@opendataloader/pdf 1.0.4 → 1.0.6
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 +25 -0
- package/dist/cli.cjs +276 -0
- package/dist/cli.cjs.map +1 -0
- package/dist/cli.d.cts +1 -0
- package/dist/cli.d.ts +1 -0
- package/dist/cli.js +252 -0
- package/dist/cli.js.map +1 -0
- package/lib/opendataloader-pdf-cli.jar +0 -0
- package/package.json +4 -1
package/README.md
CHANGED
|
@@ -87,9 +87,16 @@ opendataloader_pdf.run(
|
|
|
87
87
|
generate_markdown=True,
|
|
88
88
|
generate_html=True,
|
|
89
89
|
generate_annotated_pdf=True,
|
|
90
|
+
debug=True,
|
|
90
91
|
)
|
|
91
92
|
```
|
|
92
93
|
|
|
94
|
+
- If you want to run it via CLI, you can use the following command:
|
|
95
|
+
|
|
96
|
+
```sh
|
|
97
|
+
opendataloader-pdf path/to/document.pdf --markdown --html --pdf
|
|
98
|
+
```
|
|
99
|
+
|
|
93
100
|
### Function: run()
|
|
94
101
|
|
|
95
102
|
The main function to process PDFs.
|
|
@@ -152,6 +159,24 @@ async function main() {
|
|
|
152
159
|
main();
|
|
153
160
|
```
|
|
154
161
|
|
|
162
|
+
If you want to run it via CLI, you can use the following command:
|
|
163
|
+
|
|
164
|
+
```bash
|
|
165
|
+
npx @opendataloader/pdf path/to/document.pdf -o path/to/output --markdown --html --pdf
|
|
166
|
+
```
|
|
167
|
+
|
|
168
|
+
or you can install it globally:
|
|
169
|
+
|
|
170
|
+
```bash
|
|
171
|
+
npm install -g @opendataloader/pdf
|
|
172
|
+
```
|
|
173
|
+
|
|
174
|
+
then run:
|
|
175
|
+
|
|
176
|
+
```bash
|
|
177
|
+
opendataloader-pdf path/to/document.pdf -o path/to/output --markdown --html --pdf
|
|
178
|
+
```
|
|
179
|
+
|
|
155
180
|
### Function: run()
|
|
156
181
|
|
|
157
182
|
`run(inputPath: string, options?: RunOptions): Promise<string>`
|
package/dist/cli.cjs
ADDED
|
@@ -0,0 +1,276 @@
|
|
|
1
|
+
#!/usr/bin/env node
|
|
2
|
+
"use strict";
|
|
3
|
+
var __create = Object.create;
|
|
4
|
+
var __defProp = Object.defineProperty;
|
|
5
|
+
var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
|
|
6
|
+
var __getOwnPropNames = Object.getOwnPropertyNames;
|
|
7
|
+
var __getProtoOf = Object.getPrototypeOf;
|
|
8
|
+
var __hasOwnProp = Object.prototype.hasOwnProperty;
|
|
9
|
+
var __copyProps = (to, from, except, desc) => {
|
|
10
|
+
if (from && typeof from === "object" || typeof from === "function") {
|
|
11
|
+
for (let key of __getOwnPropNames(from))
|
|
12
|
+
if (!__hasOwnProp.call(to, key) && key !== except)
|
|
13
|
+
__defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
|
|
14
|
+
}
|
|
15
|
+
return to;
|
|
16
|
+
};
|
|
17
|
+
var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
|
|
18
|
+
// If the importer is in node compatibility mode or this is not an ESM
|
|
19
|
+
// file that has been converted to a CommonJS file using a Babel-
|
|
20
|
+
// compatible transform (i.e. "__esModule" has not been set), then set
|
|
21
|
+
// "default" to the CommonJS "module.exports" for node compatibility.
|
|
22
|
+
isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
|
|
23
|
+
mod
|
|
24
|
+
));
|
|
25
|
+
|
|
26
|
+
// src/index.ts
|
|
27
|
+
var import_child_process = require("child_process");
|
|
28
|
+
var path = __toESM(require("path"), 1);
|
|
29
|
+
var fs = __toESM(require("fs"), 1);
|
|
30
|
+
var import_url = require("url");
|
|
31
|
+
var import_meta = {};
|
|
32
|
+
var __filename = (0, import_url.fileURLToPath)(import_meta.url);
|
|
33
|
+
var __dirname = path.dirname(__filename);
|
|
34
|
+
var JAR_NAME = "opendataloader-pdf-cli.jar";
|
|
35
|
+
function getRedactedCommandString(command, commandArgs) {
|
|
36
|
+
const commandArgsForLogging = [...commandArgs];
|
|
37
|
+
const passwordIndex = commandArgsForLogging.indexOf("--password");
|
|
38
|
+
if (passwordIndex > -1 && passwordIndex + 1 < commandArgsForLogging.length) {
|
|
39
|
+
commandArgsForLogging[passwordIndex + 1] = "[REDACTED]";
|
|
40
|
+
}
|
|
41
|
+
return `${command} ${commandArgsForLogging.join(" ")}`;
|
|
42
|
+
}
|
|
43
|
+
function run(inputPath, options = {}) {
|
|
44
|
+
return new Promise((resolve, reject) => {
|
|
45
|
+
if (!fs.existsSync(inputPath)) {
|
|
46
|
+
return reject(new Error(`Input file or folder not found: ${inputPath}`));
|
|
47
|
+
}
|
|
48
|
+
const args = [];
|
|
49
|
+
if (options.outputFolder) {
|
|
50
|
+
args.push("--output-dir", options.outputFolder);
|
|
51
|
+
}
|
|
52
|
+
if (options.password) {
|
|
53
|
+
args.push("--password", options.password);
|
|
54
|
+
}
|
|
55
|
+
if (options.replaceInvalidChars) {
|
|
56
|
+
args.push("--replace-invalid-chars", options.replaceInvalidChars);
|
|
57
|
+
}
|
|
58
|
+
if (options.generateMarkdown) {
|
|
59
|
+
args.push("--markdown");
|
|
60
|
+
}
|
|
61
|
+
if (options.generateHtml) {
|
|
62
|
+
args.push("--html");
|
|
63
|
+
}
|
|
64
|
+
if (options.generateAnnotatedPdf) {
|
|
65
|
+
args.push("--pdf");
|
|
66
|
+
}
|
|
67
|
+
if (options.keepLineBreaks) {
|
|
68
|
+
args.push("--keep-line-breaks");
|
|
69
|
+
}
|
|
70
|
+
if (options.contentSafetyOff) {
|
|
71
|
+
args.push("--content-safety-off", options.contentSafetyOff);
|
|
72
|
+
}
|
|
73
|
+
if (options.htmlInMarkdown) {
|
|
74
|
+
args.push("--markdown-with-html");
|
|
75
|
+
}
|
|
76
|
+
if (options.addImageToMarkdown) {
|
|
77
|
+
args.push("--markdown-with-images");
|
|
78
|
+
}
|
|
79
|
+
if (options.noJson) {
|
|
80
|
+
args.push("--no-json");
|
|
81
|
+
}
|
|
82
|
+
args.push(inputPath);
|
|
83
|
+
const jarPath = path.join(__dirname, "..", "lib", JAR_NAME);
|
|
84
|
+
if (!fs.existsSync(jarPath)) {
|
|
85
|
+
return reject(
|
|
86
|
+
new Error(`JAR file not found at ${jarPath}. Please run the build script first.`)
|
|
87
|
+
);
|
|
88
|
+
}
|
|
89
|
+
const command = "java";
|
|
90
|
+
const commandArgs = ["-jar", jarPath, ...args];
|
|
91
|
+
if (options.debug) {
|
|
92
|
+
console.error(`Running command: ${getRedactedCommandString(command, commandArgs)}`);
|
|
93
|
+
}
|
|
94
|
+
const javaProcess = (0, import_child_process.spawn)(command, commandArgs);
|
|
95
|
+
let stdout = "";
|
|
96
|
+
let stderr = "";
|
|
97
|
+
javaProcess.stdout.on("data", (data) => {
|
|
98
|
+
const chunk = data.toString();
|
|
99
|
+
if (options.debug) {
|
|
100
|
+
process.stdout.write(chunk);
|
|
101
|
+
}
|
|
102
|
+
stdout += chunk;
|
|
103
|
+
});
|
|
104
|
+
javaProcess.stderr.on("data", (data) => {
|
|
105
|
+
const chunk = data.toString();
|
|
106
|
+
if (options.debug) {
|
|
107
|
+
process.stderr.write(chunk);
|
|
108
|
+
}
|
|
109
|
+
stderr += chunk;
|
|
110
|
+
});
|
|
111
|
+
javaProcess.on("close", (code) => {
|
|
112
|
+
if (code === 0) {
|
|
113
|
+
resolve(stdout);
|
|
114
|
+
} else {
|
|
115
|
+
const error = new Error(
|
|
116
|
+
`The opendataloader-pdf CLI exited with code ${code}.
