push('--replace-invalid-chars', options.replaceInvalidChars);\n  }\n  if (options.useStructTree) {\n    args.push('--use-struct-tree');\n  }\n  if (options.tableMethod) {\n    args.push('--table-method', options.tableMethod);\n  }\n  if (options.readingOrder) {\n    args.push('--reading-order', options.readingOrder);\n  }\n  if (options.markdownPageSeparator) {\n    args.push('--markdown-page-separator', options.markdownPageSeparator);\n  }\n  if (options.textPageSeparator) {\n    args.push('--text-page-separator', options.textPageSeparator);\n  }\n  if (options.htmlPageSeparator) {\n    args.push('--html-page-separator', options.htmlPageSeparator);\n  }\n  if (options.embedImages) {\n    args.push('--embed-images');\n  }\n  if (options.imageFormat) {\n    args.push('--image-format', options.imageFormat);\n  }\n\n  return args;\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\nimport { Command } from 'commander';\n\n/**\n * Register all CLI options on the given Commander program.\n */\nexport function registerCliOptions(program: Command): void {\n  program.option('-o, --output-dir <value>', 'Directory where output files are written. Default: input file directory');\n  program.option('-p, --password <value>', 'Password for encrypted PDF files');\n  program.option('-f, --format <value>', 'Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json');\n  program.option('-q, --quiet', 'Suppress console logging output');\n  program.option('--content-safety-off <value>', 'Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg');\n  program.option('--keep-line-breaks', 'Preserve original line breaks in extracted text');\n  program.option('--replace-invalid-chars <value>', 'Replacement character for invalid/unrecognized characters. Default: space');\n  program.option('--use-struct-tree', 'Use PDF structure tree (tagged PDF) for reading order and semantic structure');\n  program.option('--table-method <value>', 'Table detection method. Values: cluster');\n  program.option('--reading-order <value>', 'Reading order algorithm. Values: none, xycut. Default: none');\n  program.option('--markdown-page-separator <value>', 'Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none');\n  program.option('--text-page-separator <value>', 'Separator between pages in text output. Use %page-number% for page numbers. Default: none');\n  program.option('--html-page-separator <value>', 'Separator between pages in HTML output. Use %page-number% for page numbers. Default: none');\n  program.option('--embed-images', 'Embed images as Base64 data URIs instead of file path references');\n  program.option('--image-format <value>', 'Output format for extracted images. Values: png, jpeg. Default: png');\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;AACA,uBAAwC;;;ACDxC,2BAAsB;AACtB,WAAsB;AACtB,SAAoB;AACpB,iBAA8B;;;AC4DvB,SAAS,oBAAoB,YAAwC;AAC1E,QAAM,iBAAiC,CAAC;AAExC,MAAI,WAAW,WAAW;AACxB,mBAAe,YAAY,WAAW;AAAA,EACxC;AACA,MAAI,WAAW,UAAU;AACvB,mBAAe,WAAW,WAAW;AAAA,EACvC;AACA,MAAI,WAAW,QAAQ;AACrB,mBAAe,SAAS,WAAW;AAAA,EACrC;AACA,MAAI,WAAW,OAAO;AACpB,mBAAe,QAAQ;AAAA,EACzB;AACA,MAAI,WAAW,kBAAkB;AAC/B,mBAAe,mBAAmB,WAAW;AAAA,EAC/C;AACA,MAAI,WAAW,gBAAgB;AAC7B,mBAAe,iBAAiB;AAAA,EAClC;AACA,MAAI,WAAW,qBAAqB;AAClC,mBAAe,sBAAsB,WAAW;AAAA,EAClD;AACA,MAAI,WAAW,eAAe;AAC5B,mBAAe,gBAAgB;AAAA,EACjC;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc,WAAW;AAAA,EAC1C;AACA,MAAI,WAAW,cAAc;AAC3B,mBAAe,eAAe,WAAW;AAAA,EAC3C;AACA,MAAI,WAAW,uBAAuB;AACpC,mBAAe,wBAAwB,WAAW;AAAA,EACpD;AACA,MAAI,WAAW,mBAAmB;AAChC,mBAAe,oBAAoB,WAAW;AAAA,EAChD;AACA,MAAI,WAAW,mBAAmB;AAChC,mBAAe,oBAAoB,WAAW;AAAA,EAChD;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc;AAAA,EAC/B;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc,WAAW;AAAA,EAC1C;AAEA,SAAO;AACT;AAKO,SAAS,UAAU,SAAmC;AAC3D,QAAM,OAAiB,CAAC;AAExB,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,UAAU;AACpB,SAAK,KAAK,cAAc,QAAQ,QAAQ;AAAA,EAC1C;AACA,MAAI,QAAQ,QAAQ;AAClB,QAAI,MAAM,QAAQ,QAAQ,MAAM,GAAG;AACjC,UAAI,QAAQ,OAAO,SAAS,GAAG;AAC7B,aAAK,KAAK,YAAY,QAAQ,OAAO,KAAK,GAAG,CAAC;AAAA,MAChD;AAAA,IACF,OAAO;AACL,WAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,IACtC;AAAA,EACF;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,SAAS;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,MAAM,QAAQ,QAAQ,gBAAgB,GAAG;AAC3C,UAAI,QAAQ,iBAAiB,SAAS,GAAG;AACvC,aAAK,KAAK,wBAAwB,QAAQ,iBAAiB,KAAK,GAAG,CAAC;AAAA,MACtE;AAAA,IACF,OAAO;AACL,WAAK,KAAK,wBAAwB,QAAQ,gBAAgB;AAAA,IAC5D;AAAA,EACF;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,oBAAoB;AAAA,EAChC;AACA,MAAI,QAAQ,qBAAqB;AAC/B,SAAK,KAAK,2BAA2B,QAAQ,mBAAmB;AAAA,EAClE;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,cAAc;AACxB,SAAK,KAAK,mBAAmB,QAAQ,YAAY;AAAA,EACnD;AACA,MAAI,QAAQ,uBAAuB;AACjC,SAAK,KAAK,6BAA6B,QAAQ,qBAAqB;AAAA,EACtE;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,gBAAgB;AAAA,EAC5B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AAEA,SAAO;AACT;;;ADpLA;AAWA,IAAM,iBAAa,0BAAc,YAAY,GAAG;AAChD,IAAM,YAAiB,aAAQ,UAAU;AAEzC,IAAM,WAAW;AAMjB,SAAS,WAAW,MAAgB,mBAAwC,CAAC,GAAoB;AAC/F,QAAM,EAAE,eAAe,MAAM,IAAI;AAEjC,SAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,UAAM,UAAe,UAAK,WAAW,MAAM,OAAO,QAAQ;AAE1D,QAAI,CAAI,cAAW,OAAO,GAAG;AAC3B,aAAO;AAAA,QACL,IAAI,MAAM,yBAAyB,OAAO,sCAAsC;AAAA,MAClF;AAAA,IACF;AAEA,UAAM,UAAU;AAChB,UAAM,cAAc,CAAC,QAAQ,SAAS,GAAG,IAAI;AAE7C,UAAM,kBAAc,4BAAM,SAAS,WAAW;AAE9C,QAAI,SAAS;AACb,QAAI,SAAS;AAEb,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,SAAS;AAChC,UAAI,SAAS,GAAG;AACd,gBAAQ,MAAM;AAAA,MAChB,OAAO;AACL,cAAM,cAAc,UAAU;AAC9B,cAAM,QAAQ,IAAI;AAAA,UAChB,+CAA+C,IAAI;AAAA;AAAA,EAAQ,WAAW;AAAA,QACxE;AACA,eAAO,KAAK;AAAA,MACd;AAAA,IACF,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,QAAe;AACtC,UAAI,IAAI,QAAQ,SAAS,QAAQ,GAAG;AAClC;AAAA,UACE,IAAI;AAAA,YACF;AAAA,UACF;AAAA,QACF;AAAA,MACF,OAAO;AACL,eAAO,GAAG;AAAA,MACZ;AAAA,IACF,CAAC;AAAA,EACH,CAAC;AACH;AAEO,SAAS,QACd,YACA,UAA0B,CAAC,GACV;AACjB,QAAM,YAAY,MAAM,QAAQ,UAAU,IAAI,aAAa,CAAC,UAAU;AACtE,MAAI,UAAU,WAAW,GAAG;AAC1B,WAAO,QAAQ,OAAO,IAAI,MAAM,2CAA2C,CAAC;AAAA,EAC9E;AAEA,aAAW,SAAS,WAAW;AAC7B,QAAI,CAAI,cAAW,KAAK,GAAG;AACzB,aAAO,QAAQ,OAAO,IAAI,MAAM,mCAAmC,KAAK,EAAE,CAAC;AAAA,IAC7E;AAAA,EACF;AAEA,QAAM,OAAiB,CAAC,GAAG,WAAW,GAAG,UAAU,OAAO,CAAC;AAE3D,SAAO,WAAW,MAAM;AAAA,IACtB,cAAc,CAAC,QAAQ;AAAA,EACzB,CAAC;AACH;;;AE9FO,SAAS,mBAAmB,SAAwB;AACzD,UAAQ,OAAO,4BAA4B,yEAAyE;AACpH,UAAQ,OAAO,0BAA0B,kCAAkC;AAC3E,UAAQ,OAAO,wBAAwB,oIAAoI;AAC3K,UAAQ,OAAO,eAAe,iCAAiC;AAC/D,UAAQ,OAAO,gCAAgC,sFAAsF;AACrI,UAAQ,OAAO,sBAAsB,iDAAiD;AACtF,UAAQ,OAAO,mCAAmC,2EAA2E;AAC7H,UAAQ,OAAO,qBAAqB,8EAA8E;AAClH,UAAQ,OAAO,0BAA0B,yCAAyC;AAClF,UAAQ,OAAO,2BAA2B,6DAA6D;AACvG,UAAQ,OAAO,qCAAqC,+FAA+F;AACnJ,UAAQ,OAAO,iCAAiC,2FAA2F;AAC3I,UAAQ,OAAO,iCAAiC,2FAA2F;AAC3I,UAAQ,OAAO,kBAAkB,kEAAkE;AACnG,UAAQ,OAAO,0BAA0B,qEAAqE;AAChH;;;AHlBA,SAAS,gBAAyB;AAChC,QAAM,UAAU,IAAI,yBAAQ;AAE5B,UACG,KAAK,oBAAoB,EACzB,MAAM,sBAAsB,EAC5B,YAAY,4CAA4C,EACxD,mBAAmB,wCAAwC,EAC3D,yBAAyB,KAAK,EAC9B,SAAS,cAAc,uCAAuC;AAGjE,qBAAmB,OAAO;AAE1B,UAAQ,gBAAgB;AAAA,IACtB,UAAU,CAAC,QAAQ;AACjB,cAAQ,MAAM,IAAI,QAAQ,CAAC;AAAA,IAC7B;AAAA,IACA,aAAa,CAAC,KAAK,UAAU;AAC3B,YAAM,GAAG;AAAA,IACX;AAAA,EACF,CAAC;AAED,SAAO;AACT;AAEA,eAAe,OAAwB;AACrC,QAAM,UAAU,cAAc;AAE9B,UAAQ,aAAa;AAErB,MAAI;AACF,YAAQ,MAAM,QAAQ,IAAI;AAAA,EAC5B,SAAS,KAAK;AACZ,QAAI,eAAe,iCAAgB;AACjC,UAAI,IAAI,SAAS,2BAA2B;AAC1C,eAAO;AAAA,MACT;AACA,aAAO,IAAI,YAAY;AAAA,IACzB;AAEA,UAAM,UAAU,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC/D,YAAQ,MAAM,OAAO;AACrB,YAAQ,MAAM,wCAAwC;AACtD,WAAO;AAAA,EACT;AAEA,QAAM,aAAa,QAAQ,KAAiB;AAC5C,QAAM,aAAa,QAAQ;AAC3B,QAAM,iBAAiB,oBAAoB,UAAU;AAErD,MAAI;AACF,UAAM,SAAS,MAAM,QAAQ,YAAY,cAAc;AACvD,QAAI,UAAU,CAAC,eAAe,OAAO;AACnC,cAAQ,OAAO,MAAM,MAAM;AAC3B,UAAI,CAAC,OAAO,SAAS,IAAI,GAAG;AAC1B,gBAAQ,OAAO,MAAM,IAAI;AAAA,MAC3B;AAAA,IACF;AACA,WAAO;AAAA,EACT,SAAS,KAAK;AACZ,UAAM,UAAU,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC/D,YAAQ,MAAM,OAAO;AACrB,WAAO;AAAA,EACT;AACF;AAEA,KAAK,EAAE,KAAK,CAAC,SAAS;AACpB,MAAI,SAAS,GAAG;AACd,YAAQ,KAAK,IAAI;AAAA,EACnB;AACF,CAAC;","names":[]}
+{"version":3,"sources":["../src/cli.ts","../src/index.ts","../src/convert-options.generated.ts","../src/cli-options.generated.ts"],"sourcesContent":["#!/usr/bin/env node\nimport { Command, CommanderError } from 'commander';\nimport { convert } from './index.js';\nimport { CliOptions, buildConvertOptions } from './convert-options.generated.js';\nimport { registerCliOptions } from './cli-options.generated.js';\n\nfunction createProgram(): Command {\n  const program = new Command();\n\n  program\n    .name('opendataloader-pdf')\n    .usage('[options] <input...>')\n    .description('Convert PDFs using the OpenDataLoader CLI.')\n    .showHelpAfterError(\"Use '--help' to see available options.\")\n    .showSuggestionAfterError(false)\n    .argument('<input...>', 'Input files or directories to convert');\n\n  // Register CLI options from auto-generated file\n  registerCliOptions(program);\n\n  program.configureOutput({\n    writeErr: (str) => {\n      console.error(str.trimEnd());\n    },\n    outputError: (str, write) => {\n      write(str);\n    },\n  });\n\n  return program;\n}\n\nasync function main(): Promise<number> {\n  const program = createProgram();\n\n  program.exitOverride();\n\n  try {\n    program.parse(process.argv);\n  } catch (err) {\n    if (err instanceof CommanderError) {\n      if (err.code === 'commander.helpDisplayed') {\n        return 0;\n      }\n      return err.exitCode ?? 1;\n    }\n\n    const message = err instanceof Error ? err.message : String(err);\n    console.error(message);\n    console.error(\"Use '--help' to see available options.\");\n    return 1;\n  }\n\n  const cliOptions = program.opts<CliOptions>();\n  const inputPaths = program.args;\n  const convertOptions = buildConvertOptions(cliOptions);\n\n  try {\n    const output = await convert(inputPaths, convertOptions);\n    if (output && !convertOptions.quiet) {\n      process.stdout.write(output);\n      if (!output.endsWith('\\n')) {\n        process.stdout.write('\\n');\n      }\n    }\n    return 0;\n  } catch (err) {\n    const message = err instanceof Error ? err.message : String(err);\n    console.error(message);\n    return 1;\n  }\n}\n\nmain().then((code) => {\n  if (code !== 0) {\n    process.exit(code);\n  }\n});\n","import { spawn } from 'child_process';\nimport * as path from 'path';\nimport * as fs from 'fs';\nimport { fileURLToPath } from 'url';\n\n// Re-export types and utilities from auto-generated file\nexport type { ConvertOptions } from './convert-options.generated.js';\nexport { buildArgs } from './convert-options.generated.js';\nimport type { ConvertOptions } from './convert-options.generated.js';\nimport { buildArgs } from './convert-options.generated.js';\n\nconst __filename = fileURLToPath(import.meta.url);\nconst __dirname = path.dirname(__filename);\n\nconst JAR_NAME = 'opendataloader-pdf-cli.jar';\n\ninterface JarExecutionOptions {\n  streamOutput?: boolean;\n}\n\nfunction executeJar(args: string[], executionOptions: JarExecutionOptions = {}): Promise<string> {\n  const { streamOutput = false } = executionOptions;\n\n  return new Promise((resolve, reject) => {\n    const jarPath = path.join(__dirname, '..', 'lib', JAR_NAME);\n\n    if (!fs.existsSync(jarPath)) {\n      return reject(\n        new Error(`JAR file not found at ${jarPath}. Please run the build script first.`),\n      );\n    }\n\n    const command = 'java';\n    const commandArgs = ['-jar', jarPath, ...args];\n\n    const javaProcess = spawn(command, commandArgs);\n\n    let stdout = '';\n    let stderr = '';\n\n    javaProcess.stdout.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stdout.write(chunk);\n      }\n      stdout += chunk;\n    });\n\n    javaProcess.stderr.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stderr.write(chunk);\n      }\n      stderr += chunk;\n    });\n\n    javaProcess.on('close', (code) => {\n      if (code === 0) {\n        resolve(stdout);\n      } else {\n        const errorOutput = stderr || stdout;\n        const error = new Error(\n          `The opendataloader-pdf CLI exited with code ${code}.\\n\\n${errorOutput}`,\n        );\n        reject(error);\n      }\n    });\n\n    javaProcess.on('error', (err: Error) => {\n      if (err.message.includes('ENOENT')) {\n        reject(\n          new Error(\n            \"'java' command not found. Please ensure Java is installed and in your system's PATH.\",\n          ),\n        );\n      } else {\n        reject(err);\n      }\n    });\n  });\n}\n\nexport function convert(\n  inputPaths: string | string[],\n  options: ConvertOptions = {},\n): Promise<string> {\n  const inputList = Array.isArray(inputPaths) ? inputPaths : [inputPaths];\n  if (inputList.length === 0) {\n    return Promise.reject(new Error('At least one input path must be provided.'));\n  }\n\n  for (const input of inputList) {\n    if (!fs.existsSync(input)) {\n      return Promise.reject(new Error(`Input file or folder not found: ${input}`));\n    }\n  }\n\n  const args: string[] = [...inputList, ...buildArgs(options)];\n\n  return executeJar(args, {\n    streamOutput: !options.quiet,\n  });\n}\n\n/**\n * @deprecated Use `convert()` and `ConvertOptions` instead. This function will be removed in a future version.\n */\nexport interface RunOptions {\n  outputFolder?: string;\n  password?: string;\n  replaceInvalidChars?: string;\n  generateMarkdown?: boolean;\n  generateHtml?: boolean;\n  generateAnnotatedPdf?: boolean;\n  keepLineBreaks?: boolean;\n  contentSafetyOff?: string;\n  htmlInMarkdown?: boolean;\n  addImageToMarkdown?: boolean;\n  noJson?: boolean;\n  debug?: boolean;\n  useStructTree?: boolean;\n}\n\n/**\n * @deprecated Use `convert()` instead. This function will be removed in a future version.