This is mostly true, though I find issues with how StashDB makes pro-am studios have the performers name. This is problematic for auto-tag since you will often get scenes with a performer being tagged with their studio when it was done for a professional studio. FansDB has not problem with studios this way because theirs are site-specific. And for the pro-ams with their own domains I have to rename to PerformerName.com etc instead of following StashDB’s annoying way of doing it.
Autotag is really useful for pro-am performers if you name it the right way though, since they’re the least likely to be matched in a Stash Box given how much content exists compared to how few people have it.
The hardest one to use it for is tags, you need to have really specific tags for them to be correct, tags you know are only there because someone put them there with intention. People you just need to turn off auto tag for single names that aren’t unique usernames or the same as another performer.
I don’t use it at all. I am more a guy that wants full control. And I dont trust the auottagger.
I Don’t use the performace things as well.
I look that my data itself is tagged because that system also works without stash. I have different tools to Tag new files with brackets
I use my Plugin Bracket Tags
to autoupdate tags.
I’ve just recently moved onto Stash after over a decade of managing my files manually with filenames and folders. As such, I’m using autotag to gradually migrate information from the file paths into Stash.
My path structure is generally {performer}/{date} - {studio} - {title} - {tags} - {resolution}.{extension}. This works well for me because I like picking a particular performer to watch and seeing how their videos have evolved over time as they get more comfortable and try out new kinks.
Sometimes, for more amateur or scattered videos (i.e. each performer only has one video), I’ll do {tag}/{title}.{extension}. And when I’ve downloaded a large amount from one studio I’ve subscribed to, {parent studio}/{studio}/{date} - {title} - {performers} - {tags} - {resolution}.{extension}.
My workflow right now is:
- Import all scenes into Stash.
- Manually create a Stash entry for each folder (performer, tag, or studio as applicable), using scraping where possible to fill out the subfields.
- Run autotag for each performer/tag/studio I created.
- For each performer/tag/studio, use the tagger interface to match against StashDB.
- Manually fill out the unmatched videos, submitting to StashDB as applicable.
I know manually creating the entries isn’t strictly necessary, but this helps break it down into small amounts of work instead of a single list of several thousand videos. Plus, it helps keep all the unmatched videos organized.
Now that I’m using Stash, I’m getting lazier about the file naming (preferring to input it directly into Stash), but I plan to keep the same folder structure because filesystems don’t like it when you dump thousands of files into one folder.
Frictions:
- When a tag is a substring of another tag (e.g. “gangbang” vs. “reverse gangbang”), I get a lot of false matches. But at least it’s predictable, which means it’s straightforward to batch select the offenders and fix it.
- Single-name performers get a lot of false matches. I frequently have to skip autotagging them.
i have occasional new entries to my stash into a varied file structure. if i want to auto tag, i either go to each and every folder OR i have to run a scan on the entire database. it would be nice to be able to autotag ONLY new entries from the last scan.