MLproject

name: thesis

conda_env: conda.yml

# docker_env:
#   image: mlflow-docker-example-environment
#   volumes: ["/local/path:/container/mount/path"]
#   environment: [["NEW_ENV_VAR", "new_var_value"], "VAR_TO_COPY_FROM_HOST_ENVIRONMENT"]
    
entry_points:

  gridsearch:
    parameters:
      task: {type: string, default: 'asset.variety'}
      metric: {type: string, default: 'f1'}
      averaging: {type: string, default: 'macro'}
      imputer: {type: string, default: 'dropnan'}
      random_state: {type: int, default: 0}
      merge: {type: str, default: 'yes'}
      pca: {type: int, default: 0}
      n_jobs_cv: {type: int, default: 1}
      n_folds: {type: int, default: 5}
    command: |
      python3 hyperparameter_tuning/gridsearch.py --pca {pca} --merge {merge} --task {task} --imputer {imputer} --n-folds {n_folds} --n-jobs-cv {n_jobs_cv} --metric {metric} --averaging {averaging} --random-state {random_state}  

  train:
    parameters:
      task: {type: string, default: 'asset.variety'}
      target: {type: string, default: 'Server'}
      algo: {type: string, default: 'LGBM'}
      hyperparams: {type: string, default: '{}'}
      imputer: {type: string, default: 'dropnan'}
      train_size: {type: float, default: 0.8}
      split_random_state: {type: int, default: 0}
      n_folds: {type: int, default: 5}
      merge: {type: str, default: "yes"}
      pca: {type: int, default: 0}
      explain: {type: str, default: "no"}
      shap_data_percentage: {type: float, default: 0.1}
      shap_test_over_train_percentage: {type: float, default: 0.3}
    command: |
      python3 train.py --pca {pca} --merge {merge} --explain {explain} --task {task} --target {target} --algo {algo} --hyperparams {hyperparams} --imputer {imputer} --train-size {train_size} --split-random-state {split_random_state} --n-folds {n_folds} --shap-data-percentage {shap_data_percentage} --shap-test-over-train-percentage {shap_test_over_train_percentage}