\n */\nexport function run(inputPath: string, options: RunOptions = {}): Promise<string> {\n  console.warn(\n    'Warning: run() is deprecated and will be removed in a future version. Use convert() instead.',\n  );\n\n  // Build format array based on legacy boolean options\n  const formats: string[] = [];\n  if (!options.noJson) {\n    formats.push('json');\n  }\n  if (options.generateMarkdown) {\n    if (options.addImageToMarkdown) {\n      formats.push('markdown-with-images');\n    } else if (options.htmlInMarkdown) {\n      formats.push('markdown-with-html');\n    } else {\n      formats.push('markdown');\n    }\n  }\n  if (options.generateHtml) {\n    formats.push('html');\n  }\n  if (options.generateAnnotatedPdf) {\n    formats.push('pdf');\n  }\n\n  return convert(inputPath, {\n    outputDir: options.outputFolder,\n    password: options.password,\n    replaceInvalidChars: options.replaceInvalidChars,\n    keepLineBreaks: options.keepLineBreaks,\n    contentSafetyOff: options.contentSafetyOff,\n    useStructTree: options.useStructTree,\n    format: formats.length > 0 ? formats : undefined,\n    quiet: !options.debug,\n  });\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\n/**\n * Options for the convert function.\n */\nexport interface ConvertOptions {\n  /** Directory where output files are written. Default: input file directory */\n  outputDir?: string;\n  /** Password for encrypted PDF files */\n  password?: string;\n  /** Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json */\n  format?: string | string[];\n  /** Suppress console logging output */\n  quiet?: boolean;\n  /** Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg */\n  contentSafetyOff?: string | string[];\n  /** Preserve original line breaks in extracted text */\n  keepLineBreaks?: boolean;\n  /** Replacement character for invalid/unrecognized characters. Default: space */\n  replaceInvalidChars?: string;\n  /** Use PDF structure tree (tagged PDF) for reading order and semantic structure */\n  useStructTree?: boolean;\n  /** Table detection method. Values: default (border-based), cluster (border + cluster). Default: default */\n  tableMethod?: string;\n  /** Reading order algorithm. Values: off, xycut. Default: xycut */\n  readingOrder?: string;\n  /** Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none */\n  markdownPageSeparator?: string;\n  /** Separator between pages in text output. Use %page-number% for page numbers. Default: none */\n  textPageSeparator?: string;\n  /** Separator between pages in HTML output. Use %page-number% for page numbers. Default: none */\n  htmlPageSeparator?: string;\n  /** Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external */\n  imageOutput?: string;\n  /** Output format for extracted images. Values: png, jpeg. Default: png */\n  imageFormat?: string;\n}\n\n/**\n * Options as parsed from CLI (all values are strings from commander).\n */\nexport interface CliOptions {\n  outputDir?: string;\n  password?: string;\n  format?: string;\n  quiet?: boolean;\n  contentSafetyOff?: string;\n  keepLineBreaks?: boolean;\n  replaceInvalidChars?: string;\n  useStructTree?: boolean;\n  tableMethod?: string;\n  readingOrder?: string;\n  markdownPageSeparator?: string;\n  textPageSeparator?: string;\n  htmlPageSeparator?: string;\n  imageOutput?: string;\n  imageFormat?: string;\n}\n\n/**\n * Convert CLI options to ConvertOptions.\n */\nexport function buildConvertOptions(cliOptions: CliOptions): ConvertOptions {\n  const convertOptions: ConvertOptions = {};\n\n  if (cliOptions.outputDir) {\n    convertOptions.outputDir = cliOptions.outputDir;\n  }\n  if (cliOptions.password) {\n    convertOptions.password = cliOptions.password;\n  }\n  if (cliOptions.format) {\n    convertOptions.format = cliOptions.format;\n  }\n  if (cliOptions.quiet) {\n    convertOptions.quiet = true;\n  }\n  if (cliOptions.contentSafetyOff) {\n    convertOptions.contentSafetyOff = cliOptions.contentSafetyOff;\n  }\n  if (cliOptions.keepLineBreaks) {\n    convertOptions.keepLineBreaks = true;\n  }\n  if (cliOptions.replaceInvalidChars) {\n    convertOptions.replaceInvalidChars = cliOptions.replaceInvalidChars;\n  }\n  if (cliOptions.useStructTree) {\n    convertOptions.useStructTree = true;\n  }\n  if (cliOptions.tableMethod) {\n    convertOptions.tableMethod = cliOptions.tableMethod;\n  }\n  if (cliOptions.readingOrder) {\n    convertOptions.readingOrder = cliOptions.readingOrder;\n  }\n  if (cliOptions.markdownPageSeparator) {\n    convertOptions.markdownPageSeparator = cliOptions.markdownPageSeparator;\n  }\n  if (cliOptions.textPageSeparator) {\n    convertOptions.textPageSeparator = cliOptions.textPageSeparator;\n  }\n  if (cliOptions.htmlPageSeparator) {\n    convertOptions.htmlPageSeparator = cliOptions.htmlPageSeparator;\n  }\n  if (cliOptions.imageOutput) {\n    convertOptions.imageOutput = cliOptions.imageOutput;\n  }\n  if (cliOptions.imageFormat) {\n    convertOptions.imageFormat = cliOptions.imageFormat;\n  }\n\n  return convertOptions;\n}\n\n/**\n * Build CLI arguments array from ConvertOptions.\n */\nexport function buildArgs(options: ConvertOptions): string[] {\n  const args: string[] = [];\n\n  if (options.outputDir) {\n    args.push('--output-dir', options.outputDir);\n  }\n  if (options.password) {\n    args.push('--password', options.password);\n  }\n  if (options.format) {\n    if (Array.isArray(options.format)) {\n      if (options.format.length > 0) {\n        args.push('--format', options.format.join(','));\n      }\n    } else {\n      args.push('--format', options.format);\n    }\n  }\n  if (options.quiet) {\n    args.push('--quiet');\n  }\n  if (options.contentSafetyOff) {\n    if (Array.isArray(options.contentSafetyOff)) {\n      if (options.contentSafetyOff.length > 0) {\n        args.push('--content-safety-off', options.contentSafetyOff.join(','));\n      }\n    } else {\n      args.push('--content-safety-off', options.contentSafetyOff);\n    }\n  }\n  if (options.keepLineBreaks) {\n    args.push('--keep-line-breaks');\n  }\n  if (options.replaceInvalidChars) {\n    args.push('--replace-invalid-chars', options.replaceInvalidChars);\n  }\n  if (options.useStructTree) {\n    args.push('--use-struct-tree');\n  }\n  if (options.tableMethod) {\n    args.push('--table-method', options.tableMethod);\n  }\n  if (options.readingOrder) {\n    args.push('--reading-order', options.readingOrder);\n  }\n  if (options.markdownPageSeparator) {\n    args.push('--markdown-page-separator', options.markdownPageSeparator);\n  }\n  if (options.textPageSeparator) {\n    args.push('--text-page-separator', options.textPageSeparator);\n  }\n  if (options.htmlPageSeparator) {\n    args.push('--html-page-separator', options.htmlPageSeparator);\n  }\n  if (options.imageOutput) {\n    args.push('--image-output', options.imageOutput);\n  }\n  if (options.imageFormat) {\n    args.push('--image-format', options.imageFormat);\n  }\n\n  return args;\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\nimport { Command } from 'commander';\n\n/**\n * Register all CLI options on the given Commander program.\n */\nexport function registerCliOptions(program: Command): void {\n  program.option('-o, --output-dir <value>', 'Directory where output files are written. Default: input file directory');\n  program.option('-p, --password <value>', 'Password for encrypted PDF files');\n  program.option('-f, --format <value>', 'Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json');\n  program.option('-q, --quiet', 'Suppress console logging output');\n  program.option('--content-safety-off <value>', 'Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg');\n  program.option('--keep-line-breaks', 'Preserve original line breaks in extracted text');\n  program.option('--replace-invalid-chars <value>', 'Replacement character for invalid/unrecognized characters. Default: space');\n  program.option('--use-struct-tree', 'Use PDF structure tree (tagged PDF) for reading order and semantic structure');\n  program.option('--table-method <value>', 'Table detection method. Values: default (border-based), cluster (border + cluster). Default: default');\n  program.option('--reading-order <value>', 'Reading order algorithm. Values: off, xycut. Default: xycut');\n  program.option('--markdown-page-separator <value>', 'Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none');\n  program.option('--text-page-separator <value>', 'Separator between pages in text output. Use %page-number% for page numbers. Default: none');\n  program.option('--html-page-separator <value>', 'Separator between pages in HTML output. Use %page-number% for page numbers. Default: none');\n  program.option('--image-output <value>', 'Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external');\n  program.option('--image-format <value>', 'Output format for extracted images. Values: png, jpeg. Default: png');\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;AACA,uBAAwC;;;ACDxC,2BAAsB;AACtB,WAAsB;AACtB,SAAoB;AACpB,iBAA8B;;;AC4DvB,SAAS,oBAAoB,YAAwC;AAC1E,QAAM,iBAAiC,CAAC;AAExC,MAAI,WAAW,WAAW;AACxB,mBAAe,YAAY,WAAW;AAAA,EACxC;AACA,MAAI,WAAW,UAAU;AACvB,mBAAe,WAAW,WAAW;AAAA,EACvC;AACA,MAAI,WAAW,QAAQ;AACrB,mBAAe,SAAS,WAAW;AAAA,EACrC;AACA,MAAI,WAAW,OAAO;AACpB,mBAAe,QAAQ;AAAA,EACzB;AACA,MAAI,WAAW,kBAAkB;AAC/B,mBAAe,mBAAmB,WAAW;AAAA,EAC/C;AACA,MAAI,WAAW,gBAAgB;AAC7B,mBAAe,iBAAiB;AAAA,EAClC;AACA,MAAI,WAAW,qBAAqB;AAClC,mBAAe,sBAAsB,WAAW;AAAA,EAClD;AACA,MAAI,WAAW,eAAe;AAC5B,mBAAe,gBAAgB;AAAA,EACjC;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc,WAAW;AAAA,EAC1C;AACA,MAAI,WAAW,cAAc;AAC3B,mBAAe,eAAe,WAAW;AAAA,EAC3C;AACA,MAAI,WAAW,uBAAuB;AACpC,mBAAe,wBAAwB,WAAW;AAAA,EACpD;AACA,MAAI,WAAW,mBAAmB;AAChC,mBAAe,oBAAoB,WAAW;AAAA,EAChD;AACA,MAAI,WAAW,mBAAmB;AAChC,mBAAe,oBAAoB,WAAW;AAAA,EAChD;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc,WAAW;AAAA,EAC1C;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc,WAAW;AAAA,EAC1C;AAEA,SAAO;AACT;AAKO,SAAS,UAAU,SAAmC;AAC3D,QAAM,OAAiB,CAAC;AAExB,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,UAAU;AACpB,SAAK,KAAK,cAAc,QAAQ,QAAQ;AAAA,EAC1C;AACA,MAAI,QAAQ,QAAQ;AAClB,QAAI,MAAM,QAAQ,QAAQ,MAAM,GAAG;AACjC,UAAI,QAAQ,OAAO,SAAS,GAAG;AAC7B,aAAK,KAAK,YAAY,QAAQ,OAAO,KAAK,GAAG,CAAC;AAAA,MAChD;AAAA,IACF,OAAO;AACL,WAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,IACtC;AAAA,EACF;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,SAAS;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,MAAM,QAAQ,QAAQ,gBAAgB,GAAG;AAC3C,UAAI,QAAQ,iBAAiB,SAAS,GAAG;AACvC,aAAK,KAAK,wBAAwB,QAAQ,iBAAiB,KAAK,GAAG,CAAC;AAAA,MACtE;AAAA,IACF,OAAO;AACL,WAAK,KAAK,wBAAwB,QAAQ,gBAAgB;AAAA,IAC5D;AAAA,EACF;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,oBAAoB;AAAA,EAChC;AACA,MAAI,QAAQ,qBAAqB;AAC/B,SAAK,KAAK,2BAA2B,QAAQ,mBAAmB;AAAA,EAClE;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,cAAc;AACxB,SAAK,KAAK,mBAAmB,QAAQ,YAAY;AAAA,EACnD;AACA,MAAI,QAAQ,uBAAuB;AACjC,SAAK,KAAK,6BAA6B,QAAQ,qBAAqB;AAAA,EACtE;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AAEA,SAAO;AACT;;;ADpLA;AAWA,IAAM,iBAAa,0BAAc,YAAY,GAAG;AAChD,IAAM,YAAiB,aAAQ,UAAU;AAEzC,IAAM,WAAW;AAMjB,SAAS,WAAW,MAAgB,mBAAwC,CAAC,GAAoB;AAC/F,QAAM,EAAE,eAAe,MAAM,IAAI;AAEjC,SAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,UAAM,UAAe,UAAK,WAAW,MAAM,OAAO,QAAQ;AAE1D,QAAI,CAAI,cAAW,OAAO,GAAG;AAC3B,aAAO;AAAA,QACL,IAAI,MAAM,yBAAyB,OAAO,sCAAsC;AAAA,MAClF;AAAA,IACF;AAEA,UAAM,UAAU;AAChB,UAAM,cAAc,CAAC,QAAQ,SAAS,GAAG,IAAI;AAE7C,UAAM,kBAAc,4BAAM,SAAS,WAAW;AAE9C,QAAI,SAAS;AACb,QAAI,SAAS;AAEb,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,SAAS;AAChC,UAAI,SAAS,GAAG;AACd,gBAAQ,MAAM;AAAA,MAChB,OAAO;AACL,cAAM,cAAc,UAAU;AAC9B,cAAM,QAAQ,IAAI;AAAA,UAChB,+CAA+C,IAAI;AAAA;AAAA,EAAQ,WAAW;AAAA,QACxE;AACA,eAAO,KAAK;AAAA,MACd;AAAA,IACF,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,QAAe;AACtC,UAAI,IAAI,QAAQ,SAAS,QAAQ,GAAG;AAClC;AAAA,UACE,IAAI;AAAA,YACF;AAAA,UACF;AAAA,QACF;AAAA,MACF,OAAO;AACL,eAAO,GAAG;AAAA,MACZ;AAAA,IACF,CAAC;AAAA,EACH,CAAC;AACH;AAEO,SAAS,QACd,YACA,UAA0B,CAAC,GACV;AACjB,QAAM,YAAY,MAAM,QAAQ,UAAU,IAAI,aAAa,CAAC,UAAU;AACtE,MAAI,UAAU,WAAW,GAAG;AAC1B,WAAO,QAAQ,OAAO,IAAI,MAAM,2CAA2C,CAAC;AAAA,EAC9E;AAEA,aAAW,SAAS,WAAW;AAC7B,QAAI,CAAI,cAAW,KAAK,GAAG;AACzB,aAAO,QAAQ,OAAO,IAAI,MAAM,mCAAmC,KAAK,EAAE,CAAC;AAAA,IAC7E;AAAA,EACF;AAEA,QAAM,OAAiB,CAAC,GAAG,WAAW,GAAG,UAAU,OAAO,CAAC;AAE3D,SAAO,WAAW,MAAM;AAAA,IACtB,cAAc,CAAC,QAAQ;AAAA,EACzB,CAAC;AACH;;;AE9FO,SAAS,mBAAmB,SAAwB;AACzD,UAAQ,OAAO,4BAA4B,yEAAyE;AACpH,UAAQ,OAAO,0BAA0B,kCAAkC;AAC3E,UAAQ,OAAO,wBAAwB,oIAAoI;AAC3K,UAAQ,OAAO,eAAe,iCAAiC;AAC/D,UAAQ,OAAO,gCAAgC,sFAAsF;AACrI,UAAQ,OAAO,sBAAsB,iDAAiD;AACtF,UAAQ,OAAO,mCAAmC,2EAA2E;AAC7H,UAAQ,OAAO,qBAAqB,8EAA8E;AAClH,UAAQ,OAAO,0BAA0B,sGAAsG;AAC/I,UAAQ,OAAO,2BAA2B,6DAA6D;AACvG,UAAQ,OAAO,qCAAqC,+FAA+F;AACnJ,UAAQ,OAAO,iCAAiC,2FAA2F;AAC3I,UAAQ,OAAO,iCAAiC,2FAA2F;AAC3I,UAAQ,OAAO,0BAA0B,wHAAwH;AACjK,UAAQ,OAAO,0BAA0B,qEAAqE;AAChH;;;AHlBA,SAAS,gBAAyB;AAChC,QAAM,UAAU,IAAI,yBAAQ;AAE5B,UACG,KAAK,oBAAoB,EACzB,MAAM,sBAAsB,EAC5B,YAAY,4CAA4C,EACxD,mBAAmB,wCAAwC,EAC3D,yBAAyB,KAAK,EAC9B,SAAS,cAAc,uCAAuC;AAGjE,qBAAmB,OAAO;AAE1B,UAAQ,gBAAgB;AAAA,IACtB,UAAU,CAAC,QAAQ;AACjB,cAAQ,MAAM,IAAI,QAAQ,CAAC;AAAA,IAC7B;AAAA,IACA,aAAa,CAAC,KAAK,UAAU;AAC3B,YAAM,GAAG;AAAA,IACX;AAAA,EACF,CAAC;AAED,SAAO;AACT;AAEA,eAAe,OAAwB;AACrC,QAAM,UAAU,cAAc;AAE9B,UAAQ,aAAa;AAErB,MAAI;AACF,YAAQ,MAAM,QAAQ,IAAI;AAAA,EAC5B,SAAS,KAAK;AACZ,QAAI,eAAe,iCAAgB;AACjC,UAAI,IAAI,SAAS,2BAA2B;AAC1C,eAAO;AAAA,MACT;AACA,aAAO,IAAI,YAAY;AAAA,IACzB;AAEA,UAAM,UAAU,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC/D,YAAQ,MAAM,OAAO;AACrB,YAAQ,MAAM,wCAAwC;AACtD,WAAO;AAAA,EACT;AAEA,QAAM,aAAa,QAAQ,KAAiB;AAC5C,QAAM,aAAa,QAAQ;AAC3B,QAAM,iBAAiB,oBAAoB,UAAU;AAErD,MAAI;AACF,UAAM,SAAS,MAAM,QAAQ,YAAY,cAAc;AACvD,QAAI,UAAU,CAAC,eAAe,OAAO;AACnC,cAAQ,OAAO,MAAM,MAAM;AAC3B,UAAI,CAAC,OAAO,SAAS,IAAI,GAAG;AAC1B,gBAAQ,OAAO,MAAM,IAAI;AAAA,MAC3B;AAAA,IACF;AACA,WAAO;AAAA,EACT,SAAS,KAAK;AACZ,UAAM,UAAU,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC/D,YAAQ,MAAM,OAAO;AACrB,WAAO;AAAA,EACT;AACF;AAEA,KAAK,EAAE,KAAK,CAAC,SAAS;AACpB,MAAI,SAAS,GAAG;AACd,YAAQ,KAAK,IAAI;AAAA,EACnB;AACF,CAAC;","names":[]}

