Batch conversion of docx to clean HTML

Question

I'm starting to wonder if this is even possible. I've searched for solutions on Google and come up with nothing that works exactly how I'd like it to.

I think it'd benefit to explain what that entails. I work for database group at my university's IT department. My main job is to take specs of a report in a docx file, copy that over to dreamweaver, fix some formatting, and put it onto their website. My issue is that it's ridiculously tedious to do this over and over. I figured, hey, I haven't written anything in C# for some time now, perhaps I could write an application to grab a docx file, convert it to HTML, fix the CSS, stick the header, and footer from the webpage on there, and save the result. I originally planned to have it do one by one, but it probably wouldn't be difficult to have it input a list of files and batch convert.

I've found these relevant topics on how to accomplish this, but they don't fit my needs well enough.

http://www.techrepublic.com/blog/howdoi/how-do-i-modify-word-documents-using-c/190

This is probably fine for a few documents, but since it's just automating an instance of Word, I feel like it'd be slow and memory intensive. I'd prefer to avoid opening and closing an instance of Word 50+ times.

http://openxmldeveloper.org/articles/333.aspx

This is what I started using. XSLT had the benefit of not needing word to be installed nor ran for each file. After some searching I got a proof of concept working. It takes in a docx file, decompresses it, grabs the document.xml from that, and uses the DocX2Html.xsl file I scavenged from OpenXML viewer. I believe that was originally provided by MS for sharepoint servers to provide the ability to render word documents in a browser. Or something along those lines.

After adjusting that code to fit my needs, and having issues with the objXSLT.Load () method, I ended up using IlMerge to make the XSL into a DLL. No idea why I kept getting a compile error when using the plain old XSL file, but the DLL worked fine, so I was satisfied. Here (http://pastebin.com/a5HBAakJ) is my current code. It does the job of converting docx to HTML just fine (other than random spaces between some words), but the result file has ridiculously ugly HTML syntax. An example of this monstrosity can be found here (http://pastebin.com/b8sPGmFE).

Does anyone know how I could remedy this? I'm thinking perhaps I need to make a new XSL file, as the one MS provided is what's responsible for sticking all those tags and extra code in there. My issue with that is that I don't know anything about how to do that. Perhaps there's an alternative version already out there. All I'd need is one that will preserve tables and text formatting. Images aren't needed.

Answer 1

This looks like just what you need: http://msdn.microsoft.com/en-us/library/ff628051(v=office.14).aspx

The author Eric White blogged about his experiences developing that tool. You can see that list of posts on his blog here: http://blogs.msdn.com/b/ericwhite/archive/2008/10/20/eric-white-s-blog-s-table-of-contents.aspx#Open_XML_to_XHtml

Answer 2

Since I'm a big fan of Aspose.Words, a commercial library to create/process Word documents, I would do something like:

1. Open the Word document with Aspose.Words.
2. Save the Word document as HTML.
3. Use something like SgmlReader or HTML Agility Pack (or even Regular Expressions if it is suitable) to remove unwanted HTML tags/attributes.

Since you wrote you work at an university, I'm not sure whether commercial packages are an option, though.