Gru/README.md

41 lines
1.6 KiB
Markdown

# Gru
## Structure
Autofeur is composed of a few components that make up this bot
|Name|Description|
|-|-|
|autofeur_nova|This is the component for handling discord events, it uses the [nova framework](https://github.com/discordnova/nova) under the hood and is developped with TypeScript|
|deep_phonemizer|This is the component that transforms a grapheme into a phoneme using [DeepPhonemizer](https://github.com/as-ideas/DeepPhonemizer)|
|autofeur_db|This is the component used for completing te end of the words, its a DB specialized into completing this specific task|
## Running Autofeur
### Getting trained models
You'll need two files to get running with Autofeur, a trained `DeepPhonemizer` model and a IPA Dictionary file.
You can get the `DeepPhonemizer` model on the project [github page](https://github.com/as-ideas/DeepPhonemizer#pretrained-models) or follow the instructions there to create your own datasets.
You can get the IPA Dictionary on this [github page](https://github.com/open-dict-data/ipa-dict) or use your own, it's simply a CSV file with two columns, one for the word and another for the phonemized word.
### Starting `deep_phonemizer`
To run it inside docker, we recommand
`docker-compose up deep-phonemizer`
If you want to use bare metal, follow the following commands
You'll need to move your trained model into the `deep_phonemizer/assets/model.pt` file.
```sh
# Go into the folder
cd deep_phonemizer
# Create a Virtual environment with dependencies
python3 -m venv ./venv
source ./venv/bin/activate
pip install -r requirements.txt
# Run the flash application
flask run
```