Skip to main content

Browse pixiv in the terminal

Project description

koneko

GPLv3 license PyPI commits since master dev

Browse pixiv in the terminal using kitty's icat to display images (in the terminal!)

Gallery view Gallery view_square_medium1 Gallery view_square_medium2 Image view Image_view Artist search (artist profile picture on the left, 3 previews on right) artist_search View artists you're following following_users_view

Requires kitty. It uses the magical kitty +kitten icat 'kitten' to display images. For more info see the kitty documentation. Actually, lscat.py uses pixcat, which is a Python API for icat.

Why the name Koneko? Koneko (こねこ) means kitten, which is what icat is, a kitty +kitten

Features (what?)

See the manual for more details

  1. View artist illustrations (ex)
  2. View a post (ex)
    • View related images suggested by pixiv (ex: scroll down from the above example)
  3. View the artists that you are following (or any other user ID) (ex)
  4. Search for an artist/user (ex)
  5. View new illustrations from all the artists you are following (ex)
  6. View recommended illustrations (now called 'discovery') (ex)
  • Navigate between next and previous pages/images
  • Download images (PixivUtil would be more suitable for batch download) in full resolution
  • Open post in browser
  • Browse an offline cache

Why?

  • Terminal user interfaces are minimalist, fast, and doesn't load Javascript that slows down your entire browser or track you
    • Image loading is so much faster, especially if you don't delete the cache

I get 32 trackers on Pixiv. Plus, you have to disable ublock if you ever get logged out

The mobile app even directly tells you Google "and our 198 partners" "collect and use data"! See prompt 1, prompt 2 (Github can't render the images correctly for some reason) and this list

  • TUIs make you cool
  • TUIs with embedded pictures make you even cooler
  • TUIs embedded with pictures of cute anime girls make you the coolest
  • Keyboard driven
  • Familiar, vim-like key sequences
  • I use arch btw

Installation (how?)

See also: manual installation

  1. Install kitty
  2. pip install koneko (or if you use conda...):
  3. Run koneko to login and save credentials
  4. Run lscat 1 7 to help setup the recommended settings; copy to ~/.config/koneko/config.ini. (Don't skip this step! Image display in the terminal is very sensitive to your config!)
  5. See usage for how to use.

Requirements

  • Python 3.8+
  • It has been tested on kitty v0.17.2 onwards, but should work on older versions
  • Operating system: all OSes that kitty supports, which means Linux and macOS.
  • Uses xdg-open (for opening links in your browser) and curl (for safety fallback, see below)
If it crashes (it shouldn't), it might be because pip didn't 'install' the welcome pictures, *and* the script failed to download them for some reason. Try:
mkdir -p ~/.local/share/koneko/pics

curl -s https://raw.githubusercontent.com/twenty5151/koneko/master/pics/71471144_p0.png -o ~/.local/share/koneko/pics/71471144_p0.png

curl -s https://raw.githubusercontent.com/twenty5151/koneko/master/pics/79494300_p0.png -o ~/.local/share/koneko/pics/79494300_p0.png

Usage and manual

See the manual here

Upcoming changelog (in dev branch)

For full changelogs please see releases

Version 0.10.0

Features

  • Browse cache now supports navigating between pages
  • Breaking: Removed image_mode_text_offset config setting, now does nothing
  • Renewed developer guide in HACKING.md
  • Breaking: lscat app: move 'browse cache' and 'specify path' to mode 2 and 3 respectively, moving 'testgallery & 'testuser' to mode 4 and 5 respectively (swapped)
  • Add image mode preview for lscat app
  • Add related images gallery (mode 15): view related and suggested images in image view (mode 2). Press the 'r' key while viewing an image to view related images
  • Add illust_recommended_mode (mode 6): discover recommended illustrations
  • Access browse cache from koneko main; clear cache is now inside browse cache
  • Use koneko q to view frequents and launch their mode
  • Revamped user guide/manual

Bug fixes

  • Fixed browse cache crashing on empty directory
  • Fixed docs: mode 5 cannot go back to main screen
  • Fixed reloading in a gallery mode leading to user prompt
  • Browse cache now no longer accepts an invalid directory to display on
  • Fixed experimental image mode previews index errors due to it trying to download more images than possible
  • Almost fixing experimental image mode previews by saving previous cursor position and restoring it later; but still a bit unstable
  • Fixed grammar in prompt: "a image command" -> "an image command"
  • Improved stability (eg weird race conditions, files downloaded in wrong places) by not cd-ing into directories
  • Fixed logging in with a new session every time by caching the login token.
  • Stop prefetch if next_offset is invalid

Code maintenance

  • lscat show_instant() now inspects the cls passed in instead of a bool argument
  • "Inlined" some abstract methods in AbstractUI into attributes
  • Make all ui.Image methods depend on only data.ImageData like the associated free functions, so those methods can be moved to a subclass of data.ImageData (the new ui.Image class extends IO behaviours from the data class)
  • Pass in path and name to api.download instead of cd-ing into the dir and manually renaming downloads, simplifying a lot of code in the process
  • Replace pure.Map with list comprehensions. Can't force functional programming into Python after all
  • Use funcy.autocurry() instead of partial
  • Split up files.filter_dir()
  • Use ws_picker() in _pick_dirs_picker()
  • Add next_offset property to AbstractData (moved calculation from ui to data.py)
  • Change list comprehensions assigning to _ into normal for loops
  • Split up _parse_and_download()
  • Improved order of methods in GalleryData: now sorted by required methods/properties by the interface/abstract class, and unique methods/properties.
  • Rewrite UserData class to be more like GalleryData
  • Move download_path() method to AbstractData
  • Extract view_post_mode() and view_image() to ABC

