extract text from doc files(windows10 64)

Question

extract text from doc files(windows10 64)

SHocker-Yu opened this issue 8 years ago · 10 comments

"DOC extraction requires antiword be installed, link, unless on OSX in which case textutil (installed by default) is used."

OS: windows10 64
I installed antiword.exe failed,and i don't konw how to do with this problem...

Answer 1 · 2016-10-30T01:23:45.000Z

have you declared the path to your antiword.exe file in the PATH global variable ?

Answer 2 · 2016-10-31T02:20:47.000Z

@zzzwx thanks for your reply,antiword does not support Windows.

Answer 3 · 2016-10-31T18:51:26.000Z

@SHocker-Yu i am using it on windows (7 and 10)
some good fellow actually compiled it for windows, get it there : http://www-stud.rbi.informatik.uni-frankfurt.de/~markus/antiword/

Answer 4 · 2016-11-01T06:19:56.000Z

@zzzwx appreciate for your kind reply.
I loaded it at last time, but when i want to run antiword.exe, it flash back,
OS: Windows10
Have you come across this situation?
Could you tell me how to make it running success?

Answer 5 · 2016-11-01T16:48:30.000Z

@SHocker-Yu what do you mean by "flash back" ?

here are the steps I followed to make it work on windows :

0/ modify textract/lib/extractors/doc.js to fix a bug reported in a github issue

-        if ( error.toString().indexOf( 'is not a Word Document' ) ) {
+        if ( error.toString().indexOf( 'is not a Word Document' ) > 0 ) {

1/ download windows binary

2/ add antiword directory to Windows' PATH environnement variable

=> at this point it worked but only when the path to the doc file contained no spaces

3/ modify textract/lib/extractors/doc.js again to add quotes so that it reads the input path as is

-    var escapedPath = filePath.replace( /\s/g, '\\ ' );
+   var escapedPath = filePath/*.replace( /\s/g, '\\ ' )*/;

-    exec( 'antiword ' + escapedPath,
+   exec( 'antiword "' + escapedPath + '"',

=> at this point it worked for every paths

4/ modify textract/lib/extractors/doc.js one last time to manage UTF8 encoding of output text

  - exec( 'antiword "' + escapedPath + '"',
  + exec( 'antiword -m UTF-8.txt "' + escapedPath + '"',

=> and after that it worked well all the time :)

hope this helps you

Answer 6 · 2016-11-03T14:46:21.000Z

@zzzwx I really appreciate for your kind,so sorry about my pool English,'flash back' means 'crash',these days i had to work all day ,and reply you so late,really sorry, i have readed your reply,and i will try it and then tell you the result.
Best wishes.

Answer 7 · 2016-11-05T04:11:59.000Z

@zzzwx It works!Thank you so much!!!

Answer 8 · 2016-12-23T16:44:17.000Z

FYI, I've implemented the changes from above across a few different commits the last few months (sorry so slow!).

Answer 9 · 2016-12-23T16:48:24.000Z

Published as 2.1, thanks!

Answer 10 · 2016-12-23T17:28:26.000Z

Hi @dbashford , thank you for your work