hydrusnetwork / hydrusnetwork/hydrus
Filter for suggesting highly relevant tags
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 3.2k
- Forks
- 207
- PR merge metrics
- No merged PRs in 30d
Description
Filter for suggesting highly relevant tags.
Elo: For us it could be an additional system/filter where you would go through similiar like thru archive/delete filter but instead for tags.
Files would be sorted by most relatable, to lowest, so after a certain point you would stop the filter or set a min accuracy %.
A search bar would needed to specify the tag you look for, or the tag you want to search in. An entry would be created like "hash tag true/false" example "d26de... species:elf true" If true add this tag to the file, if false just save that information and skip to next file.
From what I've understood it'd work a bit more like the dupe filter rather than a/d filter, but instead of asking if these two files are the same/alternate/etc. based on file similarity, it'd ask if "this tag" belongs on this image based on how often "this tag" is used together with the already present tags.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reviewing the existing archive/delete filter and duplicate filter workflows, which the request compares against. Determine how tag co-occurrence and relevance would be calculated, how a search or minimum-accuracy threshold would work, and how true/false decisions would be stored; the feature is done when this workflow is specified and implemented for suggesting tags.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- data
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100