@opendataloader/pdf 1.5.0 → 1.6.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +57 -12
- package/dist/cli.cjs +35 -0
- package/dist/cli.cjs.map +1 -1
- package/dist/cli.js +35 -0
- package/dist/cli.js.map +1 -1
- package/dist/index.cjs +15 -0
- package/dist/index.cjs.map +1 -1
- package/dist/index.d.cts +10 -0
- package/dist/index.d.ts +10 -0
- package/dist/index.js +15 -0
- package/dist/index.js.map +1 -1
- package/lib/opendataloader-pdf-cli.jar +0 -0
- package/package.json +7 -2
package/README.md
CHANGED
|
@@ -186,6 +186,46 @@ Most PDF parsers ignore structure tags entirely. We're one of the few that fully
|
|
|
186
186
|
|
|
187
187
|
<br/>
|
|
188
188
|
|
|
189
|
+
## Hybrid Mode
|
|
190
|
+
|
|
191
|
+
For documents with complex tables or OCR needs, enable hybrid mode to route challenging pages to an AI backend while keeping simple pages fast and local.
|
|
192
|
+
|
|
193
|
+
**Results**: Table accuracy jumps from 0.49 → 0.93 (+90%) with acceptable speed trade-off.
|
|
194
|
+
|
|
195
|
+
```bash
|
|
196
|
+
pip install opendataloader-pdf[hybrid]
|
|
197
|
+
```
|
|
198
|
+
|
|
199
|
+
Terminal 1: Start the backend server
|
|
200
|
+
|
|
201
|
+
```bash
|
|
202
|
+
opendataloader-pdf-hybrid --port 5002
|
|
203
|
+
```
|
|
204
|
+
|
|
205
|
+
Terminal 2: Process PDFs with hybrid mode
|
|
206
|
+
|
|
207
|
+
```bash
|
|
208
|
+
opendataloader-pdf --hybrid docling-fast input.pdf
|
|
209
|
+
```
|
|
210
|
+
|
|
211
|
+
Or use in Python:
|
|
212
|
+
|
|
213
|
+
```python
|
|
214
|
+
opendataloader_pdf.convert(
|
|
215
|
+
input_path="complex_tables.pdf",
|
|
216
|
+
output_dir="output/",
|
|
217
|
+
hybrid="docling-fast" # Routes complex pages to AI backend
|
|
218
|
+
)
|
|
219
|
+
```
|
|
220
|
+
|
|
221
|
+
- **Local-first**: Simple pages processed locally, complex pages routed to backend
|
|
222
|
+
- **Fallback**: If backend unavailable, gracefully falls back to local processing
|
|
223
|
+
- **Privacy**: Run the backend locally in Docker for 100% on-premise
|
|
224
|
+
|
|
225
|
+
[Hybrid Mode Guide →](https://opendataloader.org/docs/hybrid-mode)
|
|
226
|
+
|
|
227
|
+
<br/>
|
|
228
|
+
|
|
189
229
|
## LangChain Integration
|
|
190
230
|
|
|
191
231
|
OpenDataLoader PDF has an official LangChain integration for seamless RAG pipeline development.
|
|
@@ -222,23 +262,24 @@ We continuously benchmark against real-world documents.
|
|
|
222
262
|
|
|
223
263
|
### Quick Comparison
|
|
224
264
|
|
|
225
|
-
| Engine
|
|
226
|
-
|
|
227
|
-
| **opendataloader**
|
|
228
|
-
|
|
|
229
|
-
|
|
|
230
|
-
|
|
|
265
|
+
| Engine | Speed (s/page) | Reading Order | Table | Heading |
|
|
266
|
+
|-----------------------------|----------------|---------------|----------|----------|
|
|
267
|
+
| **opendataloader** | **0.05** | 0.91 | 0.49 | 0.65 |
|
|
268
|
+
| **opendataloader [hybrid]** | 0.45 | **0.93** | **0.93** | 0.78 |
|
|
269
|
+
| docling | 0.73 | 0.90 | 0.89 | **0.80** |
|
|
270
|
+
| pymupdf4llm | 0.09 | 0.89 | 0.40 | 0.41 |
|
|
271
|
+
| markitdown | **0.04** | 0.88 | 0.00 | 0.00 |
|
|
231
272
|
|
|
232
273
|
> Scores are normalized to [0, 1]. Higher is better for accuracy metrics; lower is better for speed. **Bold** indicates best performance.
