Converting MS Word .doc to LaTeX by command line

Question

There are many guides to convert .doc to .tex with text editor softwares (mostly under Windows); and this popularity has made it hard to search for command-based converting methods.

I am looking for a command line to convert MS Word to LaTeX. This will be part of a PHP script running on Linux server. I mean a shell command to be executed by PHP.

Answer 1

Antiword is going to do reasonable good job converting .doc to .tex files . It makes every effort to preserve not only the content but formating as well. It is well suited for batch processing that you want to do.

Edit: Several people asked me privately about LaTeX switch in Antiword and the latest version of Antiword. The latest version is indeed 0.37. As of LaTeX output I think I mixed up things a bit. I used Antiword for formated ASCII output. I think it is capable of PostScript output but not of LaTeX output. As Jon observed you can use pandoc to convert well formated ASCII into LaTeX. However, wvWare (wv and wv2) are capable of outputting LaTeX. A bit of warning. wvWare is depreciated in favor of AbiWord but can be used for batch processing (I have no clue if AbiWord can be used from the command line). It is still a bit younger program (dormant since 2006) than Antiword(dormant since 2004).

Finally there is a tool called catdoc which is great for batch processing but will not preserve format (great for extracting content though and supports MS Excel format).

Answer 2

This answer is specific to OS X...

Command line utility textutil

There's a nice command line utility called textutil included in OS X that will convert among common document formats:

Word docx to txt

$ textutil -convert txt worddoc.docx

txt to Word docx

$ textutil -convert docx mytextdoc.txt

txt to Word, using Times New Roman 12pt

$ textutil -convert docx -font "Times New Roman" -fontsize 12 blah.txt

Also works with html, rtf, doc, odt, and others...

word2latex and latex2word by using textutil with Pandoc

If you use Pandoc in combination with textutil you have can have a decent Word-to-LaTeX and LaTeX-to-Word roundtrip. For docx support you need the latest version of Pandoc (1.9+).

word2latex

$ textutil -convert html worddoc.docx -stdout | pandoc -s -f html -t latex -o latexdoc.tex

latex2word

$ pandoc -t docx -f latex -o backtoword.docx latexdoc.tex

Answer 3

A lot depends on how complicated the Word document formatting is. I have had very good success with rtf2latex2e, which converts RTF formatted text to LaTeX. It has various levels of matching the RTF formatting. I have mainly used its "minimal LaTeX markup mode", which is ideal for a document that will be subsequently hand-edited (which I understand is not the same conditions as you require.)

Answer 4

I have tried most of the suggestions mentioned here at some point. The best workflow I have developed for this starts from first converting the MS Word document to an ODT file using Libre Office on the command line:

loffice --headless --convert-to odt msword.doc && cp msword.odt loffice.odt

After that, Writer2LaTeX does a pretty good job of converting the basic structure into TeX. I almost always still need to tweak some things manually, so I go with the ultraclean formatting option in order to minimize the output:

w2l -ultraclean loffice.odt tweak_this.tex

At this point, I normally convert the file to a PDF and visually inspect it, fixing anything that is wrong. When I convert many similar files, I often write short scripts in Python to deal with the most common tweaks.

If you are using Debian/Ubuntu, you can install Writer2LaTeX with apt-get install writer2latex, of course. Otherwise, see their website for installation and usage details.

Answer 5

A multiplatform solution is in the works. Rob Oakes is implementing this feature in LyX. Once implemented, I think that using it on the commandline would be straightforward. However, (1) I'm not sure you want to install a somewhat large application for just this conversion and (2) the work is in alpha stages right now. Rob is looking for testers. He's mainly looking for .doc files to test on. He just gave an update on the features that his tool supports: http://marc.info/?l=lyx-users&m=133070969217214&w=2 Here is a list that of features already implemented (copied from the url):

1. Translating Word paragraph and character styles to LyX paragraph and character styles. In the case of character styles that aren't defined, it will write entries for them into the local layout (including basic LaTeX commands).
2. Importing Word tables, including those with merged rows or columns. It will also do its best with the table borders.
3. Enumerated and itemized lists.
4. Importing images from the Word document. (It skips over embedded objects, such as charts from Excel.)
5. The use of custom templates, which allows you to fine tune importing your documents. I've created templates for article.cls and book.cls. I'll also probably create one for memoir.cls as well.

Answer 6

I was dismayed that the Linux user above expressed envy of textutil on Mac OS X. Although I am on Mac OS X, I don't use textutil. As far as I know, it is not capable of reading any formatting information from files anyway. I believe it just extracts text. There are many utilities for that purpose. For instance antiword does it for pre-docx MSWord files and is cross platform. catdoc, mentioned above, works better on some files. I'm sure I'm forgetting a few good ones ...

Instead of textutil, I use a shell script I inherited (I don't know who posted it originally or where) to decode .docx files. Here it is:

#!/bin/sh
unzip -p "$1" word/document.xml \
| sed -e 's/<\/w:p>/\
/g' -e 's/<[^>]*>//g'

I named it docx, put it in /usr/local/bin and said sudo chmod 700 docx to make it executable. Then I say

docx mystupidlittlewordfile.docx | less

to read the contents of any .docx file quickly and conveniently.

All it does is three things one per line (not counting the first line that tells which shell to use). First, it unzips the xml document containing text. This xml document is currently usually word/document.xml in the .docx file which, if you have not guessed, is a structured .zip file. If this script does not work, the first thing I do is to say

unzip -l mystupidwordfile.docx

so I can see where the text is. You should look inside word/document.xml anyway, to understand the next two lines of the shell script above. The second line pipes the output to a sed script that replaces MSWord's symbol for newline with an actual newline. The third line pipes this output to another sed script that sends everything to stdout except whatever is between angle brackets.

There are actually a number of other useful items between angle brackets, such as table layouts, that would be easy to add with intermediate sed scripts. You could replace many of these useful items, for instance, with markdown equivalents. I just haven't gotten around to it. The main reason I use this script is so that I can see the content of an average .docx file on a network volume to see if I can understand its content without having to run MSWord to read it. Nine out of ten times, this is all I need to read and understand .docx files. Formatting in average .docx files is usually decorative rather than meaningful.

Presumably, Microsoft will change these angle bracket directives with successive versions of MSWord to make it harder to use any non-Microsoft tools on them. I have heard that Microsoft's original attempt at an open format was essentially a MSWord document surrounded by angle brackets.

Nevertheless, as long as public pressure remains for them to use XML, at least some of the directives can be read and modified from version to version. My view is that the end user should be able to modify the tool when it fails because intentional incompatibility has been introduced by the monopolist. I'm not sure how to do that. My solution requires that the user know sed and regular expressions and be less lazy than, evidently, I am. This will only fit a small subset of people who would like, perhaps for ethical reasons, to avoid owning a copy of MSWord.

Answer 7

I have had a lot of success with an XSL stylesheet called WordML2Latex. It does an excellent of converting formatting to sensible equivalents. Much more than your one-liner. You first save your Word doc as word XML2003, then apply the stylesheet to it.

Being XSL it is cross platform and you can use it from the command line :-). It is also open source ad modifiable.

It's on CTAN http://www.ctan.org/tex-archive/support/WordML2LaTeX

Answer 8

I think this site can help you.

https://www.docx2latex.com/