I made a native Python module for MS Word docx

http://github.com/mikemaccana/python-docx

by nailer • 17 years ago

144 25 17 years ago

nailer5• 17 years ago

Hi Hacker News,

I was recently wandering around StackOverflow and PyPI and wondering why the only solutions to make MS Office documents from Python seemed to be based around either COM automation, OpenOffice automation, or calling Java or .net libraries.

So I made this module, which reads and writes Microsoft Office Word 2007 docx files. These are referred to as 'WordML', 'Office Open XML' and 'Open XML' by Microsoft.

They can be opened in Microsoft Office 2007, Microsoft Mac Office 2008, OpenOffice.org 2.2, and Apple iWork 08.

The docx module has the following features:

Making documents

• Paragraphs

• Bullets

• Numbered lists

• Multiple levels of headings

• Tables

Editing documents

Thanks to lxml module, we can:

• Search and replace

• Extract plain text of document

• Add and delete items anywhere within the document

• Run xpath queries against particular locations in the document - useful for retrieving data from user-completed templates.

It's only a couple of hundred lines - lxml does most of the heavy work - but it's incredibly simple to use, check out example.py in the link.

Hope you find it useful.

Mike

coderdude• 17 years ago

This is great! I'm already using lxml for my project and using ReportLab to create pdf reports of data I'm mining from the Web. This will be perfect for me and now I can generate Microsoft Word reports.

nailer5• 17 years ago

Glad you like it. A great deal of the credit belongs to the lxml authors (Stephan Behnel in particular). Though this only took a day or do, I've been using lxml for projects for years and always loved it - in particular, the xpath makes modifying and querying trees really easy.

If more people used LXML and saw how easy it was, the whole idea of modifying serialized XML 'Cthulu style' would be seen as being as odd as modifying other data by manipulating their pickled forms.

bigsassy2• 17 years ago

Thank you, nailer! I've been putting up with Word randomly crashing when using it's COM object in long running processes. Now I can skip Word altogether and work on the file level. THANKS!

marcinw• 17 years ago

Thanks Mike for providing such a useful module!

Some plans I have to add to this module (if you don't plan too already):

Add support for:

* Pass a font-size, bold, italic, bold+italic and spacing to paragraph method to modify font

* Pass a table row header to table method

* Specify table column widths

* Change text alignment for table rows/headers, etc.

* Change text style in table cells

* Ability to change background color for every other table row

* Create method to insert a page/section break

* Create a method to add an image at a specific position in document page

On one other note:

I'm not 100% positive, but thumbnail.jpg/thumbnail.wmf may result in inadvertent disclosure of sensitive data if using this to generate report documents...

nailer5• 17 years ago

Awesome! I welcome contributions. Could you send me something for page & section breaks?

I'm already working on images and 100% nose coverage. I'm intending to do document properties after that.

Re: zebra striping for tables, we already have that via the inbuilt styles. But more options to control styles would be useful.

Coding style is:

* Functional

* Google style - http://code.google.com/p/soc/wiki/PythonStyleGuide

* Unit tests are handled with nose / coverage

omouse• 17 years ago

Which license is it under? GPL I hope?

nailer5• 17 years ago

MIT

gaius• 17 years ago

Cheers dude :-)

dirkstoop• 17 years ago

Awesome, thanks. Are you putting this in the public domain, or are you licensing it under any particular license agreement?

I couldn't find anything about a license or copyright on your project's github page.

nailer5• 17 years ago

Good point. I've licensed it under MIT (originally I thought github made me pick a license, but it seems not), see the updated README.

elblanco• 17 years ago

docx has been a real boon to people writing things like search engines and want to use the text inside of the files.

We've been using http://b2xtranslator.sourceforge.net/blog/ to convert legacy binary office formats to openXML, then using XSLT to transform the XML into our preferred schema for text extraction.

bigsassy2• 17 years ago

Python also has a native Python module for reading and writing Excel files. It's pretty simple to use and also doesn't require any COM nastiness. http://www.python-excel.org/

anon42389475• 17 years ago

this is great. thanks.

k_shehadeh• 17 years ago

Just wanted to add my thanks to the list. This was one of the bigger holes in document processing and now it's been filled for the most part.

est• 17 years ago

A quick hack for if you need .doc not .docx format: generate .rtf and rename it to .doc :)

PS, .eml and .mht are actually the same format.

stevenbedrick• 17 years ago

An even worse hack, that unfortunately has worked really well for me in the past: generate a simple HTML file (em, b, u, ul/ol, etc.), rename it to .doc, and Word will open it without appearing to notice or comment. This saved my butt once when I discovered mid-project that the Textile library I was using couldn't generate RTF... I felt more than a little bit dirty, but it worked without a hitch.

pyre• 17 years ago

> .eml and .mht are actually the same format.

.mht uses MIME/Base64 to store all of the HTML and external files together, but IIRC it's not exactly the same as an email because the headers are different/missing (i.e. I wouldn't make sense to have a To:/Reply-To:/Cc:/etc header)

benatkin• 17 years ago

Thanks! I've had to deal with docx files from time to time. I'm going to try this out.

Does anyone know of a good way to read xlsx files with Python? I think I tried one library and it didn't quite work for this 20,000 row file, and I wound up using OpenOffice to convert it to csv. If not, hopefully this can be used as a starting point for developing a nice xlsx library.

brendano2• 17 years ago

I wrote the following python script to read xlsx. It's extremely basic, but works very well for me.

http://github.com/brendano/tsvutils/blob/master/xlsx2tsv

I've never tried it on a 20,000 row file. I suspect it would work. A really large file might need to switch to a streaming XML parser, but probably Excel itself wouldn't handle that use case too well.

pmorici• 17 years ago

Is there something similar that will support .doc and / or .ppt files that doesn't rely on com or the open office Uno bridge?

gus_massa• 17 years ago

I think that is will be easier to support .rtf files.

cschneid• 17 years ago

About the closest you'll get is the apache POI library. It's mostly an excel 97 lib, but it supports word to some extent.

brendano2• 17 years ago

Great, thanks for doing this!

A while ago, I wrote a little Python script to read from xlsx Excel files. It's nice that these Microsoft Office XML files can be processed in pure Python.

http://github.com/brendano/tsvutils/blob/master/xlsx2tsv

wenbert• 17 years ago

Thanks for this. This will make other people's lives so much better. This is good for everyone.

I made a native Python module for MS Word docx

Related Stories