Fw: PDF library for reading PDF files

[Peter Galfi]

Hi!

I am looking for a library in Python that would read PDF files and I could extract information from the PDF with it. I have searched with google, but only found libraries that can be used to write PDF files.

Any ideas?

Peter

 

[Harald Massa]

I am looking for a library in Python that would read PDF files and I

could extract information from the PDF with it. I have searched with
google, but only found libraries that can be used to write PDF files.

reportlab has a lib called pagecatcher; it is fully supported with python,
it is not free.

Harald

 

[David Boddie]

I am looking for a library in Python that would read PDF files and I
could extract information from the PDF with it. I have searched with
google, but only found libraries that can be used to write PDF files.

Any ideas?

I quickly searched back through Google, but I knew exactly what I was
looking for: ;-)

http://groups.google.com/[email protected]

The page referred to is here:

http://www.boddie.org.uk/david/Projects/Python/pdftools/

The module is very much a "work in progress". You can probably get
some text and bitmap images out of a few documents, but that's
probably all you can expect unless you want to improve it (and
submit patches).

Good luck!

David

 

[Cameron Laird]

reportlab has a lib called pagecatcher; it is fully supported with python,
it is not free.

Harald

ReportLab's libraries are great things--but they do not "extract
information from the PDF" in the sense I believe the original
questioner intended. As Andreas suggested, he's probably best
off using existing stand-alone applications as separate processes,
controlled from Python.

 

[Robert Kern]

reportlab has a lib called pagecatcher; it is fully supported with python,
it is not free.

Harald

ReportLab's libraries are great things--but they do not "extract
information from the PDF" in the sense I believe the original
questioner intended. [/QUOTE]

No, but ReportLab (the company) has a product separate from reportlab
(the package) called PageCatcher that does exactly what the OP asked
for. It is not open source, however, and costs a chunk of change.

 

[Cameron Laird]

No, but ReportLab (the company) has a product separate from reportlab
(the package) called PageCatcher that does exactly what the OP asked
for. It is not open source, however, and costs a chunk of change.

Let's take this one step farther. Two posts now have
quite clearly recommended ReportLab's PageCatcher <URL:
http://reportlab.com/docs/pagecatcher-ds.pdf >. I
completely understand and agree that ReportLab supports
a mix of open-source, no-fee, and for-fee products, and
that PageCatcher carries a significant license fee. I
entirely agree that PageCatcher "read PDF files ...
and ... extract information from the PDF with it."

HOWEVER, I suspect that what the original questioner
meant by his words was some sort of PDF-to-text "extrac-
tion" (true?) and, unless PageCatcher has changed a lot
since I got my last copy, PDF-to-text is NOT one of its
functions.

 

[Robin Becker]

No, but ReportLab (the company) has a product separate from reportlab
(the package) called PageCatcher that does exactly what the OP asked
for. It is not open source, however, and costs a chunk of change.

Let's take this one step farther. Two posts now have
quite clearly recommended ReportLab's PageCatcher <URL:
http://reportlab.com/docs/pagecatcher-ds.pdf >. I
completely understand and agree that ReportLab supports
a mix of open-source, no-fee, and for-fee products, and
that PageCatcher carries a significant license fee. I
entirely agree that PageCatcher "read PDF files ...
and ... extract information from the PDF with it."

HOWEVER, I suspect that what the original questioner
meant by his words was some sort of PDF-to-text "extrac-
tion" (true?) and, unless PageCatcher has changed a lot
since I got my last copy, PDF-to-text is NOT one of its
functions. [/QUOTE]
I suspect Cameron is right. ReportLab does have a product called
pageCatcher, but its main function is to grab individual pages for
reuse. I believe it could be extended to go deeper and mess about with
text streams, but it certainly doesn't do that now and would take some
effort to do properly as text can be complicated in PDF (or postscript).

 

[Andreas Lobinger]

Aloha,

Peter Galfi schrieb:
I am looking for a library in Python that would read PDF files and I
could extract information from the PDF with it. I have searched with
google, but only found libraries that can be used to write PDF files.
Any ideas?

Use file, split, zlib and a broad knowledge of the PDF-spec...

Accessing certain objects in the .pdf is not that complicated if
you f.e. try to read the /Info dictionary. Getting text from
actual page content could be very complicated.

Can you explain your 'information' further?

Wishing a happy day
LOBI

 

[Robert Kern]

No, but ReportLab (the company) has a product separate from reportlab
(the package) called PageCatcher that does exactly what the OP asked
for. It is not open source, however, and costs a chunk of change.

