Create a virtual environment.
conda create -n eppi_text python=3.11
conda activate eppi_text
Install.
pip3 install -e .
python3 -m spacy download en_core_web_sm
If you wish to run tests, you will need to install the test dependencies.
pip3 install -e ".[test]"
The workbench uses a database to track the hyperparameters and results of the hyperparameter search. To use this feature, you must have a database in an appropriate location.
On a local system, the OptunaHyperparameterOptimisation object will automatically find the optuna.db in this repo and use it.
On Azure ML Studio, the database must be set more carefully to be on the same storage device as the compute instance. For example, when using notebooks, create a database at /mnt/tmp and set the db_url of OptunaHyperaparmeterOptimisation object appropriately:
cd /mnt/tmp
touch optuna.db
And set db_url appropriately in your script/notebook. In this case:
optimiser = OptunaHyperparameterOptimisation(
db_url=f"sqlite:////mnt/tmp/optuna.db",
)
.
├── README.md
├── data
│ └── raw
│ └── debunking_review.tsv
├── eppi_text_classification
│ ├── __init__.py
│ ├── opt.py
│ ├── plotly_confusion.py
│ ├── plotly_roc.py
│ ├── plots.py
│ ├── predict.py
│ ├── save_features_labels.py
│ ├── shap_colors
│ │ ├── __init__.py
│ │ ├── _colorconv.py
│ │ └── _colors.py
│ ├── shap_plotter.py
│ ├── utils.py
│ └── validation.py
├── notebooks
│ ├── lgbm
│ │ └── lgbm_binary.ipynb
│ ├── random_forest
│ │ └── random_forest_binary.ipynb
│ ├── svm
│ │ └── svm_binary.ipynb
│ └── xgboost
│ └── xgboost_binary.ipynb
├── optuna.db
├── pyproject.toml
├── setup.py
├── tests
│ ├── check_install.py
│ └── test_00_smoke.py
└── tox.ini