@opendataloader/pdf 0.0.0

package/dist/index.js

@@ -0,0 +1,102 @@
+// src/index.ts
+import { spawn } from "child_process";
+import * as path from "path";
+import * as fs from "fs";
+import { fileURLToPath } from "url";
+var __filename = fileURLToPath(import.meta.url);
+var __dirname = path.dirname(__filename);
+var JAR_NAME = "opendataloader-pdf-cli.jar";
+function run(inputPath, options = {}) {
+  return new Promise((resolve, reject) => {
+    if (!fs.existsSync(inputPath)) {
+      return reject(new Error(`Input file or folder not found: ${inputPath}`));
+    }
+    const args = [];
+    if (options.outputFolder) {
+      args.push("--output-dir", options.outputFolder);
+    }
+    if (options.password) {
+      args.push("--password", options.password);
+    }
+    if (options.replaceInvalidChars) {
+      args.push("--replace-invalid-chars", options.replaceInvalidChars);
+    }
+    if (options.generateMarkdown) {
+      args.push("--markdown");
+    }
+    if (options.generateHtml) {
+      args.push("--html");
+    }
+    if (options.generateAnnotatedPdf) {
+      args.push("--pdf");
+    }
+    if (options.keepLineBreaks) {
+      args.push("--keep-line-breaks");
+    }
+    if (options.findHiddenText) {
+      args.push("--findhiddentext");
+    }
+    if (options.htmlInMarkdown) {
+      args.push("--markdown-with-html");
+    }
+    if (options.addImageToMarkdown) {
+      args.push("--markdown-with-images");
+    }
+    args.push(inputPath);
+    const jarPath = path.join(__dirname, "..", "lib", JAR_NAME);
+    if (!fs.existsSync(jarPath)) {
+      return reject(
+        new Error(`JAR file not found at ${jarPath}. Please run the build script first.`)
+      );
+    }
+    const command = "java";
+    const commandArgs = ["-jar", jarPath, ...args];
+    if (options.debug) {
+      console.error(`Running command: ${command} ${commandArgs.join(" ")}`);
+    }
+    const javaProcess = spawn(command, commandArgs);
+    let stdout = "";
+    let stderr = "";
+    javaProcess.stdout.on("data", (data) => {
+      const chunk = data.toString();
+      if (options.debug) {
+        process.stdout.write(chunk);
+      }
+      stdout += chunk;
+    });
+    javaProcess.stderr.on("data", (data) => {
+      const chunk = data.toString();
+      if (options.debug) {
+        process.stderr.write(chunk);
+      }
+      stderr += chunk;
+    });
+    javaProcess.on("close", (code) => {
+      if (code === 0) {
+        resolve(stdout);
+      } else {
+        const error = new Error(
+          `The opendataloader-pdf CLI exited with code ${code}.
+
+${stderr}`
+        );
+        reject(error);
+      }
+    });
+    javaProcess.on("error", (err) => {
+      if (err.message.includes("ENOENT")) {
+        reject(
+          new Error(
+            "'java' command not found. Please ensure Java is installed and in your system's PATH."
+          )
+        );
+      } else {
+        reject(err);
+      }
+    });
+  });
+}
+export {
+  run
+};
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/index.ts"],"sourcesContent":["import { spawn } from 'child_process';\nimport * as path from 'path';\nimport * as fs from 'fs';\nimport { fileURLToPath } from 'url';\n\nconst __filename = fileURLToPath(import.meta.url);\nconst __dirname = path.dirname(__filename);\n\nconst JAR_NAME = 'opendataloader-pdf-cli.jar';\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  findHiddenText?: boolean;\n  htmlInMarkdown?: boolean;\n  addImageToMarkdown?: boolean;\n  debug?: boolean;\n}\n\nexport function run(inputPath: string, options: RunOptions = {}): Promise<string> {\n  return new Promise((resolve, reject) => {\n    if (!fs.existsSync(inputPath)) {\n      return reject(new Error(`Input file or folder not found: ${inputPath}`));\n    }\n\n    const args: string[] = [];\n    if (options.outputFolder) {\n      args.push('--output-dir', options.outputFolder);\n    }\n    if (options.password) {\n      args.push('--password', options.password);\n    }\n    if (options.replaceInvalidChars) {\n      args.push('--replace-invalid-chars', options.replaceInvalidChars);\n    }\n    if (options.generateMarkdown) {\n      args.push('--markdown');\n    }\n    if (options.generateHtml) {\n      args.push('--html');\n    }\n    if (options.generateAnnotatedPdf) {\n      args.push('--pdf');\n    }\n    if (options.keepLineBreaks) {\n      args.push('--keep-line-breaks');\n    }\n    if (options.findHiddenText) {\n      args.push('--findhiddentext');\n    }\n    if (options.htmlInMarkdown) {\n      args.push('--markdown-with-html');\n    }\n    if (options.addImageToMarkdown) {\n      args.push('--markdown-with-images');\n    }\n\n    args.push(inputPath);\n\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    if (options.debug) {\n      console.error(`Running command: ${command} ${commandArgs.join(' ')}`);\n    }\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 (options.debug) {\n        process.stdout.write(chunk);\n      }\n      stdout += chunk;\n    });\n\n    javaProcess.stderr.on('data', (data) => {\n      const chunk = data.toString();\n      if (options.debug) {\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 error = new Error(\n          `The opendataloader-pdf CLI exited with code ${code}.\\n\\n${stderr}`,\n        );\n        reject(error);\n      }\n    });\n\n    javaProcess.on('error', (err) => {\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"],"mappings":";AAAA,SAAS,aAAa;AACtB,YAAY,UAAU;AACtB,YAAY,QAAQ;AACpB,SAAS,qBAAqB;AAE9B,IAAM,aAAa,cAAc,YAAY,GAAG;AAChD,IAAM,YAAiB,aAAQ,UAAU;AAEzC,IAAM,WAAW;AAgBV,SAAS,IAAI,WAAmB,UAAsB,CAAC,GAAoB;AAChF,SAAO,IAAI,QAAQ,CAAC,SAAS,WAAW;AACtC,QAAI,CAAI,cAAW,SAAS,GAAG;AAC7B,aAAO,OAAO,IAAI,MAAM,mCAAmC,SAAS,EAAE,CAAC;AAAA,IACzE;AAEA,UAAM,OAAiB,CAAC;AACxB,QAAI,QAAQ,cAAc;AACxB,WAAK,KAAK,gBAAgB,QAAQ,YAAY;AAAA,IAChD;AACA,QAAI,QAAQ,UAAU;AACpB,WAAK,KAAK,cAAc,QAAQ,QAAQ;AAAA,IAC1C;AACA,QAAI,QAAQ,qBAAqB;AAC/B,WAAK,KAAK,2BAA2B,QAAQ,mBAAmB;AAAA,IAClE;AACA,QAAI,QAAQ,kBAAkB;AAC5B,WAAK,KAAK,YAAY;AAAA,IACxB;AACA,QAAI,QAAQ,cAAc;AACxB,WAAK,KAAK,QAAQ;AAAA,IACpB;AACA,QAAI,QAAQ,sBAAsB;AAChC,WAAK,KAAK,OAAO;AAAA,IACnB;AACA,QAAI,QAAQ,gBAAgB;AAC1B,WAAK,KAAK,oBAAoB;AAAA,IAChC;AACA,QAAI,QAAQ,gBAAgB;AAC1B,WAAK,KAAK,kBAAkB;AAAA,IAC9B;AACA,QAAI,QAAQ,gBAAgB;AAC1B,WAAK,KAAK,sBAAsB;AAAA,IAClC;AACA,QAAI,QAAQ,oBAAoB;AAC9B,WAAK,KAAK,wBAAwB;AAAA,IACpC;AAEA,SAAK,KAAK,SAAS;AAEnB,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,QAAI,QAAQ,OAAO;AACjB,cAAQ,MAAM,oBAAoB,OAAO,IAAI,YAAY,KAAK,GAAG,CAAC,EAAE;AAAA,IACtE;AAEA,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,QAAQ,OAAO;AACjB,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,QAAQ,OAAO;AACjB,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,QAAQ,IAAI;AAAA,UAChB,+CAA+C,IAAI;AAAA;AAAA,EAAQ,MAAM;AAAA,QACnE;AACA,eAAO,KAAK;AAAA,MACd;AAAA,IACF,CAAC;AAED,gBAAY,GAAG,SAAS,CAAC,QAAQ;AAC/B,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;","names":[]}