|
|
233
274
|
|
|
234
275
|
### When to Use Each Engine
|
|
235
276
|
|
|
236
|
-
| Use Case
|
|
237
|
-
|
|
238
|
-
| Best overall balance
|
|
239
|
-
|
|
|
240
|
-
| Speed-critical pipelines
|
|
241
|
-
| PyMuPDF ecosystem
|
|
277
|
+
| Use Case | Recommended Engine | Why |
|
|
278
|
+
|-------------------------------|-----------------------------|--------------------------------------------------------|
|
|
279
|
+
| Best overall balance | **opendataloader** | Fast with high reading order accuracy |
|
|
280
|
+
| Complex tables & max accuracy | **opendataloader [hybrid]** | Docling-level accuracy, 40% faster |
|
|
281
|
+
| Speed-critical pipelines | markitdown | Fastest, but no table/heading extraction |
|
|
282
|
+
| PyMuPDF ecosystem | pymupdf4llm | Good balance if already using PyMuPDF |
|
|
242
283
|
|
|
243
284
|
### Visual Comparison
|
|
244
285
|
|
|
@@ -289,6 +330,10 @@ OpenDataLoader takes a different approach from many PDF parsers:
|
|
|
289
330
|
|
|
290
331
|
This means: consistent output (same input = same output), no GPU required, faster processing, and no model hallucinations.
|
|
291
332
|
|
|
333
|
+
### How do I get better accuracy for complex tables?
|
|
334
|
+
|
|
335
|
+
Enable hybrid mode with `pip install opendataloader-pdf[hybrid]`. This routes pages with complex tables to an AI backend (like docling-serve) while keeping simple pages fast and local. Table accuracy improves from 0.49 to 0.93 — matching or exceeding dedicated AI parsers while remaining faster and more cost-effective.
|
|
336
|
+
|
|
292
337
|
<br/>
|
|
293
338
|
|
|
294
339
|
## Contributing
|
package/dist/cli.cjs
CHANGED
|
@@ -80,6 +80,21 @@ function buildConvertOptions(cliOptions) {
|
|
|
80
80
|
if (cliOptions.imageFormat) {
|
|
81
81
|
convertOptions.imageFormat = cliOptions.imageFormat;
|
|
82
82
|
}
|
|
83
|
+
if (cliOptions.pages) {
|
|
84
|
+
convertOptions.pages = cliOptions.pages;
|
|
85
|
+
}
|
|
86
|
+
if (cliOptions.hybrid) {
|
|
87
|
+
convertOptions.hybrid = cliOptions.hybrid;
|
|
88
|
+
}
|
|
89
|
+
if (cliOptions.hybridUrl) {
|
|
90
|
+
convertOptions.hybridUrl = cliOptions.hybridUrl;
|
|
91
|
+
}
|
|
92
|
+
if (cliOptions.hybridTimeout) {
|
|
93
|
+
convertOptions.hybridTimeout = cliOptions.hybridTimeout;
|
|
94
|
+
}
|
|
95
|
+
if (cliOptions.hybridFallback) {
|
|
96
|
+
convertOptions.hybridFallback = true;
|
|
97
|
+
}
|
|
83
98
|
return convertOptions;
|
|
84
99
|
}
|
|
85
100
|
function buildArgs(options) {
|
|
@@ -141,6 +156,21 @@ function buildArgs(options) {
|
|
|
141
156
|
if (options.imageFormat) {
|
|
142
157
|
args.push("--image-format", options.imageFormat);
|
|
143
158
|
}
|
|
159
|
+
if (options.pages) {
|
|
160
|
+
args.push("--pages", options.pages);
|
|
161
|
+
}
|
|
162
|
+
if (options.hybrid) {
|
|
163
|
+
args.push("--hybrid", options.hybrid);
|
|
164
|
+
}
|
|
165
|
+
if (options.hybridUrl) {
|
|
166
|
+
args.push("--hybrid-url", options.hybridUrl);
|
|
167
|
+
}
|
|
168
|
+
if (options.hybridTimeout) {
|
|
169
|
+
args.push("--hybrid-timeout", options.hybridTimeout);
|
|
170
|
+
}
|
|
171
|
+
if (options.hybridFallback) {
|
|
172
|
+
args.push("--hybrid-fallback");
|
|
173
|
+
}
|
|
144
174
|
return args;
|
|
145
175
|
}
|
|
146
176
|
|
|
@@ -236,6 +266,11 @@ function registerCliOptions(program) {