|
|
117
|
+
|
|
118
|
+
${stderr}`
|
|
119
|
+
);
|
|
120
|
+
reject(error);
|
|
121
|
+
}
|
|
122
|
+
});
|
|
123
|
+
javaProcess.on("error", (err) => {
|
|
124
|
+
if (err.message.includes("ENOENT")) {
|
|
125
|
+
reject(
|
|
126
|
+
new Error(
|
|
127
|
+
"'java' command not found. Please ensure Java is installed and in your system's PATH."
|
|
128
|
+
)
|
|
129
|
+
);
|
|
130
|
+
} else {
|
|
131
|
+
reject(err);
|
|
132
|
+
}
|
|
133
|
+
});
|
|
134
|
+
});
|
|
135
|
+
}
|
|
136
|
+
|
|
137
|
+
// src/cli.ts
|
|
138
|
+
function printHelp() {
|
|
139
|
+
console.log(`Usage: opendataloader-pdf [options] <input>`);
|
|
140
|
+
console.log("");
|
|
141
|
+
console.log("Options:");
|
|
142
|
+
console.log(" -o, --output-dir <path> Directory where outputs are written");
|
|
143
|
+
console.log(" -p, --password <password> Password for encrypted PDFs");
|
|
144
|
+
console.log(" --replace-invalid-chars <c> Replacement character for invalid characters");
|
|
145
|
+
console.log(" --content-safety-off <mode> Disable content safety filtering (provide mode)");
|
|
146
|
+
console.log(" --markdown Generate Markdown output");
|
|
147
|
+
console.log(" --html Generate HTML output");
|
|
148
|
+
console.log(" --pdf Generate annotated PDF output");
|
|
149
|
+
console.log(" --keep-line-breaks Preserve line breaks in text output");
|
|
150
|
+
console.log(" --markdown-with-html Allow raw HTML within Markdown output");
|
|
151
|
+
console.log(" --markdown-with-images Embed images in Markdown output");
|
|
152
|
+
console.log(" --no-json Disable JSON output generation");
|
|
153
|
+
console.log(" --debug Stream CLI logs directly to stdout/stderr");
|
|
154
|
+
console.log(" -h, --help Show this message and exit");
|
|
155
|
+
}
|
|
156
|
+
function parseArgs(argv) {
|
|
157
|
+
const options = {};
|
|
158
|
+
let inputPath;
|
|
159
|
+
let showHelp = false;
|
|
160
|
+
const readValue = (currentIndex, option) => {
|
|
161
|
+
const nextValue = argv[currentIndex + 1];
|
|
162
|
+
if (!nextValue || nextValue.startsWith("-")) {
|
|
163
|
+
throw new Error(`Option ${option} requires a value.`);
|
|
164
|
+
}
|
|
165
|
+
return { value: nextValue, nextIndex: currentIndex + 1 };
|
|
166
|
+
};
|
|
167
|
+
for (let i = 0; i < argv.length; i += 1) {
|
|
168
|
+
const arg = argv[i];
|
|
169
|
+
switch (arg) {
|
|
170
|
+
case "--help":
|
|
171
|
+
case "-h":
|
|
172
|
+
showHelp = true;
|
|
173
|
+
i = argv.length;
|
|
174
|
+
break;
|
|
175
|
+
case "--output-dir":
|
|
176
|
+
case "-o": {
|
|
177
|
+
const { value, nextIndex } = readValue(i, arg);
|
|
178
|
+
options.outputFolder = value;
|
|
179
|
+
i = nextIndex;
|
|
180
|
+
break;
|
|
181
|
+
}
|
|
182
|
+
case "--password":
|
|
183
|
+
case "-p": {
|
|
184
|
+
const { value, nextIndex } = readValue(i, arg);
|
|
185
|
+
options.password = value;
|
|
186
|
+
i = nextIndex;
|
|
187
|
+
break;
|
|
188
|
+
}
|
|
189
|
+
case "--replace-invalid-chars": {
|
|
190
|
+
const { value, nextIndex } = readValue(i, arg);
|
|
191
|
+
options.replaceInvalidChars = value;
|
|
192
|
+
i = nextIndex;
|
|
193
|
+
break;
|
|
194
|
+
}
|
|
195
|
+
case "--content-safety-off": {
|
|
196
|
+
const { value, nextIndex } = readValue(i, arg);
|
|
197
|
+
options.contentSafetyOff = value;
|
|
198
|
+
i = nextIndex;
|
|
199
|
+
break;
|
|
200
|
+
}
|
|
201
|
+
case "--markdown":
|
|
202
|
+
options.generateMarkdown = true;
|
|
203
|
+
break;
|
|
204
|
+
case "--html":
|
|
205
|
+
options.generateHtml = true;
|
|
206
|
+
break;
|
|
207
|
+
case "--pdf":
|
|
208
|
+
options.generateAnnotatedPdf = true;
|
|
209
|
+
break;
|
|
210
|
+
case "--keep-line-breaks":
|
|
211
|
+
options.keepLineBreaks = true;
|
|
212
|
+
break;
|
|
213
|
+
case "--markdown-with-html":
|
|
214
|
+
options.htmlInMarkdown = true;
|
|
215
|
+
break;
|
|
216
|
+
case "--markdown-with-images":
|
|
217
|
+
options.addImageToMarkdown = true;
|
|
218
|
+
break;
|
|
219
|
+
case "--no-json":
|
|
220
|
+
options.noJson = true;
|
|
221
|
+
break;
|
|
222
|
+
case "--debug":
|
|
223
|
+
options.debug = true;
|
|
224
|
+
break;
|
|
225
|
+
default:
|
|
226
|
+
if (arg.startsWith("-")) {
|
|
227
|
+
throw new Error(`Unknown option: ${arg}`);
|
|
228
|
+
}
|
|
229
|
+
if (inputPath) {
|
|
230
|
+
throw new Error("Multiple input paths provided. Only one input path is allowed.");
|
|
231
|
+
}
|
|
232
|
+
inputPath = arg;
|
|
233
|
+
}
|
|
234
|
+
}
|
|
235
|
+
return { inputPath, options, showHelp };
|
|
236
|
+
}
|
|
237
|
+
async function main() {
|
|
238
|
+
let parsed;
