pdfbox-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Maruan Sahyoun (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (PDFBOX-4210) Unable to extract the text from a PDF ("No Unicode mapping.." warnings)
Date Tue, 08 May 2018 12:05:00 GMT

    [ https://issues.apache.org/jira/browse/PDFBOX-4210?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16467306#comment-16467306

Maruan Sahyoun commented on PDFBOX-4210:

So both Reader and PDFBox have the same text extraction result as there is no text to extract.

But what about the Copy & Paste result. This invisible text is meant for interactive search
and was generated using OCR but is not part of the text content of this PDF. We could look
at using such text - if it exists - as a fallback if, and only if, there is no text to extract,
but strictly speaking this PDF has no extractable text (according to the PDF spec).

If we do this I would suggest that it doesn't become a part of the 'core' text extraction
but should be implemented as a utility/application as this is more application oriented.

[~aputnik] You've written in the other issue that you are using OCR as a fallback. Which tool
are you using for that? 

> Unable to extract the text from a PDF ("No Unicode mapping.." warnings)
> -----------------------------------------------------------------------
>                 Key: PDFBOX-4210
>                 URL: https://issues.apache.org/jira/browse/PDFBOX-4210
>             Project: PDFBox
>          Issue Type: Bug
>          Components: Text extraction
>    Affects Versions: 2.0.9
>            Reporter: Aleksandar Putnik
>            Priority: Major
>         Attachments: Testdokument.pdf
> I'm using Tika (v1.18 , which means pdfbox 2.0.9) to extract the text from PDF.
> I have a document from which the Acrobat Reader (Adobe Acrobat Reader DC) can extract
the text (although not with a 100% precision).
> Besides warnings "WARNING: No Unicode mapping for ... in font ArialMT" pdfbox 2.0.9
doesn't return anything.
> As you can see from the warning, the font in question is ArialMT. It is custom encoding
and the pdf doesn't include toUnicode mapping. Font type is CID TrueType (this info is provided
by "pdffonts")
> "pdftotext" also can't extract anything but only shows an error `Syntax Error: Unknown
character collection 'Adobe-ArialMT'`
> The pdf producer (used by the customer) is VintaSoft PDF .NET Plug-in v5.5.
> I would like to determine whether there is a bug in pdfbox or the pdf producer has to
adjust and improve the "readability" of pdf.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org

View raw message