|
|
|
236
266
|
program.option("--html-page-separator <value>", "Separator between pages in HTML output. Use %page-number% for page numbers. Default: none");
|
|
237
267
|
program.option("--image-output <value>", "Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external");
|
|
238
268
|
program.option("--image-format <value>", "Output format for extracted images. Values: png, jpeg. Default: png");
|
|
269
|
+
program.option("--pages <value>", 'Pages to extract (e.g., "1,3,5-7"). Default: all pages');
|
|
270
|
+
program.option("--hybrid <value>", "Hybrid backend for AI processing. Values: off (default), docling-fast");
|
|
271
|
+
program.option("--hybrid-url <value>", "Hybrid backend server URL (overrides default)");
|
|
272
|
+
program.option("--hybrid-timeout <value>", "Hybrid backend request timeout in milliseconds. Default: 30000");
|
|
273
|
+
program.option("--hybrid-fallback", "Fallback to Java processing on hybrid backend error. Default: true");
|
|
239
274
|
}
|
|
240
275
|
|
|
241
276
|
// src/cli.ts
|
package/dist/cli.cjs.map
CHANGED
|
@@ -1 +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: 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":[]}
|
|
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: 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 /** Pages to extract (e.g., \"1,3,5-7\"). Default: all pages */\n pages?: string;\n /** Hybrid backend for AI processing. Values: off (default), docling-fast */\n hybrid?: string;\n /** Hybrid backend server URL (overrides default) */\n hybridUrl?: string;\n /** Hybrid backend request timeout in milliseconds. Default: 30000 */\n hybridTimeout?: string;\n /** Fallback to Java processing on hybrid backend error. Default: true */\n hybridFallback?: boolean;\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 pages?: string;\n hybrid?: string;\n hybridUrl?: string;\n hybridTimeout?: string;\n hybridFallback?: boolean;\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 if (cliOptions.pages) {\n convertOptions.pages = cliOptions.pages;\n }\n if (cliOptions.hybrid) {\n convertOptions.hybrid = cliOptions.hybrid;\n }\n if (cliOptions.hybridUrl) {\n convertOptions.hybridUrl = cliOptions.hybridUrl;\n }\n if (cliOptions.hybridTimeout) {\n convertOptions.hybridTimeout = cliOptions.hybridTimeout;\n }\n if (cliOptions.hybridFallback) {\n convertOptions.hybridFallback = true;\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 if (options.pages) {\n args.push('--pages', options.pages);\n }\n if (options.hybrid) {\n args.push('--hybrid', options.hybrid);\n }\n if (options.hybridUrl) {\n args.push('--hybrid-url', options.hybridUrl);\n }\n if (options.hybridTimeout) {\n args.push('--hybrid-timeout', options.hybridTimeout);\n }\n if (options.hybridFallback) {\n args.push('--hybrid-fallback');\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 program.option('--pages <value>', 'Pages to extract (e.g., \"1,3,5-7\"). Default: all pages');\n program.option('--hybrid <value>', 'Hybrid backend for AI processing. Values: off (default), docling-fast');\n program.option('--hybrid-url <value>', 'Hybrid backend server URL (overrides default)');\n program.option('--hybrid-timeout <value>', 'Hybrid backend request timeout in milliseconds. Default: 30000');\n