A PDF text parser for Scriptable App (iOS).
iOS shortcuts PDF conversion is not always the good tool. Due to the "intelligent" column recognition of PDFKit, text output is sometime in total disorder. I needed a parser which always output text in a predictable order, and this tool does that using the pdfjs text extraction.
First install dependencies:
npm install
To create a build:
npm run build
Copy the file dist/scriptable-pdfjs.html
into Scriptable Document Folder.
(or you may simply download the latest version from here)
and use the gist scriptable-pdfjs-demo to play.
In this example we also expect to find scriptable-pdfjs.html` in Scriptable documents folder
- Pass a PDF using the sharesheet or an Open File action
- Save a copy of that file in the Shortcut Folder to allow the next bookmark creation action
- Add a Scriptable/Create bookmark action
- Add a Scriptable/Run Inline action containing :
let fm;
try {
fm = FileManager.iCloud();
} catch(e) {
fm = FileManager.local();
}
// use the same bookmark name as in the action above
const filePath = fm.bookmarkedPath("ShortcutPDF");
// We execute pdfjs in a WebView
const wv = new WebView();
const htmlFileUrl = fm.joinPath(fm.documentsDirectory(), "scriptable-pdfjs.html");
await fm.downloadFileFromiCloud(htmlFileUrl);
await wv.loadFile(htmlFileUrl);
let javascript = 'pdfjs.getText(';
javascript += '"' + fm.read(filePath).toBase64String() + '"';
//javascript += ', (pageText) => pageText.includes("(Long copy #1)")';
//javascript += ', true';
javascript += ');'
let result = "";
try {
result = await wv.evaluateJavaScript(javascript, true);
} catch (e) {
result = "";
}
return result;
For some reasons... (bug in shortcuts or scriptable ?) You cannot convert the PDF to Base64 and pass it as an argument to the script. You have to use the bookmark trick and make the base64 conversion in Scriptable.
For file larger than 2.5 Mo, you can not run this script inline and you have to modify the script to get results by using the clipboard.