#3443 NORM Untriag: Create bindings to use libabiword for file conversion

[Zarro Boogs per Child]

#3443: Create bindings to use libabiword for file conversion
---------------------------------------+------------------------------------
  Reporter:  uwog                      |       Owner:  uwog     
      Type:  defect                    |      Status:  new      
  Priority:  normal                    |   Milestone:  Untriaged
 Component:  write-activity (abiword)  |     Version:           
Resolution:                            |    Keywords:           
  Verified:  0                         |  
---------------------------------------+------------------------------------

Comment(by tomeu):

 Extracting plain text from various formats will surely be important when
 we restore indexing of document contents and improve our fulltext search
 capabilities.

 Although we should compare the reliability of abiword filters against
 alternatives like pdftotext, odt2txt, etc, I see advantages in using
 abiword for all of them.

 Regarding the impl of python bindings for this functionality, lately I
 have thought it could be better if the datastore executed the conversion
 out of process, as abiword is a big piece of non-python code and if
 executed in-process by the DS, could bring down the whole DS process in
 case of a segmentation fault.

 So perhaps we don't need to add pyabiword bindings for content indexing,
 just bundle the abiword binary so it can have piped in docs and piped out
 plain text.

-- 
Ticket URL: <https://dev.laptop.org/ticket/3443#comment:1>
One Laptop Per Child <https://dev.laptop.org>
OLPC bug tracking system