Danbooru? Is it something you eat?

Well, aside from the Wikipedia link above, I think a small introduction is in order, at least for the readers coming from PlanetKDE. Danbooru is a kind of image board which structures its data semantically, by having tags attached to images (along with other things, such as favorites, rating, etc.). It can be browsed normally (newest posts, etc.) or by using tags and other properties. Some boards are quite popular in the anime-viewing community. The neat thing about Danbooru (which is by itself a Ruby on Rails application) is the fact that it can provide a REST and POST API to access data. So. it’s technically possible to access such boards programmatically: the API permits not only retrieving posts, but also upload, tag, and perform other operations.

The API could be technically used also for client applications, in order to free the user from using a browser. That is what Danbooru Client is aiming to do.

Introducing Danbooru Client

Danbooru Client fits exactly these needs by providing a GUI to (part of) the Danbooru API.

Features:

• Connect to any Danbooru board (three predefined)
• Download up to 100 images with selectable tags;
• Download or view images with the KDE preferred image viewer;
• Tag semantically the images using Nepomuk.

Requirements:

• PyQt (at least version 4.5)
• PyKDE4 (tested with PyKDE 4.3 only)
• (optional) Nepomuk
• Python (at least version 2.5)

Screenshots

(click to enlarge)

Download & Installation

You can obtain Danbooru Client from kde-apps.org. For the bleeding edge people (but are there such users for such an application?) there is a git repository set up at Gitorious. Once downloaded, you need to use CMake to install the files. Unfortunately due to the way CMake is set up, you’ll need the KDE development headers and a working C++ compiler, even though you won’t need them for the installation.

The installation process is very straightforward:

cd /path/to/source
mkdir build; cd build
cmake -DCMAKE_INSTALL_PREFIX=`kde4-config --prefix` ../
make # This just byte-compiles Python files
sudo make install

Then, just launch “danbooru_client”.

Known limitations

There are plenty for now, it’s just version 0.1:

• Zero documentation (although it’s kind of straightforward to use)
• Empty cells are created when a row is not filled with images
• No support for multi-download
• Untested login/password access
• The interface may be horrid
• Danbooru does not support rating filtering via API, so it’s not currently possible to do so

The client is licensed under the GPL v2 or later. The artwork for the splash screen is also under the GPL and was made by Melissa Adkins.

Once you save the file, however, all the tags are just present in your file name, and nowhere else. LIke that, they’re not that informative. That’s where danbooru2nepomuk comes into play. danbooru2nepomuk is a small Python program that can turn the tags present into the filenames into real semantic tags.

Requirements

As of this post, danbooru2nepomuk works only on Linux, so if you are a Windows user, you’re out of luck. Also, it requires a KDE (tested with version 4.3.1).

danbooru2nepomuk is a Python program, so it requires first of all the Python interpreter, version 2.5 or later. It has been developed to use the Nepomuk semantic desktop framework present in KDE, so you’ll also need PyQt4 and PyKDE4, along with a working Nepomuk installation. Most distributions use the broken Soprano redland backend, which will not work, so I suggest you to switch to the sesame2 backend, which (although dependent on Java) works reasonably well.

Download and installation

Simply get danbooru2nepomuk.zip, rename it to .py from .zip, save it in your PATH, and make it executable. Nothing more than that.

Usage

danbooru2nepomuk is a command line appplication. Its syntax is simple:

danbooru2nepomuk.py [-r] <file or directory>

If you specify a file, it will be tagged directly; if you specify a directory, it will be scanned for files, and those in turn will be tagged. If you add the -r switch to a directory, it will be scanned recursively, while it will be simply ignored if you use it with a file.

Configuration

You can specify a tag blacklist for tags in the filename that you don’t want to get in. To do so, edit the TAGS_BLACKLIST variable on line 41, and add more tags you don’t want Nepomuk to pick up.

Known issues

None that I know of, at least! If you find any, let me know.

danbooru2nepomuk is licensed under the GNU General Public License, version 2.