program.option('--hybrid-fallback', 'Fallback to Java processing on hybrid backend error. Default: true');\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;AACA,uBAAwC;;;ACDxC,2BAAsB;AACtB,WAAsB;AACtB,SAAoB;AACpB,iBAA8B;;;AC2EvB,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;AACA,MAAI,WAAW,OAAO;AACpB,mBAAe,QAAQ,WAAW;AAAA,EACpC;AACA,MAAI,WAAW,QAAQ;AACrB,mBAAe,SAAS,WAAW;AAAA,EACrC;AACA,MAAI,WAAW,WAAW;AACxB,mBAAe,YAAY,WAAW;AAAA,EACxC;AACA,MAAI,WAAW,eAAe;AAC5B,mBAAe,gBAAgB,WAAW;AAAA,EAC5C;AACA,MAAI,WAAW,gBAAgB;AAC7B,mBAAe,iBAAiB;AAAA,EAClC;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;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,WAAW,QAAQ,KAAK;AAAA,EACpC;AACA,MAAI,QAAQ,QAAQ;AAClB,SAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,EACtC;AACA,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,oBAAoB,QAAQ,aAAa;AAAA,EACrD;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AAEA,SAAO;AACT;;;ADjOA;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;AAC9G,UAAQ,OAAO,mBAAmB,wDAAwD;AAC1F,UAAQ,OAAO,oBAAoB,uEAAuE;AAC1G,UAAQ,OAAO,wBAAwB,+CAA+C;AACtF,UAAQ,OAAO,4BAA4B,gEAAgE;AAC3G,UAAQ,OAAO,qBAAqB,oEAAoE;AAC1G;;;AHvBA,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
CHANGED
|
@@ -57,6 +57,21 @@ function buildConvertOptions(cliOptions) {
|
|
|
57
57
|
if (cliOptions.imageFormat) {
|
|
58
58
|
convertOptions.imageFormat = cliOptions.imageFormat;
|
|
59
59
|
}
|
|
60
|
+
if (cliOptions.pages) {
|
|
61
|
+
convertOptions.pages = cliOptions.pages;
|
|
62
|
+
}
|
|
63
|
+
if (cliOptions.hybrid) {
|
|
64
|
+
convertOptions.hybrid = cliOptions.hybrid;
|
|
65
|
+
}
|
|
66
|
+
if (cliOptions.hybridUrl) {
|
|
67
|
+
convertOptions.hybridUrl = cliOptions.hybridUrl;
|
|
68
|
+
}
|
|
69
|
+
if (cliOptions.hybridTimeout) {
|
|
70
|
+
convertOptions.hybridTimeout = cliOptions.hybridTimeout;
|
|
71
|
+
}
|
|
72
|
+
if (cliOptions.hybridFallback) {
|
|
73
|
+
convertOptions.hybridFallback = true;
|
|
74
|
+
}
|
|
60
75
|
return convertOptions;
|
|
61
76
|
}
|
|
62
77
|
function buildArgs(options) {
|
|
@@ -118,6 +133,21 @@ function buildArgs(options) {
|
|
|
118
133
|
if (options.imageFormat) {
|
|
119
134
|
args.push("--image-format", options.imageFormat);
|
|
120
135
|
}
|
|
136
|
+
if (options.pages) {
|
|
137
|
+
args.push("--pages", options.pages);
|
|
138
|
+
}
|
|
139
|
+
if (options.hybrid) {
|
|
140
|
+
args.push("--hybrid", options.hybrid);
|
|
141
|
+
}
|
|
142
|
+
if (options.hybridUrl) {
|
|
143
|
+
args.push("--hybrid-url", options.hybridUrl);
|
|
144
|
+
}
|
|
145
|
+
if (options.hybridTimeout) {
|
|
146
|
+
args.push("--hybrid-timeout", options.hybridTimeout);
|
|
147
|
+
}
|
|
148
|
+
if (options.hybridFallback) {
|
|
149
|
+
args.push("--hybrid-fallback");
|
|
150
|
+
}
|
|
121
151
|
return args;
|
|
122
152
|
}
|
|
123
153
|
|
|
@@ -212,6 +242,11 @@ function registerCliOptions(program) {
|
|
|
212
242
|
program.option("--html-page-separator <value>", "Separator between pages in HTML output. Use %page-number% for page numbers. Default: none");
|
|
213
243
|
program.option("--image-output <value>", "Image output mode. Values: off (no images), embedded (Base64 data URIs), external (file references). Default: external");
|
|
214
244
|
program.option("--image-format <value>", "Output format for extracted images. Values: png, jpeg. Default: png");
