Showcase visualizations about the common Japanese morphemes in YOASOBI's songs' lyrics.
Morphemes are the smallest units of meaning in a language.
- Japanese morphemes were extracted using SudachiPy and romanized using Cutlet.
- Lyrics were based on genius.com
- Lyrics of 29 YOASOBI songs
Common Japanese Morphemes in YOASOBI Lyrics:
- Download Docker Desktop
- Download docker-compose.yml file from this repo.
- Place the Docker Compose file in a directory of your choice.
- Create
lyrics
directory in the same directory that you place the Docker Compose file.
-
Below is a template of the JSON file.
{ "songs": [ { "title": "", "romanji_title": "", "lyrics": "" } ] }
- The JSON file contains an array of
songs
. - Each song in the array has three fields:
title
: The original Japanese title of the song.romanji_title
: The romanized version of the title.lyrics
: The full lyrics of the song in Japanese.
- Create a new JSON file in the
lyrics
directory. - Follow the structure of the JSON file as explained in the above sections.
- For each new song, add a new object to the
songs
array with the required fields. - Make sure to separate multiple song objects with commas.
- Save the file with a meaningful name (e.g.,
yoasobi_songs.json
). - The script will automatically process all JSON files in the lyrics directory.
You can seperate songs into it own JSON file, but the JSON structure should be as instructed
{
"songs": [
{
"title": "夜に駆ける",
"romanji_title": "Yoru ni Kakeru",
"lyrics": "ふざけあっても 分かり合えた気がした..."
},
{
"title": "群青",
"romanji_title": "Gunjou",
"lyrics": "嗚呼いつもの様に 過ぎる日々にあくびが出る..."
}
]
}
-
Make sure that Docker Desktop is running.
-
Run:
docker compose up -d
-
Make sure that Dock Desktop and the
morphemes-extractor
container are running. -
Run:
docker exec morphemes-extractor python main.py
-
Data is saved to a Postgres database.