- Add very quick pixel-wise image comparison in beginning
- Perform structural similarity check only if pixel-wise comparison failed
- Improve Watermark check
v.0.4.1
Improved (experimental) text detection with `east` when using argument `ocr_engine=east`
Allow usage of `dict` , `list` and `strings` for `mask` argument, to ignore parts from comparison without additional placeholder_file
Allow comparing if images with different sizes (they will be scaled to match before comparison) with argument `resize_candidate`
Better keyword documentation of `Compare Images` Keyword
v.0.3.1
Cleanup and refactor to use poetry for dependencies, build and publish.
Removed some unneeded files and dependencies like pdfminer.six.
Minor changes to functionality.
- PDF Should Contain Strings now also accepts single string as an argument (before it only accepted Lists)
- Removed infinite recursion error when using deepdiff to check [mediabox] object in PDF Content check
v.0.2.0
- Use PyMuPDF (instead of GhostScript/ImageMagick) for rendering PDF files before visual comparison.
- Use PyMuPDF (instead of pdfminer) to parse PDF content like metadata, text, fonts, images, signatures.
- Option to provide watermark_file as parameter for visual comparison
- More content checks on PDF files possible
- Checking move_tolerance for visual comparison is more accurate
Bug fixes for
- Images with different dimensions shwing up as equal
- move_tolerance checks at the edge of image