Search, filter, inspect, download posts from Danbooru and local folders.
Search, filter, inspect, download posts from Danbooru/Safebooru and find them back from the command line without setting up any service or database.
Suitable for daily terminal usage, usable for scripts and as a library for Python 3.6+.
See also lunasync to automatically download and keep in sync particular tags using lunafind, similar to Danbooru tag subscriptions or saved searches.
- Operate on tag searches, URLs or file paths
- Combine results from multiple searches
- Specify custom page ranges, or just get everything
- Filter and order booru results to work around the two tags search limit
- Optional partial/fuzzy tag matching for filter and local searches
- Search downloaded posts by tags as if they were on a booru, without the hassle of setting up one
- Instant results from local searches in most cases after indexing
- Fast multithreaded downloads; 8 downloads in parallel by default
- Supports operating on post media (image, ugoira WebM, etc), info, notes,
- Getting the URLs or file/folder paths
- Printing on standard output
Local searches performance
The first time a local post search is done, an index file to speed up future searches will be automatically created and updated when new post directories exist or are removed.
Test with ~165 000 posts
AMD FX-8300 (8 cores, 3.3GHz), TOSHIBA DT01ACA2 7200 RPM HDD, BTRFS file system, Void Linux 4.18.14 x86_64:
- It takes about 2m30s - 3m to index everything from scratch
- After this, search results start coming instantly unless
- Searches finish completely in 8-20s
Test with ~60 000 posts
Celeron B815 (2 cores, 1.60GHz), 5400 RPM HDD, BTRFS file system, Void Linux 4.18.20 x86_64:
- About 4mn to index everything from scratch
- Results come instantly
- Searches finishes in ~20s
Command line usage
Downloading to the current folder every post tagged blonde and 2girls (default booru is https://danbooru.donmai.us):
lunafind "blonde 2girls" --limit 200 --pages all --download .
Searching through the posts we just downloaded, printing image paths for the results:
lunafind "blonde blue_eyes rating:s score:>5" --source . --show-location media
lunafind --help for all options and examples.
No real documentation yet. Three main classes are provided:
Post: represents a local or remote single post, with its info, media, notes and artcom (artist commentary).
Album: works like a dictionary of
Post, where keys are the post IDs. Has magic methods and operators to facilitate working with them. Can be filtered, ordered, downloaded, and more.
Stream: an efficiant lazy iterator yielding posts. Can be filtered and multithread-downloaded.
Reproducing the command line examples in the section above:
from lunafind import Stream Stream("blonde 2girls", limit=200, pages="all").download() for post in Stream("blonde blue_eyes rating:s score:>5", client="."): print(post.get_location("media"))
Requires Python 3.6+ and pip (for automatic easy install).
Tested on GNU/Linux and Windows 7, probably works on OSX and other POSIX systems.
pip3 install -U lunafind
Release history Release notifications | RSS feed
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.