一右二烏

Purpose

This is a tool to generate printable kanji flashcards, because the free existing sets I found were not satisfying to me. The goal is to have the flashcards:

• Good looking (design was inspired by White Rabbit flashcards)
• Automatically generated

Multilingual support was an early goal, but the current state of dictionaries for language other than English makes it less promising. French is supported, with results not on par with the English version (fewer compounds, more ‘things bigger than their boxes’ cases, etc.).

Example

Generated decks

• https://github.com/makemeunsee/ichimiginikarasu/tree/results/examples/en
• https://github.com/makemeunsee/ichimiginikarasu/tree/results/examples/fr

Samples

Requirements

• stack
• inkscape (to generate pdf/tex from svg strokes diagrams)
• xelatex
• a Kanjidic2 (xml) copy
• a JMdict (xml) copy

Usage

Build

stack build

Create .tex flashcard file for the kanji of a text file

stack exec ichimiginikarasu-exe — -f resources/jpn_words_girardi_kelly -j resources/JMdict -k resources/kanjidic2.xml -n mydeck -i inputfile > mydeck.tex

where ‘inputfile’ is the text file to get kanjis from, and ‘mydeck’ is the name of the deck (appears top right of the flashcards front).

Other parameters can be changed to load custom resources.

Generate .pdf from .tex

xelatex mydeck.tex

Advanced

Launch options are available using:

stack exec ichimiginikarasu-exe — -h

Grepping ‘error’ from the latex compilation can help identify potential typesetting problems (content too large and overlapping typically):

xelatex mydeck.tex | grep ‘error’

Issues

• The issues file lists kanjis that are rendered somewhat off
• tex & pdf files are generated during the execution in the ‘resources/kanji_vg directory’ and are not cleaned
• Some interesting compounds readings and translations are missing, as no good strategy was found to select them. E.g.: 空’s reading ‘そら - sky’ is missing, whereas ‘から - empty’ is present.

Acknowledgements

Lots of existing resources were necessary to create this tool and flashcards. In no particular order:

• The kanji stroke diagrams by Ulrich Apel
• The kanji stroke diagrams converter by Kim Ahlström of jisho.org
• The kradfile, JMdict, kanjidic2 by the Electronic Dictionary Research & Development Group at Monash university
• The Kelly & Girardi Japanese words frequency list
• The kanji similarity matrix, by Lars Yencken

See the LICENSE file for details, and these sites for similar resources:

http://ftp.monash.edu.au/pub/nihongo/00INDEX.html

http://lars.yencken.org/datasets/phd/

Future features

• Multi language: the first, easy step to have a language supported is to provide a localized “radicals_haskelled”.
• Custom input (ie. kanji content specified in individual files, not loaded from dictionaries)
• Web port? To allow non haskellers to use the tool…