Doc support?

[dridk]

Hi,

I just discovered this tool.
Please, tell me that you support the old doc ( not docx ) format.
Actually, there is only antiword and Apache tikka that work with it.

[dridk]

I've just tested, the doc format is not supported.

I think it's really important to have this support, because even if the format is no longer used, the number of documents existing with this format is very large, if not the largest.
For example, in hospital, there are a lot of trouble retrieving the content of these old documents, impacting research and care.

Having explored everything, here are some things that may help:

Using openoffice or unoconv

https://github.com/unoconv/unoconv
https://www.openoffice.org/fr/

Openoffice doesn't handle parallel execution very well. You can't run the conversion of millions of documents in parallel. Even with unoconv, which tries to remedy this.

antiword

https://github.com/grobian/antiword
This tool does not work on all doc format. The code no longer seems to be maintained.

Apache Tika

https://tika.apache.org/

Clearly the best tool and the one I use today to extract text from doc format. It handles this very well, work in parallel and it is written in Java. It is a simple command line to run.

Creating a markitdown plugin

However, these tools lack functionality, such as the ability to retrieve formatting. So it would be really nice if a microsoft tool could support the historical microsoft format.

Microsoft has published the format specifications here . So there is no need for retro engineering.
https://learn.microsoft.com/en-us/openspecs/office_file_formats/ms-doc/ccd7b486-7881-484c-a137-51170af7cc22#published-version

I don't think it would be that difficult to make a minimalist parser in python based on the apache Tika code:
https://github.com/apache/tika/blob/2c9b6ab48a169f850bdfe2f4676afa6af2c1c540/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java

[dridk]

Hi,

I finally succeed to create a greate reader for the old legacy word doc format.
It works like a charm ( thanks Claude and Microsoft to open the standard specification )

It is written in rust
https://github.com/dridk/unword

I m going to create a PR for markitdown.