package/dist/cli.js

@@ -51,8 +51,8 @@ function buildConvertOptions(cliOptions) {
   if (cliOptions.htmlPageSeparator) {
     convertOptions.htmlPageSeparator = cliOptions.htmlPageSeparator;
   }
-  if (cliOptions.embedImages) {
-    convertOptions.embedImages = true;
+  if (cliOptions.imageOutput) {
+    convertOptions.imageOutput = cliOptions.imageOutput;
   }
   if (cliOptions.imageFormat) {
     convertOptions.imageFormat = cliOptions.imageFormat;
@@ -112,8 +112,8 @@ function buildArgs(options) {
   if (options.htmlPageSeparator) {
     args.push("--html-page-separator", options.htmlPageSeparator);
   }
-  if (options.embedImages) {
-    args.push("--embed-images");
+  if (options.imageOutput) {
+    args.push("--image-output", options.imageOutput);
   }
   if (options.imageFormat) {
     args.push("--image-format", options.imageFormat);
@@ -205,12 +205,12 @@ function registerCliOptions(program) {
   program.option("--keep-line-breaks", "Preserve original line breaks in extracted text");
   program.option("--replace-invalid-chars <value>", "Replacement character for invalid/unrecognized characters. Default: space");
   program.option("--use-struct-tree", "Use PDF structure tree (tagged PDF) for reading order and semantic structure");
-  program.option("--table-method <value>", "Table detection method. Values: cluster");
-  program.option("--reading-order <value>", "Reading order algorithm. Values: none, xycut. Default: none");
+  program.option("--table-method <value>", "Table detection method. Values: default (border-based), cluster (border + cluster). Default: default");
+  program.option("--reading-order <value>", "Reading order algorithm. Values: off, xycut. Default: xycut");
   program.option("--markdown-page-separator <value>", "Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none");
   program.option("--text-page-separator <value>", "Separator between pages in text output. Use %page-number% for page numbers. Default: none");
   program.option("--html-page-separator <value>", "Separator between pages in HTML output. Use %page-number% for page numbers. Default: none");
-  program.option("--embed-images", "Embed images as Base64 data URIs instead of file path references");
+  program.option("--image-output <value>", "Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external");
   program.option("--image-format <value>", "Output format for extracted images. Values: png, jpeg. Default: png");
 }
 

