Title and other properties not read with getDetails for some files

[dbarron]

• PHP Version: 8.1
• PDFParser Version: 2.10.0

Description:

I recently used pdfparser to load 2500 documents in to Drupal. I pulled the properties with getDetails() to populate the Drupal fields.

About 50 documents did not get the right title and other properties, even though I can see them when I look at the files with a text editor and every other PDF tool I used saw them (eCopy, Chrome, Linux document viewer, Funnelback search engine).

I've attached a zip with some files that failed, and a few that succeeded, in case that helps track this down.

PDF input

Attached.

Expected output & actual output

Input is in attached file. Output expected to include all properties, but many are missing.

I have never seen an error of any kind.

Code

public function getMetadataFromFile($file) {
  $metadata = "";
  $parser = new Parser();

  $pdf = $parser->parseFile($file);
  $metadata = $pdf->getDetails();

  return $metadata;
}

I have as similar function that uses parseContent and the problem is there, too. (I used parseConent to import existing files from an old system, and parseFile for new uploads.) Documents.zip

[GreyWyvern]

It looks like the Title metadata for these "bad" files is located in the XMP data only. However, extractXMPMetadata() expects there to be a dc:format tag that tells the PdfParser that the metadata is referring to the PDF, and not an embedded file. In these PDF files there is no dc:format tag.

The fix for this is to merge the found XMP metadata if a dc:format tag doesn't exist, and if it does exist, only merge it if the MIME-type is application/pdf.

Updated line 290 from Document.php:

            if (!isset($metadata['dc:format']) || 'application/pdf' == $metadata['dc:format']) {