Sandro Mani
Sandro Mani
The contrast/brightness spinboxes could be bound to settings using the `ADD_SETTING` macro, so that the values are persistently stored in the application settings. Question though is whether having them persistently...
Thanks a lot! The Fedora dependency issues are being resolved. Regarding the appimage workflow, would it be possible to also build the qt variant instead of only gtk? Which qt6...
Sorry this fell of my radar. If you have no objections I would try to use a fedora base image to get the qt6 variant building, since qt5 is pretty...
See [1], the outer container is indeed ubuntu, but you can run a fedora container within that container. [1] https://github.com/manisandro/gImageReader/blob/master/.github/workflows/build.yml#L7
This is expected - in hOCR mode, the layout detection is performed automatically and indeed the resulting document tree is based on the layout detection. Currently, hOCR will always process...
You need to recognize in hOCR + PDF mode
You need to choose the PDF with invisible text overlay output mode:
Hmm, I guess this is releated to qt6? There is no light/dark mode logic in gImageReader itself.
Thanks, Note that there are many other locations in the code where `html` is the expected extension for hOCR files.
Does it help to select this page segmentation option: ?