radtoo / gazer

experimental local booru and booru client

Gazer

This is an experimental booru client and image archive tool I made to consolidate my own image browsing and archiving habits. It combines a simple locally running image booru with a client for browsing some of the common booru sites. (gelbooru, yandere, konachan, etc) and has a simple scraper for automatic downloading.

Features

Application runs in a local web server rather than as a traditional gui interface. Can be used either locally on a single computer or accessed over a local area network (for example from a homeserver or media pc).

Local Booru

• mirrors all tags from source
• tag and multitag search
• viewing statistics (by image and by tag)
• sort by date created/score/local views
• adjust thumbnail size
• adjust thumbnails per page

Booru Client:

• supports gelbooru and moebooru based sites
• uses locally cached images when possible (saves bandwidth)
• 1 click archiving in client

Scraper

• supports gelbooru and moebooru based sites
• scrape and archive posts based on any tag or set of tags

Deep Danbooru Integration

• local archive has extra tags inferred from image using machine learning black magic
• inferred tags fully searchable, enabled via toggle

Why not use hydrus or x other software

I found hydrus to have a lot more features than I needed. What I really wanted was just a personally curated booru with an easy way to browse and archive new images.

Running

Code is still very experimental so I don't have a nice packaging solution ready, needs to be run from commandline and accessed via browser.

Linux/Unix/Mac

Install any version of python3 then navigate to the gazer root directory and run the following commands in your terminal client.

python -m venv venv
source venv/bin/activate
python setup.py install
python gazer

Open a browser and go to localhost:5000 or 127.0.0.1:5000

Enable DeepDanbooru Integration

chmod +x enable_dd.sh
./enable_dd.sh

Windows

Install any version of python3 then navigate to the gazer root directory and run the following commands in your terminal client.

python -m venv venv
.\venv\Scripts\activate
python setup.py install
python gazer

Open a browser and go to localhost:5000 or 127.0.0.1:5000

Docker

If you have docker set up you can use the included dockerfile to run the service. Note by default this dockerfile has deepdanbooru integration enabled.

docker build -t gazer .
docker run -p 5000:5000 gazer

Open a browser and go to localhost:5000 or 127.0.0.1:5000

Getting Started

Setup Auto Scraper

Find the configuration file at gazer/gazer/scraper_data/tag_set.json

[
{"name":"Cute Autist", "source":"safebooru", "tags": ["constanze_amalie_von_braunschbank-albrechtsberger", "1girl"]}
]

You can edit this file to have any number of "tag sets". Each tag set must have 1 or more tags and list a supported booru to download these images from as its source ie.

[
{"name":"Cute Autist", "source":"safebooru", "tags": ["constanze_amalie_von_braunschbank-albrechtsberger", "1girl"]},
{"name":"Cute Maid Autist with gun", "source":"safebooru", "tags": ["constanze_amalie_von_braunschbank-albrechtsberger", "maid", "gun"]}
]

Scraper will attempt to download all images from the source that match your tags. The scraper will automatically start with gazer. If you update your tags in the config file while gazer is running or want to scrape any new images: go to 127.0.0.1:5000/scraper and press the 'restart' button.