[Share] Remove the first page of a PDF file

[raffaelemancuso]

Description

Remove the first page from a PDF file

Event

None

Operation

Script

Data

/**
 * Remove the first page of a PDF attachment (in place, via pdfcpu)
 * @author Raffaele Mancuso
 * @usage Select a PDF attachment (or a regular item with exactly one PDF attachment) and trigger the action
 * @link https://github.com/windingwind/zotero-actions-tags/discussions/623
 */

// EDIT THESE SETTINGS

/** @type {string} Full path to the pdfcpu executable. On Windows escape backslashes ("\\"). */
const pdfcpuPath = "C:\\data\\software\\pdfcpu.exe";

/** @type {boolean} If true, keep a copy of the original as "<name>.pdf.bak" next to it. */
const keepBackup = false;

// END OF EDITABLE SETTINGS

// The action runs once per selected item, plus once with `items` and no `item`:
// skip that batch pass silently instead of reporting an error
if (!item) return;

// Resolve a regular item to its PDF attachment: require exactly one, so the action
// never silently picks one of several PDFs (trashed attachments are ignored)
let att = item;
if (att.isRegularItem()) {
	const pdfs = Zotero.Items.get(item.getAttachments()).filter((a) => a.isPDFAttachment());
	if (pdfs.length === 0) {
		return `[ERROR] "${item.getField("title")}" has no PDF attachment.`;
	}
	if (pdfs.length > 1) {
		return `[ERROR] "${item.getField("title")}" has ${pdfs.length} PDF attachments; select the one to trim directly.`;
	}
	att = pdfs[0];
}
if (!att.isPDFAttachment()) {
	return "[ERROR] Selected item is not a PDF attachment.";
}

// Get input file path
const originalPath = await att.getFilePathAsync();
if (!originalPath) {
	return "[ERROR] Cannot get file path (attachment file may be missing).";
}
if (!(await IOUtils.exists(originalPath))) {
	return "[ERROR] Input file doesn't exist: " + originalPath;
}

if (!(await IOUtils.exists(pdfcpuPath))) {
	return "[ERROR] pdfcpu not found at: " + pdfcpuPath;
}

// Trim into a temporary file in the same directory, so the final move stays on one volume.
// Match the extension case-insensitively: with a .PDF input a non-matching replace would
// leave outputPath === originalPath and pdfcpu would read and write the same file.
const filename = PathUtils.filename(originalPath);
const outputPath = PathUtils.join(
	PathUtils.parent(originalPath),
	filename.replace(/\.pdf$/i, "") + "_nofirstpage.pdf"
);

// Keep pages 2..end
try {
	await Zotero.Utilities.Internal.exec(pdfcpuPath, ["trim", "--pages", "2-", originalPath, outputPath]);
}
catch (e) {
	await IOUtils.remove(outputPath, { ignoreAbsent: true });
	return "[ERROR] pdfcpu failed (single-page or corrupt PDF?): " + e;
}
if (!(await IOUtils.exists(outputPath))) {
	return "[ERROR] pdfcpu did not produce an output file.";
}

if (keepBackup) {
	try {
		await IOUtils.copy(originalPath, originalPath + ".bak");
	}
	catch (e) {
		await IOUtils.remove(outputPath, { ignoreAbsent: true });
		return "[ERROR] While creating backup: " + e;
	}
}

// Substitute the original file with the trimmed one
try {
	await IOUtils.move(outputPath, originalPath);
}
catch (e) {
	await IOUtils.remove(outputPath, { ignoreAbsent: true });
	return "[ERROR] While replacing original file: " + e;
}

// Best effort: refresh the full-text index, since page 1's text is gone
try {
	await Zotero.FullText.indexItems([att.id], { complete: true });
}
catch (e) {}

return `First page removed from "${filename}"`;

Anything else

You need pdfcpu

[raffaelemancuso]

Revised the script to make it more resistant to errors (in pdfcpu or in copying the temp file over) and possibility to keep a backup copy of the file

[raffaelemancuso]

Updated the script again so that it errors out if you select an item with more than one PDF attachment (selecting directly the PDF attachment works as expected)