solution to convert PDFs, DOCs, DOCXs into a textual format with python

Question

I am developing a full text search engine for indexing popular binary formats. I know that there are hundereds of such questions (and solutions) already, but I found it tough to find one:

• cross platform
• supports DOC, DOCX and PDF formats at once
• easy to use with python
• can be set up in a major shared host

Answer 1

• For PDFs, I recommend PDFminer.
• Try the docx module (I have not used it myself)
• I am not aware of any pure python module that can read .doc files.
• There are command-line tools to extract text from .doc files: antiword and catdoc (and probably others). If the packages are installed on your shared host, you could use subprocess to shell out to these tools. Available on Windows via Cygwin.
• Apache POI is a Java library that can extract text from Office documents. If your shared host has Java installed, you could write a bit of Java (or Jython) code and execute using subprocess.

Answer 2

Textract uses the default tools for every kind of file.

https://github.com/deanmalmgren/textract

Answer 3

If at server side you can use OpenOffice then you can use unoconv: Convert between any document format supported by OpenOffice

Answer 4

One possible solution is to use google documents to extract the text contents from binary .doc-files. You upload the document to google docs and then download the text contents. It is a fairly slow process, but it is the only "pure Python" solution I know of since it doesn't require any external tools except for network access. An external tool such as catdoc or antiword is a much better solution if you are allowed to install it on your host.