package/dist/cli.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../src/cli.ts","../src/index.ts","../src/convert-options.generated.ts","../src/cli-options.generated.ts"],"sourcesContent":["#!/usr/bin/env node\nimport { Command, CommanderError } from 'commander';\nimport { convert } from './index.js';\nimport { CliOptions, buildConvertOptions } from './convert-options.generated.js';\nimport { registerCliOptions } from './cli-options.generated.js';\n\nfunction createProgram(): Command {\n  const program = new Command();\n\n  program\n    .name('opendataloader-pdf')\n    .usage('[options] <input...>')\n    .description('Convert PDFs using the OpenDataLoader CLI.')\n    .showHelpAfterError(\"Use '--help' to see available options.\")\n    .showSuggestionAfterError(false)\n    .argument('<input...>', 'Input files or directories to convert');\n\n  // Register CLI options from auto-generated file\n  registerCliOptions(program);\n\n  program.configureOutput({\n    writeErr: (str) => {\n      console.error(str.trimEnd());\n    },\n    outputError: (str, write) => {\n      write(str);\n    },\n  });\n\n  return program;\n}\n\nasync function main(): Promise<number> {\n  const program = createProgram();\n\n  program.exitOverride();\n\n  try {\n    program.parse(process.argv);\n  } catch (err) {\n    if (err instanceof CommanderError) {\n      if (err.code === 'commander.helpDisplayed') {\n        return 0;\n      }\n      return err.exitCode ?? 1;\n    }\n\n    const message = err instanceof Error ? err.message : String(err);\n    console.error(message);\n    console.error(\"Use '--help' to see available options.\");\n    return 1;\n  }\n\n  const cliOptions = program.opts<CliOptions>();\n  const inputPaths = program.args;\n  const convertOptions = buildConvertOptions(cliOptions);\n\n  try {\n    const output = await convert(inputPaths, convertOptions);\n    if (output && !convertOptions.quiet) {\n      process.stdout.write(output);\n      if (!output.endsWith('\\n')) {\n        process.stdout.write('\\n');\n      }\n    }\n    return 0;\n  } catch (err) {\n    const message = err instanceof Error ? err.message : String(err);\n    console.error(message);\n    return 1;\n  }\n}\n\nmain().then((code) => {\n  if (code !== 0) {\n    process.exit(code);\n  }\n});\n","import { spawn } from 'child_process';\nimport * as path from 'path';\nimport * as fs from 'fs';\nimport { fileURLToPath } from 'url';\n\n// Re-export types and utilities from auto-generated file\nexport type { ConvertOptions } from './convert-options.generated.js';\nexport { buildArgs } from './convert-options.generated.js';\nimport type { ConvertOptions } from './convert-options.generated.js';\nimport { buildArgs } from './convert-options.generated.js';\n\nconst __filename = fileURLToPath(import.meta.url);\nconst __dirname = path.dirname(__filename);\n\nconst JAR_NAME = 'opendataloader-pdf-cli.jar';\n\ninterface JarExecutionOptions {\n  streamOutput?: boolean;\n}\n\nfunction executeJar(args: string[], executionOptions: JarExecutionOptions = {}): Promise<string> {\n  const { streamOutput = false } = executionOptions;\n\n  return new Promise((resolve, reject) => {\n    const jarPath = path.join(__dirname, '..', 'lib', JAR_NAME);\n\n    if (!fs.existsSync(jarPath)) {\n      return reject(\n        new Error(`JAR file not found at ${jarPath}. Please run the build script first.`),\n      );\n    }\n\n    const command = 'java';\n    const commandArgs = ['-jar', jarPath, ...args];\n\n    const javaProcess = spawn(command, commandArgs);\n\n    let stdout = '';\n    let stderr = '';\n\n    javaProcess.stdout.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stdout.write(chunk);\n      }\n      stdout += chunk;\n    });\n\n    javaProcess.stderr.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stderr.write(chunk);\n      }\n      stderr += chunk;\n    });\n\n    javaProcess.on('close', (code) => {\n      if (code === 0) {\n        resolve(stdout);\n      } else {\n        const errorOutput = stderr || stdout;\n        const error = new Error(\n          `The opendataloader-pdf CLI exited with code ${code}.\\n\\n${errorOutput}`,\n        );\n        reject(error);\n      }\n    });\n\n    javaProcess.on('error', (err: Error) => {\n      if (err.message.includes('ENOENT')) {\n        reject(\n          new Error(\n            \"'java' command not found. Please ensure Java is installed and in your system's PATH.\",\n          ),\n        );\n      } else {\n        reject(err);\n      }\n    });\n  });\n}\n\nexport function convert(\n  inputPaths: string | string[],\n  options: ConvertOptions = {},\n): Promise<string> {\n  const inputList = Array.isArray(inputPaths) ? inputPaths : [inputPaths];\n  if (inputList.length === 0) {\n    return Promise.reject(new Error('At least one input path must be provided.'));\n  }\n\n  for (const input of inputList) {\n    if (!fs.existsSync(input)) {\n      return Promise.reject(new Error(`Input file or folder not found: ${input}`));\n    }\n  }\n\n  const args: string[] = [...inputList, ...buildArgs(options)];\n\n  return executeJar(args, {\n    streamOutput: !options.quiet,\n  });\n}\n\n/**\n * @deprecated Use `convert()` and `ConvertOptions` instead. This function will be removed in a future version.\n */\nexport interface RunOptions {\n  outputFolder?: string;\n  password?: string;\n  replaceInvalidChars?: string;\n  generateMarkdown?: boolean;\n  generateHtml?: boolean;\n  generateAnnotatedPdf?: boolean;\n  keepLineBreaks?: boolean;\n  contentSafetyOff?: string;\n  htmlInMarkdown?: boolean;\n  addImageToMarkdown?: boolean;\n  noJson?: boolean;\n  debug?: boolean;\n  useStructTree?: boolean;\n}\n\n/**\n * @deprecated Use `convert()` instead. This function will be removed in a future version.\n */\nexport function run(inputPath: string, options: RunOptions = {}): Promise<string> {\n  console.warn(\n    'Warning: run() is deprecated and will be removed in a future version. Use convert() instead.',\n  );\n\n  // Build format array based on legacy boolean options\n  const formats: string[] = [];\n  if (!options.noJson) {\n    formats.push('json');\n  }\n  if (options.generateMarkdown) {\n    if (options.addImageToMarkdown) {\n      formats.push('markdown-with-images');\n    } else if (options.htmlInMarkdown) {\n      formats.push('markdown-with-html');\n    } else {\n      formats.push('markdown');\n    }\n  }\n  if (options.generateHtml) {\n    formats.push('html');\n  }\n  if (options.generateAnnotatedPdf) {\n    formats.push('pdf');\n  }\n\n  return convert(inputPath, {\n    outputDir: options.outputFolder,\n    password: options.password,\n    replaceInvalidChars: options.replaceInvalidChars,\n    keepLineBreaks: options.keepLineBreaks,\n    contentSafetyOff: options.contentSafetyOff,\n    useStructTree: options.useStructTree,\n    format: formats.length > 0 ? formats : undefined,\n    quiet: !options.debug,\n  });\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\n/**\n * Options for the convert function.\n */\nexport interface ConvertOptions {\n  /** Directory where output files are written. Default: input file directory */\n  outputDir?: string;\n  /** Password for encrypted PDF files */\n  password?: string;\n  /** Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json */\n  format?: string | string[];\n  /** Suppress console logging output */\n  quiet?: boolean;\n  /** Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg */\n  contentSafetyOff?: string | string[];\n  /** Preserve original line breaks in extracted text */\n  keepLineBreaks?: boolean;\n  /** Replacement character for invalid/unrecognized characters. Default: space */\n  replaceInvalidChars?: string;\n  /** Use PDF structure tree (tagged PDF) for reading order and semantic structure */\n  useStructTree?: boolean;\n  /** Table detection method. Values: cluster */\n  tableMethod?: string;\n  /** Reading order algorithm. Values: none, xycut. Default: none */\n  readingOrder?: string;\n  /** Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none */\n  markdownPageSeparator?: string;\n  /** Separator between pages in text output. Use %page-number% for page numbers. Default: none */\n  textPageSeparator?: string;\n  /** Separator between pages in HTML output. Use %page-number% for page numbers. Default: none */\n  htmlPageSeparator?: string;\n  /** Embed images as Base64 data URIs instead of file path references */\n  embedImages?: boolean;\n  /** Output format for extracted images. Values: png, jpeg. Default: png */\n  imageFormat?: string;\n}\n\n/**\n * Options as parsed from CLI (all values are strings from commander).\n */\nexport interface CliOptions {\n  outputDir?: string;\n  password?: string;\n  format?: string;\n  quiet?: boolean;\n  contentSafetyOff?: string;\n  keepLineBreaks?: boolean;\n  replaceInvalidChars?: string;\n  useStructTree?: boolean;\n  tableMethod?: string;\n  readingOrder?: string;\n  markdownPageSeparator?: string;\n  textPageSeparator?: string;\n  htmlPageSeparator?: string;\n  embedImages?: boolean;\n  imageFormat?: string;\n}\n\n/**\n * Convert CLI options to ConvertOptions.\n */\nexport function buildConvertOptions(cliOptions: CliOptions): ConvertOptions {\n  const convertOptions: ConvertOptions = {};\n\n  if (cliOptions.outputDir) {\n    convertOptions.outputDir = cliOptions.outputDir;\n  }\n  if (cliOptions.password) {\n    convertOptions.password = cliOptions.password;\n  }\n  if (cliOptions.format) {\n    convertOptions.format = cliOptions.format;\n  }\n  if (cliOptions.quiet) {\n    convertOptions.quiet = true;\n  }\n  if (cliOptions.contentSafetyOff) {\n    convertOptions.contentSafetyOff = cliOptions.contentSafetyOff;\n  }\n  if (cliOptions.keepLineBreaks) {\n    convertOptions.keepLineBreaks = true;\n  }\n  if (cliOptions.replaceInvalidChars) {\n    convertOptions.replaceInvalidChars = cliOptions.replaceInvalidChars;\n  }\n  if (cliOptions.useStructTree) {\n    convertOptions.useStructTree = true;\n  }\n  if (cliOptions.tableMethod) {\n    convertOptions.tableMethod = cliOptions.tableMethod;\n  }\n  if (cliOptions.readingOrder) {\n    convertOptions.readingOrder = cliOptions.readingOrder;\n  }\n  if (cliOptions.markdownPageSeparator) {\n    convertOptions.markdownPageSeparator = cliOptions.markdownPageSeparator;\n  }\n  if (cliOptions.textPageSeparator) {\n    convertOptions.textPageSeparator = cliOptions.textPageSeparator;\n  }\n  if (cliOptions.htmlPageSeparator) {\n    convertOptions.htmlPageSeparator = cliOptions.htmlPageSeparator;\n  }\n  if (cliOptions.embedImages) {\n    convertOptions.embedImages = true;\n  }\n  if (cliOptions.imageFormat) {\n    convertOptions.imageFormat = cliOptions.imageFormat;\n  }\n\n  return convertOptions;\n}\n\n/**\n * Build CLI arguments array from ConvertOptions.\n */\nexport function buildArgs(options: ConvertOptions): string[] {\n  const args: string[] = [];\n\n  if (options.outputDir) {\n    args.push('--output-dir', options.outputDir);\n  }\n  if (options.password) {\n    args.push('--password', options.password);\n  }\n  if (options.format) {\n    if (Array.isArray(options.format)) {\n      if (options.format.length > 0) {\n        args.push('--format', options.format.join(','));\n      }\n    } else {\n      args.push('--format', options.format);\n    }\n  }\n  if (options.quiet) {\n    args.push('--quiet');\n  }\n  if (options.contentSafetyOff) {\n    if (Array.isArray(options.contentSafetyOff)) {\n      if (options.contentSafetyOff.length > 0) {\n        args.push('--content-safety-off', options.contentSafetyOff.join(','));\n      }\n    } else {\n      args.push('--content-safety-off', options.contentSafetyOff);\n    }\n  }\n  if (options.keepLineBreaks) {\n    args.push('--keep-line-breaks');\n  }\n  if (options.replaceInvalidChars) {\n    args.push('--replace-invalid-chars', options.replaceInvalidChars);\n  }\n  if (options.useStructTree) {\n    args.push('--use-struct-tree');\n  }\n  if (options.tableMethod) {\n    args.push('--table-method', options.tableMethod);\n  }\n  if (options.readingOrder) {\n    args.push('--reading-order', options.readingOrder);\n  }\n  if (options.markdownPageSeparator) {\n    args.push('--markdown-page-separator', options.markdownPageSeparator);\n  }\n  if (options.textPageSeparator) {\n    args.push('--text-page-separator', options.textPageSeparator);\n  }\n  if (options.htmlPageSeparator) {\n    args.push('--html-page-separator', options.htmlPageSeparator);\n  }\n  if (options.embedImages) {\n    args.push('--embed-images');\n  }\n  if (options.imageFormat) {\n    args.push('--image-format', options.imageFormat);\n  }\n\n  return args;\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\nimport { Command } from 'commander';\n\n/**\n * Register all CLI options on the given Commander program.\n */\nexport function registerCliOptions(program: Command): void {\n  program.option('-o, --output-dir <value>', 'Directory where output files are written. Default: input file directory');\n  program.option('-p, --password <value>', 'Password for encrypted PDF files');\n  program.option('-f, --format <value>', 'Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json');\n  program.option('-q, --quiet', 'Suppress console logging output');\n  program.option('--content-safety-off <value>', 'Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg');\n  program.option('--keep-line-breaks', 'Preserve original line breaks in extracted text');\n  program.option('--replace-invalid-chars <value>', 'Replacement character for invalid/unrecognized characters. Default: space');\n  program.option('--use-struct-tree', 'Use PDF structure tree (tagged PDF) for reading order and semantic structure');\n  program.option('--table-method <value>', 'Table detection method. Values: cluster');\n  program.option('--reading-order <value>', 'Reading order algorithm. Values: none, xycut. Default: none');\n  program.option('--markdown-page-separator <value>', 'Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none');\n  program.option('--text-page-separator <value>', 'Separator between pages in text output. Use %page-number% for page numbers. Default: none');\n  program.option('--html-page-separator <value>', 'Separator between pages in HTML output. Use %page-number% for page numbers. Default: none');\n  program.option('--embed-images', 'Embed images as Base64 data URIs instead of file path references');\n  program.option('--image-format <value>', 'Output format for extracted images. Values: png, jpeg. Default: png');\n}\n"],"mappings":";;;AACA,SAAS,SAAS,sBAAsB;;;ACDxC,SAAS,aAAa;AACtB,YAAY,UAAU;AACtB,YAAY,QAAQ;AACpB,SAAS,qBAAqB;;;AC4DvB,SAAS,oBAAoB,YAAwC;AAC1E,QAAM,iBAAiC,CAAC;AAExC,MAAI,WAAW,WAAW;AACxB,mBAAe,YAAY,WAAW;AAAA,EACxC;AACA,MAAI,WAAW,UAAU;AACvB,mBAAe,WAAW,WAAW;AAAA,EACvC;AACA,MAAI,WAAW,QAAQ;AACrB,mBAAe,SAAS,WAAW;AAAA,EACrC;AACA,MAAI,WAAW,OAAO;AACpB,mBAAe,QAAQ;AAAA,EACzB;AACA,MAAI,WAAW,kBAAkB;AAC/B,mBAAe,mBAAmB,WAAW;AAAA,EAC/C;AACA,MAAI,WAAW,gBAAgB;AAC7B,mBAAe,iBAAiB;AAAA,EAClC;AACA,MAAI,WAAW,qBAAqB;AAClC,mBAAe,sBAAsB,WAAW;AAAA,EAClD;AACA,MAAI,WAAW,eAAe;AAC5B,mBAAe,gBAAgB;AAAA,EACjC;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc,WAAW;AAAA,EAC1C;AACA,MAAI,WAAW,cAAc;AAC3B,mBAAe,eAAe,WAAW;AAAA,EAC3C;AACA,MAAI,WAAW,uBAAuB;AACpC,mBAAe,wBAAwB,WAAW;AAAA,EACpD;AACA,MAAI,WAAW,mBAAmB;AAChC,mBAAe,oBAAoB,WAAW;AAAA,EAChD;AACA,MAAI,WAAW,mBAAmB;AAChC,mBAAe,oBAAoB,WAAW;AAAA,EAChD;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc;AAAA,EAC/B;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc,WAAW;AAAA,EAC1C;AAEA,SAAO;AACT;AAKO,SAAS,UAAU,SAAmC;AAC3D,QAAM,OAAiB,CAAC;AAExB,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,UAAU;AACpB,SAAK,KAAK,cAAc,QAAQ,QAAQ;AAAA,EAC1C;AACA,MAAI,QAAQ,QAAQ;AAClB,QAAI,MAAM,QAAQ,QAAQ,MAAM,GAAG;AACjC,UAAI,QAAQ,OAAO,SAAS,GAAG;AAC7B,aAAK,KAAK,YAAY,QAAQ,OAAO,KAAK,GAAG,CAAC;AAAA,MAChD;AAAA,IACF,OAAO;AACL,WAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,IACtC;AAAA,EACF;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,SAAS;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,MAAM,QAAQ,QAAQ,gBAAgB,GAAG;AAC3C,UAAI,QAAQ,iBAAiB,SAAS,GAAG;AACvC,aAAK,KAAK,wBAAwB,QAAQ,iBAAiB,KAAK,GAAG,CAAC;AAAA,MACtE;AAAA,IACF,OAAO;AACL,WAAK,KAAK,wBAAwB,QAAQ,gBAAgB;AAAA,IAC5D;AAAA,EACF;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,oBAAoB;AAAA,EAChC;AACA,MAAI,QAAQ,qBAAqB;AAC/B,SAAK,KAAK,2BAA2B,QAAQ,mBAAmB;AAAA,EAClE;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,cAAc;AACxB,SAAK,KAAK,mBAAmB,QAAQ,YAAY;AAAA,EACnD;AACA,MAAI,QAAQ,uBAAuB;AACjC,SAAK,KAAK,6BAA6B,QAAQ,qBAAqB;AAAA,EACtE;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,gBAAgB;AAAA,EAC5B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AAEA,SAAO;AACT;;;ADzKA,IAAM,aAAa,cAAc,YAAY,GAAG;AAChD,IAAM,YAAiB,aAAQ,UAAU;AAEzC,IAAM,WAAW;AAMjB,SAAS,WAAW,MAAgB,mBAAwC,CAAC,GAAoB;AAC/F,QAAM,EAAE,eAAe,MAAM,IAAI;AAEjC,SAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,UAAM,UAAe,UAAK,WAAW,MAAM,OAAO,QAAQ;AAE1D,QAAI,CAAI,cAAW,OAAO,GAAG;AAC3B,aAAO;AAAA,QACL,IAAI,MAAM,yBAAyB,OAAO,sCAAsC;AAAA,MAClF;AAAA,IACF;AAEA,UAAM,UAAU;AAChB,UAAM,cAAc,CAAC,QAAQ,SAAS,GAAG,IAAI;AAE7C,UAAM,cAAc,MAAM,SAAS,WAAW;AAE9C,QAAI,SAAS;AACb,QAAI,SAAS;AAEb,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,SAAS;AAChC,UAAI,SAAS,GAAG;AACd,gBAAQ,MAAM;AAAA,MAChB,OAAO;AACL,cAAM,cAAc,UAAU;AAC9B,cAAM,QAAQ,IAAI;AAAA,UAChB,+CAA+C,IAAI;AAAA;AAAA,EAAQ,WAAW;AAAA,QACxE;AACA,eAAO,KAAK;AAAA,MACd;AAAA,IACF,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,QAAe;AACtC,UAAI,IAAI,QAAQ,SAAS,QAAQ,GAAG;AAClC;AAAA,UACE,IAAI;AAAA,YACF;AAAA,UACF;AAAA,QACF;AAAA,MACF,OAAO;AACL,eAAO,GAAG;AAAA,MACZ;AAAA,IACF,CAAC;AAAA,EACH,CAAC;AACH;AAEO,SAAS,QACd,YACA,UAA0B,CAAC,GACV;AACjB,QAAM,YAAY,MAAM,QAAQ,UAAU,IAAI,aAAa,CAAC,UAAU;AACtE,MAAI,UAAU,WAAW,GAAG;AAC1B,WAAO,QAAQ,OAAO,IAAI,MAAM,2CAA2C,CAAC;AAAA,EAC9E;AAEA,aAAW,SAAS,WAAW;AAC7B,QAAI,CAAI,cAAW,KAAK,GAAG;AACzB,aAAO,QAAQ,OAAO,IAAI,MAAM,mCAAmC,KAAK,EAAE,CAAC;AAAA,IAC7E;AAAA,EACF;AAEA,QAAM,OAAiB,CAAC,GAAG,WAAW,GAAG,UAAU,OAAO,CAAC;AAE3D,SAAO,WAAW,MAAM;AAAA,IACtB,cAAc,CAAC,QAAQ;AAAA,EACzB,CAAC;AACH;;;AE9FO,SAAS,mBAAmB,SAAwB;AACzD,UAAQ,OAAO,4BAA4B,yEAAyE;AACpH,UAAQ,OAAO,0BAA0B,kCAAkC;AAC3E,UAAQ,OAAO,wBAAwB,oIAAoI;AAC3K,UAAQ,OAAO,eAAe,iCAAiC;AAC/D,UAAQ,OAAO,gCAAgC,sFAAsF;AACrI,UAAQ,OAAO,sBAAsB,iDAAiD;AACtF,UAAQ,OAAO,mCAAmC,2EAA2E;AAC7H,UAAQ,OAAO,qBAAqB,8EAA8E;AAClH,UAAQ,OAAO,0BAA0B,yCAAyC;AAClF,UAAQ,OAAO,2BAA2B,6DAA6D;AACvG,UAAQ,OAAO,qCAAqC,+FAA+F;AACnJ,UAAQ,OAAO,iCAAiC,2FAA2F;AAC3I,UAAQ,OAAO,iCAAiC,2FAA2F;AAC3I,UAAQ,OAAO,kBAAkB,kEAAkE;AACnG,UAAQ,OAAO,0BAA0B,qEAAqE;AAChH;;;AHlBA,SAAS,gBAAyB;AAChC,QAAM,UAAU,IAAI,QAAQ;AAE5B,UACG,KAAK,oBAAoB,EACzB,MAAM,sBAAsB,EAC5B,YAAY,4CAA4C,EACxD,mBAAmB,wCAAwC,EAC3D,yBAAyB,KAAK,EAC9B,SAAS,cAAc,uCAAuC;AAGjE,qBAAmB,OAAO;AAE1B,UAAQ,gBAAgB;AAAA,IACtB,UAAU,CAAC,QAAQ;AACjB,cAAQ,MAAM,IAAI,QAAQ,CAAC;AAAA,IAC7B;AAAA,IACA,aAAa,CAAC,KAAK,UAAU;AAC3B,YAAM,GAAG;AAAA,IACX;AAAA,EACF,CAAC;AAED,SAAO;AACT;AAEA,eAAe,OAAwB;AACrC,QAAM,UAAU,cAAc;AAE9B,UAAQ,aAAa;AAErB,MAAI;AACF,YAAQ,MAAM,QAAQ,IAAI;AAAA,EAC5B,SAAS,KAAK;AACZ,QAAI,eAAe,gBAAgB;AACjC,UAAI,IAAI,SAAS,2BAA2B;AAC1C,eAAO;AAAA,MACT;AACA,aAAO,IAAI,YAAY;AAAA,IACzB;AAEA,UAAM,UAAU,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC/D,YAAQ,MAAM,OAAO;AACrB,YAAQ,MAAM,wCAAwC;AACtD,WAAO;AAAA,EACT;AAEA,QAAM,aAAa,QAAQ,KAAiB;AAC5C,QAAM,aAAa,QAAQ;AAC3B,QAAM,iBAAiB,oBAAoB,UAAU;AAErD,MAAI;AACF,UAAM,SAAS,MAAM,QAAQ,YAAY,cAAc;AACvD,QAAI,UAAU,CAAC,eAAe,OAAO;AACnC,cAAQ,OAAO,MAAM,MAAM;AAC3B,UAAI,CAAC,OAAO,SAAS,IAAI,GAAG;AAC1B,gBAAQ,OAAO,MAAM,IAAI;AAAA,MAC3B;AAAA,IACF;AACA,WAAO;AAAA,EACT,SAAS,KAAK;AACZ,UAAM,UAAU,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC/D,YAAQ,MAAM,OAAO;AACrB,WAAO;AAAA,EACT;AACF;AAEA,KAAK,EAAE,KAAK,CAAC,SAAS;AACpB,MAAI,SAAS,GAAG;AACd,YAAQ,KAAK,IAAI;AAAA,EACnB;AACF,CAAC;","names":[]}