|
|
239
|
+
try {
|
|
240
|
+
parsed = parseArgs(process.argv.slice(2));
|
|
241
|
+
} catch (err) {
|
|
242
|
+
const message = err instanceof Error ? err.message : String(err);
|
|
243
|
+
console.error(message);
|
|
244
|
+
console.error("Use '--help' to see available options.");
|
|
245
|
+
return 1;
|
|
246
|
+
}
|
|
247
|
+
if (parsed.showHelp) {
|
|
248
|
+
printHelp();
|
|
249
|
+
return 0;
|
|
250
|
+
}
|
|
251
|
+
if (!parsed.inputPath) {
|
|
252
|
+
console.error("Missing required input path.");
|
|
253
|
+
console.error("Use '--help' to see usage information.");
|
|
254
|
+
return 1;
|
|
255
|
+
}
|
|
256
|
+
try {
|
|
257
|
+
const output = await run(parsed.inputPath, parsed.options);
|
|
258
|
+
if (output && !parsed.options.debug) {
|
|
259
|
+
process.stdout.write(output);
|
|
260
|
+
}
|
|
261
|
+
if (output && !output.endsWith("\n") && !parsed.options.debug) {
|
|
262
|
+
process.stdout.write("\n");
|
|
263
|
+
}
|
|
264
|
+
return 0;
|
|
265
|
+
} catch (err) {
|
|
266
|
+
const message = err instanceof Error ? err.message : String(err);
|
|
267
|
+
console.error(message);
|
|
268
|
+
return 1;
|
|
269
|
+
}
|
|
270
|
+
}
|
|
271
|
+
main().then((code) => {
|
|
272
|
+
if (code !== 0) {
|
|
273
|
+
process.exit(code);
|
|
274
|
+
}
|
|
275
|
+
});
|
|
276
|
+
//# sourceMappingURL=cli.cjs.map
|
package/dist/cli.cjs.map
ADDED
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"sources":["../src/index.ts","../src/cli.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\nfunction getRedactedCommandString(command: string, commandArgs: string[]): string {\n const commandArgsForLogging = [...commandArgs];\n const passwordIndex = commandArgsForLogging.indexOf('--password');\n if (passwordIndex > -1 && passwordIndex + 1 < commandArgsForLogging.length) {\n commandArgsForLogging[passwordIndex + 1] = '[REDACTED]';\n }\n return `${command} ${commandArgsForLogging.join(' ')}`;\n}\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}\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.contentSafetyOff) {\n args.push('--content-safety-off', options.contentSafetyOff);\n }\n if (options.htmlInMarkdown) {\n args.push('--markdown-with-html');\n }\n if (options.addImageToMarkdown) {\n args.push('--markdown-with-images');\n }\n if (options.noJson) {\n args.push('--no-json');\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: ${getRedactedCommandString(command, commandArgs)}`);\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","#!/usr/bin/env node\nimport { run, RunOptions } from './index.js';\n\ninterface ParsedArgs {\n inputPath?: string;\n options: RunOptions;\n showHelp: boolean;\n}\n\nfunction printHelp(): void {\n console.log(`Usage: opendataloader-pdf [options] <input>`);\n console.log('');\n console.log('Options:');\n console.log(' -o, --output-dir <path> Directory where outputs are written');\n console.log(' -p, --password <password> Password for encrypted PDFs');\n console.log(' --replace-invalid-chars <c> Replacement character for invalid characters');\n console.log(' --content-safety-off <mode> Disable content safety filtering (provide mode)');\n console.log(' --markdown Generate Markdown output');\n console.log(' --html Generate HTML output');\n console.log(' --pdf Generate annotated PDF output');\n console.log(' --keep-line-breaks Preserve line breaks in text output');\n console.log(' --markdown-with-html Allow raw HTML within Markdown output');\n console.log(' --markdown-with-images Embed images in Markdown output');\n console.log(' --no-json Disable JSON output generation');\n console.log(' --debug Stream CLI logs directly to stdout/stderr');\n console.log(' -h, --help Show this message and exit');\n}\n\nfunction parseArgs(argv: string[]): ParsedArgs {\n const options: RunOptions = {};\n let inputPath: string | undefined;\n let showHelp = false;\n\n const readValue = (currentIndex: number, option: string): { value: string; nextIndex: number } => {\n const nextValue = argv[currentIndex + 1];\n if (!nextValue || nextValue.startsWith('-')) {\n throw new Error(`Option ${option} requires a value.