Unstructured File Converter: maintenance and refactoring

Open anakin87 opened this issue 9 months ago • 1 comments

This component was developed a while ago when the Unstructured ecosystem was smaller and simpler. It evolved over time and now includes: the open-source library, free and paid APIs, API clients, the Docker image (for running the API locally).

TODO

~Ensure compatibility with unstructured-client>=0.30.0 (see #1416).~ This was magically fixed in #1841
Evaluate whether we can remove the dependency on the unstructured library. Initially, this was the only way to programmatically query self-hosted APIs, but we should explore if the client alone is sufficient.
Verify that our integration correctly works with APIs hosted by Unstructured (review URLs, etc...) or fix any issues.

Feb 18 '25 13:02 anakin87

One thing to have in mind also: renaming the paths argument to sources with the same type as other converters to be in line with what other converters expect :)

Mar 21 '25 08:03 lambda-science