+{"version":3,"sources":["../src/cli.ts","../src/index.ts","../src/convert-options.generated.ts","../src/cli-options.generated.ts"],"sourcesContent":["#!/usr/bin/env node\nimport { Command, CommanderError } from 'commander';\nimport { convert } from './index.js';\nimport { CliOptions, buildConvertOptions } from './convert-options.generated.js';\nimport { registerCliOptions } from './cli-options.generated.js';\n\nfunction createProgram(): Command {\n  const program = new Command();\n\n  program\n    .name('opendataloader-pdf')\n    .usage('[options] <input...>')\n    .description('Convert PDFs using the OpenDataLoader CLI.')\n    .showHelpAfterError(\"Use '--help' to see available options.\")\n    .showSuggestionAfterError(false)\n    .argument('<input...>', 'Input files or directories to convert');\n\n  // Register CLI options from auto-generated file\n  registerCliOptions(program);\n\n  program.configureOutput({\n    writeErr: (str) => {\n      console.error(str.trimEnd());\n    },\n    outputError: (str, write) => {\n      write(str);\n    },\n  });\n\n  return program;\n}\n\nasync function main(): Promise<number> {\n  const program = createProgram();\n\n  program.exitOverride();\n\n  try {\n    program.parse(process.argv);\n  } catch (err) {\n    if (err instanceof CommanderError) {\n      if (err.code === 'commander.helpDisplayed') {\n        return 0;\n      }\n      return err.exitCode ?? 1;\n    }\n\n    const message = err instanceof Error ? err.message : String(err);\n    console.error(message);\n    console.error(\"Use '--help' to see available options.\");\n    return 1;\n  }\n\n  const cliOptions = program.opts<CliOptions>();\n  const inputPaths = program.args;\n  const convertOptions = buildConvertOptions(cliOptions);\n\n  try {\n    const output = await convert(inputPaths, convertOptions);\n    if (output && !convertOptions.quiet) {\n      process.stdout.write(output);\n      if (!output.endsWith('\\n')) {\n        process.stdout.write('\\n');\n      }\n    }\n    return 0;\n  } catch (err) {\n    const message = err instanceof Error ? err.message : String(err);\n    console.error(message);\n    return 1;\n  }\n}\n\nmain().then((code) => {\n  if (code !== 0) {\n    process.exit(code);\n  }\n});\n","import { spawn } from 'child_process';\nimport * as path from 'path';\nimport * as fs from 'fs';\nimport { fileURLToPath } from 'url';\n\n// Re-export types and utilities from auto-generated file\nexport type { ConvertOptions } from './convert-options.generated.js';\nexport { buildArgs } from './convert-options.generated.js';\nimport type { ConvertOptions } from './convert-options.generated.js';\nimport { buildArgs } from './convert-options.generated.js';\n\nconst __filename = fileURLToPath(import.meta.url);\nconst __dirname = path.dirname(__filename);\n\nconst JAR_NAME = 'opendataloader-pdf-cli.jar';\n\ninterface JarExecutionOptions {\n  streamOutput?: boolean;\n}\n\nfunction executeJar(args: string[], executionOptions: JarExecutionOptions = {}): Promise<string> {\n  const { streamOutput = false } = executionOptions;\n\n  return new Promise((resolve, reject) => {\n    const jarPath = path.join(__dirname, '..', 'lib', JAR_NAME);\n\n    if (!fs.existsSync(jarPath)) {\n      return reject(\n        new Error(`JAR file not found at ${jarPath}. Please run the build script first.`),\n      );\n    }\n\n    const command = 'java';\n    const commandArgs = ['-jar', jarPath, ...args];\n\n    const javaProcess = spawn(command, commandArgs);\n\n    let stdout = '';\n    let stderr = '';\n\n    javaProcess.stdout.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stdout.write(chunk);\n      }\n      stdout += chunk;\n    });\n\n    javaProcess.stderr.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stderr.write(chunk);\n      }\n      stderr += chunk;\n    });\n\n    javaProcess.on('close', (code) => {\n      if (code === 0) {\n        resolve(stdout);\n      } else {\n        const errorOutput = stderr || stdout;\n        const error = new Error(\n          `The opendataloader-pdf CLI exited with code ${code}.\\n\\n${errorOutput}`,\n        );\n        reject(error);\n      }\n    });\n\n    javaProcess.on('error', (err: Error) => {\n      if (err.message.includes('ENOENT')) {\n        reject(\n          new Error(\n            \"'java' command not found. Please ensure Java is installed and in your system's PATH.\",\n          ),\n        );\n      } else {\n        reject(err);\n      }\n    });\n  });\n}\n\nexport function convert(\n  inputPaths: string | string[],\n  options: ConvertOptions = {},\n): Promise<string> {\n  const inputList = Array.isArray(inputPaths) ? inputPaths : [inputPaths];\n  if (inputList.length === 0) {\n    return Promise.reject(new Error('At least one input path must be provided.'));\n  }\n\n  for (const input of inputList) {\n    if (!fs.existsSync(input)) {\n      return Promise.reject(new Error(`Input file or folder not found: ${input}`));\n    }\n  }\n\n  const args: string[] = [...inputList, ...buildArgs(options)];\n\n  return executeJar(args, {\n    streamOutput: !options.quiet,\n  });\n}\n\n/**\n * @deprecated Use `convert()` and `ConvertOptions` instead. This function will be removed in a future version.\n */\nexport interface RunOptions {\n  outputFolder?: string;\n  password?: string;\n  replaceInvalidChars?: string;\n  generateMarkdown?: boolean;\n  generateHtml?: boolean;\n  generateAnnotatedPdf?: boolean;\n  keepLineBreaks?: boolean;\n  contentSafetyOff?: string;\n  htmlInMarkdown?: boolean;\n  addImageToMarkdown?: boolean;\n  noJson?: boolean;\n  debug?: boolean;\n  useStructTree?: boolean;\n}\n\n/**\n * @deprecated Use `convert()` instead. This function will be removed in a future version.\n */\nexport function run(inputPath: string, options: RunOptions = {}): Promise<string> {\n  console.warn(\n    'Warning: run() is deprecated and will be removed in a future version. Use convert() instead.',\n  );\n\n  // Build format array based on legacy boolean options\n  const formats: string[] = [];\n  if (!options.noJson) {\n    formats.push('json');\n  }\n  if (options.generateMarkdown) {\n    if (options.addImageToMarkdown) {\n      formats.push('markdown-with-images');\n    } else if (options.htmlInMarkdown) {\n      formats.push('markdown-with-html');\n    } else {\n      formats.push('markdown');\n    }\n  }\n  if (options.generateHtml) {\n    formats.push('html');\n  }\n  if (options.generateAnnotatedPdf) {\n    formats.push('pdf');\n  }\n\n  return convert(inputPath, {\n    outputDir: options.outputFolder,\n    password: options.password,\n    replaceInvalidChars: options.replaceInvalidChars,\n    keepLineBreaks: options.keepLineBreaks,\n    contentSafetyOff: options.contentSafetyOff,\n    useStructTree: options.useStructTree,\n    format: formats.length > 0 ? formats : undefined,\n    quiet: !options.debug,\n  });\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\n/**\n * Options for the convert function.\n */\nexport interface ConvertOptions {\n  /** Directory where output files are written. Default: input file directory */\n  outputDir?: string;\n  /** Password for encrypted PDF files */\n  password?: string;\n  /** Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json */\n  format?: string | string[];\n  /** Suppress console logging output */\n  quiet?: boolean;\n  /** Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg */\n  contentSafetyOff?: string | string[];\n  /** Preserve original line breaks in extracted text */\n  keepLineBreaks?: boolean;\n  /** Replacement character for invalid/unrecognized characters. Default: space */\n  replaceInvalidChars?: string;\n  /** Use PDF structure tree (tagged PDF) for reading order and semantic structure */\n  useStructTree?: boolean;\n  /** Table detection method. Values: default (border-based), cluster (border + cluster). Default: default */\n  tableMethod?: string;\n  /** Reading order algorithm. Values: off, xycut. Default: xycut */\n  readingOrder?: string;\n  /** Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none */\n  markdownPageSeparator?: string;\n  /** Separator between pages in text output. Use %page-number% for page numbers. Default: none */\n  textPageSeparator?: string;\n  /** Separator between pages in HTML output. Use %page-number% for page numbers. Default: none */\n  htmlPageSeparator?: string;\n  /** Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external */\n  imageOutput?: string;\n  /** Output format for extracted images. Values: png, jpeg. Default: png */\n  imageFormat?: string;\n}\n\n/**\n * Options as parsed from CLI (all values are strings from commander).\n */\nexport interface CliOptions {\n  outputDir?: string;\n  password?: string;\n  format?: string;\n  quiet?: boolean;\n  contentSafetyOff?: string;\n  keepLineBreaks?: boolean;\n  replaceInvalidChars?: string;\n  useStructTree?: boolean;\n  tableMethod?: string;\n  readingOrder?: string;\n  markdownPageSeparator?: string;\n  textPageSeparator?: string;\n  htmlPageSeparator?: string;\n  imageOutput?: string;\n  imageFormat?: string;\n}\n\n/**\n * Convert CLI options to ConvertOptions.\n */\nexport function buildConvertOptions(cliOptions: CliOptions): ConvertOptions {\n  const convertOptions: ConvertOptions = {};\n\n  if (cliOptions.outputDir) {\n    convertOptions.outputDir = cliOptions.outputDir;\n  }\n  if (cliOptions.password) {\n    convertOptions.password = cliOptions.password;\n  }\n  if (cliOptions.format) {\n    convertOptions.format = cliOptions.format;\n  }\n  if (cliOptions.quiet) {\n    convertOptions.quiet = true;\n  }\n  if (cliOptions.contentSafetyOff) {\n    convertOptions.contentSafetyOff = cliOptions.contentSafetyOff;\n  }\n  if (cliOptions.keepLineBreaks) {\n    convertOptions.keepLineBreaks = true;\n  }\n  if (cliOptions.replaceInvalidChars) {\n    convertOptions.replaceInvalidChars = cliOptions.replaceInvalidChars;\n  }\n  if (cliOptions.useStructTree) {\n    convertOptions.useStructTree = true;\n  }\n  if (cliOptions.tableMethod) {\n    convertOptions.tableMethod = cliOptions.tableMethod;\n  }\n  if (cliOptions.readingOrder) {\n    convertOptions.readingOrder = cliOptions.readingOrder;\n  }\n  if (cliOptions.markdownPageSeparator) {\n    convertOptions.markdownPageSeparator = cliOptions.markdownPageSeparator;\n  }\n  if (cliOptions.textPageSeparator) {\n    convertOptions.textPageSeparator = cliOptions.textPageSeparator;\n  }\n  if (cliOptions.htmlPageSeparator) {\n    convertOptions.htmlPageSeparator = cliOptions.htmlPageSeparator;\n  }\n  if (cliOptions.imageOutput) {\n    convertOptions.imageOutput = cliOptions.imageOutput;\n  }\n  if (cliOptions.imageFormat) {\n    convertOptions.imageFormat = cliOptions.imageFormat;\n  }\n\n  return convertOptions;\n}\n\n/**\n * Build CLI arguments array from ConvertOptions.\n */\nexport function buildArgs(options: ConvertOptions): string[] {\n  const args: string[] = [];\n\n  if (options.outputDir) {\n    args.push('--output-dir', options.outputDir);\n  }\n  if (options.password) {\n    args.push('--password', options.password);\n  }\n  if (options.format) {\n    if (Array.isArray(options.format)) {\n      if (options.format.length > 0) {\n        args.push('--format', options.format.join(','));\n      }\n    } else {\n      args.push('--format', options.format);\n    }\n  }\n  if (options.quiet) {\n    args.push('--quiet');\n  }\n  if (options.contentSafetyOff) {\n    if (Array.isArray(options.contentSafetyOff)) {\n      if (options.contentSafetyOff.length > 0) {\n        args.push('--content-safety-off', options.contentSafetyOff.join(','));\n      }\n    } else {\n      args.push('--content-safety-off', options.contentSafetyOff);\n    }\n  }\n  if (options.keepLineBreaks) {\n    args.push('--keep-line-breaks');\n  }\n  if (options.replaceInvalidChars) {\n    args.push('--replace-invalid-chars', options.replaceInvalidChars);\n  }\n  if (options.useStructTree) {\n    args.push('--use-struct-tree');\n  }\n  if (options.tableMethod) {\n    args.push('--table-method', options.tableMethod);\n  }\n  if (options.readingOrder) {\n    args.push('--reading-order', options.readingOrder);\n  }\n  if (options.markdownPageSeparator) {\n    args.push('--markdown-page-separator', options.markdownPageSeparator);\n  }\n  if (options.textPageSeparator) {\n    args.push('--text-page-separator', options.textPageSeparator);\n  }\n  if (options.htmlPageSeparator) {\n    args.push('--html-page-separator', options.htmlPageSeparator);\n  }\n  if (options.imageOutput) {\n    args.push('--image-output', options.imageOutput);\n  }\n  if (options.imageFormat) {\n    args.push('--image-format', options.imageFormat);\n  }\n\n  return args;\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\nimport { Command } from 'commander';\n\n/**\n * Register all CLI options on the given Commander program.\n */\nexport function registerCliOptions(program: Command): void {\n  program.option('-o, --output-dir <value>', 'Directory where output files are written. Default: input file directory');\n  program.option('-p, --password <value>', 'Password for encrypted PDF files');\n  program.option('-f, --format <value>', 'Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json');\n  program.option('-q, --quiet', 'Suppress console logging output');\n  program.option('--content-safety-off <value>', 'Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg');\n  program.option('--keep-line-breaks', 'Preserve original line breaks in extracted text');\n  program.option('--replace-invalid-chars <value>', 'Replacement character for invalid/unrecognized characters. Default: space');\n  program.option('--use-struct-tree', 'Use PDF structure tree (tagged PDF) for reading order and semantic structure');\n  program.option('--table-method <value>', 'Table detection method. Values: default (border-based), cluster (border + cluster). Default: default');\n  program.option('--reading-order <value>', 'Reading order algorithm. Values: off, xycut. Default: xycut');\n  program.option('--markdown-page-separator <value>', 'Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none');\n  program.option('--text-page-separator <value>', 'Separator between pages in text output. Use %page-number% for page numbers. Default: none');\n  program.option('--html-page-separator <value>', 'Separator between pages in HTML output. Use %page-number% for page numbers. Default: none');\n  program.option('--image-output <value>', 'Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external');\n  program.option('--image-format <value>', 'Output format for extracted images. Values: png, jpeg. Default: png');\n}\n"],"mappings":";;;AACA,SAAS,SAAS,sBAAsB;;;ACDxC,SAAS,aAAa;AACtB,YAAY,UAAU;AACtB,YAAY,QAAQ;AACpB,SAAS,qBAAqB;;;AC4DvB,SAAS,oBAAoB,YAAwC;AAC1E,QAAM,iBAAiC,CAAC;AAExC,MAAI,WAAW,WAAW;AACxB,mBAAe,YAAY,WAAW;AAAA,EACxC;AACA,MAAI,WAAW,UAAU;AACvB,mBAAe,WAAW,WAAW;AAAA,EACvC;AACA,MAAI,WAAW,QAAQ;AACrB,mBAAe,SAAS,WAAW;AAAA,EACrC;AACA,MAAI,WAAW,OAAO;AACpB,mBAAe,QAAQ;AAAA,EACzB;AACA,MAAI,WAAW,kBAAkB;AAC/B,mBAAe,mBAAmB,WAAW;AAAA,EAC/C;AACA,MAAI,WAAW,gBAAgB;AAC7B,mBAAe,iBAAiB;AAAA,EAClC;AACA,MAAI,WAAW,qBAAqB;AAClC,mBAAe,sBAAsB,WAAW;AAAA,EAClD;AACA,MAAI,WAAW,eAAe;AAC5B,mBAAe,gBAAgB;AAAA,EACjC;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc,WAAW;AAAA,EAC1C;AACA,MAAI,WAAW,cAAc;AAC3B,mBAAe,eAAe,WAAW;AAAA,EAC3C;AACA,MAAI,WAAW,uBAAuB;AACpC,mBAAe,wBAAwB,WAAW;AAAA,EACpD;AACA,MAAI,WAAW,mBAAmB;AAChC,mBAAe,oBAAoB,WAAW;AAAA,EAChD;AACA,MAAI,WAAW,mBAAmB;AAChC,mBAAe,oBAAoB,WAAW;AAAA,EAChD;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc,WAAW;AAAA,EAC1C;AACA,MAAI,WAAW,aAAa;AAC1B,mBAAe,cAAc,WAAW;AAAA,EAC1C;AAEA,SAAO;AACT;AAKO,SAAS,UAAU,SAAmC;AAC3D,QAAM,OAAiB,CAAC;AAExB,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,UAAU;AACpB,SAAK,KAAK,cAAc,QAAQ,QAAQ;AAAA,EAC1C;AACA,MAAI,QAAQ,QAAQ;AAClB,QAAI,MAAM,QAAQ,QAAQ,MAAM,GAAG;AACjC,UAAI,QAAQ,OAAO,SAAS,GAAG;AAC7B,aAAK,KAAK,YAAY,QAAQ,OAAO,KAAK,GAAG,CAAC;AAAA,MAChD;AAAA,IACF,OAAO;AACL,WAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,IACtC;AAAA,EACF;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,SAAS;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,MAAM,QAAQ,QAAQ,gBAAgB,GAAG;AAC3C,UAAI,QAAQ,iBAAiB,SAAS,GAAG;AACvC,aAAK,KAAK,wBAAwB,QAAQ,iBAAiB,KAAK,GAAG,CAAC;AAAA,MACtE;AAAA,IACF,OAAO;AACL,WAAK,KAAK,wBAAwB,QAAQ,gBAAgB;AAAA,IAC5D;AAAA,EACF;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,oBAAoB;AAAA,EAChC;AACA,MAAI,QAAQ,qBAAqB;AAC/B,SAAK,KAAK,2BAA2B,QAAQ,mBAAmB;AAAA,EAClE;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,cAAc;AACxB,SAAK,KAAK,mBAAmB,QAAQ,YAAY;AAAA,EACnD;AACA,MAAI,QAAQ,uBAAuB;AACjC,SAAK,KAAK,6BAA6B,QAAQ,qBAAqB;AAAA,EACtE;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AAEA,SAAO;AACT;;;ADzKA,IAAM,aAAa,cAAc,YAAY,GAAG;AAChD,IAAM,YAAiB,aAAQ,UAAU;AAEzC,IAAM,WAAW;AAMjB,SAAS,WAAW,MAAgB,mBAAwC,CAAC,GAAoB;AAC/F,QAAM,EAAE,eAAe,MAAM,IAAI;AAEjC,SAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,UAAM,UAAe,UAAK,WAAW,MAAM,OAAO,QAAQ;AAE1D,QAAI,CAAI,cAAW,OAAO,GAAG;AAC3B,aAAO;AAAA,QACL,IAAI,MAAM,yBAAyB,OAAO,sCAAsC;AAAA,MAClF;AAAA,IACF;AAEA,UAAM,UAAU;AAChB,UAAM,cAAc,CAAC,QAAQ,SAAS,GAAG,IAAI;AAE7C,UAAM,cAAc,MAAM,SAAS,WAAW;AAE9C,QAAI,SAAS;AACb,QAAI,SAAS;AAEb,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,SAAS;AAChC,UAAI,SAAS,GAAG;AACd,gBAAQ,MAAM;AAAA,MAChB,OAAO;AACL,cAAM,cAAc,UAAU;AAC9B,cAAM,QAAQ,IAAI;AAAA,UAChB,+CAA+C,IAAI;AAAA;AAAA,EAAQ,WAAW;AAAA,QACxE;AACA,eAAO,KAAK;AAAA,MACd;AAAA,IACF,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,QAAe;AACtC,UAAI,IAAI,QAAQ,SAAS,QAAQ,GAAG;AAClC;AAAA,UACE,IAAI;AAAA,YACF;AAAA,UACF;AAAA,QACF;AAAA,MACF,OAAO;AACL,eAAO,GAAG;AAAA,MACZ;AAAA,IACF,CAAC;AAAA,EACH,CAAC;AACH;AAEO,SAAS,QACd,YACA,UAA0B,CAAC,GACV;AACjB,QAAM,YAAY,MAAM,QAAQ,UAAU,IAAI,aAAa,CAAC,UAAU;AACtE,MAAI,UAAU,WAAW,GAAG;AAC1B,WAAO,QAAQ,OAAO,IAAI,MAAM,2CAA2C,CAAC;AAAA,EAC9E;AAEA,aAAW,SAAS,WAAW;AAC7B,QAAI,CAAI,cAAW,KAAK,GAAG;AACzB,aAAO,QAAQ,OAAO,IAAI,MAAM,mCAAmC,KAAK,EAAE,CAAC;AAAA,IAC7E;AAAA,EACF;AAEA,QAAM,OAAiB,CAAC,GAAG,WAAW,GAAG,UAAU,OAAO,CAAC;AAE3D,SAAO,WAAW,MAAM;AAAA,IACtB,cAAc,CAAC,QAAQ;AAAA,EACzB,CAAC;AACH;;;AE9FO,SAAS,mBAAmB,SAAwB;AACzD,UAAQ,OAAO,4BAA4B,yEAAyE;AACpH,UAAQ,OAAO,0BAA0B,kCAAkC;AAC3E,UAAQ,OAAO,wBAAwB,oIAAoI;AAC3K,UAAQ,OAAO,eAAe,iCAAiC;AAC/D,UAAQ,OAAO,gCAAgC,sFAAsF;AACrI,UAAQ,OAAO,sBAAsB,iDAAiD;AACtF,UAAQ,OAAO,mCAAmC,2EAA2E;AAC7H,UAAQ,OAAO,qBAAqB,8EAA8E;AAClH,UAAQ,OAAO,0BAA0B,sGAAsG;AAC/I,UAAQ,OAAO,2BAA2B,6DAA6D;AACvG,UAAQ,OAAO,qCAAqC,+FAA+F;AACnJ,UAAQ,OAAO,iCAAiC,2FAA2F;AAC3I,UAAQ,OAAO,iCAAiC,2FAA2F;AAC3I,UAAQ,OAAO,0BAA0B,wHAAwH;AACjK,UAAQ,OAAO,0BAA0B,qEAAqE;AAChH;;;AHlBA,SAAS,gBAAyB;AAChC,QAAM,UAAU,IAAI,QAAQ;AAE5B,UACG,KAAK,oBAAoB,EACzB,MAAM,sBAAsB,EAC5B,YAAY,4CAA4C,EACxD,mBAAmB,wCAAwC,EAC3D,yBAAyB,KAAK,EAC9B,SAAS,cAAc,uCAAuC;AAGjE,qBAAmB,OAAO;AAE1B,UAAQ,gBAAgB;AAAA,IACtB,UAAU,CAAC,QAAQ;AACjB,cAAQ,MAAM,IAAI,QAAQ,CAAC;AAAA,IAC7B;AAAA,IACA,aAAa,CAAC,KAAK,UAAU;AAC3B,YAAM,GAAG;AAAA,IACX;AAAA,EACF,CAAC;AAED,SAAO;AACT;AAEA,eAAe,OAAwB;AACrC,QAAM,UAAU,cAAc;AAE9B,UAAQ,aAAa;AAErB,MAAI;AACF,YAAQ,MAAM,QAAQ,IAAI;AAAA,EAC5B,SAAS,KAAK;AACZ,QAAI,eAAe,gBAAgB;AACjC,UAAI,IAAI,SAAS,2BAA2B;AAC1C,eAAO;AAAA,MACT;AACA,aAAO,IAAI,YAAY;AAAA,IACzB;AAEA,UAAM,UAAU,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC/D,YAAQ,MAAM,OAAO;AACrB,YAAQ,MAAM,wCAAwC;AACtD,WAAO;AAAA,EACT;AAEA,QAAM,aAAa,QAAQ,KAAiB;AAC5C,QAAM,aAAa,QAAQ;AAC3B,QAAM,iBAAiB,oBAAoB,UAAU;AAErD,MAAI;AACF,UAAM,SAAS,MAAM,QAAQ,YAAY,cAAc;AACvD,QAAI,UAAU,CAAC,eAAe,OAAO;AACnC,cAAQ,OAAO,MAAM,MAAM;AAC3B,UAAI,CAAC,OAAO,SAAS,IAAI,GAAG;AAC1B,gBAAQ,OAAO,MAAM,IAAI;AAAA,MAC3B;AAAA,IACF;AACA,WAAO;AAAA,EACT,SAAS,KAAK;AACZ,UAAM,UAAU,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC/D,YAAQ,MAAM,OAAO;AACrB,WAAO;AAAA,EACT;AACF;AAEA,KAAK,EAAE,KAAK,CAAC,SAAS;AACpB,MAAI,SAAS,GAAG;AACd,YAAQ,KAAK,IAAI;AAAA,EACnB;AACF,CAAC;","names":[]}