`);\n }\n return { value: nextValue, nextIndex: currentIndex + 1 };\n };\n\n for (let i = 0; i < argv.length; i += 1) {\n const arg = argv[i];\n\n switch (arg) {\n case '--help':\n case '-h':\n showHelp = true;\n i = argv.length; // exit loop\n break;\n case '--output-dir':\n case '-o': {\n const { value, nextIndex } = readValue(i, arg);\n options.outputFolder = value;\n i = nextIndex;\n break;\n }\n case '--password':\n case '-p': {\n const { value, nextIndex } = readValue(i, arg);\n options.password = value;\n i = nextIndex;\n break;\n }\n case '--replace-invalid-chars': {\n const { value, nextIndex } = readValue(i, arg);\n options.replaceInvalidChars = value;\n i = nextIndex;\n break;\n }\n case '--content-safety-off': {\n const { value, nextIndex } = readValue(i, arg);\n options.contentSafetyOff = value;\n i = nextIndex;\n break;\n }\n case '--markdown':\n options.generateMarkdown = true;\n break;\n case '--html':\n options.generateHtml = true;\n break;\n case '--pdf':\n options.generateAnnotatedPdf = true;\n break;\n case '--keep-line-breaks':\n options.keepLineBreaks = true;\n break;\n case '--markdown-with-html':\n options.htmlInMarkdown = true;\n break;\n case '--markdown-with-images':\n options.addImageToMarkdown = true;\n break;\n case '--no-json':\n options.noJson = true;\n break;\n case '--debug':\n options.debug = true;\n break;\n default:\n if (arg.startsWith('-')) {\n throw new Error(`Unknown option: ${arg}`);\n }\n\n if (inputPath) {\n throw new Error('Multiple input paths provided. Only one input path is allowed.');\n }\n inputPath = arg;\n }\n }\n\n return { inputPath, options, showHelp };\n}\n\nasync function main(): Promise<number> {\n let parsed: ParsedArgs;\n\n try {\n parsed = parseArgs(process.argv.slice(2));\n } catch (err) {\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 if (parsed.showHelp) {\n printHelp();\n return 0;\n }\n\n if (!parsed.inputPath) {\n console.error('Missing required input path.');\n console.error(\"Use '--help' to see usage information.\");\n return 1;\n }\n\n try {\n const output = await run(parsed.inputPath, parsed.options);\n if (output && !parsed.options.debug) {\n process.stdout.write(output);\n }\n if (output && !output.endsWith('\\n') && !parsed.options.debug) {\n process.stdout.write('\\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"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;AAAA,2BAAsB;AACtB,WAAsB;AACtB,SAAoB;AACpB,iBAA8B;AAH9B;AAKA,IAAM,iBAAa,0BAAc,YAAY,GAAG;AAChD,IAAM,YAAiB,aAAQ,UAAU;AAEzC,IAAM,WAAW;AAEjB,SAAS,yBAAyB,SAAiB,aAA+B;AAChF,QAAM,wBAAwB,CAAC,GAAG,WAAW;AAC7C,QAAM,gBAAgB,sBAAsB,QAAQ,YAAY;AAChE,MAAI,gBAAgB,MAAM,gBAAgB,IAAI,sBAAsB,QAAQ;AAC1E,0BAAsB,gBAAgB,CAAC,IAAI;AAAA,EAC7C;AACA,SAAO,GAAG,OAAO,IAAI,sBAAsB,KAAK,GAAG,CAAC;AACtD;AAiBO,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,kBAAkB;AAC5B,WAAK,KAAK,wBAAwB,QAAQ,gBAAgB;AAAA,IAC5D;AACA,QAAI,QAAQ,gBAAgB;AAC1B,WAAK,KAAK,sBAAsB;AAAA,IAClC;AACA,QAAI,QAAQ,oBAAoB;AAC9B,WAAK,KAAK,wBAAwB;AAAA,IACpC;AACA,QAAI,QAAQ,QAAQ;AAClB,WAAK,KAAK,WAAW;AAAA,IACvB;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,yBAAyB,SAAS,WAAW,CAAC,EAAE;AAAA,IACpF;AAEA,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,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;;;AC/HA,SAAS,YAAkB;AACzB,UAAQ,IAAI,6CAA6C;AACzD,UAAQ,IAAI,EAAE;AACd,UAAQ,IAAI,UAAU;AACtB,UAAQ,IAAI,oEAAoE;AAChF,UAAQ,IAAI,4DAA4D;AACxE,UAAQ,IAAI,6EAA6E;AACzF,UAAQ,IAAI,gFAAgF;AAC5F,UAAQ,IAAI,yDAAyD;AACrE,UAAQ,IAAI,qDAAqD;AACjE,UAAQ,IAAI,8DAA8D;AAC1E,UAAQ,IAAI,oEAAoE;AAChF,UAAQ,IAAI,sEAAsE;AAClF,UAAQ,IAAI,gEAAgE;AAC5E,UAAQ,IAAI,+DAA+D;AAC3E,UAAQ,IAAI,0EAA0E;AACtF,UAAQ,IAAI,2DAA2D;AACzE;AAEA,SAAS,UAAU,MAA4B;AAC7C,QAAM,UAAsB,CAAC;AAC7B,MAAI;AACJ,MAAI,WAAW;AAEf,QAAM,YAAY,CAAC,cAAsB,WAAyD;AAChG,UAAM,YAAY,KAAK,eAAe,CAAC;AACvC,QAAI,CAAC,aAAa,UAAU,WAAW,GAAG,GAAG;AAC3C,YAAM,IAAI,MAAM,UAAU,MAAM,oBAAoB;AAAA,IACtD;AACA,WAAO,EAAE,OAAO,WAAW,WAAW,eAAe,EAAE;AAAA,EACzD;AAEA,WAAS,IAAI,GAAG,IAAI,KAAK,QAAQ,KAAK,GAAG;AACvC,UAAM,MAAM,KAAK,CAAC;AAElB,YAAQ,KAAK;AAAA,MACX,KAAK;AAAA,MACL,KAAK;AACH,mBAAW;AACX,YAAI,KAAK;AACT;AAAA,MACF,KAAK;AAAA,MACL,KAAK,MAAM;AACT,cAAM,EAAE,OAAO,UAAU,IAAI,UAAU,GAAG,GAAG;AAC7C,gBAAQ,eAAe;AACvB,YAAI;AACJ;AAAA,MACF;AAAA,MACA,KAAK;AAAA,MACL,KAAK,MAAM;AACT,cAAM,EAAE,OAAO,UAAU,IAAI,UAAU,GAAG,GAAG;AAC7C,gBAAQ,WAAW;AACnB,YAAI;AACJ;AAAA,MACF;AAAA,MACA,KAAK,2BAA2B;AAC9B,cAAM,EAAE,OAAO,UAAU,IAAI,UAAU,GAAG,GAAG;AAC7C,gBAAQ,sBAAsB;AAC9B,YAAI;AACJ;AAAA,MACF;AAAA,MACA,KAAK,wBAAwB;AAC3B,cAAM,EAAE,OAAO,UAAU,IAAI,UAAU,GAAG,GAAG;AAC7C,gBAAQ,mBAAmB;AAC3B,YAAI;AACJ;AAAA,MACF;AAAA,MACA,KAAK;AACH,gBAAQ,mBAAmB;AAC3B;AAAA,MACF,KAAK;AACH,gBAAQ,eAAe;AACvB;AAAA,MACF,KAAK;AACH,gBAAQ,uBAAuB;AAC/B;AAAA,MACF,KAAK;AACH,gBAAQ,iBAAiB;AACzB;AAAA,MACF,KAAK;AACH,gBAAQ,iBAAiB;AACzB;AAAA,MACF,KAAK;AACH,gBAAQ,qBAAqB;AAC7B;AAAA,MACF,KAAK;AACH,gBAAQ,SAAS;AACjB;AAAA,MACF,KAAK;AACH,gBAAQ,QAAQ;AAChB;AAAA,MACF;AACE,YAAI,IAAI,WAAW,GAAG,GAAG;AACvB,gBAAM,IAAI,MAAM,mBAAmB,GAAG,EAAE;AAAA,QAC1C;AAEA,YAAI,WAAW;AACb,gBAAM,IAAI,MAAM,gEAAgE;AAAA,QAClF;AACA,oBAAY;AAAA,IAChB;AAAA,EACF;AAEA,SAAO,EAAE,WAAW,SAAS,SAAS;AACxC;AAEA,eAAe,OAAwB;AACrC,MAAI;AAEJ,MAAI;AACF,aAAS,UAAU,QAAQ,KAAK,MAAM,CAAC,CAAC;AAAA,EAC1C,SAAS,KAAK;AACZ,UAAM,UAAU,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC/D,YAAQ,MAAM,OAAO;AACrB,YAAQ,MAAM,wCAAwC;AACtD,WAAO;AAAA,EACT;AAEA,MAAI,OAAO,UAAU;AACnB,cAAU;AACV,WAAO;AAAA,EACT;AAEA,MAAI,CAAC,OAAO,WAAW;AACrB,YAAQ,MAAM,8BAA8B;AAC5C,YAAQ,MAAM,wCAAwC;AACtD,WAAO;AAAA,EACT;AAEA,MAAI;AACF,UAAM,SAAS,MAAM,IAAI,OAAO,WAAW,OAAO,OAAO;AACzD,QAAI,UAAU,CAAC,OAAO,QAAQ,OAAO;AACnC,cAAQ,OAAO,MAAM,MAAM;AAAA,IAC7B;AACA,QAAI,UAAU,CAAC,OAAO,SAAS,IAAI,KAAK,CAAC,OAAO,QAAQ,OAAO;AAC7D,cAAQ,OAAO,MAAM,IAAI;AAAA,IAC3B;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.d.cts
ADDED
|
@@ -0,0 +1 @@
|
|
|
1
|
+
#!/usr/bin/env node
|
package/dist/cli.d.ts
ADDED
|
@@ -0,0 +1 @@
|
|
|
1
|
+
#!/usr/bin/env node
|
package/dist/cli.js
ADDED
|
@@ -0,0 +1,252 @@
|
|
|
1
|
+
#!/usr/bin/env node
|
|
2
|
+
|
|
3
|
+
// src/index.ts
|
|
4
|
+
import { spawn } from "child_process";
|
|
5
|
+
import * as path from "path";
|
|
6
|
+
import * as fs from "fs";
|
|
7
|
+
import { fileURLToPath } from "url";
|
|
8
|
+
var __filename = fileURLToPath(import.meta.url);
|
|
9
|
+
var __dirname = path.dirname(__filename);
|
|
10
|
+
var JAR_NAME = "opendataloader-pdf-cli.jar";
|
|
11
|
+
function getRedactedCommandString(command, commandArgs) {
|
|
12
|
+
const commandArgsForLogging = [...commandArgs];
|
|
13
|
+
const passwordIndex = commandArgsForLogging.indexOf("--password");
|
|
14
|
+
if (passwordIndex > -1 && passwordIndex + 1 < commandArgsForLogging.length) {
|
|
15
|
+
commandArgsForLogging[passwordIndex + 1] = "[REDACTED]";
|
|
16
|
+
}
|
|
17
|
+
return `${command} ${commandArgsForLogging.join(" ")}`;
|
|
18
|
+
}
|
|
19
|
+
function run(inputPath, options = {}) {
|
|
20
|
+
return new Promise((resolve, reject) => {
|
|
21
|
+
if (!fs.existsSync(inputPath)) {
|
|
22
|
+
return reject(new Error(`Input file or folder not found: ${inputPath}`));
|
|
23
|
+
}
|
|
24
|
+
const args = [];
|
|
25
|
+
if (options.outputFolder) {
|
|
26
|
+
args.push("--output-dir", options.outputFolder);
|
|
27
|
+
}
|
|
28
|
+
if (options.password) {
|
|
29
|
+
args.push("--password", options.password);
|
|
30
|
+
}
|
|
31
|
+
if (options.replaceInvalidChars) {
|
|
32
|
+
args.push("--replace-invalid-chars", options.replaceInvalidChars);
|
|
33
|
+
}
|
|
34
|
+
if (options.generateMarkdown) {
|
|
35
|
+
args.push("--markdown");
|
|
36
|
+
}
|
|
37
|
+
if (options.generateHtml) {
|
|
38
|
+
args.push("--html");
|
|
39
|
+
}
|
|
40
|
+
if (options.generateAnnotatedPdf) {
|
|
41
|
+
args.push("--pdf");
|
|
42
|
+
}
|
|
43
|
+
if (options.keepLineBreaks) {
|
|
44
|
+
args.push("--keep-line-breaks");
|
|
45
|
+
}
|
|
46
|
+
if (options.contentSafetyOff) {
|
|
47
|
+
args.push("--content-safety-off", options.contentSafetyOff);
|
|
48
|
+
}
|
|
49
|
+
if (options.htmlInMarkdown) {
|
|
50
|
+
args.push("--markdown-with-html");
|
|
51
|
+
}
|
|
52
|
+
if (options.addImageToMarkdown) {
|
|
53
|
+
args.push("--markdown-with-images");
|
|
54
|
+
}
|
|
55
|
+
if (options.noJson) {
|
|
56
|
+
args.push("--no-json");
|
|
57
|
+
}
|
|
58
|
+
args.push(inputPath);
|
|
59
|
+
const jarPath = path.join(__dirname, "..", "lib", JAR_NAME);
|
|
60
|
+
if (!fs.existsSync(jarPath)) {
|
|
61
|
+
return reject(
|
|
62
|
+
new Error(`JAR file not found at ${jarPath}. Please run the build script first.`)
|
|
63
|
+
);
|
|
64
|
+
}
|
|
65
|
+
const command = "java";
|
|
66
|
+
const commandArgs = ["-jar", jarPath, ...args];
|
|
67
|
+
if (options.debug) {
|
|
68
|
+
console.error(`Running command: ${getRedactedCommandString(command, commandArgs)}`);
|
|
69
|
+
}
|
|
70
|
+
const javaProcess = spawn(command, commandArgs);
|
|
71
|
+
let stdout = "";
|
|
72
|
+
let stderr = "";
|
|
73
|
+
javaProcess.stdout.on("data", (data) => {
|
|
74
|
+
const chunk = data.toString();
|
|
75
|
+
if (options.debug) {
|
|
76
|
+
process.stdout.write(chunk);
|
|
77
|
+
}
|
|
78
|
+
stdout += chunk;
|
|
79
|
+
});
|
|
80
|
+
javaProcess.stderr.on("data", (data) => {
|
|
81
|
+
const chunk = data.toString();
|
|
82
|
+
if (options.debug) {
|
|
83
|
+
process.stderr.write(chunk);
|
|
84
|
+
}
|
|
85
|
+
stderr += chunk;
|
|
86
|
+
});
|
|
87
|
+
javaProcess.on("close", (code) => {
|
|
88
|
+
if (code === 0) {
|
|
89
|
+
resolve(stdout);
|
|
90
|
+
} else {
|
|
91
|
+
const error = new Error(
|
|
92
|
+
`The opendataloader-pdf CLI exited with code ${code}.
