mikan 0.2.0

A Japanese language manipulation library

Mikan

Mikan is a Japanese language manipulation library.

import mikan

eat = mikan.IchidanVerb('食べる', 'たべる')

want_to_eat = eat.conjugate(mikan.VerbForm.TAI)

did_not_want_to_eat = want_to_eat.conjugate(mikan.AdjectiveForm.PAST, negative=True)

three_small_animals = mikan.Number(3) + mikan.Counter('匹', 'ひき')

cat = mikan.Word('猫', 'ねこ')

sentence = cat + 'が' + three_small_animals + did_not_want_to_eat + 'です'

reading = sentence.readings.pop()

print(sentence) # 猫が3匹食べたくなかったです
print(reading) # ねこがさんびきたべたくなかったです
print(mikan.to_romaji(reading)) # nekogasanbikitabetakunakattadesu

Motivation

I wanted to learn japanese, but it proved to be hard. So I wrote some Python app mixing a dictionary and a SRS system, with generated items (like inflected verbs), to help me. But the code became messy and hard to manage, so I moved the generation code into a Python library. I named it Mikan, because I like how the みかん hiragana looks like (though fruit names tends to be written with katakana...). Now I'm working on this library and my app is kind of stale. And I'm still not speaking japanese.

Words, writings and readings

In Mikan, most class classes inherits from the Word class. A word is a list of writings. Each writing can also be a reading.

For example:

import mikan

# create a word with three writings: 日本, にほん and にっぽん
# にほん and にっぽん are readings too
japan = mikan.Word('日本', 'にほん', 'にっぽん')

print(japan.writings) # ['日本', 'にほん', 'にっぽん']
print(japan.readings) # ['にほん', 'にっぽん']

Words can be combined to create compounds:

import mikan

japan = mikan.Word('日本', 'にほん')
cooking = mikan.Word('料理', 'りょうり')

japanese_cooking = japan + cooking
print(japanese_cooking.writings) # ['日本料理', 'にほん料理', '日本りょうり', 'にほんりょうり']
print(japanese_cooking.readings) # ['にほんりょうり']

Verbs

Mikan provides 5 verb classes:

• GodanVerb for class 1 verbs
• IchidanVerb for class 2 verbs
• KuruVerb for 来る (to come) and 来る-based compounds (ie: 持ってくる)
• SuruVerb for する and する-verbs (ie: 勉強する)
• IkuVerb for 行く and 行く-based compounds (ie: 持っていく)

Verbs have a conjugate method that returns the inflected form of a verb.

conjugate(forms, negative=False, polite=False)

The available verb forms are:

• VerbForm.PRESENT
• VerbForm.PAST
• VerbForm.IMPERATIVE
• VerbForm.TE_FORM
• VerbForm.CONDITIONAL_EBA
• VerbForm.CONDITIONAL_RA
• VerbForm.PRESUMPTIVE
• VerbForm.VOLITIONAL
• VerbForm.POTENTIAL
• VerbForm.PASSIVE
• VerbForm.CAUSATIVE
• VerbForm.TAI

Adjectives

Mikan only has support for i-adjectives yet, available using the IAdjective class.

They have the same interface than verbs, with the following forms:

• PRESENT
• PAST
• TE_FORM
• ADVERB

Numbers

Numbers are created with the Number and can be created from python integers or from kanjis:

import mikan

a1 = mikan.Number(42)
a2 = mikan.Number('四十二')

assert a1 == a2

Counters

Mikan includes a generic class Counter and some specific classes for reading exceptions:

• DayHourCounter for hours of day counter (よじ, くじ)
• MonthDayCounter for days of month counter (ついたち, ふつか, ...)
• MonthCounter for months counter (しがつ, しちがつ, ...)
• TsuCounter for the generic つ counter (ひとつ, ふたつ, ...)

Combining numbers and counters produce the required reading changes:

import mikan

number = mikan.Number(1)
hon = mikan.Counter('本', 'ほん')

assert 'いっぽん' in (number + hon).readings

Datetime

Mikan has a wrapper for Python datetime.date class:

import datetime
import mikan

mikan_date = mikan.Date(datetime.date(year=2020, month=3, day=16))

print(mikan_date) # 2020年3月16日

Romaji

Some functions are provided to convert between romaji and kana:

• to_romaji converts string in kana to romaji/hepburn
• to_hiragana converts string in romaji/hepburn to hiragana
• to_katakana converts string in romaji/hepburn to katakana

import mikan

friday = mikan.to_romaji('きんようび')
kinyoubi = mikan.to_hiragana(friday)

print(friday) # kin'youbi
print(kinyoubi) # きんようび