| // Extracts each PDF's text page by page, using macOS's built-in PDFKit. | |
| // usage: osascript -l JavaScript rag/extract.js out.json a.pdf b.pdf ... | |
| ObjC.import('Foundation'); | |
| ObjC.import('PDFKit'); | |
| function run(argv) { | |
| const out = argv[0]; | |
| const pages = []; | |
| for (const path of argv.slice(1)) { | |
| const doc = $.PDFDocument.alloc.initWithURL($.NSURL.fileURLWithPath(path)); | |
| if (!doc.js) throw new Error('Could not open ' + path); | |
| const book = path.split('/').pop(); | |
| const n = doc.pageCount; | |
| for (let i = 0; i < n; i++) { | |
| const s = doc.pageAtIndex(i).string; | |
| pages.push({ book, page: i + 1, label: ObjC.unwrap(doc.pageAtIndex(i).label) || '', text: s.js ? ObjC.unwrap(s) : '' }); | |
| } | |
| } | |
| $.NSString.alloc.initWithUTF8String(JSON.stringify(pages)).writeToFileAtomicallyEncodingError(out, true, $.NSUTF8StringEncoding, null); | |
| return pages.length + ' pages'; | |
| } | |