Let's take this one step farther. Two posts now have
quite clearly recommended ReportLab's PageCatcher <URL:
http://reportlab.com/docs/pagecatcher-ds.pdf >. I
completely understand and agree that ReportLab supports
a mix of open-source, no-fee, and for-fee products, and
that PageCatcher carries a significant license fee. I
entirely agree that PageCatcher "read PDF files ...
and ... extract information from the PDF with it."

HOWEVER, I suspect that what the original questioner
meant by his words was some sort of PDF-to-text "extrac-
tion" (true?) and, unless PageCatcher has changed a lot
since I got my last copy, PDF-to-text is NOT one of its
functions.

Rereading http://www.reportlab.com/PageCatchIntro.html , you're right.
My apologies. I thought you were talking about the open source reportlab
package and not PageCatcher specifically.

 

[Peter Galfi]

Thanks. I am studying the PDF spec, it just does not seem to be that easy
having to implement all the decompressions, etc. The "information" I am
trying to extract from the PDF file is the text, specifically in a way to
keep the original paragraphs of the text. I have seen so far one shareware
standalone tool that extracts the text (and a lot of other formatting
garbage) into an RTF document keeping the paragraphs as well. I would need
only the text.

Any suggestions?

Peter

----- Original Message -----
From: "Andreas Lobinger" <[email protected]>
Newsgroups: comp.lang.python
To: <[email protected]>
Sent: Monday, January 19, 2004 5:02 PM
Subject: Re: Fw: PDF library for reading PDF files


Aloha,

Peter Galfi schrieb:
I am looking for a library in Python that would read PDF files and I
could extract information from the PDF with it. I have searched with
google, but only found libraries that can be used to write PDF files.
Any ideas?

Use file, split, zlib and a broad knowledge of the PDF-spec...

Accessing certain objects in the .pdf is not that complicated if
you f.e. try to read the /Info dictionary. Getting text from
actual page content could be very complicated.

Can you explain your 'information' further?

Wishing a happy day
LOBI

 

[Josiah Carlson]

Thanks. I am studying the PDF spec, it just does not seem to be that easy

having to implement all the decompressions, etc. The "information" I am
trying to extract from the PDF file is the text, specifically in a way to
keep the original paragraphs of the text. I have seen so far one shareware
standalone tool that extracts the text (and a lot of other formatting
garbage) into an RTF document keeping the paragraphs as well. I would need
only the text.

Any suggestions?

Peter,

Suggestion: extract the document to RTF using that other tool, then use
any one of the few dozen RTF parsers to convert them into plaintext.

- Josiah

 

[Andreas Lobinger]

Aloha,

Thanks. I am studying the PDF spec, it just does not seem to be that easy
having to implement all the decompressions, etc. The "information" I am
trying to extract from the PDF file is the text, specifically in a way to
keep the original paragraphs of the text. I have seen so far one shareware
standalone tool that extracts the text (and a lot of other formatting
garbage) into an RTF document keeping the paragraphs as well. I would need
only the text.

As others wrote here, the simplest solution is to use a external
pdf-2-text programm and postprocess the data. Read comp.text.pdf

There is no simple and consistent way to extract text from a .pdf
because there are many ways to set text. The optical impression
of a paragraph may not be represented by a similar command structure
in the .pdf.

Adobe recognized the difficulties for document reuse and introduced
tagged .pdf in 1.4. With tagged-pdf it is possible to insert
structural information in the .pdf. If you are interested in
using this, contact me.

Wishing a happy day
LOBI

 

[Cameron Laird]

Aloha,

Peter Galfi schrieb: .
.
. .
.
.
As others wrote here, the simplest solution is to use a external
pdf-2-text programm and postprocess the data. Read comp.text.pdf

There is no simple and consistent way to extract text from a .pdf
because there are many ways to set text. The optical impression

.
.
.
I want to emphasize that final sentence. If you insist on pursuing
this, though, refer to <URL:
http://phaseit.net/claird/comp.text.pdf/PDF_converters.html#pdf2txt >.

 

[Dennis Lee Bieber]

Any suggestions?

Configure a text-only printer, with "print to file" capability,
and "print" the PDF file to it... Then read the print-out...


--

 

[Jeff Sandys]

....
The "information" I am trying to extract from the PDF file is the text,
specifically in a way to keep the original paragraphs of the text. ....

Any suggestions?

Ghostscript has an Extract Text capability that I have used
successfully on some pdf files (but not on some others):
http://www.cs.wisc.edu/~ghost/

Thanks,
Jeff Sandys