| Summary: | pdftotext cannot extract text correctly from specific pdf | ||
|---|---|---|---|
| Product: | poppler | Reporter: | Derrick <ls89822> |
| Component: | general | Assignee: | poppler-bugs <poppler-bugs> |
| Status: | RESOLVED NOTABUG | QA Contact: | |
| Severity: | normal | ||
| Priority: | medium | CC: | stefan.bruens |
| Version: | unspecified | ||
| Hardware: | x86-64 (AMD64) | ||
| OS: | Windows (All) | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Attachments: |
source pdf file
data debug information of pdf file by poppler |
||
As far as i can see there's no correct text info in that pdf (at least Acrobat can't seem to extract the text either) Created attachment 139365 [details]
data
The data is chinese， i can view the pdf data via the microsoft edge browser. My computer system language is chinese. You can view the PDF too using a poppler based viewer (okular, evince, pdftoppm). That doesn't mean you can extract the text. It's two completely different things I download the evince, the file could be viewed by evince. Could you tell me why it could not be extracted by pdftotext? I know the file can't be extracted under the following scenes: 1. The word is drawed by line, not word. 2. The font type is fontTrueType or fontTrueTypeOT, the character set is same as system character set. But in this case, the font type is "fontCIDType2 Identity-H BFAIEF+SimSun-GBK-EUC-H", I think it should by extracted correctlly. ==================================================== gs /GS1 exec op gs gfx state dict: << /OP false /OPM 1 /SA false /SM 0.02 /Type /ExtGState /op false >> BT exec op BT Tf /TT2 1 exec op Tf opSetFont args[0].getName=TT2 font: tag=TT2 name='BFAIEF+SimSun-GBK-EUC-H' 1 Tm 16.0751 0 0 16.0751 167.949 734.655 exec op Tm cs /Cs6 exec op cs scn 0.019608 0.003922 0 exec op scn Tc 0 exec op Tc Tw 0 exec op Tw Tj (ɩ exec op Tj font doShowText key=TT2 name=10 Identity-H BFAIEF+SimSun-GBK-EUC-H ET exec op ET q exec op q i 1 exec op i =============================================================== By default, poppler think the character is UCS-4 in operation TJ, maybe in this case， the character is other, cause extracted uncorrectlly. The attachment is debug information of the pdf file by poppler. Created attachment 139392 [details]
debug information of pdf file by poppler
Sorry, I don't really have time to explain to you how pdf files work. I've not been able to find any other PDF viewer (neither the one from the people that invented PDF itself) out there that can extract the text from that file, so i'd say it's just the file being broken. If you can extract the text with some PDF tool or even better provide a patch i'll be happy to review it. This bug report contains markup which cannot be properly exported by Bugzilla, making it generate invalid XML. As it seems the result is that this bug will not be taken forward, I'm closing it now rather than migrating it to GitLab. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.
Created attachment 139361 [details] source pdf file Hi, pdftotext fails to extract text correctly from specific pdf (see attachment). exit status is 0 and no warnings or errors are reported. command: pdftotext.exe -bbox-layout q7.pdf product version: poppler 0.62.0 x86_64 poppler-data 0.4.8 best regards Derrick