Added transactions to bulk import
Renamed beginTransaction to begin_transaction for naming consistency
Updated cron uploader to handle bulk import transactions
Changed ffmpeg thumbnailer to instead output a full-size png which is forwarded to the image thumbnailer, to allow it to take advantage of all available scaling options
Looks like github updated their default container, so now it contains
postgres and postgres-client by default - we just need to manually
launch the installed-but-disabled daemon
This change implements a simple category-based alphabetical sorting system that puts all tags containing ':' in front of general tabs. It also groups them together for easier styling into columns, grids, or even opening the door for drop-down categories in the edit menu. A much needed feature for me since I have hundreds of tags and I manage them all by my own.
Removed count limit, the cron job now checks the max PH execution time and auto-stops itself at 80% of that value.
Now skips os-specific image cache files like thumbs.db and the __macosx folder.
Changed failed folder re-deployment to allow re-deploying to populated queue, making it easier to re-process lots of failed batches all at once.
Changed page to output as a stream, allowing a long-running process to provide output as it runs rather than just at the very end.
Changed import loop to use the yield convention, allowing faster consumption of found files and lower memory use overall.
Changed mime type map to deal with the reality that certain file types have multiple extensions and/or multiple mime types, as well as constants supporting all of the data. Created new functions using the updated mime type map to resolve mime types and extensions. Updated various items around the project that determine mime/extension to take advantage of the new functions.