|
|
93
|
+
|
|
94
|
+
${stderr}`
|
|
95
|
+
);
|
|
96
|
+
reject(error);
|
|
97
|
+
}
|
|
98
|
+
});
|
|
99
|
+
javaProcess.on("error", (err) => {
|
|
100
|
+
if (err.message.includes("ENOENT")) {
|
|
101
|
+
reject(
|
|
102
|
+
new Error(
|
|
103
|
+
"'java' command not found. Please ensure Java is installed and in your system's PATH."
|
|
104
|
+
)
|
|
105
|
+
);
|
|
106
|
+
} else {
|
|
107
|
+
reject(err);
|
|
108
|
+
}
|
|
109
|
+
});
|
|
110
|
+
});
|
|
111
|
+
}
|
|
112
|
+
|
|
113
|
+
// src/cli.ts
|
|
114
|
+
function printHelp() {
|
|
115
|
+
console.log(`Usage: opendataloader-pdf [options] <input>`);
|
|
116
|
+
console.log("");
|
|
117
|
+
console.log("Options:");
|
|
118
|
+
console.log(" -o, --output-dir <path> Directory where outputs are written");
|
|
119
|
+
console.log(" -p, --password <password> Password for encrypted PDFs");
|
|
120
|
+
console.log(" --replace-invalid-chars <c> Replacement character for invalid characters");
|
|
121
|
+
console.log(" --content-safety-off <mode> Disable content safety filtering (provide mode)");
|
|
122
|
+
console.log(" --markdown Generate Markdown output");
|
|
123
|
+
console.log(" --html Generate HTML output");
|
|
124
|
+
console.log(" --pdf Generate annotated PDF output");
|
|
125
|
+
console.log(" --keep-line-breaks Preserve line breaks in text output");
|
|
126
|
+
console.log(" --markdown-with-html Allow raw HTML within Markdown output");
|
|
127
|
+
console.log(" --markdown-with-images Embed images in Markdown output");
|
|
128
|
+
console.log(" --no-json Disable JSON output generation");
|
|
129
|
+
console.log(" --debug Stream CLI logs directly to stdout/stderr");
|
|
130
|
+
console.log(" -h, --help Show this message and exit");
|
|
131
|
+
}
|
|
132
|
+
function parseArgs(argv) {
|
|
133
|
+
const options = {};
|
|
134
|
+
let inputPath;
|
|
135
|
+
let showHelp = false;
|
|
136
|
+
const readValue = (currentIndex, option) => {
|
|
137
|
+
const nextValue = argv[currentIndex + 1];
|
|
138
|
+
if (!nextValue || nextValue.startsWith("-")) {
|
|
139
|
+
throw new Error(`Option ${option} requires a value.`);
|
|
140
|
+
}
|
|
141
|
+
return { value: nextValue, nextIndex: currentIndex + 1 };
|
|
142
|
+
};
|
|
143
|
+
for (let i = 0; i < argv.length; i += 1) {
|
|
144
|
+
const arg = argv[i];
|
|
145
|
+
switch (arg) {
|
|
146
|
+
case "--help":
|
|
147
|
+
case "-h":
|
|
148
|
+
showHelp = true;
|
|
149
|
+
i = argv.length;
|
|
150
|
+
break;
|
|
151
|
+
case "--output-dir":
|
|
152
|
+
case "-o": {
|
|
153
|
+
const { value, nextIndex } = readValue(i, arg);
|
|
154
|
+
options.outputFolder = value;
|
|
155
|
+
i = nextIndex;
|
|
156
|
+
break;
|
|
157
|
+
}
|
|
158
|
+
case "--password":
|
|
159
|
+
case "-p": {
|
|
160
|
+
const { value, nextIndex } = readValue(i, arg);
|
|
161
|
+
options.password = value;
|
|
162
|
+
i = nextIndex;
|
|
163
|
+
break;
|
|
164
|
+
}
|
|
165
|
+
case "--replace-invalid-chars": {
|
|
166
|
+
const { value, nextIndex } = readValue(i, arg);
|
|
167
|
+
options.replaceInvalidChars = value;
|
|
168
|
+
i = nextIndex;
|
|
169
|
+
break;
|
|
170
|
+
}
|
|
171
|
+
case "--content-safety-off": {
|
|
172
|
+
const { value, nextIndex } = readValue(i, arg);
|
|
173
|
+
options.contentSafetyOff = value;
|
|
174
|
+
i = nextIndex;
|
|
175
|
+
break;
|
|
176
|
+
}
|
|
177
|
+
case "--markdown":
|
|
178
|
+
options.generateMarkdown = true;
|
|
179
|
+
break;
|
|
180
|
+
case "--html":
|
|
181
|
+
options.generateHtml = true;
|
|
182
|
+
break;
|
|
183
|
+
case "--pdf":
|
|
184
|
+
options.generateAnnotatedPdf = true;
|
|
185
|
+
break;
|
|
186
|
+
case "--keep-line-breaks":
|
|
187
|
+
options.keepLineBreaks = true;
|
|
188
|
+
break;
|
|
189
|
+
case "--markdown-with-html":
|
|
190
|
+
options.htmlInMarkdown = true;
|
|
191
|
+
break;
|
|
192
|
+
case "--markdown-with-images":
|
|
193
|
+
options.addImageToMarkdown = true;
|
|
194
|
+
break;
|
|
195
|
+
case "--no-json":
|
|
196
|
+
options.noJson = true;
|
|
197
|
+
break;
|
|
198
|
+
case "--debug":
|
|
199
|
+
options.debug = true;
|
|
200
|
+
break;
|
|
201
|
+
default:
|
|
202
|
+
if (arg.startsWith("-")) {
|
|
203
|
+
throw new Error(`Unknown option: ${arg}`);
|
|
204
|
+
}
|
|
205
|
+
if (inputPath) {
|
|
206
|
+
throw new Error("Multiple input paths provided. Only one input path is allowed.");
|
|
207
|
+
}
|
|
208
|
+
inputPath = arg;
|
|
209
|
+
}
|
|
210
|
+
}
|
|
211
|
+
return { inputPath, options, showHelp };
|
|
212
|
+
}
|
|
213
|
+
async function main() {
|
|
214
|
+
let parsed;
|
|
215
|
+
try {
|
|
216
|
+
parsed = parseArgs(process.argv.slice(2));
|
|
217
|
+
} catch (err) {
|
|
218
|
+
const message = err instanceof Error ? err.message : String(err);
|
|
219
|
+
console.error(message);
|
|
220
|
+
console.error("Use '--help' to see available options.");
|
|
221
|
+
return 1;
|
|
222
|
+
}
|
|
223
|
+
if (parsed.showHelp) {
