Thank you very much for your interest.

Will this library support the newer versions of DocTR?

Yes.

As an integration project deepdoctection will always lag behind the cutting edge development of libraries it integrates. The version I have pinned in setup are those that have tested against in integration tests. This pin will ensure that deepdoctection in version xx will run with e.g. DocTr with version yy. I try to avoid upper version bounds but if there is a breaking change I am forced to pin a specific version.

With respect to DocTr I had to pin to to 0.90 because 0.91.0 requires numpy >=2.0 and some tests failed (where I was not able to fix them quickly). Later versions of DocTr will also require Python 3.10 whereas deepdoctection will support 3.9 at least for some more month. DocTr versions higher than 1.0 will drop Tensorflow support whereas deepdoctection will support Tensorflow until Python 3.10.

To conclude:
It is possible that DocTr will already work with a higher version for your purpose.
I do not check if all DocTr run with deepdoctection even though they should (and feel free to open an issue if you run into a problem).
I cannot give a definitive timeline but checking compatibility with higher DocTr versions is regularly on the bucket list.