https://youtube.com/watch?v=tigTaObQORM
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v476/Hydrus.Network.476.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v476/Hydrus.Network.476.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v476/Hydrus.Network.476.-.macOS.-.App.dmg
linux
tar.gz: https://github.com/hydrusnetwork/hydrus/releases/download/v476/Hydrus.Network.476.-.Linux.-.Executable.tar.gz

I had a good week integrating two new features: autocomplete tag search in the client api, and saved 'post times' from downloaders.

post times

'Modified time' is neat, but it isn't super useful for downloaded files--since the file was only just added to your hard drive, it'll always be the same as import time. This week I integrate the 'source time' we parse from various websites to improve the modified time for downloaded files. The objective is to make 'modified' a fairly decent 'this file was completed around this time' number for searching and sorting purposes.

I'm being careful not to overwrite anything. The client now saves its best 'source time' for every different site it downloads from and then the earliest of those + modified date is used as the aggregate modified date. You don't have to do anything, but with luck you will see your new watcher and gallery files start to get some nicer modified times in the media viewer and thumbnail right-click menu.

There are many potential future expansions here. I can grab better post times from sites, show and edit every stored timestamp in UI, allow clever search and sort of those specifically (e.g. 'sort all these files by their danbooru post time), and most importantly make some sort of maintenance system to retroactively fetch a good post time for all the files we downloaded before post times were saved. This is just a first step.

I integrated the new 'archive time' too during this work. This now shows in the media viewer and thumbnail right-click menu similarly and can be sorted by. Search will come soon. I also want to think about optionally filling in some estimate dummy data here for all the files we archived before timestamps were tracked.

client api autocomplete search

A user has helpfully written Client API routines for autocomplete tag search, which is something I have had trouble fitting in. I appreciate the work. This should let the various tools that use the Client API do more tag browsing in future.

The documentation is here: https://hydrusnetwork.github.io/hydrus/developer_api.html#add_tags_search_tags

There are several ways to expand this too, so if you are an API dev interested in it, let me know how it goes.