Kotobase is a Japanese language Python package which provides simple programmatic access to various data sources via a pre-built database which is updated weekly via a GitHub action.

Detailed documentation is avilable at the GitHub Pages Site

Kotobase uses data from these sources to build its Database.

• JMDict : Japanese-Multilingual Dictionary.

• JMnedict : A dictionary of Japanese proper names.

• KanjiDic2 : A comprehensive kanji dictionary.

• Tatoeba : A large database of example sentences.

• JLPT Lists : Curated list of Grammar, Vocabulary and Kanji separated by Japanese Language Proficiency Test levels, made available on Jonathan Weller's website.

The licenses of these data sources and the NOTICE is available at docs/licenses in this repository.

• Comprehensive Lookups → Search for words (kanji, kana, or romaji), kanji, and proper names.

• Organized Data → Get detailed information including readings, senses, parts of speech, kanji stroke counts, meanings, and JLPT levels formatted into Python Data Objects.

• Example Sentences → Find example sentences from Tatoeba that contain the searched query.

• Wildcard Search → Use * or % for wildcard searches.

• Command-Line Interface → User-friendly CLI for quick lookups from the terminal.

• Self-Contained → All data is stored in a local SQLite database, so it's fast and works offline.

• Easy Database Management → Includes commands to automatically download the latest pre-built database from the public Drive or download source files and build the database locally.

• Install the package

pip install kotobase

This will install the kotobase package and its dependencies, and it will also make the kotobase command-line tool available in your shell.

• Pull the Database from Drive or Build it locally by running of the commands below in the environment you installed kotobase

# Pull from Drive
kotobase pull-db
# Build locally
kotobase build

The database will be downloaded or built internally in the package at kotobase/src/db/kotobase.db and will be available for use.

Kotobase can be used as a command-line tool or as a Python library.

The kotobase command provides several subcommands for different types of lookups.

The lookup command is the most comprehensive way to search for a word.

kotobase lookup 日本語

This will show you dictionary entries, kanji information, JLPT levels, and example sentences for the word "日本語".

Options:

• -n, --names: Include proper names from JMnedict in the search.
• -w, --wildcard: Treat * or % as wildcards in the search term.
• -s, --sentences: Specify the number of example sentences to show.
• --json-out: Output the full results as a JSON object.

To get information about a specific kanji character:

kotobase kanji 語

This will display the kanji's grade, stroke count, meanings, on'yomi, and kun'yomi readings, and JLPT level.

To check the JLPT level for a word or kanji:

kotobase jlpt 勉強

You can also use Kotobase in your own Python code.

from kotobase import Kotobase

kb = Kotobase()

# Comprehensive lookup
result = kb.lookup("日本語")
print(result.to_json())

# Get info for a single kanji
kanji_info = kb.kanji("語")
print(kanji_info)

# Get example sentences
sentences = kb.sentences("勉強")
for sentence in sentences:
    print(sentence.text)

Kotobase relies on a local SQLite database.

You can also build it from the source files yourself.

The following commands are available for managing the database:

• kotobase pull-db: Downloads the pre-built SQLite database from a public Google Drive Folder. This file is overwritten every week with a rebuilt database from updated sources. The rebuilding and overwriting is managed by a GitHub action in this repository.

• kotobase build: Builds the SQLite database from the raw source files. This will download the latest version of the source files (Except Tanos JLPT lists which are shipped with the package itself.) and build the database locally.