baskaufs
commented
2 years ago Relevant scripts:
- [x] The script get_image_dimensions.ipynb was used to get the dimensions of the lower-res images to check for dimension reversals. Its output was the file image_dimensions.csv, which has the filename, accession number, pixel dimensions, and subdirectory name. It probably needs to be re-run on the highres images, which I now have access to. It probably should also grab the file extension (to get the image type: tif, jpg, png, etc.) and the file size since those factors might also be useful for determining which images are "high res".
- [x] image_reconciliation.ipynb - used to determine copyright status of images. Output: items_status_abbrev.csv and items_status_abbrev.xlsx, which contain only the copyright-relevant fields and a status assessment for each image and items_with_status.csv, which contains all metadata (including image URLs) plus the status assessment.
- [x] The script determine_item_type.ipynb determines the "dimension" of the work using the following categories: 3D, 2D, poster, and ambiguous. The output is in works_classification.csv.
The "highres" images need to be sorted by three criteria:
The first category to be uploaded would be really highres, 2D images that are public domain. That will be the simplest case.
The second category will be highres, 3D images that are public domain.
After that we would need to decide whether to move towards making highres copyrighted images available via IIIF (but not Commons) if that's allowed, or towards uploading lower resolution images and planning to replace them later with better images. Or to do nothing and wait for the higher res images to be available.