package/dist/index.cjs

@@ -94,8 +94,8 @@ function buildArgs(options) {
   if (options.htmlPageSeparator) {
     args.push("--html-page-separator", options.htmlPageSeparator);
   }
-  if (options.embedImages) {
-    args.push("--embed-images");
+  if (options.imageOutput) {
+    args.push("--image-output", options.imageOutput);
   }
   if (options.imageFormat) {
     args.push("--image-format", options.imageFormat);

package/dist/index.cjs.map

@@ -1 +1 @@
-{"version":3,"sources":["../src/index.ts","../src/convert-options.generated.ts"],"sourcesContent":["import { spawn } from 'child_process';\nimport * as path from 'path';\nimport * as fs from 'fs';\nimport { fileURLToPath } from 'url';\n\n// Re-export types and utilities from auto-generated file\nexport type { ConvertOptions } from './convert-options.generated.js';\nexport { buildArgs } from './convert-options.generated.js';\nimport type { ConvertOptions } from './convert-options.generated.js';\nimport { buildArgs } from './convert-options.generated.js';\n\nconst __filename = fileURLToPath(import.meta.url);\nconst __dirname = path.dirname(__filename);\n\nconst JAR_NAME = 'opendataloader-pdf-cli.jar';\n\ninterface JarExecutionOptions {\n  streamOutput?: boolean;\n}\n\nfunction executeJar(args: string[], executionOptions: JarExecutionOptions = {}): Promise<string> {\n  const { streamOutput = false } = executionOptions;\n\n  return new Promise((resolve, reject) => {\n    const jarPath = path.join(__dirname, '..', 'lib', JAR_NAME);\n\n    if (!fs.existsSync(jarPath)) {\n      return reject(\n        new Error(`JAR file not found at ${jarPath}. Please run the build script first.`),\n      );\n    }\n\n    const command = 'java';\n    const commandArgs = ['-jar', jarPath, ...args];\n\n    const javaProcess = spawn(command, commandArgs);\n\n    let stdout = '';\n    let stderr = '';\n\n    javaProcess.stdout.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stdout.write(chunk);\n      }\n      stdout += chunk;\n    });\n\n    javaProcess.stderr.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stderr.write(chunk);\n      }\n      stderr += chunk;\n    });\n\n    javaProcess.on('close', (code) => {\n      if (code === 0) {\n        resolve(stdout);\n      } else {\n        const errorOutput = stderr || stdout;\n        const error = new Error(\n          `The opendataloader-pdf CLI exited with code ${code}.\\n\\n${errorOutput}`,\n        );\n        reject(error);\n      }\n    });\n\n    javaProcess.on('error', (err: Error) => {\n      if (err.message.includes('ENOENT')) {\n        reject(\n          new Error(\n            \"'java' command not found. Please ensure Java is installed and in your system's PATH.\",\n          ),\n        );\n      } else {\n        reject(err);\n      }\n    });\n  });\n}\n\nexport function convert(\n  inputPaths: string | string[],\n  options: ConvertOptions = {},\n): Promise<string> {\n  const inputList = Array.isArray(inputPaths) ? inputPaths : [inputPaths];\n  if (inputList.length === 0) {\n    return Promise.reject(new Error('At least one input path must be provided.'));\n  }\n\n  for (const input of inputList) {\n    if (!fs.existsSync(input)) {\n      return Promise.reject(new Error(`Input file or folder not found: ${input}`));\n    }\n  }\n\n  const args: string[] = [...inputList, ...buildArgs(options)];\n\n  return executeJar(args, {\n    streamOutput: !options.quiet,\n  });\n}\n\n/**\n * @deprecated Use `convert()` and `ConvertOptions` instead. This function will be removed in a future version.\n */\nexport interface RunOptions {\n  outputFolder?: string;\n  password?: string;\n  replaceInvalidChars?: string;\n  generateMarkdown?: boolean;\n  generateHtml?: boolean;\n  generateAnnotatedPdf?: boolean;\n  keepLineBreaks?: boolean;\n  contentSafetyOff?: string;\n  htmlInMarkdown?: boolean;\n  addImageToMarkdown?: boolean;\n  noJson?: boolean;\n  debug?: boolean;\n  useStructTree?: boolean;\n}\n\n/**\n * @deprecated Use `convert()` instead. This function will be removed in a future version.\n */\nexport function run(inputPath: string, options: RunOptions = {}): Promise<string> {\n  console.warn(\n    'Warning: run() is deprecated and will be removed in a future version. Use convert() instead.',\n  );\n\n  // Build format array based on legacy boolean options\n  const formats: string[] = [];\n  if (!options.noJson) {\n    formats.push('json');\n  }\n  if (options.generateMarkdown) {\n    if (options.addImageToMarkdown) {\n      formats.push('markdown-with-images');\n    } else if (options.htmlInMarkdown) {\n      formats.push('markdown-with-html');\n    } else {\n      formats.push('markdown');\n    }\n  }\n  if (options.generateHtml) {\n    formats.push('html');\n  }\n  if (options.generateAnnotatedPdf) {\n    formats.push('pdf');\n  }\n\n  return convert(inputPath, {\n    outputDir: options.outputFolder,\n    password: options.password,\n    replaceInvalidChars: options.replaceInvalidChars,\n    keepLineBreaks: options.keepLineBreaks,\n    contentSafetyOff: options.contentSafetyOff,\n    useStructTree: options.useStructTree,\n    format: formats.length > 0 ? formats : undefined,\n    quiet: !options.debug,\n  });\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\n/**\n * Options for the convert function.\n */\nexport interface ConvertOptions {\n  /** Directory where output files are written. Default: input file directory */\n  outputDir?: string;\n  /** Password for encrypted PDF files */\n  password?: string;\n  /** Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json */\n  format?: string | string[];\n  /** Suppress console logging output */\n  quiet?: boolean;\n  /** Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg */\n  contentSafetyOff?: string | string[];\n  /** Preserve original line breaks in extracted text */\n  keepLineBreaks?: boolean;\n  /** Replacement character for invalid/unrecognized characters. Default: space */\n  replaceInvalidChars?: string;\n  /** Use PDF structure tree (tagged PDF) for reading order and semantic structure */\n  useStructTree?: boolean;\n  /** Table detection method. Values: cluster */\n  tableMethod?: string;\n  /** Reading order algorithm. Values: none, xycut. Default: none */\n  readingOrder?: string;\n  /** Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none */\n  markdownPageSeparator?: string;\n  /** Separator between pages in text output. Use %page-number% for page numbers. Default: none */\n  textPageSeparator?: string;\n  /** Separator between pages in HTML output. Use %page-number% for page numbers. Default: none */\n  htmlPageSeparator?: string;\n  /** Embed images as Base64 data URIs instead of file path references */\n  embedImages?: boolean;\n  /** Output format for extracted images. Values: png, jpeg. Default: png */\n  imageFormat?: string;\n}\n\n/**\n * Options as parsed from CLI (all values are strings from commander).\n */\nexport interface CliOptions {\n  outputDir?: string;\n  password?: string;\n  format?: string;\n  quiet?: boolean;\n  contentSafetyOff?: string;\n  keepLineBreaks?: boolean;\n  replaceInvalidChars?: string;\n  useStructTree?: boolean;\n  tableMethod?: string;\n  readingOrder?: string;\n  markdownPageSeparator?: string;\n  textPageSeparator?: string;\n  htmlPageSeparator?: string;\n  embedImages?: boolean;\n  imageFormat?: string;\n}\n\n/**\n * Convert CLI options to ConvertOptions.\n */\nexport function buildConvertOptions(cliOptions: CliOptions): ConvertOptions {\n  const convertOptions: ConvertOptions = {};\n\n  if (cliOptions.outputDir) {\n    convertOptions.outputDir = cliOptions.outputDir;\n  }\n  if (cliOptions.password) {\n    convertOptions.password = cliOptions.password;\n  }\n  if (cliOptions.format) {\n    convertOptions.format = cliOptions.format;\n  }\n  if (cliOptions.quiet) {\n    convertOptions.quiet = true;\n  }\n  if (cliOptions.contentSafetyOff) {\n    convertOptions.contentSafetyOff = cliOptions.contentSafetyOff;\n  }\n  if (cliOptions.keepLineBreaks) {\n    convertOptions.keepLineBreaks = true;\n  }\n  if (cliOptions.replaceInvalidChars) {\n    convertOptions.replaceInvalidChars = cliOptions.replaceInvalidChars;\n  }\n  if (cliOptions.useStructTree) {\n    convertOptions.useStructTree = true;\n  }\n  if (cliOptions.tableMethod) {\n    convertOptions.tableMethod = cliOptions.tableMethod;\n  }\n  if (cliOptions.readingOrder) {\n    convertOptions.readingOrder = cliOptions.readingOrder;\n  }\n  if (cliOptions.markdownPageSeparator) {\n    convertOptions.markdownPageSeparator = cliOptions.markdownPageSeparator;\n  }\n  if (cliOptions.textPageSeparator) {\n    convertOptions.textPageSeparator = cliOptions.textPageSeparator;\n  }\n  if (cliOptions.htmlPageSeparator) {\n    convertOptions.htmlPageSeparator = cliOptions.htmlPageSeparator;\n  }\n  if (cliOptions.embedImages) {\n    convertOptions.embedImages = true;\n  }\n  if (cliOptions.imageFormat) {\n    convertOptions.imageFormat = cliOptions.imageFormat;\n  }\n\n  return convertOptions;\n}\n\n/**\n * Build CLI arguments array from ConvertOptions.\n */\nexport function buildArgs(options: ConvertOptions): string[] {\n  const args: string[] = [];\n\n  if (options.outputDir) {\n    args.push('--output-dir', options.outputDir);\n  }\n  if (options.password) {\n    args.push('--password', options.password);\n  }\n  if (options.format) {\n    if (Array.isArray(options.format)) {\n      if (options.format.length > 0) {\n        args.push('--format', options.format.join(','));\n      }\n    } else {\n      args.push('--format', options.format);\n    }\n  }\n  if (options.quiet) {\n    args.push('--quiet');\n  }\n  if (options.contentSafetyOff) {\n    if (Array.isArray(options.contentSafetyOff)) {\n      if (options.contentSafetyOff.length > 0) {\n        args.push('--content-safety-off', options.contentSafetyOff.join(','));\n      }\n    } else {\n      args.push('--content-safety-off', options.contentSafetyOff);\n    }\n  }\n  if (options.keepLineBreaks) {\n    args.push('--keep-line-breaks');\n  }\n  if (options.replaceInvalidChars) {\n    args.push('--replace-invalid-chars', options.replaceInvalidChars);\n  }\n  if (options.useStructTree) {\n    args.push('--use-struct-tree');\n  }\n  if (options.tableMethod) {\n    args.push('--table-method', options.tableMethod);\n  }\n  if (options.readingOrder) {\n    args.push('--reading-order', options.readingOrder);\n  }\n  if (options.markdownPageSeparator) {\n    args.push('--markdown-page-separator', options.markdownPageSeparator);\n  }\n  if (options.textPageSeparator) {\n    args.push('--text-page-separator', options.textPageSeparator);\n  }\n  if (options.htmlPageSeparator) {\n    args.push('--html-page-separator', options.htmlPageSeparator);\n  }\n  if (options.embedImages) {\n    args.push('--embed-images');\n  }\n  if (options.imageFormat) {\n    args.push('--image-format', options.imageFormat);\n  }\n\n  return args;\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,2BAAsB;AACtB,WAAsB;AACtB,SAAoB;AACpB,iBAA8B;;;ACmHvB,SAAS,UAAU,SAAmC;AAC3D,QAAM,OAAiB,CAAC;AAExB,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,UAAU;AACpB,SAAK,KAAK,cAAc,QAAQ,QAAQ;AAAA,EAC1C;AACA,MAAI,QAAQ,QAAQ;AAClB,QAAI,MAAM,QAAQ,QAAQ,MAAM,GAAG;AACjC,UAAI,QAAQ,OAAO,SAAS,GAAG;AAC7B,aAAK,KAAK,YAAY,QAAQ,OAAO,KAAK,GAAG,CAAC;AAAA,MAChD;AAAA,IACF,OAAO;AACL,WAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,IACtC;AAAA,EACF;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,SAAS;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,MAAM,QAAQ,QAAQ,gBAAgB,GAAG;AAC3C,UAAI,QAAQ,iBAAiB,SAAS,GAAG;AACvC,aAAK,KAAK,wBAAwB,QAAQ,iBAAiB,KAAK,GAAG,CAAC;AAAA,MACtE;AAAA,IACF,OAAO;AACL,WAAK,KAAK,wBAAwB,QAAQ,gBAAgB;AAAA,IAC5D;AAAA,EACF;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,oBAAoB;AAAA,EAChC;AACA,MAAI,QAAQ,qBAAqB;AAC/B,SAAK,KAAK,2BAA2B,QAAQ,mBAAmB;AAAA,EAClE;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,cAAc;AACxB,SAAK,KAAK,mBAAmB,QAAQ,YAAY;AAAA,EACnD;AACA,MAAI,QAAQ,uBAAuB;AACjC,SAAK,KAAK,6BAA6B,QAAQ,qBAAqB;AAAA,EACtE;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,gBAAgB;AAAA,EAC5B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AAEA,SAAO;AACT;;;ADpLA;AAWA,IAAM,iBAAa,0BAAc,YAAY,GAAG;AAChD,IAAM,YAAiB,aAAQ,UAAU;AAEzC,IAAM,WAAW;AAMjB,SAAS,WAAW,MAAgB,mBAAwC,CAAC,GAAoB;AAC/F,QAAM,EAAE,eAAe,MAAM,IAAI;AAEjC,SAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,UAAM,UAAe,UAAK,WAAW,MAAM,OAAO,QAAQ;AAE1D,QAAI,CAAI,cAAW,OAAO,GAAG;AAC3B,aAAO;AAAA,QACL,IAAI,MAAM,yBAAyB,OAAO,sCAAsC;AAAA,MAClF;AAAA,IACF;AAEA,UAAM,UAAU;AAChB,UAAM,cAAc,CAAC,QAAQ,SAAS,GAAG,IAAI;AAE7C,UAAM,kBAAc,4BAAM,SAAS,WAAW;AAE9C,QAAI,SAAS;AACb,QAAI,SAAS;AAEb,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,SAAS;AAChC,UAAI,SAAS,GAAG;AACd,gBAAQ,MAAM;AAAA,MAChB,OAAO;AACL,cAAM,cAAc,UAAU;AAC9B,cAAM,QAAQ,IAAI;AAAA,UAChB,+CAA+C,IAAI;AAAA;AAAA,EAAQ,WAAW;AAAA,QACxE;AACA,eAAO,KAAK;AAAA,MACd;AAAA,IACF,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,QAAe;AACtC,UAAI,IAAI,QAAQ,SAAS,QAAQ,GAAG;AAClC;AAAA,UACE,IAAI;AAAA,YACF;AAAA,UACF;AAAA,QACF;AAAA,MACF,OAAO;AACL,eAAO,GAAG;AAAA,MACZ;AAAA,IACF,CAAC;AAAA,EACH,CAAC;AACH;AAEO,SAAS,QACd,YACA,UAA0B,CAAC,GACV;AACjB,QAAM,YAAY,MAAM,QAAQ,UAAU,IAAI,aAAa,CAAC,UAAU;AACtE,MAAI,UAAU,WAAW,GAAG;AAC1B,WAAO,QAAQ,OAAO,IAAI,MAAM,2CAA2C,CAAC;AAAA,EAC9E;AAEA,aAAW,SAAS,WAAW;AAC7B,QAAI,CAAI,cAAW,KAAK,GAAG;AACzB,aAAO,QAAQ,OAAO,IAAI,MAAM,mCAAmC,KAAK,EAAE,CAAC;AAAA,IAC7E;AAAA,EACF;AAEA,QAAM,OAAiB,CAAC,GAAG,WAAW,GAAG,UAAU,OAAO,CAAC;AAE3D,SAAO,WAAW,MAAM;AAAA,IACtB,cAAc,CAAC,QAAQ;AAAA,EACzB,CAAC;AACH;AAwBO,SAAS,IAAI,WAAmB,UAAsB,CAAC,GAAoB;AAChF,UAAQ;AAAA,IACN;AAAA,EACF;AAGA,QAAM,UAAoB,CAAC;AAC3B,MAAI,CAAC,QAAQ,QAAQ;AACnB,YAAQ,KAAK,MAAM;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,QAAQ,oBAAoB;AAC9B,cAAQ,KAAK,sBAAsB;AAAA,IACrC,WAAW,QAAQ,gBAAgB;AACjC,cAAQ,KAAK,oBAAoB;AAAA,IACnC,OAAO;AACL,cAAQ,KAAK,UAAU;AAAA,IACzB;AAAA,EACF;AACA,MAAI,QAAQ,cAAc;AACxB,YAAQ,KAAK,MAAM;AAAA,EACrB;AACA,MAAI,QAAQ,sBAAsB;AAChC,YAAQ,KAAK,KAAK;AAAA,EACpB;AAEA,SAAO,QAAQ,WAAW;AAAA,IACxB,WAAW,QAAQ;AAAA,IACnB,UAAU,QAAQ;AAAA,IAClB,qBAAqB,QAAQ;AAAA,IAC7B,gBAAgB,QAAQ;AAAA,IACxB,kBAAkB,QAAQ;AAAA,IAC1B,eAAe,QAAQ;AAAA,IACvB,QAAQ,QAAQ,SAAS,IAAI,UAAU;AAAA,IACvC,OAAO,CAAC,QAAQ;AAAA,EAClB,CAAC;AACH;","names":[]}
+{"version":3,"sources":["../src/index.ts","../src/convert-options.generated.ts"],"sourcesContent":["import { spawn } from 'child_process';\nimport * as path from 'path';\nimport * as fs from 'fs';\nimport { fileURLToPath } from 'url';\n\n// Re-export types and utilities from auto-generated file\nexport type { ConvertOptions } from './convert-options.generated.js';\nexport { buildArgs } from './convert-options.generated.js';\nimport type { ConvertOptions } from './convert-options.generated.js';\nimport { buildArgs } from './convert-options.generated.js';\n\nconst __filename = fileURLToPath(import.meta.url);\nconst __dirname = path.dirname(__filename);\n\nconst JAR_NAME = 'opendataloader-pdf-cli.jar';\n\ninterface JarExecutionOptions {\n  streamOutput?: boolean;\n}\n\nfunction executeJar(args: string[], executionOptions: JarExecutionOptions = {}): Promise<string> {\n  const { streamOutput = false } = executionOptions;\n\n  return new Promise((resolve, reject) => {\n    const jarPath = path.join(__dirname, '..', 'lib', JAR_NAME);\n\n    if (!fs.existsSync(jarPath)) {\n      return reject(\n        new Error(`JAR file not found at ${jarPath}. Please run the build script first.`),\n      );\n    }\n\n    const command = 'java';\n    const commandArgs = ['-jar', jarPath, ...args];\n\n    const javaProcess = spawn(command, commandArgs);\n\n    let stdout = '';\n    let stderr = '';\n\n    javaProcess.stdout.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stdout.write(chunk);\n      }\n      stdout += chunk;\n    });\n\n    javaProcess.stderr.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stderr.write(chunk);\n      }\n      stderr += chunk;\n    });\n\n    javaProcess.on('close', (code) => {\n      if (code === 0) {\n        resolve(stdout);\n      } else {\n        const errorOutput = stderr || stdout;\n        const error = new Error(\n          `The opendataloader-pdf CLI exited with code ${code}.\\n\\n${errorOutput}`,\n        );\n        reject(error);\n      }\n    });\n\n    javaProcess.on('error', (err: Error) => {\n      if (err.message.includes('ENOENT')) {\n        reject(\n          new Error(\n            \"'java' command not found. Please ensure Java is installed and in your system's PATH.\",\n          ),\n        );\n      } else {\n        reject(err);\n      }\n    });\n  });\n}\n\nexport function convert(\n  inputPaths: string | string[],\n  options: ConvertOptions = {},\n): Promise<string> {\n  const inputList = Array.isArray(inputPaths) ? inputPaths : [inputPaths];\n  if (inputList.length === 0) {\n    return Promise.reject(new Error('At least one input path must be provided.'));\n  }\n\n  for (const input of inputList) {\n    if (!fs.existsSync(input)) {\n      return Promise.reject(new Error(`Input file or folder not found: ${input}`));\n    }\n  }\n\n  const args: string[] = [...inputList, ...buildArgs(options)];\n\n  return executeJar(args, {\n    streamOutput: !options.quiet,\n  });\n}\n\n/**\n * @deprecated Use `convert()` and `ConvertOptions` instead. This function will be removed in a future version.\n */\nexport interface RunOptions {\n  outputFolder?: string;\n  password?: string;\n  replaceInvalidChars?: string;\n  generateMarkdown?: boolean;\n  generateHtml?: boolean;\n  generateAnnotatedPdf?: boolean;\n  keepLineBreaks?: boolean;\n  contentSafetyOff?: string;\n  htmlInMarkdown?: boolean;\n  addImageToMarkdown?: boolean;\n  noJson?: boolean;\n  debug?: boolean;\n  useStructTree?: boolean;\n}\n\n/**\n * @deprecated Use `convert()` instead. This function will be removed in a future version.\n */\nexport function run(inputPath: string, options: RunOptions = {}): Promise<string> {\n  console.warn(\n    'Warning: run() is deprecated and will be removed in a future version. Use convert() instead.',\n  );\n\n  // Build format array based on legacy boolean options\n  const formats: string[] = [];\n  if (!options.noJson) {\n    formats.push('json');\n  }\n  if (options.generateMarkdown) {\n    if (options.addImageToMarkdown) {\n      formats.push('markdown-with-images');\n    } else if (options.htmlInMarkdown) {\n      formats.push('markdown-with-html');\n    } else {\n      formats.push('markdown');\n    }\n  }\n  if (options.generateHtml) {\n    formats.push('html');\n  }\n  if (options.generateAnnotatedPdf) {\n    formats.push('pdf');\n  }\n\n  return convert(inputPath, {\n    outputDir: options.outputFolder,\n    password: options.password,\n    replaceInvalidChars: options.replaceInvalidChars,\n    keepLineBreaks: options.keepLineBreaks,\n    contentSafetyOff: options.contentSafetyOff,\n    useStructTree: options.useStructTree,\n    format: formats.length > 0 ? formats : undefined,\n    quiet: !options.debug,\n  });\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\n/**\n * Options for the convert function.\n */\nexport interface ConvertOptions {\n  /** Directory where output files are written. Default: input file directory */\n  outputDir?: string;\n  /** Password for encrypted PDF files */\n  password?: string;\n  /** Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json */\n  format?: string | string[];\n  /** Suppress console logging output */\n  quiet?: boolean;\n  /** Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg */\n  contentSafetyOff?: string | string[];\n  /** Preserve original line breaks in extracted text */\n  keepLineBreaks?: boolean;\n  /** Replacement character for invalid/unrecognized characters. Default: space */\n  replaceInvalidChars?: string;\n  /** Use PDF structure tree (tagged PDF) for reading order and semantic structure */\n  useStructTree?: boolean;\n  /** Table detection method. Values: default (border-based), cluster (border + cluster). Default: default */\n  tableMethod?: string;\n  /** Reading order algorithm. Values: off, xycut. Default: xycut */\n  readingOrder?: string;\n  /** Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none */\n  markdownPageSeparator?: string;\n  /** Separator between pages in text output. Use %page-number% for page numbers. Default: none */\n  textPageSeparator?: string;\n  /** Separator between pages in HTML output. Use %page-number% for page numbers. Default: none */\n  htmlPageSeparator?: string;\n  /** Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external */\n  imageOutput?: string;\n  /** Output format for extracted images. Values: png, jpeg. Default: png */\n  imageFormat?: string;\n}\n\n/**\n * Options as parsed from CLI (all values are strings from commander).\n */\nexport interface CliOptions {\n  outputDir?: string;\n  password?: string;\n  format?: string;\n  quiet?: boolean;\n  contentSafetyOff?: string;\n  keepLineBreaks?: boolean;\n  replaceInvalidChars?: string;\n  useStructTree?: boolean;\n  tableMethod?: string;\n  readingOrder?: string;\n  markdownPageSeparator?: string;\n  textPageSeparator?: string;\n  htmlPageSeparator?: string;\n  imageOutput?: string;\n  imageFormat?: string;\n}\n\n/**\n * Convert CLI options to ConvertOptions.\n */\nexport function buildConvertOptions(cliOptions: CliOptions): ConvertOptions {\n  const convertOptions: ConvertOptions = {};\n\n  if (cliOptions.outputDir) {\n    convertOptions.outputDir = cliOptions.outputDir;\n  }\n  if (cliOptions.password) {\n    convertOptions.password = cliOptions.password;\n  }\n  if (cliOptions.format) {\n    convertOptions.format = cliOptions.format;\n  }\n  if (cliOptions.quiet) {\n    convertOptions.quiet = true;\n  }\n  if (cliOptions.contentSafetyOff) {\n    convertOptions.contentSafetyOff = cliOptions.contentSafetyOff;\n  }\n  if (cliOptions.keepLineBreaks) {\n    convertOptions.keepLineBreaks = true;\n  }\n  if (cliOptions.replaceInvalidChars) {\n    convertOptions.replaceInvalidChars = cliOptions.replaceInvalidChars;\n  }\n  if (cliOptions.useStructTree) {\n    convertOptions.useStructTree = true;\n  }\n  if (cliOptions.tableMethod) {\n    convertOptions.tableMethod = cliOptions.tableMethod;\n  }\n  if (cliOptions.readingOrder) {\n    convertOptions.readingOrder = cliOptions.readingOrder;\n  }\n  if (cliOptions.markdownPageSeparator) {\n    convertOptions.markdownPageSeparator = cliOptions.markdownPageSeparator;\n  }\n  if (cliOptions.textPageSeparator) {\n    convertOptions.textPageSeparator = cliOptions.textPageSeparator;\n  }\n  if (cliOptions.htmlPageSeparator) {\n    convertOptions.htmlPageSeparator = cliOptions.htmlPageSeparator;\n  }\n  if (cliOptions.imageOutput) {\n    convertOptions.imageOutput = cliOptions.imageOutput;\n  }\n  if (cliOptions.imageFormat) {\n    convertOptions.imageFormat = cliOptions.imageFormat;\n  }\n\n  return convertOptions;\n}\n\n/**\n * Build CLI arguments array from ConvertOptions.\n */\nexport function buildArgs(options: ConvertOptions): string[] {\n  const args: string[] = [];\n\n  if (options.outputDir) {\n    args.push('--output-dir', options.outputDir);\n  }\n  if (options.password) {\n    args.push('--password', options.password);\n  }\n  if (options.format) {\n    if (Array.isArray(options.format)) {\n      if (options.format.length > 0) {\n        args.push('--format', options.format.join(','));\n      }\n    } else {\n      args.push('--format', options.format);\n    }\n  }\n  if (options.quiet) {\n    args.push('--quiet');\n  }\n  if (options.contentSafetyOff) {\n    if (Array.isArray(options.contentSafetyOff)) {\n      if (options.contentSafetyOff.length > 0) {\n        args.push('--content-safety-off', options.contentSafetyOff.join(','));\n      }\n    } else {\n      args.push('--content-safety-off', options.contentSafetyOff);\n    }\n  }\n  if (options.keepLineBreaks) {\n    args.push('--keep-line-breaks');\n  }\n  if (options.replaceInvalidChars) {\n    args.push('--replace-invalid-chars', options.replaceInvalidChars);\n  }\n  if (options.useStructTree) {\n    args.push('--use-struct-tree');\n  }\n  if (options.tableMethod) {\n    args.push('--table-method', options.tableMethod);\n  }\n  if (options.readingOrder) {\n    args.push('--reading-order', options.readingOrder);\n  }\n  if (options.markdownPageSeparator) {\n    args.push('--markdown-page-separator', options.markdownPageSeparator);\n  }\n  if (options.textPageSeparator) {\n    args.push('--text-page-separator', options.textPageSeparator);\n  }\n  if (options.htmlPageSeparator) {\n    args.push('--html-page-separator', options.htmlPageSeparator);\n  }\n  if (options.imageOutput) {\n    args.push('--image-output', options.imageOutput);\n  }\n  if (options.imageFormat) {\n    args.push('--image-format', options.imageFormat);\n  }\n\n  return args;\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,2BAAsB;AACtB,WAAsB;AACtB,SAAoB;AACpB,iBAA8B;;;ACmHvB,SAAS,UAAU,SAAmC;AAC3D,QAAM,OAAiB,CAAC;AAExB,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,UAAU;AACpB,SAAK,KAAK,cAAc,QAAQ,QAAQ;AAAA,EAC1C;AACA,MAAI,QAAQ,QAAQ;AAClB,QAAI,MAAM,QAAQ,QAAQ,MAAM,GAAG;AACjC,UAAI,QAAQ,OAAO,SAAS,GAAG;AAC7B,aAAK,KAAK,YAAY,QAAQ,OAAO,KAAK,GAAG,CAAC;AAAA,MAChD;AAAA,IACF,OAAO;AACL,WAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,IACtC;AAAA,EACF;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,SAAS;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,MAAM,QAAQ,QAAQ,gBAAgB,GAAG;AAC3C,UAAI,QAAQ,iBAAiB,SAAS,GAAG;AACvC,aAAK,KAAK,wBAAwB,QAAQ,iBAAiB,KAAK,GAAG,CAAC;AAAA,MACtE;AAAA,IACF,OAAO;AACL,WAAK,KAAK,wBAAwB,QAAQ,gBAAgB;AAAA,IAC5D;AAAA,EACF;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,oBAAoB;AAAA,EAChC;AACA,MAAI,QAAQ,qBAAqB;AAC/B,SAAK,KAAK,2BAA2B,QAAQ,mBAAmB;AAAA,EAClE;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,cAAc;AACxB,SAAK,KAAK,mBAAmB,QAAQ,YAAY;AAAA,EACnD;AACA,MAAI,QAAQ,uBAAuB;AACjC,SAAK,KAAK,6BAA6B,QAAQ,qBAAqB;AAAA,EACtE;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AAEA,SAAO;AACT;;;ADpLA;AAWA,IAAM,iBAAa,0BAAc,YAAY,GAAG;AAChD,IAAM,YAAiB,aAAQ,UAAU;AAEzC,IAAM,WAAW;AAMjB,SAAS,WAAW,MAAgB,mBAAwC,CAAC,GAAoB;AAC/F,QAAM,EAAE,eAAe,MAAM,IAAI;AAEjC,SAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,UAAM,UAAe,UAAK,WAAW,MAAM,OAAO,QAAQ;AAE1D,QAAI,CAAI,cAAW,OAAO,GAAG;AAC3B,aAAO;AAAA,QACL,IAAI,MAAM,yBAAyB,OAAO,sCAAsC;AAAA,MAClF;AAAA,IACF;AAEA,UAAM,UAAU;AAChB,UAAM,cAAc,CAAC,QAAQ,SAAS,GAAG,IAAI;AAE7C,UAAM,kBAAc,4BAAM,SAAS,WAAW;AAE9C,QAAI,SAAS;AACb,QAAI,SAAS;AAEb,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,SAAS;AAChC,UAAI,SAAS,GAAG;AACd,gBAAQ,MAAM;AAAA,MAChB,OAAO;AACL,cAAM,cAAc,UAAU;AAC9B,cAAM,QAAQ,IAAI;AAAA,UAChB,+CAA+C,IAAI;AAAA;AAAA,EAAQ,WAAW;AAAA,QACxE;AACA,eAAO,KAAK;AAAA,MACd;AAAA,IACF,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,QAAe;AACtC,UAAI,IAAI,QAAQ,SAAS,QAAQ,GAAG;AAClC;AAAA,UACE,IAAI;AAAA,YACF;AAAA,UACF;AAAA,QACF;AAAA,MACF,OAAO;AACL,eAAO,GAAG;AAAA,MACZ;AAAA,IACF,CAAC;AAAA,EACH,CAAC;AACH;AAEO,SAAS,QACd,YACA,UAA0B,CAAC,GACV;AACjB,QAAM,YAAY,MAAM,QAAQ,UAAU,IAAI,aAAa,CAAC,UAAU;AACtE,MAAI,UAAU,WAAW,GAAG;AAC1B,WAAO,QAAQ,OAAO,IAAI,MAAM,2CAA2C,CAAC;AAAA,EAC9E;AAEA,aAAW,SAAS,WAAW;AAC7B,QAAI,CAAI,cAAW,KAAK,GAAG;AACzB,aAAO,QAAQ,OAAO,IAAI,MAAM,mCAAmC,KAAK,EAAE,CAAC;AAAA,IAC7E;AAAA,EACF;AAEA,QAAM,OAAiB,CAAC,GAAG,WAAW,GAAG,UAAU,OAAO,CAAC;AAE3D,SAAO,WAAW,MAAM;AAAA,IACtB,cAAc,CAAC,QAAQ;AAAA,EACzB,CAAC;AACH;AAwBO,SAAS,IAAI,WAAmB,UAAsB,CAAC,GAAoB;AAChF,UAAQ;AAAA,IACN;AAAA,EACF;AAGA,QAAM,UAAoB,CAAC;AAC3B,MAAI,CAAC,QAAQ,QAAQ;AACnB,YAAQ,KAAK,MAAM;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,QAAQ,oBAAoB;AAC9B,cAAQ,KAAK,sBAAsB;AAAA,IACrC,WAAW,QAAQ,gBAAgB;AACjC,cAAQ,KAAK,oBAAoB;AAAA,IACnC,OAAO;AACL,cAAQ,KAAK,UAAU;AAAA,IACzB;AAAA,EACF;AACA,MAAI,QAAQ,cAAc;AACxB,YAAQ,KAAK,MAAM;AAAA,EACrB;AACA,MAAI,QAAQ,sBAAsB;AAChC,YAAQ,KAAK,KAAK;AAAA,EACpB;AAEA,SAAO,QAAQ,WAAW;AAAA,IACxB,WAAW,QAAQ;AAAA,IACnB,UAAU,QAAQ;AAAA,IAClB,qBAAqB,QAAQ;AAAA,IAC7B,gBAAgB,QAAQ;AAAA,IACxB,kBAAkB,QAAQ;AAAA,IAC1B,eAAe,QAAQ;AAAA,IACvB,QAAQ,QAAQ,SAAS,IAAI,UAAU;AAAA,IACvC,OAAO,CAAC,QAAQ;AAAA,EAClB,CAAC;AACH;","names":[]}