|
|
245
|
+
program.option("--pages <value>", 'Pages to extract (e.g., "1,3,5-7"). Default: all pages');
|
|
246
|
+
program.option("--hybrid <value>", "Hybrid backend for AI processing. Values: off (default), docling-fast");
|
|
247
|
+
program.option("--hybrid-url <value>", "Hybrid backend server URL (overrides default)");
|
|
248
|
+
program.option("--hybrid-timeout <value>", "Hybrid backend request timeout in milliseconds. Default: 30000");
|
|
249
|
+
program.option("--hybrid-fallback", "Fallback to Java processing on hybrid backend error. Default: true");
|
|
215
250
|
}
|
|
216
251
|
|
|
217
252
|
// src/cli.ts
|
package/dist/cli.js.map
CHANGED
|
@@ -1 +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: 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":[]}
|
|
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: 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 /** Pages to extract (e.g., \"1,3,5-7\"). Default: all pages */\n pages?: string;\n /** Hybrid backend for AI processing. Values: off (default), docling-fast */\n hybrid?: string;\n /** Hybrid backend server URL (overrides default) */\n hybridUrl?: string;\n /** Hybrid backend request timeout in milliseconds. Default: 30000 */\n hybridTimeout?: string;\n /** Fallback to Java processing on hybrid backend error. Default: true */\n hybridFallback?: boolean;\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 pages?: string;\n hybrid?: string;\n hybridUrl?: string;\n hybridTimeout?: string;\n hybridFallback?: boolean;\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 if (cliOptions.pages) {\n convertOptions.pages = cliOptions.pages;\n }\n if (cliOptions.hybrid) {\n convertOptions.hybrid = cliOptions.hybrid;\n }\n if (cliOptions.hybridUrl) {\n convertOptions.hybridUrl = cliOptions.hybridUrl;\n }\n if (cliOptions.hybridTimeout) {\n convertOptions.hybridTimeout = cliOptions.hybridTimeout;\n }\n if (cliOptions.hybridFallback) {\n convertOptions.hybridFallback = true;\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 if (options.pages) {\n args.push('--pages', options.pages);\n }\n if (options.hybrid) {\n args.push('--hybrid', options.hybrid);\n }\n if (options.hybridUrl) {\n args.push('--hybrid-url', options.hybridUrl);\n }\n if (options.hybridTimeout) {\n args.push('--hybrid-timeout', options.hybridTimeout);\n }\n if (options.hybridFallback) {\n args.push('--hybrid-fallback');\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 program.option('--pages <value>', 'Pages to extract (e.g., \"1,3,5-7\"). Default: all pages');\n program.option('--hybrid <value>', 'Hybrid backend for AI processing. Values: off (default), docling-fast');\n program.option('--hybrid-url <value>', 'Hybrid backend server URL (overrides default)');\n program.option('--hybrid-timeout <value>', 'Hybrid backend request timeout in milliseconds. Default: 30000');\n program.option('--hybrid-fallback', 'Fallback to Java processing on hybrid backend error. Default: