JMdictParser
JMdictParser is a script for parsing through JMdict data http://www.edrdg.org/jmdict/edict_doc.html for Japanese words that are made up entirely of kanji and are of lengths greater than 1 character. The script returns a Python dictionary containing readings, definitions, and JLPT data for each word and exports that dictionary to compound_dict.json.
Usage
compound_dict.json is already included along with a copy of the JMdict_e file needed to build it.
Simply run the script in the same directory as the JMdict file in order to build a new dictionary.
About the compound_dict dictionary
The current version of the dictionary contains 96390 entries, all of which have relevant definitions, reading, and JLPT data. Regarding usage, kanji words are keys that return sub-dictionaries that use 'reading', 'meaning', 'jlpt' as keys to access the data.
License information
JMdictParser.py is free to use and modify. kanji_dict.json is built using the script found at https://github.com/nramkissoon/Kanjidicparser and utilizes data from the KANJIDIC project. JLPT data from the KANJIDIC projects is subject to conditions found at https://www.edrdg.org/edrdg/licence.html. JMdict data is subject to conditions found at http://www.edrdg.org/.