PDF only being partly read.

[bclarkson72]

Have many PDF invoices that are able to be read fine by the library. But the odd one just does not return the full document

Reads Fine.pdf
Bad Read.pdf

I cannot see any difference in the files at all.

[k00ni]

Maybe related to #473 and #471.

Was it always this way or did it work in the past?

[bclarkson72]

TBH only started with these invoice PDFs recently - previous files have been fine but noticed issues when trying to parse these particular client files. Was using an old version but upgraded to recent build and issues the same.

[bclarkson72]

Found problem - ET present in textCleaned so screwing with preg_match_all:

The following change gets around the problem but don't think it is a suitable fix.

public function getSectionsText(?string $content): array
{
    $sections = [];
    $content = ' '.$content.' ';
    $textCleaned = $this->cleanContent($content, '_');
	$textCleaned = str_replace("PET","TEP",$textCleaned); //Added this as ET was present in other lines so screwed up the preg_match_all
	//echo $textCleaned;
    // Extract text blocks.
    if (preg_match_all('/(\sQ)?\s+BT[\s|\(|\[]+(.*?)\s*ET(\sq)?/s', $textCleaned, $matches, \PREG_OFFSET_CAPTURE)) {

[bclarkson72]

Output from cleanContent which shows ET :

58.03 288.9 485.01 10.01 re
f
58.03 242.8 485.01 10.01 re
f
58.03 242.8 485.01 10.01 re
f
BT
0 g
0 Tr
/FTxkPETkkj 8 Tf
1 0 0 1 535.55 627.4 Tm
[_________________________________________________________________________]TJ
ET
q
0 38.88 -24 0 42.1 452.40002 cm
/IMcGhHwtqz Do

[bclarkson72]

This change seems to be the solution as ET should always be followed by newline?

if (preg_match_all('/\s+BT[\s|(|[]+(.?)\sET\n/s', $textCleaned, $matches, PREG_OFFSET_CAPTURE)) {

[k00ni]

If you think its a sustainable solution please send us a pull request and we can discuss details there.

[k00ni]

@bclarkson72 can you please test if #533 fixes your problem?