true');\n}\n"],"mappings":";;;AACA,SAAS,SAAS,sBAAsB;;;ACDxC,SAAS,aAAa;AACtB,YAAY,UAAU;AACtB,YAAY,QAAQ;AACpB,SAAS,qBAAqB;;;AC2EvB,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;AACA,MAAI,WAAW,OAAO;AACpB,mBAAe,QAAQ,WAAW;AAAA,EACpC;AACA,MAAI,WAAW,QAAQ;AACrB,mBAAe,SAAS,WAAW;AAAA,EACrC;AACA,MAAI,WAAW,WAAW;AACxB,mBAAe,YAAY,WAAW;AAAA,EACxC;AACA,MAAI,WAAW,eAAe;AAC5B,mBAAe,gBAAgB,WAAW;AAAA,EAC5C;AACA,MAAI,WAAW,gBAAgB;AAC7B,mBAAe,iBAAiB;AAAA,EAClC;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;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,WAAW,QAAQ,KAAK;AAAA,EACpC;AACA,MAAI,QAAQ,QAAQ;AAClB,SAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,EACtC;AACA,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,oBAAoB,QAAQ,aAAa;AAAA,EACrD;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AAEA,SAAO;AACT;;;ADtNA,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;AAC9G,UAAQ,OAAO,mBAAmB,wDAAwD;AAC1F,UAAQ,OAAO,oBAAoB,uEAAuE;AAC1G,UAAQ,OAAO,wBAAwB,+CAA+C;AACtF,UAAQ,OAAO,4BAA4B,gEAAgE;AAC3G,UAAQ,OAAO,qBAAqB,oEAAoE;AAC1G;;;AHvBA,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
CHANGED
|
@@ -100,6 +100,21 @@ function buildArgs(options) {
|
|
|
100
100
|
if (options.imageFormat) {
|
|
101
101
|
args.push("--image-format", options.imageFormat);
|
|
102
102
|
}
|
|
103
|
+
if (options.pages) {
|
|
104
|
+
args.push("--pages", options.pages);
|
|
105
|
+
}
|
|
106
|
+
if (options.hybrid) {
|
|
107
|
+
args.push("--hybrid", options.hybrid);
|
|
108
|
+
}
|
|
109
|
+
if (options.hybridUrl) {
|
|
110
|
+
args.push("--hybrid-url", options.hybridUrl);
|
|
111
|
+
}
|
|
112
|
+
if (options.hybridTimeout) {
|
|
113
|
+
args.push("--hybrid-timeout", options.hybridTimeout);
|
|
114
|
+
}
|
|
115
|
+
if (options.hybridFallback) {
|
|
116
|
+
args.push("--hybrid-fallback");
|
|
117
|
+
}
|
|
103
118
|
return args;
|
|
104
119
|
}
|
|
105
120
|
|
package/dist/index.cjs.map
CHANGED
|
@@ -1 +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: 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":[]}
|
|
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: 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 /** Pages to extract (e.g., \"1,3,5-7\"). Default: all pages */\n pages?: string;\n /** Hybrid backend for AI processing. Values: off (default), docling-fast */\n hybrid?: string;\n /** Hybrid backend server URL (overrides default) */\n hybridUrl?: string;\n /** Hybrid backend request timeout in milliseconds. Default: 30000 */\n hybridTimeout?: string;\n /** Fallback to Java processing on hybrid backend error. Default: true */\n hybridFallback?: boolean;\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 pages?: string;\n hybrid?: string;\n hybridUrl?: string;\n hybridTimeout?: string;\n hybridFallback?: boolean;\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 if (cliOptions.pages) {\n convertOptions.pages = cliOptions.pages;\n }\n if (cliOptions.hybrid) {\n convertOptions.hybrid = cliOptions.hybrid;\n }\n if (cliOptions.hybridUrl) {\n convertOptions.hybridUrl = cliOptions.hybridUrl;\n }\n if (cliOptions.hybridTimeout) {\n convertOptions.hybridTimeout = cliOptions.hybridTimeout;\n }\n if (cliOptions.hybridFallback) {\n convertOptions.hybridFallback = true;\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 if (options.pages) {\n args.push('--pages', options.pages);\n }\n if (options.hybrid) {\n args.push('--hybrid', options.hybrid);\n }\n if (options.hybridUrl) {\n args.push('--hybrid-url', options.hybridUrl);\n }\n if (options.hybridTimeout) {\n