|
|
224
|
+
printHelp();
|
|
225
|
+
return 0;
|
|
226
|
+
}
|
|
227
|
+
if (!parsed.inputPath) {
|
|
228
|
+
console.error("Missing required input path.");
|
|
229
|
+
console.error("Use '--help' to see usage information.");
|
|
230
|
+
return 1;
|
|
231
|
+
}
|
|
232
|
+
try {
|
|
233
|
+
const output = await run(parsed.inputPath, parsed.options);
|
|
234
|
+
if (output && !parsed.options.debug) {
|
|
235
|
+
process.stdout.write(output);
|
|
236
|
+
}
|
|
237
|
+
if (output && !output.endsWith("\n") && !parsed.options.debug) {
|
|
238
|
+
process.stdout.write("\n");
|
|
239
|
+
}
|
|
240
|
+
return 0;
|
|
241
|
+
} catch (err) {
|
|
242
|
+
const message = err instanceof Error ? err.message : String(err);
|
|
243
|
+
console.error(message);
|
|
244
|
+
return 1;
|
|
245
|
+
}
|
|
246
|
+
}
|
|
247
|
+
main().then((code) => {
|
|
248
|
+
if (code !== 0) {
|
|
249
|
+
process.exit(code);
|
|
250
|
+
}
|
|
251
|
+
});
|
|
252
|
+
//# sourceMappingURL=cli.js.map
|
package/dist/cli.js.map
ADDED
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"sources":["../src/index.ts","../src/cli.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\nfunction getRedactedCommandString(command: string, commandArgs: string[]): string {\n const commandArgsForLogging = [...commandArgs];\n const passwordIndex = commandArgsForLogging.indexOf('--password');\n if (passwordIndex > -1 && passwordIndex + 1 < commandArgsForLogging.length) {\n commandArgsForLogging[passwordIndex + 1] = '[REDACTED]';\n }\n return `${command} ${commandArgsForLogging.join(' ')}`;\n}\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}\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.contentSafetyOff) {\n args.push('--content-safety-off', options.contentSafetyOff);\n }\n if (options.htmlInMarkdown) {\n args.push('--markdown-with-html');\n }\n if (options.addImageToMarkdown) {\n args.push('--markdown-with-images');\n }\n if (options.noJson) {\n args.push('--no-json');\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: ${getRedactedCommandString(command, commandArgs)}`);\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","#!/usr/bin/env node\nimport { run, RunOptions } from './index.js';\n\ninterface ParsedArgs {\n inputPath?: string;\n options: RunOptions;\n showHelp: boolean;\n}\n\nfunction printHelp(): void {\n console.log(`Usage: opendataloader-pdf [options] <input>`);\n console.log('');\n console.log('Options:');\n console.log(' -o, --output-dir <path> Directory where outputs are written');\n console.log(' -p, --password <password> Password for encrypted PDFs');\n console.log(' --replace-invalid-chars <c> Replacement character for invalid characters');\n console.log(' --content-safety-off <mode> Disable content safety filtering (provide mode)');\n console.log(' --markdown Generate Markdown output');\n console.log(' --html Generate HTML output');\n console.log(' --pdf Generate annotated PDF output');\n console.log(' --keep-line-breaks Preserve line breaks in text output');\n console.log(' --markdown-with-html Allow raw HTML within Markdown output');\n console.log(' --markdown-with-images Embed images in Markdown output');\n console.log(' --no-json Disable JSON output generation');\n console.log(' --debug Stream CLI logs directly to stdout/stderr');\n console.log(' -h, --help Show this message and exit');\n}\n\nfunction parseArgs(argv: string[]): ParsedArgs {\n const options: RunOptions = {};\n let inputPath: string | undefined;\n let showHelp = false;\n\n const readValue = (currentIndex: number, option: string): { value: string; nextIndex: number } => {\n const nextValue = argv[currentIndex + 1];\n if (!nextValue || nextValue.startsWith('-')) {\n throw new Error(`Option ${option} requires a value.`);\n }\n return { value: nextValue, nextIndex: currentIndex + 1 };\n };\n\n for (let i = 0; i < argv.length; i += 1) {\n const arg = argv[i];\n\n switch (arg) {\n case '--help':\n case '-h':\n showHelp = true;\n i = argv.length; // exit loop\n break;\n case '--output-dir':\n case '-o': {\n const { value, nextIndex } = readValue(i, arg);\n options.outputFolder = value;\n i = nextIndex;\n break;\n }\n case '--password':\n case '-p': {\n const { value, nextIndex } = readValue(i, arg);\n options.password = value;\n i = nextIndex;\n break;\n }\n case '--replace-invalid-chars': {\n const { value, nextIndex } = readValue(i, arg);\n options.replaceInvalidChars = value;\n i = nextIndex;\n break;\n }\n case '--content-safety-off': {\n const { value, nextIndex } = readValue(i, arg);\n options.contentSafetyOff = value;\n i = nextIndex;\n break;\n }\n case '--markdown':\n options.generateMarkdown = true;\n break;\n case '--html':\n options.generateHtml = true;\n break;\n case '--pdf':\n options.generateAnnotatedPdf = true;\n break;\n case '--keep-line-breaks':\n options.keepLineBreaks = true;\n break;\n case '--markdown-with-html':\n options.htmlInMarkdown = true;\n break;\n case '--markdown-with-images':\n options.addImageToMarkdown = true;\n break;\n case '--no-json':\n options.noJson = true;\n break;\n case '--debug':\n options.debug = true;\n break;\n default:\n if (arg.startsWith('-')) {\n throw new Error(`Unknown option: ${arg}`);\n }\n\n if (inputPath) {\n throw new Error('Multiple input paths provided. Only one input path is allowed.');\n }\n inputPath = arg;\n }\n }\n\n return { inputPath, options, showHelp };\n}\n\nasync function main(): Promise<number> {\n let parsed: ParsedArgs;\n\n try {\n parsed = parseArgs(process.argv.slice(2));\n } catch (err) {\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 if (parsed.showHelp) {\n printHelp();\n return 0;\n }\n\n if (!parsed.inputPath) {\n console.error('Missing required input path.');\n console.error(\"Use '--help' to see usage information.\");\n return 1;\n }\n\n try {\n const output = await run(parsed.inputPath, parsed.options);\n if (output && !parsed.options.debug) {\n process.stdout.write(output);\n }\n if (output && !output.endsWith('\\n') && !parsed.options.debug) {\n process.stdout.write('\\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"],"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;AAEjB,SAAS,yBAAyB,SAAiB,aAA+B;AAChF,QAAM,wBAAwB,CAAC,GAAG,WAAW;AAC7C,QAAM,gBAAgB,sBAAsB,QAAQ,YAAY;AAChE,MAAI,gBAAgB,MAAM,gBAAgB,IAAI,sBAAsB,QAAQ;AAC1E,0BAAsB,gBAAgB,CAAC,IAAI;AAAA,EAC7C;AACA,SAAO,GAAG,OAAO,IAAI,sBAAsB,KAAK,GAAG,CAAC;AACtD;AAiBO,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,kBAAkB;AAC5B,WAAK,KAAK,wBAAwB,QAAQ,gBAAgB;AAAA,IAC5D;AACA,QAAI,QAAQ,gBAAgB;AAC1B,WAAK,KAAK,sBAAsB;AAAA,IAClC;AACA,QAAI,QAAQ,oBAAoB;AAC9B,WAAK,KAAK,wBAAwB;AAAA,IACpC;AACA,QAAI,QAAQ,QAAQ;AAClB,WAAK,KAAK,WAAW;AAAA,IACvB;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,yBAAyB,SAAS,WAAW,CAAC,EAAE;AAAA,IACpF;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;;;AC/HA,SAAS,YAAkB;AACzB,UAAQ,IAAI,6CAA6C;AACzD,UAAQ,IAAI,EAAE;AACd,UAAQ,IAAI,UAAU;AACtB,UAAQ,IAAI,oEAAoE;AAChF,UAAQ,IAAI,4DAA4D;AACxE,UAAQ,IAAI,6EAA6E;AACzF,UAAQ,IAAI,gFAAgF;AAC5F,UAAQ,IAAI,yDAAyD;AACrE,UAAQ,IAAI,qDAAqD;AACjE,UAAQ,IAAI,8DAA8D;AAC1E,UAAQ,IAAI,oEAAoE;AAChF,UAAQ,IAAI,sEAAsE;AAClF,UAAQ,IAAI,gEAAgE;AAC5E,UAAQ,IAAI,+DAA+D;AAC3E,UAAQ,IAAI,0EAA0E;AACtF,UAAQ,IAAI,2DAA2D;AACzE;AAEA,SAAS,UAAU,MAA4B;AAC7C,QAAM,UAAsB,CAAC;AAC7B,MAAI;AACJ,MAAI,WAAW;AAEf,QAAM,YAAY,CAAC,cAAsB,WAAyD;AAChG,UAAM,YAAY,KAAK,eAAe,CAAC;AACvC,QAAI,CAAC,aAAa,UAAU,WAAW,GAAG,GAAG;AAC3C,YAAM,IAAI,MAAM,UAAU,MAAM,oBAAoB;AAAA,IACtD;AACA,WAAO,EAAE,OAAO,WAAW,WAAW,eAAe,EAAE;AAAA,EACzD;AAEA,WAAS,IAAI,GAAG,IAAI,KAAK,QAAQ,KAAK,GAAG;AACvC,UAAM,MAAM,KAAK,CAAC;AAElB,YAAQ,KAAK;AAAA,MACX,KAAK;AAAA,MACL,KAAK;AACH,mBAAW;AACX,YAAI,KAAK;AACT;AAAA,MACF,KAAK;AAAA,MACL,KAAK,MAAM;AACT,cAAM,EAAE,OAAO,UAAU,IAAI,UAAU,GAAG,GAAG;AAC7C,gBAAQ,eAAe;AACvB,YAAI;AACJ;AAAA,MACF;AAAA,MACA,KAAK;AAAA,MACL,KAAK,MAAM;AACT,cAAM,EAAE,OAAO,UAAU,IAAI,UAAU,GAAG,GAAG;AAC7C,gBAAQ,WAAW;AACnB,YAAI;AACJ;AAAA,MACF;AAAA,MACA,KAAK,2BAA2B;AAC9B,cAAM,EAAE,OAAO,UAAU,IAAI,UAAU,GAAG,GAAG;AAC7C,gBAAQ,sBAAsB;AAC9B,YAAI;AACJ;AAAA,MACF;AAAA,MACA,KAAK,wBAAwB;AAC3B,cAAM,EAAE,OAAO,UAAU,IAAI,UAAU,GAAG,GAAG;AAC7C,gBAAQ,mBAAmB;AAC3B,YAAI;AACJ;AAAA,MACF;AAAA,MACA,KAAK;AACH,gBAAQ,mBAAmB;AAC3B;AAAA,MACF,KAAK;AACH,gBAAQ,eAAe;AACvB;AAAA,MACF,KAAK;AACH,gBAAQ,uBAAuB;AAC/B;AAAA,MACF,KAAK;AACH,gBAAQ,iBAAiB;AACzB;AAAA,MACF,KAAK;AACH,gBAAQ,iBAAiB;AACzB;AAAA,MACF,KAAK;AACH,gBAAQ,qBAAqB;AAC7B;AAAA,MACF,KAAK;AACH,gBAAQ,SAAS;AACjB;AAAA,MACF,KAAK;AACH,gBAAQ,QAAQ;AAChB;AAAA,MACF;AACE,YAAI,IAAI,WAAW,GAAG,GAAG;AACvB,gBAAM,IAAI,MAAM,mBAAmB,GAAG,EAAE;AAAA,QAC1C;AAEA,YAAI,WAAW;AACb,gBAAM,IAAI,MAAM,gEAAgE;AAAA,QAClF;AACA,oBAAY;AAAA,IAChB;AAAA,EACF;AAEA,SAAO,EAAE,WAAW,SAAS,SAAS;AACxC;AAEA,eAAe,OAAwB;AACrC,MAAI;AAEJ,MAAI;AACF,aAAS,UAAU,QAAQ,KAAK,MAAM,CAAC,CAAC;AAAA,EAC1C,SAAS,KAAK;AACZ,UAAM,UAAU,eAAe,QAAQ,IAAI,UAAU,OAAO,GAAG;AAC/D,YAAQ,MAAM,OAAO;AACrB,YAAQ,MAAM,wCAAwC;AACtD,WAAO;AAAA,EACT;AAEA,MAAI,OAAO,UAAU;AACnB,cAAU;AACV,WAAO;AAAA,EACT;AAEA,MAAI,CAAC,OAAO,WAAW;AACrB,YAAQ,MAAM,8BAA8B;AAC5C,YAAQ,MAAM,wCAAwC;AACtD,WAAO;AAAA,EACT;AAEA,MAAI;AACF,UAAM,SAAS,MAAM,IAAI,OAAO,WAAW,OAAO,OAAO;AACzD,QAAI,UAAU,CAAC,OAAO,QAAQ,OAAO;AACnC,cAAQ,OAAO,MAAM,MAAM;AAAA,IAC7B;AACA,QAAI,UAAU,CAAC,OAAO,SAAS,IAAI,KAAK,CAAC,OAAO,QAAQ,OAAO;AAC7D,cAAQ,OAAO,MAAM,IAAI;AAAA,IAC3B;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":[]}
|
|
Binary file
|
package/package.json
CHANGED
|
@@ -1,11 +1,14 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@opendataloader/pdf",
|
|
3
|
-
"version": "1.0.
|
|
3
|
+
"version": "1.0.6",
|
|
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",
|
|
7
7
|
"types": "./dist/index.d.ts",
|
|
8
8
|
"type": "module",
|
|
9
|
+
"bin": {
|
|
10
|
+
"opendataloader-pdf": "./dist/cli.js"
|
|
11
|
+
},
|
|
9
12
|
"exports": {
|
|
10
13
|
".": {
|
|
11
14
|
"import": "./dist/index.js",
|