parsed incorectly, wrong symbols

[NazarSolovei]

• PHP Version: 8.0
• PDFParser Version: Latest: 2.5.5

Description: pdf file parsed and outputted as unreadable characters.

PDF input: inv.pdf

Expected output & actual output

EXPECTED:
Buyer Receipt
Receipt # 19509334
Receipt Date 10/26/2022
Received By AC/ASAP S...... (and so on as in pdf file)

ACTUAL:
%X\HU�5HFHLSW 5HFHLSW�� �������� 5HFHLSW�'DWH ���������� 5HFHLYHG�%\ $&�$6$3�6 6ROG�$W�%UDQFK ������'UHD

Code

include "vendor/autoload.php";
$parser = new \Smalot\PdfParser\Parser();
$pdf = $parser->parseFile("inv.pdf");
$text = $pdf->getText();
echo $text;

[NazarSolovei]

@k00ni have you tried to fix it?
Thank you for your answer.

[k00ni]

@NazarSolovei No, my spare time is very limited so I have to triage what issues I focus on.

[GreyWyvern]

There are actually two issues here. One is that the font CIDMaps are gzcompressed with a CRC32 checksum instead of the Adler-32 PHP expects. See: https://www.php.net/manual/en/function.gzuncompress.php#79042

The code at the link above fixes the majority of the PdfParser extracted output by at least setting the default page font (F14). However, there is one section of the file in a different font (F17) which still displays undecoded bytes, and the reason F17 is not getting used is because the change font commands (/F17 9.744095 Tf for example) have been placed outside of the text blocks. Between blocks, basically. See this example bit of stream:

S
/F17 9.744095 Tf
/GS0 gs
0 0 0 rg
BT
54.035439 -109.771729 Td
[______________________________________________________________________________________________] TJ
ET
/F17 9.744095 Tf
/GS0 gs
0 0 0 rg
BT
236.515762 -110.657532 Td
[________________________] TJ
ET
/GS0 gs
0 0 0 RG 0.885827 w 0 J 0 j 1.414 M [] 0 d
121.358276 -110.657532 m
224.114197 -110.657532 l
S
/GS0 gs
0 0 0 RG 0.885827 w 0 J 0 j 1.414 M [] 0 d
271.948822 -110.657532 m
403.937042 -110.657532 l
S
/F14 7.972442 Tf
/GS0 gs
0 0 0 rg
BT
53.149609 -128.374084 Td
[_____________________________________________________________________________] TJ
ET
/F14 7.972442 Tf
/GS0 gs
0 0 0 rg
BT
318.011841 -130.145752 Td
[_________________________________________________________________________________] TJ
ET

PdfParser considers everything between a BT and an ET as a "text block". Normally font commands Tf only occur inside these blocks. But in the stream above, they don't.

The current getSectionsText() function in PDFObject.php only extracts commands that are within these text blocks and ignores all other lines of the stream. I think this function will need to be heavily overhauled to account for commands that potentially happen outside of text blocks BT ... ET.

[GreyWyvern]

Hi @NazarSolovei. May I use your inv.pdf as an addition to the PdfParser test suite? Is it free to use? Thanks.

[NazarSolovei]

Hello, Yes it’s free. Can I ask you to notify me once you solve the issue? Пн, 31 июля 2023 г. в 18:51, Brian Huisman ***@***.***>:

Hi @NazarSolovei <https://github.com/NazarSolovei>. May I use your inv.pdf as an addition to the PdfParser test suite? Is it free to use? Thanks. — Reply to this email directly, view it on GitHub <#592 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A65NKCQUT66U3JMEVWHSJ63XS7IBZANCNFSM6AAAAAAWQPHWBU> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- -- Nazar Solovei