package/package.json

@@ -0,0 +1,63 @@
+{
+  "name": "@opendataloader/pdf",
+  "version": "0.0.0",
+  "description": "A Node.js wrapper for the opendataloader-pdf Java CLI.",
+  "main": "./dist/index.cjs",
+  "module": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "type": "module",
+  "exports": {
+    ".": {
+      "import": "./dist/index.js",
+      "require": "./dist/index.cjs"
+    }
+  },
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/opendataloader-project/opendataloader-pdf.git"
+  },
+  "keywords": [
+    "pdf",
+    "markdown",
+    "html",
+    "convert",
+    "pdf-convert",
+    "pdf-parser",
+    "pdf-parsing",
+    "pdf-to-json",
+    "pdf-to-markdown",
+    "pdf-to-html"
+  ],
+  "author": "opendataloader-project",
+  "license": "MPL-2.0",
+  "bugs": {
+    "url": "https://github.com/opendataloader-project/opendataloader-pdf/issues"
+  },
+  "homepage": "https://github.com/opendataloader-project/opendataloader-pdf#readme",
+  "publishConfig": {
+    "access": "public"
+  },
+  "devDependencies": {
+    "@types/glob": "^8.1.0",
+    "@types/node": "^24.3.3",
+    "glob": "^11.0.3",
+    "prettier": "^3.6.2",
+    "tsup": "^8.5.0",
+    "typescript": "^5.9.2",
+    "vitest": "^3.2.4"
+  },
+  "files": [
+    "dist",
+    "lib",
+    "LICENSE",
+    "NOTICE.md",
+    "README.md",
+    "THIRD_PARTY"
+  ],
+  "scripts": {
+    "setup": "node ./scripts/setup.cjs",
+    "build": "pnpm run setup && tsup",
+    "test": "vitest",
+    "format": "prettier --write \"**/*.{ts,js,json,md}\""
+  }
+}