args.push('--hybrid-timeout', options.hybridTimeout);\n }\n if (options.hybridFallback) {\n args.push('--hybrid-fallback');\n }\n\n return args;\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,2BAAsB;AACtB,WAAsB;AACtB,SAAoB;AACpB,iBAA8B;;;ACiJvB,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;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,WAAW,QAAQ,KAAK;AAAA,EACpC;AACA,MAAI,QAAQ,QAAQ;AAClB,SAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,EACtC;AACA,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,oBAAoB,QAAQ,aAAa;AAAA,EACrD;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AAEA,SAAO;AACT;;;ADjOA;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
CHANGED
|
@@ -32,6 +32,16 @@ interface ConvertOptions {
|
|
|
32
32
|
imageOutput?: string;
|
|
33
33
|
/** Output format for extracted images. Values: png, jpeg. Default: png */
|
|
34
34
|
imageFormat?: string;
|
|
35
|
+
/** Pages to extract (e.g., "1,3,5-7"). Default: all pages */
|
|
36
|
+
pages?: string;
|
|
37
|
+
/** Hybrid backend for AI processing. Values: off (default), docling-fast */
|
|
38
|
+
hybrid?: string;
|
|
39
|
+
/** Hybrid backend server URL (overrides default) */
|
|
40
|
+
hybridUrl?: string;
|
|
41
|
+
/** Hybrid backend request timeout in milliseconds. Default: 30000 */
|
|
42
|
+
hybridTimeout?: string;
|
|
43
|
+
/** Fallback to Java processing on hybrid backend error. Default: true */
|
|
44
|
+
hybridFallback?: boolean;
|
|
35
45
|
}
|
|
36
46
|
/**
|
|
37
47
|
* Build CLI arguments array from ConvertOptions.
|
package/dist/index.d.ts
CHANGED
|
@@ -32,6 +32,16 @@ interface ConvertOptions {
|
|
|
32
32
|
imageOutput?: string;
|
|
33
33
|
/** Output format for extracted images. Values: png, jpeg. Default: png */
|
|
34
34
|
imageFormat?: string;
|
|
35
|
+
/** Pages to extract (e.g., "1,3,5-7"). Default: all pages */
|
|
36
|
+
pages?: string;
|
|
37
|
+
/** Hybrid backend for AI processing. Values: off (default), docling-fast */
|
|
38
|
+
hybrid?: string;
|
|
39
|
+
/** Hybrid backend server URL (overrides default) */
|
|
40
|
+
hybridUrl?: string;
|
|
41
|
+
/** Hybrid backend request timeout in milliseconds. Default: 30000 */
|
|
42
|
+
hybridTimeout?: string;
|
|
43
|
+
/** Fallback to Java processing on hybrid backend error. Default: true */
|
|
44
|
+
hybridFallback?: boolean;
|
|
35
45
|
}
|
|
36
46
|
/**
|
|
37
47
|
* Build CLI arguments array from ConvertOptions.
|
package/dist/index.js
CHANGED
|
@@ -64,6 +64,21 @@ function buildArgs(options) {
|
|
|
64
64
|
if (options.imageFormat) {
|
|
65
65
|
args.push("--image-format", options.imageFormat);
|
|
66
66
|
}
|
|
67
|
+
if (options.pages) {
|
|
68
|
+
args.push("--pages", options.pages);
|
|
69
|
+
}
|
|
70
|
+
if (options.hybrid) {
|
|
71
|
+
args.push("--hybrid", options.hybrid);
|
|
72
|
+
}
|
|
73
|
+
if (options.hybridUrl) {
|
|
74
|
+
args.push("--hybrid-url", options.hybridUrl);
|
|
75
|
+
}
|
|
76
|
+
if (options.hybridTimeout) {
|
|
77
|
+
args.push("--hybrid-timeout", options.hybridTimeout);
|
|
78
|
+
}
|
|
79
|
+
if (options.hybridFallback) {
|
|
80
|
+
args.push("--hybrid-fallback");
|
|
81
|
+
}
|
|
67
82
|
return args;
|
|
68
83
|
}
|
|
69
84
|
|
package/dist/index.js.map
CHANGED
|
@@ -1 +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: 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":[]}