package/dist/index.d.cts

@@ -18,9 +18,9 @@ interface ConvertOptions {
     replaceInvalidChars?: string;
     /** Use PDF structure tree (tagged PDF) for reading order and semantic structure */
     useStructTree?: boolean;
-    /** Table detection method. Values: cluster */
+    /** Table detection method. Values: default (border-based), cluster (border + cluster). Default: default */
     tableMethod?: string;
-    /** Reading order algorithm. Values: none, xycut. Default: none */
+    /** Reading order algorithm. Values: off, xycut. Default: xycut */
     readingOrder?: string;
     /** Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none */
     markdownPageSeparator?: string;
@@ -28,8 +28,8 @@ interface ConvertOptions {
     textPageSeparator?: string;
     /** Separator between pages in HTML output. Use %page-number% for page numbers. Default: none */
     htmlPageSeparator?: string;
-    /** Embed images as Base64 data URIs instead of file path references */
-    embedImages?: boolean;
+    /** Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external */
+    imageOutput?: string;
     /** Output format for extracted images. Values: png, jpeg. Default: png */
     imageFormat?: string;
 }

package/dist/index.d.ts

@@ -18,9 +18,9 @@ interface ConvertOptions {
     replaceInvalidChars?: string;
     /** Use PDF structure tree (tagged PDF) for reading order and semantic structure */
     useStructTree?: boolean;
-    /** Table detection method. Values: cluster */
+    /** Table detection method. Values: default (border-based), cluster (border + cluster). Default: default */
     tableMethod?: string;
-    /** Reading order algorithm. Values: none, xycut. Default: none */
+    /** Reading order algorithm. Values: off, xycut. Default: xycut */
     readingOrder?: string;
     /** Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none */
     markdownPageSeparator?: string;
@@ -28,8 +28,8 @@ interface ConvertOptions {
     textPageSeparator?: string;
     /** Separator between pages in HTML output. Use %page-number% for page numbers. Default: none */
     htmlPageSeparator?: string;
-    /** Embed images as Base64 data URIs instead of file path references */
-    embedImages?: boolean;
+    /** Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external */
+    imageOutput?: string;
     /** Output format for extracted images. Values: png, jpeg. Default: png */
     imageFormat?: string;
 }

package/dist/index.js

@@ -58,8 +58,8 @@ function buildArgs(options) {
   if (options.htmlPageSeparator) {
     args.push("--html-page-separator", options.htmlPageSeparator);
   }
-  if (options.embedImages) {
-    args.push("--embed-images");
+  if (options.imageOutput) {
+    args.push("--image-output", options.imageOutput);
   }
   if (options.imageFormat) {
     args.push("--image-format", options.imageFormat);

