The latest pyocr version now allows running it with the latest tesseract
version. Hopefully this means better OCR results.
I am not sure about whether there are binary packages for the latest
tesseract. But on my setup it was simply a case of checking out the
master branch [0] and compiling + installing from there. It seems to work
fine with paperless as well.
[0] https://github.com/tesseract-ocr/tesseract
This change includes a filthy hack around how Django handles
change_list_results.html -- I'm not thrilled with it, but it's as
elegant as I could come up with. I'm happy to field alternative ideas.
More details can be found in `documents/templatetags/hacks.py`
Specifically, this merge includes a significant facelift to the
documents listing page, moving away from the tabular layout and toward a
tileset look. I tried fiddling with the colours, but I just don't have
any skills in that area, so we're all stuck with Django'd default
colours until someone with an eye for colour can submit a better CSS.
There appears to be quite the mess out there with regard to how DRF
handles filtering. DRF has its own built-in stuff, but recommends
django_filter for the advanced stuff, which has its own overriding
module that explodes with this message when used as per the
documentation:
AttributeError: 'NoneType' object has no attribute 'DjangoFilterBackend'
Then there's djangorestframework-filter, another package that claims to
do the same thing, that does everything just differently enough that
nothing worked while I had it enabled.
I ended up using django_filter, but doing so importing each element
explicitly, rather than just using the recommended (and broken, at least
in this project) method of:
import django_filter.restframework as fitlers
Anyway, this should bring the dependencies up to date, and strips out a
lot of redundant code.