|
|
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: 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 /** Pages to extract (e.g., \"1,3,5-7\"). Default: all pages */\n pages?: string;\n /** Hybrid backend for AI processing. Values: off (default), docling-fast */\n hybrid?: string;\n /** Hybrid backend server URL (overrides default) */\n hybridUrl?: string;\n /** Hybrid backend request timeout in milliseconds. Default: 30000 */\n hybridTimeout?: string;\n /** Fallback to Java processing on hybrid backend error. Default: true */\n hybridFallback?: boolean;\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 pages?: string;\n hybrid?: string;\n hybridUrl?: string;\n hybridTimeout?: string;\n hybridFallback?: boolean;\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 if (cliOptions.pages) {\n convertOptions.pages = cliOptions.pages;\n }\n if (cliOptions.hybrid) {\n convertOptions.hybrid = cliOptions.hybrid;\n }\n if (cliOptions.hybridUrl) {\n convertOptions.hybridUrl = cliOptions.hybridUrl;\n }\n if (cliOptions.hybridTimeout) {\n convertOptions.hybridTimeout = cliOptions.hybridTimeout;\n }\n if (cliOptions.hybridFallback) {\n convertOptions.hybridFallback = true;\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 if (options.pages) {\n args.push('--pages', options.pages);\n }\n if (options.hybrid) {\n args.push('--hybrid', options.hybrid);\n }\n if (options.hybridUrl) {\n args.push('--hybrid-url', options.hybridUrl);\n }\n if (options.hybridTimeout) {\n args.push('--hybrid-timeout', options.hybridTimeout);\n }\n if (options.hybridFallback) {\n args.push('--hybrid-fallback');\n }\n\n return args;\n}\n"],"mappings":";AAAA,SAAS,aAAa;AACtB,YAAY,UAAU;AACtB,YAAY,QAAQ;AACpB,SAAS,qBAAqB;;;ACiJvB,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;AACA,MAAI,QAAQ,OAAO;AACjB,SAAK,KAAK,WAAW,QAAQ,KAAK;AAAA,EACpC;AACA,MAAI,QAAQ,QAAQ;AAClB,SAAK,KAAK,YAAY,QAAQ,MAAM;AAAA,EACtC;AACA,MAAI,QAAQ,WAAW;AACrB,SAAK,KAAK,gBAAgB,QAAQ,SAAS;AAAA,EAC7C;AACA,MAAI,QAAQ,eAAe;AACzB,SAAK,KAAK,oBAAoB,QAAQ,aAAa;AAAA,EACrD;AACA,MAAI,QAAQ,gBAAgB;AAC1B,SAAK,KAAK,mBAAmB;AAAA,EAC/B;AAEA,SAAO;AACT;;;ADtNA,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":[]}
|
|
Binary file
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@opendataloader/pdf",
|
|
3
|
-
"version": "1.
|
|
3
|
+
"version": "1.6.0",
|
|
4
4
|
"description": "A Node.js wrapper for the opendataloader-pdf Java CLI.",
|
|
5
5
|
"main": "./dist/index.cjs",
|
|
6
6
|
"module": "./dist/index.js",
|
|
@@ -45,6 +45,9 @@
|
|
|
45
45
|
},
|
|
46
46
|
"devDependencies": {
|
|
47
47
|
"@types/node": "^24.10.1",
|
|
48
|
+
"@typescript-eslint/eslint-plugin": "^8.33.0",
|
|
49
|
+
"@typescript-eslint/parser": "^8.33.0",
|
|
50
|
+
"eslint": "^9.28.0",
|
|
48
51
|
"glob": "^12.0.0",
|
|
49
52
|
"prettier": "^3.6.2",
|
|
50
53
|
"tsup": "^8.5.1",
|
|
@@ -64,6 +67,8 @@
|
|
|
64
67
|
"setup": "node ./scripts/setup.cjs",
|
|
65
68
|
"build": "pnpm run setup && tsup",
|
|
66
69
|
"test": "vitest --run",
|
|
67
|
-
"format": "prettier --write \"**/*.{ts,js,json,md}\""
|
|
70
|
+
"format": "prettier --write \"**/*.{ts,js,json,md}\"",
|
|
71
|
+
"lint": "eslint \"src/**/*.ts\"",
|
|
72
|
+
"lint:fix": "eslint \"src/**/*.ts\" --fix"
|
|
68
73
|
}
|
|
69
74
|
}
|