読み
よみ ⋅ yomi
(noun)
- reading
- reading (of a kanji, esp. kun reading)
Mate
/meɪt/
- (colloquial, Britain, Australia, New Zealand, sometimes elsewhere in the Commonwealth) A friend
An Optical Character Recognition driven dictionary. Inpired by the pop-up dictionary browser extension Yomichan (https://github.com/FooSoft/yomichan)
Created as an entry for the UQ Computing Society Hackathon 2022
Choose an image for Yomi-mate to scan through and it'll show you any Japanese text it detects. Then, you can highlight a particular word or character and it'll show you a dictionary entry for it.
Shows you details like the JLPT grade (if any), radicals, readings, as well as its meanings.
It also checks if the selection contains a name, and shows its variations.
Linux / MacOS (untested): Python 3.10 or later
Windows: Untested
In the project's root folder:
# Create and activate virtual environment
python -m venv env
source env/bin/activate
# Install requirements
pip install -r requirements.txt
Then, to run the application:
# Activate virtual environment
source env/bin/activate
python ./frontend.py
- Long & unfriendly installation process
- Untested on Windows
- Inaccurate readings
- Unreliable GPU support
- UI could be nicer
Frankly, this project is a bit of a mess. It's built on spaghetti hackathon code and is a hassle to get working for most users. I plan on reworking the entire thing to deal with these issues ~some~ time in the future.
EasyOCR (Optical Character Recognition)
JamDict (JMDict Parsing & Querying)
Pykakasi (Kana/Hiragana/Kanji to Romaji conversion)