Roadmap

Features

  • In-depth usage documentation; use letters to represent modes (at least in public docs) rather than numbers
    • Make sure the diagrams in HACKING.md is up-to-date
    • Offline mode: Using lscat_app as a base for offline mode, rather than koneko
      • Handle moving around pages
  • Option to save username, but prompt for password (and not save it) every time

Known bugs

  • Prefetch thread still running (downloading) hangs the entire app, even when user quits. Cannot use daemon threads as it still hangs then noisly aborts. Changing prompt.ask_quit() into a UI method so that it can pass a threading.Event() to downloads, doesn't work either as all the downloads has already been submitted to the ThreadPoolExecutor before the user is quick enough to send 'q'. The only way is to interrupt the urllib download process, which is going to be unsafe if you don't know what you're doing.
  • In the logs, urllib3 warns that Connection pool is full, discarding connection: i.pximg.net. See customising pool behaviour from urllib3.
  • There seems to be a delay between entering koneko and startup, but the delay is before the first line of the script even executes. Import time is fast. pip install using the wheel seems to reduce the delay. Directly running the script using python koneko/main.py or python koneko/lscat_app.py is faster as well. Seems like it's a delay in going to miniconda/lib/python3.8/site-packages/koneko.egg-link (installed with python setup.py develop)?

FAQ

  • I'm having problems with lscat

For the best experience use the terminal in full screen, unless your screen is big enough. Moving and resizing it abruptly will not be good for icat, which is really kitty's problem not mine. Extra information can be disabled from being printed.

You can also use versions less than v0.5.1, which retains legacy support for the original lsix shell script. Note that I've never really tested it, which is why I decided to be honest and depreciated legacy support from v0.6 onwards. The current lscat API has matured to the point where it's simple to write a replacement anyway.

Contributing

  • Fork it
  • Edit the files on your fork/branch
  • Run tests with pytest testing/ -vvvv -l -s (--inte for integration tests)
  • Submit a pull request

Tips:

  • See HACKING.md to understand the code.
  • If your git client complains about committing to master, just remove .pre-commit-config.yaml
  • If you want to, you can create an issue first. Ask any questions by opening a new issue.
  • If you're encountering/fixing a bug and you're stuck, try clearing the cache. For example, a bug might have downloaded to the wrong folder, but after fixing the bug, you need to clear the cache, otherwise it would not download anything and display the wrong contents.

See the rest in CONTRIBUTING.md

Unit tests

Run pytest testing/ -vvvv -l -s. Add --inte for integration testing, but don't be surprised if it fails

Build and upload to PyPI

When test installing with pip, don't forget to use pip install . or python setup.py install, not pip install koneko (which will grab from latest stable version). (Yes, I made the same mistake again)

Test installing with pip install ., python setup.py install, python setup.py develop, and python -m koneko.main is now automated.

Bump version info in __init__.py, setup.py, and CONTRIBUTING.md

python setup.py sdist bdist_wheel
twine upload dist/*
pip install koneko --upgrade

Trackers avoided

This is a list of trackers present when you use the official pixiv website or app. koneko frees you from them.

Nine trackers in the Android app, according to exodus:

  • Amazon Advertisement
  • AMoAd
  • Google Ads
  • Google CrashLytics
  • Google DoubleClick
  • Google Firebase Analytics
  • Integral Ad Science
  • Moat
  • Twitter MoPub

Advertisers from pixiv's privacy policy:

  • Looker
  • Repro
  • Qualaroo
  • DDAI(Date Driven Advertising Initiative)
  • YourAdChoices
  • Rubicon Project
  • i-Mobile Co., Ltd.
  • Akinasista Corporation
  • Axel Mark Inc.
  • AppLovin
  • Amazon Japan G.K.
  • AmoAd Inc.
  • AOL Platforms Japan K.K.
  • OpenX
  • Google Inc.
  • CRITEO K.K.
  • CyberAgent, Inc.
  • Geniee, Inc.
  • Supership Inc.
  • GMO AD Marketing Inc.
  • F@N Communications, Inc.
  • Facebook Inc.
  • Fluct, Inc.
  • Platform One Inc.
  • MicroAd Inc.
  • MoPub Inc.
  • Yahoo! Japan Corporation
  • United, Inc.
  • 株式会社Zucks
  • PubMatic, Inc.
  • Liftoff Mobile, Inc.
  • Mobfox US LLC
  • OneSignal
  • Smaato, Inc.
  • SMN株式会社
  • 株式会社アドインテ

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

koneko-0.10.0.tar.gz (10.0 MB view hashes)

Uploaded Source

Built Distribution

koneko-0.10.0-py3-none-any.whl (60.1 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page