You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@pdfbox.apache.org by "John Hewson (JIRA)" <ji...@apache.org> on 2016/02/23 20:20:19 UTC
[jira] [Comment Edited] (PDFBOX-2740) Text extraction failed on
Korean PDF
[ https://issues.apache.org/jira/browse/PDFBOX-2740?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15159473#comment-15159473 ]
John Hewson edited comment on PDFBOX-2740 at 2/23/16 7:20 PM:
--------------------------------------------------------------
The difference between copy & paste and save as in Acrobat which Maruan noticed reminded me of an issue we had before where the same thing was happening. In that case it turned out that the text which was extracted via copy and paste was the accessible text in the "marked content", rather than the text from the content stream.
Looking at this PDF, we see the same thing:
{code}
/Span << /ActualText (\376\377\263\304) >> BDC
BT
/F33 1 Tf
8.4647 0 0 8.4647 392.0664 324.946 Tm
(^) Tj
ET
EMC
{code}
We've seen cases in the past where marked content actually contains bad text, while the content stream contains good text, so we don't extract text from marked content, neither do most PDF viewers. We do provide PDFMarkedContentExtractor for those wanting to extract marked content only.
[~chengas123], your problem may or may not be the same - open your problem PDF with PDFDebugger and search the Contents stream for {{/ActualText}}. If you see it, then you have the same problem. Otherwise, feel free to open a new issue if/when you have a PDF which you can post.
was (Author: jahewson):
The difference between copy & paste and save as in Acrobat which Maruan noticed reminded me of an issue we had before where the same thing was happening. In that case it turned out that the text which was extracted via copy and paste was the accessible text in the "marked content", rather than the text from the content stream.
Looking at this PDF, we see the same thing:
{code}
/Span << /ActualText (\376\377\263\304) >> BDC
BT
/F33 1 Tf
8.4647 0 0 8.4647 392.0664 324.946 Tm
(^) Tj
ET
EMC
{code}
We've seen cases in the past where marked content actually contains bad text, while the content stream contains good text, so we don't extract text from marked content, neither do most PDF viewers. We do provide PDFMarkedContentExtractor for those wanting to extract marked content only.
[~chengas123], your problem may or may not be the same - open your problem PDF with PDFDebugger and search the Contents stream for {/ActualText}.
> Text extraction failed on Korean PDF
> ------------------------------------
>
> Key: PDFBOX-2740
> URL: https://issues.apache.org/jira/browse/PDFBOX-2740
> Project: PDFBox
> Issue Type: Bug
> Components: Text extraction
> Affects Versions: 1.8.7, 1.8.8, 1.8.9, 2.0.0
> Reporter: Julien Ortega
> Attachments: g_KO_201506-ReaderDC-cutAndPaste.txt, g_KO_201506-ReaderDC-saveAsText.txt, g_KO_201506.pdf, g_KO_201506.txt
>
>
> Trying to extract text on a Korean PDF gives me a lot of warnings :
> WARNING: No Unicode mapping for US (33) in font DVCAYA+WtKoBaeumMyungjoL063zb4?Pw
> avr. 01, 2015 12:05:32 PM org.apache.pdfbox.pdmodel.font.PDSimpleFont toUnicode
> WARNING: No Unicode mapping for NAK (33) in font JYLDGG+WtKoBaeumMyungjoL053zb4?Pw
> avr. 01, 2015 12:05:32 PM org.apache.pdfbox.pdmodel.font.PDSimpleFont toUnicode
> WARNING: No Unicode mapping for RS (38) in font WRYULE+WtKoBaeumMyungjoL013zb4?Pw
> avr. 01, 2015 12:05:32 PM org.apache.pdfbox.pdmodel.font.PDFont <init>
> WARNING: Invalid ToUnicode CMap in font FZEFOY+WtKoBaeumGothicL0422b4?Pw
> avr. 01, 2015 12:05:32 PM org.apache.pdfbox.pdmodel.font.PDSimpleFont toUnicode
> WARNING: No Unicode mapping for DEL (33) in font FZEFOY+WtKoBaeumGothicL0422b4?Pw
> avr. 01, 2015 12:05:32 PM org.apache.pdfbox.pdmodel.font.PDFont <init>
> WARNING: Invalid ToUnicode CMap in font OOLNBG+WtKoBaeumGothicL0122b4?Pw
> avr. 01, 2015 12:05:32 PM org.apache.pdfbox.pdmodel.font.PDSimpleFont toUnicode
> WARNING: No Unicode mapping for SOH (33) in font OOLNBG+WtKoBaeumGothicL0122b4?Pw
> and the result is not readable. The pdf is containing the necessary conversion table because every pdf reader (Desktop or Mobile) let me copy and past the text without problem.
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org