package/dist/index.js.map

@@ -1 +1 @@
-{"version":3,"sources":["../src/index.ts","../src/convert-options.generated.ts"],"sourcesContent":["import { spawn } from 'child_process';\nimport * as path from 'path';\nimport * as fs from 'fs';\nimport { fileURLToPath } from 'url';\n\n// Re-export types and utilities from auto-generated file\nexport type { ConvertOptions } from './convert-options.generated.js';\nexport { buildArgs } from './convert-options.generated.js';\nimport type { ConvertOptions } from './convert-options.generated.js';\nimport { buildArgs } from './convert-options.generated.js';\n\nconst __filename = fileURLToPath(import.meta.url);\nconst __dirname = path.dirname(__filename);\n\nconst JAR_NAME = 'opendataloader-pdf-cli.jar';\n\ninterface JarExecutionOptions {\n  streamOutput?: boolean;\n}\n\nfunction executeJar(args: string[], executionOptions: JarExecutionOptions = {}): Promise<string> {\n  const { streamOutput = false } = executionOptions;\n\n  return new Promise((resolve, reject) => {\n    const jarPath = path.join(__dirname, '..', 'lib', JAR_NAME);\n\n    if (!fs.existsSync(jarPath)) {\n      return reject(\n        new Error(`JAR file not found at ${jarPath}. Please run the build script first.`),\n      );\n    }\n\n    const command = 'java';\n    const commandArgs = ['-jar', jarPath, ...args];\n\n    const javaProcess = spawn(command, commandArgs);\n\n    let stdout = '';\n    let stderr = '';\n\n    javaProcess.stdout.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stdout.write(chunk);\n      }\n      stdout += chunk;\n    });\n\n    javaProcess.stderr.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stderr.write(chunk);\n      }\n      stderr += chunk;\n    });\n\n    javaProcess.on('close', (code) => {\n      if (code === 0) {\n        resolve(stdout);\n      } else {\n        const errorOutput = stderr || stdout;\n        const error = new Error(\n          `The opendataloader-pdf CLI exited with code ${code}.\\n\\n${errorOutput}`,\n        );\n        reject(error);\n      }\n    });\n\n    javaProcess.on('error', (err: Error) => {\n      if (err.message.includes('ENOENT')) {\n        reject(\n          new Error(\n            \"'java' command not found. Please ensure Java is installed and in your system's PATH.\",\n          ),\n        );\n      } else {\n        reject(err);\n      }\n    });\n  });\n}\n\nexport function convert(\n  inputPaths: string | string[],\n  options: ConvertOptions = {},\n): Promise<string> {\n  const inputList = Array.isArray(inputPaths) ? inputPaths : [inputPaths];\n  if (inputList.length === 0) {\n    return Promise.reject(new Error('At least one input path must be provided.'));\n  }\n\n  for (const input of inputList) {\n    if (!fs.existsSync(input)) {\n      return Promise.reject(new Error(`Input file or folder not found: ${input}`));\n    }\n  }\n\n  const args: string[] = [...inputList, ...buildArgs(options)];\n\n  return executeJar(args, {\n    streamOutput: !options.quiet,\n  });\n}\n\n/**\n * @deprecated Use `convert()` and `ConvertOptions` instead. This function will be removed in a future version.\n */\nexport interface RunOptions {\n  outputFolder?: string;\n  password?: string;\n  replaceInvalidChars?: string;\n  generateMarkdown?: boolean;\n  generateHtml?: boolean;\n  generateAnnotatedPdf?: boolean;\n  keepLineBreaks?: boolean;\n  contentSafetyOff?: string;\n  htmlInMarkdown?: boolean;\n  addImageToMarkdown?: boolean;\n  noJson?: boolean;\n  debug?: boolean;\n  useStructTree?: boolean;\n}\n\n/**\n * @deprecated Use `convert()` instead. This function will be removed in a future version.\n */\nexport function run(inputPath: string, options: RunOptions = {}): Promise<string> {\n  console.warn(\n    'Warning: run() is deprecated and will be removed in a future version. Use convert() instead.',\n  );\n\n  // Build format array based on legacy boolean options\n  const formats: string[] = [];\n  if (!options.noJson) {\n    formats.push('json');\n  }\n  if (options.generateMarkdown) {\n    if (options.addImageToMarkdown) {\n      formats.push('markdown-with-images');\n    } else if (options.htmlInMarkdown) {\n      formats.push('markdown-with-html');\n    } else {\n      formats.push('markdown');\n    }\n  }\n  if (options.generateHtml) {\n    formats.push('html');\n  }\n  if (options.generateAnnotatedPdf) {\n    formats.push('pdf');\n  }\n\n  return convert(inputPath, {\n    outputDir: options.outputFolder,\n    password: options.password,\n    replaceInvalidChars: options.replaceInvalidChars,\n    keepLineBreaks: options.keepLineBreaks,\n    contentSafetyOff: options.contentSafetyOff,\n    useStructTree: options.useStructTree,\n    format: formats.length > 0 ? formats : undefined,\n    quiet: !options.debug,\n  });\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\n/**\n * Options for the convert function.\n */\nexport interface ConvertOptions {\n  /** Directory where output files are written. Default: input file directory */\n  outputDir?: string;\n  /** Password for encrypted PDF files */\n  password?: string;\n  /** Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json */\n  format?: string | string[];\n  /** Suppress console logging output */\n  quiet?: boolean;\n  /** Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg */\n  contentSafetyOff?: string | string[];\n  /** Preserve original line breaks in extracted text */\n  keepLineBreaks?: boolean;\n  /** Replacement character for invalid/unrecognized characters. Default: space */\n  replaceInvalidChars?: string;\n  /** Use PDF structure tree (tagged PDF) for reading order and semantic structure */\n  useStructTree?: boolean;\n  /** Table detection method. Values: cluster */\n  tableMethod?: string;\n  /** Reading order algorithm. Values: none, xycut. Default: none */\n  readingOrder?: string;\n  /** Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none */\n  markdownPageSeparator?: string;\n  /** Separator between pages in text output. Use %page-number% for page numbers. Default: none */\n  textPageSeparator?: string;\n  /** Separator between pages in HTML output. Use %page-number% for page numbers. Default: none */\n  htmlPageSeparator?: string;\n  /** Embed images as Base64 data URIs instead of file path references */\n  embedImages?: boolean;\n  /** Output format for extracted images. Values: png, jpeg. Default: png */\n  imageFormat?: string;\n}\n\n/**\n * Options as parsed from CLI (all values are strings from commander).\n */\nexport interface CliOptions {\n  outputDir?: string;\n  password?: string;\n  format?: string;\n  quiet?: boolean;\n  contentSafetyOff?: string;\n  keepLineBreaks?: boolean;\n  replaceInvalidChars?: string;\n  useStructTree?: boolean;\n  tableMethod?: string;\n  readingOrder?: string;\n  markdownPageSeparator?: string;\n  textPageSeparator?: string;\n  htmlPageSeparator?: string;\n  embedImages?: boolean;\n  imageFormat?: string;\n}\n\n/**\n * Convert CLI options to ConvertOptions.\n */\nexport function buildConvertOptions(cliOptions: CliOptions): ConvertOptions {\n  const convertOptions: ConvertOptions = {};\n\n  if (cliOptions.outputDir) {\n    convertOptions.outputDir = cliOptions.outputDir;\n  }\n  if (cliOptions.password) {\n    convertOptions.password = cliOptions.password;\n  }\n  if (cliOptions.format) {\n    convertOptions.format = cliOptions.format;\n  }\n  if (cliOptions.quiet) {\n    convertOptions.quiet = true;\n  }\n  if (cliOptions.contentSafetyOff) {\n    convertOptions.contentSafetyOff = cliOptions.contentSafetyOff;\n  }\n  if (cliOptions.keepLineBreaks) {\n    convertOptions.keepLineBreaks = true;\n  }\n  if (cliOptions.replaceInvalidChars) {\n    convertOptions.replaceInvalidChars = cliOptions.replaceInvalidChars;\n  }\n  if (cliOptions.useStructTree) {\n    convertOptions.useStructTree = true;\n  }\n  if (cliOptions.tableMethod) {\n    convertOptions.tableMethod = cliOptions.tableMethod;\n  }\n  if (cliOptions.readingOrder) {\n    convertOptions.readingOrder = cliOptions.readingOrder;\n  }\n  if (cliOptions.markdownPageSeparator) {\n    convertOptions.markdownPageSeparator = cliOptions.markdownPageSeparator;\n  }\n  if (cliOptions.textPageSeparator) {\n    convertOptions.textPageSeparator = cliOptions.textPageSeparator;\n  }\n  if (cliOptions.htmlPageSeparator) {\n    convertOptions.htmlPageSeparator = cliOptions.htmlPageSeparator;\n  }\n  if (cliOptions.embedImages) {\n    convertOptions.embedImages = true;\n  }\n  if (cliOptions.imageFormat) {\n    convertOptions.imageFormat = cliOptions.imageFormat;\n  }\n\n  return convertOptions;\n}\n\n/**\n * Build CLI arguments array from ConvertOptions.\n */\nexport function buildArgs(options: ConvertOptions): string[] {\n  const args: string[] = [];\n\n  if (options.outputDir) {\n    args.push('--output-dir', options.outputDir);\n  }\n  if (options.password) {\n    args.push('--password', options.password);\n  }\n  if (options.format) {\n    if (Array.isArray(options.format)) {\n      if (options.format.length > 0) {\n        args.push('--format', options.format.join(','));\n      }\n    } else {\n      args.push('--format', options.format);\n    }\n  }\n  if (options.quiet) {\n    args.push('--quiet');\n  }\n  if (options.contentSafetyOff) {\n    if (Array.isArray(options.contentSafetyOff)) {\n      if (options.contentSafetyOff.length > 0) {\n        args.push('--content-safety-off', options.contentSafetyOff.join(','));\n      }\n    } else {\n      args.push('--content-safety-off', options.contentSafetyOff);\n    }\n  }\n  if (options.keepLineBreaks) {\n    args.push('--keep-line-breaks');\n  }\n  if (options.replaceInvalidChars) {\n    args.push('--replace-invalid-chars', options.replaceInvalidChars);\n  }\n  if (options.useStructTree) {\n    args.push('--use-struct-tree');\n  }\n  if (options.tableMethod) {\n    args.push('--table-method', options.tableMethod);\n  }\n  if (options.readingOrder) {\n    args.push('--reading-order', options.readingOrder);\n  }\n  if (options.markdownPageSeparator) {\n    args.push('--markdown-page-separator', options.markdownPageSeparator);\n  }\n  if (options.textPageSeparator) {\n    args.push('--text-page-separator', options.textPageSeparator);\n  }\n  if (options.htmlPageSeparator) {\n    args.push('--html-page-separator', options.htmlPageSeparator);\n  }\n  if (options.embedImages) {\n    args.push('--embed-images');\n  }\n  if (options.imageFormat) {\n    args.push('--image-format', options.imageFormat);\n  }\n\n  return args;\n}\n"],"mappings":";AAAA,SAAS,aAAa;AACtB,YAAY,UAAU;AACtB,YAAY,QAAQ;AACpB,SAAS,qBAAqB;;;ACmHvB,SAAS,UAAU,SAAmC;AAC3D,QAAM,OAAiB,CAAC;AAExB,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,UAAU;AACpB,SAAK,KAAK,cAAc,QAAQ,QAAQ;AAAA,EAC1C;AACA,MAAI,QAAQ,QAAQ;AAClB,QAAI,MAAM,QAAQ,QAAQ,MAAM,GAAG;AACjC,UAAI,QAAQ,OAAO,SAAS,GAAG;AAC7B,aAAK,KAAK,YAAY,QAAQ,OAAO,KAAK,GAAG,CAAC;AAAA,MAChD;AAAA,IACF,OAAO;AACL,WAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,IACtC;AAAA,EACF;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,SAAS;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,MAAM,QAAQ,QAAQ,gBAAgB,GAAG;AAC3C,UAAI,QAAQ,iBAAiB,SAAS,GAAG;AACvC,aAAK,KAAK,wBAAwB,QAAQ,iBAAiB,KAAK,GAAG,CAAC;AAAA,MACtE;AAAA,IACF,OAAO;AACL,WAAK,KAAK,wBAAwB,QAAQ,gBAAgB;AAAA,IAC5D;AAAA,EACF;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,oBAAoB;AAAA,EAChC;AACA,MAAI,QAAQ,qBAAqB;AAC/B,SAAK,KAAK,2BAA2B,QAAQ,mBAAmB;AAAA,EAClE;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,cAAc;AACxB,SAAK,KAAK,mBAAmB,QAAQ,YAAY;AAAA,EACnD;AACA,MAAI,QAAQ,uBAAuB;AACjC,SAAK,KAAK,6BAA6B,QAAQ,qBAAqB;AAAA,EACtE;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,gBAAgB;AAAA,EAC5B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AAEA,SAAO;AACT;;;ADzKA,IAAM,aAAa,cAAc,YAAY,GAAG;AAChD,IAAM,YAAiB,aAAQ,UAAU;AAEzC,IAAM,WAAW;AAMjB,SAAS,WAAW,MAAgB,mBAAwC,CAAC,GAAoB;AAC/F,QAAM,EAAE,eAAe,MAAM,IAAI;AAEjC,SAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,UAAM,UAAe,UAAK,WAAW,MAAM,OAAO,QAAQ;AAE1D,QAAI,CAAI,cAAW,OAAO,GAAG;AAC3B,aAAO;AAAA,QACL,IAAI,MAAM,yBAAyB,OAAO,sCAAsC;AAAA,MAClF;AAAA,IACF;AAEA,UAAM,UAAU;AAChB,UAAM,cAAc,CAAC,QAAQ,SAAS,GAAG,IAAI;AAE7C,UAAM,cAAc,MAAM,SAAS,WAAW;AAE9C,QAAI,SAAS;AACb,QAAI,SAAS;AAEb,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,SAAS;AAChC,UAAI,SAAS,GAAG;AACd,gBAAQ,MAAM;AAAA,MAChB,OAAO;AACL,cAAM,cAAc,UAAU;AAC9B,cAAM,QAAQ,IAAI;AAAA,UAChB,+CAA+C,IAAI;AAAA;AAAA,EAAQ,WAAW;AAAA,QACxE;AACA,eAAO,KAAK;AAAA,MACd;AAAA,IACF,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,QAAe;AACtC,UAAI,IAAI,QAAQ,SAAS,QAAQ,GAAG;AAClC;AAAA,UACE,IAAI;AAAA,YACF;AAAA,UACF;AAAA,QACF;AAAA,MACF,OAAO;AACL,eAAO,GAAG;AAAA,MACZ;AAAA,IACF,CAAC;AAAA,EACH,CAAC;AACH;AAEO,SAAS,QACd,YACA,UAA0B,CAAC,GACV;AACjB,QAAM,YAAY,MAAM,QAAQ,UAAU,IAAI,aAAa,CAAC,UAAU;AACtE,MAAI,UAAU,WAAW,GAAG;AAC1B,WAAO,QAAQ,OAAO,IAAI,MAAM,2CAA2C,CAAC;AAAA,EAC9E;AAEA,aAAW,SAAS,WAAW;AAC7B,QAAI,CAAI,cAAW,KAAK,GAAG;AACzB,aAAO,QAAQ,OAAO,IAAI,MAAM,mCAAmC,KAAK,EAAE,CAAC;AAAA,IAC7E;AAAA,EACF;AAEA,QAAM,OAAiB,CAAC,GAAG,WAAW,GAAG,UAAU,OAAO,CAAC;AAE3D,SAAO,WAAW,MAAM;AAAA,IACtB,cAAc,CAAC,QAAQ;AAAA,EACzB,CAAC;AACH;AAwBO,SAAS,IAAI,WAAmB,UAAsB,CAAC,GAAoB;AAChF,UAAQ;AAAA,IACN;AAAA,EACF;AAGA,QAAM,UAAoB,CAAC;AAC3B,MAAI,CAAC,QAAQ,QAAQ;AACnB,YAAQ,KAAK,MAAM;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,QAAQ,oBAAoB;AAC9B,cAAQ,KAAK,sBAAsB;AAAA,IACrC,WAAW,QAAQ,gBAAgB;AACjC,cAAQ,KAAK,oBAAoB;AAAA,IACnC,OAAO;AACL,cAAQ,KAAK,UAAU;AAAA,IACzB;AAAA,EACF;AACA,MAAI,QAAQ,cAAc;AACxB,YAAQ,KAAK,MAAM;AAAA,EACrB;AACA,MAAI,QAAQ,sBAAsB;AAChC,YAAQ,KAAK,KAAK;AAAA,EACpB;AAEA,SAAO,QAAQ,WAAW;AAAA,IACxB,WAAW,QAAQ;AAAA,IACnB,UAAU,QAAQ;AAAA,IAClB,qBAAqB,QAAQ;AAAA,IAC7B,gBAAgB,QAAQ;AAAA,IACxB,kBAAkB,QAAQ;AAAA,IAC1B,eAAe,QAAQ;AAAA,IACvB,QAAQ,QAAQ,SAAS,IAAI,UAAU;AAAA,IACvC,OAAO,CAAC,QAAQ;AAAA,EAClB,CAAC;AACH;","names":[]}
+{"version":3,"sources":["../src/index.ts","../src/convert-options.generated.ts"],"sourcesContent":["import { spawn } from 'child_process';\nimport * as path from 'path';\nimport * as fs from 'fs';\nimport { fileURLToPath } from 'url';\n\n// Re-export types and utilities from auto-generated file\nexport type { ConvertOptions } from './convert-options.generated.js';\nexport { buildArgs } from './convert-options.generated.js';\nimport type { ConvertOptions } from './convert-options.generated.js';\nimport { buildArgs } from './convert-options.generated.js';\n\nconst __filename = fileURLToPath(import.meta.url);\nconst __dirname = path.dirname(__filename);\n\nconst JAR_NAME = 'opendataloader-pdf-cli.jar';\n\ninterface JarExecutionOptions {\n  streamOutput?: boolean;\n}\n\nfunction executeJar(args: string[], executionOptions: JarExecutionOptions = {}): Promise<string> {\n  const { streamOutput = false } = executionOptions;\n\n  return new Promise((resolve, reject) => {\n    const jarPath = path.join(__dirname, '..', 'lib', JAR_NAME);\n\n    if (!fs.existsSync(jarPath)) {\n      return reject(\n        new Error(`JAR file not found at ${jarPath}. Please run the build script first.`),\n      );\n    }\n\n    const command = 'java';\n    const commandArgs = ['-jar', jarPath, ...args];\n\n    const javaProcess = spawn(command, commandArgs);\n\n    let stdout = '';\n    let stderr = '';\n\n    javaProcess.stdout.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stdout.write(chunk);\n      }\n      stdout += chunk;\n    });\n\n    javaProcess.stderr.on('data', (data) => {\n      const chunk = data.toString();\n      if (streamOutput) {\n        process.stderr.write(chunk);\n      }\n      stderr += chunk;\n    });\n\n    javaProcess.on('close', (code) => {\n      if (code === 0) {\n        resolve(stdout);\n      } else {\n        const errorOutput = stderr || stdout;\n        const error = new Error(\n          `The opendataloader-pdf CLI exited with code ${code}.\\n\\n${errorOutput}`,\n        );\n        reject(error);\n      }\n    });\n\n    javaProcess.on('error', (err: Error) => {\n      if (err.message.includes('ENOENT')) {\n        reject(\n          new Error(\n            \"'java' command not found. Please ensure Java is installed and in your system's PATH.\",\n          ),\n        );\n      } else {\n        reject(err);\n      }\n    });\n  });\n}\n\nexport function convert(\n  inputPaths: string | string[],\n  options: ConvertOptions = {},\n): Promise<string> {\n  const inputList = Array.isArray(inputPaths) ? inputPaths : [inputPaths];\n  if (inputList.length === 0) {\n    return Promise.reject(new Error('At least one input path must be provided.'));\n  }\n\n  for (const input of inputList) {\n    if (!fs.existsSync(input)) {\n      return Promise.reject(new Error(`Input file or folder not found: ${input}`));\n    }\n  }\n\n  const args: string[] = [...inputList, ...buildArgs(options)];\n\n  return executeJar(args, {\n    streamOutput: !options.quiet,\n  });\n}\n\n/**\n * @deprecated Use `convert()` and `ConvertOptions` instead. This function will be removed in a future version.\n */\nexport interface RunOptions {\n  outputFolder?: string;\n  password?: string;\n  replaceInvalidChars?: string;\n  generateMarkdown?: boolean;\n  generateHtml?: boolean;\n  generateAnnotatedPdf?: boolean;\n  keepLineBreaks?: boolean;\n  contentSafetyOff?: string;\n  htmlInMarkdown?: boolean;\n  addImageToMarkdown?: boolean;\n  noJson?: boolean;\n  debug?: boolean;\n  useStructTree?: boolean;\n}\n\n/**\n * @deprecated Use `convert()` instead. This function will be removed in a future version.\n */\nexport function run(inputPath: string, options: RunOptions = {}): Promise<string> {\n  console.warn(\n    'Warning: run() is deprecated and will be removed in a future version. Use convert() instead.',\n  );\n\n  // Build format array based on legacy boolean options\n  const formats: string[] = [];\n  if (!options.noJson) {\n    formats.push('json');\n  }\n  if (options.generateMarkdown) {\n    if (options.addImageToMarkdown) {\n      formats.push('markdown-with-images');\n    } else if (options.htmlInMarkdown) {\n      formats.push('markdown-with-html');\n    } else {\n      formats.push('markdown');\n    }\n  }\n  if (options.generateHtml) {\n    formats.push('html');\n  }\n  if (options.generateAnnotatedPdf) {\n    formats.push('pdf');\n  }\n\n  return convert(inputPath, {\n    outputDir: options.outputFolder,\n    password: options.password,\n    replaceInvalidChars: options.replaceInvalidChars,\n    keepLineBreaks: options.keepLineBreaks,\n    contentSafetyOff: options.contentSafetyOff,\n    useStructTree: options.useStructTree,\n    format: formats.length > 0 ? formats : undefined,\n    quiet: !options.debug,\n  });\n}\n","// AUTO-GENERATED FROM options.json - DO NOT EDIT DIRECTLY\n// Run `npm run generate-options` to regenerate\n\n/**\n * Options for the convert function.\n */\nexport interface ConvertOptions {\n  /** Directory where output files are written. Default: input file directory */\n  outputDir?: string;\n  /** Password for encrypted PDF files */\n  password?: string;\n  /** Output formats (comma-separated). Values: json, text, html, pdf, markdown, markdown-with-html, markdown-with-images. Default: json */\n  format?: string | string[];\n  /** Suppress console logging output */\n  quiet?: boolean;\n  /** Disable content safety filters. Values: all, hidden-text, off-page, tiny, hidden-ocg */\n  contentSafetyOff?: string | string[];\n  /** Preserve original line breaks in extracted text */\n  keepLineBreaks?: boolean;\n  /** Replacement character for invalid/unrecognized characters. Default: space */\n  replaceInvalidChars?: string;\n  /** Use PDF structure tree (tagged PDF) for reading order and semantic structure */\n  useStructTree?: boolean;\n  /** Table detection method. Values: default (border-based), cluster (border + cluster). Default: default */\n  tableMethod?: string;\n  /** Reading order algorithm. Values: off, xycut. Default: xycut */\n  readingOrder?: string;\n  /** Separator between pages in Markdown output. Use %page-number% for page numbers. Default: none */\n  markdownPageSeparator?: string;\n  /** Separator between pages in text output. Use %page-number% for page numbers. Default: none */\n  textPageSeparator?: string;\n  /** Separator between pages in HTML output. Use %page-number% for page numbers. Default: none */\n  htmlPageSeparator?: string;\n  /** Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external */\n  imageOutput?: string;\n  /** Output format for extracted images. Values: png, jpeg. Default: png */\n  imageFormat?: string;\n}\n\n/**\n * Options as parsed from CLI (all values are strings from commander).\n */\nexport interface CliOptions {\n  outputDir?: string;\n  password?: string;\n  format?: string;\n  quiet?: boolean;\n  contentSafetyOff?: string;\n  keepLineBreaks?: boolean;\n  replaceInvalidChars?: string;\n  useStructTree?: boolean;\n  tableMethod?: string;\n  readingOrder?: string;\n  markdownPageSeparator?: string;\n  textPageSeparator?: string;\n  htmlPageSeparator?: string;\n  imageOutput?: string;\n  imageFormat?: string;\n}\n\n/**\n * Convert CLI options to ConvertOptions.\n */\nexport function buildConvertOptions(cliOptions: CliOptions): ConvertOptions {\n  const convertOptions: ConvertOptions = {};\n\n  if (cliOptions.outputDir) {\n    convertOptions.outputDir = cliOptions.outputDir;\n  }\n  if (cliOptions.password) {\n    convertOptions.password = cliOptions.password;\n  }\n  if (cliOptions.format) {\n    convertOptions.format = cliOptions.format;\n  }\n  if (cliOptions.quiet) {\n    convertOptions.quiet = true;\n  }\n  if (cliOptions.contentSafetyOff) {\n    convertOptions.contentSafetyOff = cliOptions.contentSafetyOff;\n  }\n  if (cliOptions.keepLineBreaks) {\n    convertOptions.keepLineBreaks = true;\n  }\n  if (cliOptions.replaceInvalidChars) {\n    convertOptions.replaceInvalidChars = cliOptions.replaceInvalidChars;\n  }\n  if (cliOptions.useStructTree) {\n    convertOptions.useStructTree = true;\n  }\n  if (cliOptions.tableMethod) {\n    convertOptions.tableMethod = cliOptions.tableMethod;\n  }\n  if (cliOptions.readingOrder) {\n    convertOptions.readingOrder = cliOptions.readingOrder;\n  }\n  if (cliOptions.markdownPageSeparator) {\n    convertOptions.markdownPageSeparator = cliOptions.markdownPageSeparator;\n  }\n  if (cliOptions.textPageSeparator) {\n    convertOptions.textPageSeparator = cliOptions.textPageSeparator;\n  }\n  if (cliOptions.htmlPageSeparator) {\n    convertOptions.htmlPageSeparator = cliOptions.htmlPageSeparator;\n  }\n  if (cliOptions.imageOutput) {\n    convertOptions.imageOutput = cliOptions.imageOutput;\n  }\n  if (cliOptions.imageFormat) {\n    convertOptions.imageFormat = cliOptions.imageFormat;\n  }\n\n  return convertOptions;\n}\n\n/**\n * Build CLI arguments array from ConvertOptions.\n */\nexport function buildArgs(options: ConvertOptions): string[] {\n  const args: string[] = [];\n\n  if (options.outputDir) {\n    args.push('--output-dir', options.outputDir);\n  }\n  if (options.password) {\n    args.push('--password', options.password);\n  }\n  if (options.format) {\n    if (Array.isArray(options.format)) {\n      if (options.format.length > 0) {\n        args.push('--format', options.format.join(','));\n      }\n    } else {\n      args.push('--format', options.format);\n    }\n  }\n  if (options.quiet) {\n    args.push('--quiet');\n  }\n  if (options.contentSafetyOff) {\n    if (Array.isArray(options.contentSafetyOff)) {\n      if (options.contentSafetyOff.length > 0) {\n        args.push('--content-safety-off', options.contentSafetyOff.join(','));\n      }\n    } else {\n      args.push('--content-safety-off', options.contentSafetyOff);\n    }\n  }\n  if (options.keepLineBreaks) {\n    args.push('--keep-line-breaks');\n  }\n  if (options.replaceInvalidChars) {\n    args.push('--replace-invalid-chars', options.replaceInvalidChars);\n  }\n  if (options.useStructTree) {\n    args.push('--use-struct-tree');\n  }\n  if (options.tableMethod) {\n    args.push('--table-method', options.tableMethod);\n  }\n  if (options.readingOrder) {\n    args.push('--reading-order', options.readingOrder);\n  }\n  if (options.markdownPageSeparator) {\n    args.push('--markdown-page-separator', options.markdownPageSeparator);\n  }\n  if (options.textPageSeparator) {\n    args.push('--text-page-separator', options.textPageSeparator);\n  }\n  if (options.htmlPageSeparator) {\n    args.push('--html-page-separator', options.htmlPageSeparator);\n  }\n  if (options.imageOutput) {\n    args.push('--image-output', options.imageOutput);\n  }\n  if (options.imageFormat) {\n    args.push('--image-format', options.imageFormat);\n  }\n\n  return args;\n}\n"],"mappings":";AAAA,SAAS,aAAa;AACtB,YAAY,UAAU;AACtB,YAAY,QAAQ;AACpB,SAAS,qBAAqB;;;ACmHvB,SAAS,UAAU,SAAmC;AAC3D,QAAM,OAAiB,CAAC;AAExB,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,UAAU;AACpB,SAAK,KAAK,cAAc,QAAQ,QAAQ;AAAA,EAC1C;AACA,MAAI,QAAQ,QAAQ;AAClB,QAAI,MAAM,QAAQ,QAAQ,MAAM,GAAG;AACjC,UAAI,QAAQ,OAAO,SAAS,GAAG;AAC7B,aAAK,KAAK,YAAY,QAAQ,OAAO,KAAK,GAAG,CAAC;AAAA,MAChD;AAAA,IACF,OAAO;AACL,WAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,IACtC;AAAA,EACF;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,SAAS;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,MAAM,QAAQ,QAAQ,gBAAgB,GAAG;AAC3C,UAAI,QAAQ,iBAAiB,SAAS,GAAG;AACvC,aAAK,KAAK,wBAAwB,QAAQ,iBAAiB,KAAK,GAAG,CAAC;AAAA,MACtE;AAAA,IACF,OAAO;AACL,WAAK,KAAK,wBAAwB,QAAQ,gBAAgB;AAAA,IAC5D;AAAA,EACF;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,oBAAoB;AAAA,EAChC;AACA,MAAI,QAAQ,qBAAqB;AAC/B,SAAK,KAAK,2BAA2B,QAAQ,mBAAmB;AAAA,EAClE;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,cAAc;AACxB,SAAK,KAAK,mBAAmB,QAAQ,YAAY;AAAA,EACnD;AACA,MAAI,QAAQ,uBAAuB;AACjC,SAAK,KAAK,6BAA6B,QAAQ,qBAAqB;AAAA,EACtE;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,mBAAmB;AAC7B,SAAK,KAAK,yBAAyB,QAAQ,iBAAiB;AAAA,EAC9D;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AACA,MAAI,QAAQ,aAAa;AACvB,SAAK,KAAK,kBAAkB,QAAQ,WAAW;AAAA,EACjD;AAEA,SAAO;AACT;;;ADzKA,IAAM,aAAa,cAAc,YAAY,GAAG;AAChD,IAAM,YAAiB,aAAQ,UAAU;AAEzC,IAAM,WAAW;AAMjB,SAAS,WAAW,MAAgB,mBAAwC,CAAC,GAAoB;AAC/F,QAAM,EAAE,eAAe,MAAM,IAAI;AAEjC,SAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,UAAM,UAAe,UAAK,WAAW,MAAM,OAAO,QAAQ;AAE1D,QAAI,CAAI,cAAW,OAAO,GAAG;AAC3B,aAAO;AAAA,QACL,IAAI,MAAM,yBAAyB,OAAO,sCAAsC;AAAA,MAClF;AAAA,IACF;AAEA,UAAM,UAAU;AAChB,UAAM,cAAc,CAAC,QAAQ,SAAS,GAAG,IAAI;AAE7C,UAAM,cAAc,MAAM,SAAS,WAAW;AAE9C,QAAI,SAAS;AACb,QAAI,SAAS;AAEb,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,OAAO,GAAG,QAAQ,CAAC,SAAS;AACtC,YAAM,QAAQ,KAAK,SAAS;AAC5B,UAAI,cAAc;AAChB,gBAAQ,OAAO,MAAM,KAAK;AAAA,MAC5B;AACA,gBAAU;AAAA,IACZ,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,SAAS;AAChC,UAAI,SAAS,GAAG;AACd,gBAAQ,MAAM;AAAA,MAChB,OAAO;AACL,cAAM,cAAc,UAAU;AAC9B,cAAM,QAAQ,IAAI;AAAA,UAChB,+CAA+C,IAAI;AAAA;AAAA,EAAQ,WAAW;AAAA,QACxE;AACA,eAAO,KAAK;AAAA,MACd;AAAA,IACF,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,QAAe;AACtC,UAAI,IAAI,QAAQ,SAAS,QAAQ,GAAG;AAClC;AAAA,UACE,IAAI;AAAA,YACF;AAAA,UACF;AAAA,QACF;AAAA,MACF,OAAO;AACL,eAAO,GAAG;AAAA,MACZ;AAAA,IACF,CAAC;AAAA,EACH,CAAC;AACH;AAEO,SAAS,QACd,YACA,UAA0B,CAAC,GACV;AACjB,QAAM,YAAY,MAAM,QAAQ,UAAU,IAAI,aAAa,CAAC,UAAU;AACtE,MAAI,UAAU,WAAW,GAAG;AAC1B,WAAO,QAAQ,OAAO,IAAI,MAAM,2CAA2C,CAAC;AAAA,EAC9E;AAEA,aAAW,SAAS,WAAW;AAC7B,QAAI,CAAI,cAAW,KAAK,GAAG;AACzB,aAAO,QAAQ,OAAO,IAAI,MAAM,mCAAmC,KAAK,EAAE,CAAC;AAAA,IAC7E;AAAA,EACF;AAEA,QAAM,OAAiB,CAAC,GAAG,WAAW,GAAG,UAAU,OAAO,CAAC;AAE3D,SAAO,WAAW,MAAM;AAAA,IACtB,cAAc,CAAC,QAAQ;AAAA,EACzB,CAAC;AACH;AAwBO,SAAS,IAAI,WAAmB,UAAsB,CAAC,GAAoB;AAChF,UAAQ;AAAA,IACN;AAAA,EACF;AAGA,QAAM,UAAoB,CAAC;AAC3B,MAAI,CAAC,QAAQ,QAAQ;AACnB,YAAQ,KAAK,MAAM;AAAA,EACrB;AACA,MAAI,QAAQ,kBAAkB;AAC5B,QAAI,QAAQ,oBAAoB;AAC9B,cAAQ,KAAK,sBAAsB;AAAA,IACrC,WAAW,QAAQ,gBAAgB;AACjC,cAAQ,KAAK,oBAAoB;AAAA,IACnC,OAAO;AACL,cAAQ,KAAK,UAAU;AAAA,IACzB;AAAA,EACF;AACA,MAAI,QAAQ,cAAc;AACxB,YAAQ,KAAK,MAAM;AAAA,EACrB;AACA,MAAI,QAAQ,sBAAsB;AAChC,YAAQ,KAAK,KAAK;AAAA,EACpB;AAEA,SAAO,QAAQ,WAAW;AAAA,IACxB,WAAW,QAAQ;AAAA,IACnB,UAAU,QAAQ;AAAA,IAClB,qBAAqB,QAAQ;AAAA,IAC7B,gBAAgB,QAAQ;AAAA,IACxB,kBAAkB,QAAQ;AAAA,IAC1B,eAAe,QAAQ;AAAA,IACvB,QAAQ,QAAQ,SAAS,IAAI,UAAU;AAAA,IACvC,OAAO,CAAC,QAAQ;AAAA,EAClB,CAAC;AACH;","names":[]}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@opendataloader/pdf",
-  "version": "1.4.3",
+  "version": "1.5.1",
   "description": "A Node.js wrapper for the opendataloader-pdf Java CLI.",
   "main": "./dist/index.cjs",
   "module": "./dist/index.js",