dvc.yaml

stages:
  data_cleaning:
    cmd: python src/data_preprocessing.py
    deps:
      - src/data_preprocessing.py
      - data/raw/Dev_data_to_be_shared.csv
      - data/raw/validation_data_to_be_shared.csv
    params:
      - experiment.EXPERIMENT_NAME
      - data_cleaning.ENCODING_TECHNIQUE
      - data_cleaning.DEV_DATA_PATH
      - data_cleaning.VAL_DATA_PATH
      - data_cleaning.NULL_PERCENTAGE
      - data_cleaning.VARIANCE_THRESHOLD
      - data_cleaning.IMPUTATION_TECHNIQUE
      - data_cleaning.KNN_IMPUTER_N_NEIGHBORS
      - data_cleaning.TEST_SIZE
      - data_cleaning.RANDOM_STATE
    outs:
      - data/interim/train.csv
      - data/interim/test.csv
      - data/interim/validation.csv

  feature_engineering:
    cmd: python src/feature_engineering.py
    deps:
      - src/feature_engineering.py
      - data/interim/train.csv
      - data/interim/test.csv
      - data/interim/validation.csv
    params:
      - experiment.EXPERIMENT_NAME
      - feature_engineering.FEATURE_SELECTION_TECHNIQUE
      - feature_engineering.ANOVA_K
      - feature_engineering.SFS_FORWARD
      - feature_engineering.N_ESTIMATORS
      - feature_engineering.MAX_DEPTH
      - feature_engineering.MIN_SAMPLES_SPLIT
      - feature_engineering.MIN_SAMPLES_LEAF
      - feature_engineering.MAX_FEATURES
      - feature_engineering.RANDOM_STATE
      - feature_engineering.NO_FEATURES
      - feature_engineering.FEATURE_IMPORTANCE_TECHNIQUE
    outs:
      - data/processed/train.csv
      - data/processed/test.csv
      - data/processed/val.csv

  model_training:
    cmd: python src/Model_Building.py
    deps:
      - src/Model_Building.py
      - data/processed/train.csv
      - data/processed/test.csv
      - data/processed/val.csv
    params:
      - experiment.EXPERIMENT_NAME
      - model_training.MODEL
      - model_training.MAX_DEPTH 
      - model_training.OBJECTIVE
      - model_training.EVAL_METRIC
      - model_training.ETA  
      - model_training.SUB_SAMPLE
      - model_training.COL_SAMPLE_BY_TREE 
      - model_training.MIN_CHILD_WEIGHT
      - model_training.GAMMA_XGB 
      - model_training.LAMBDA 
      - model_training.ALPHA
      - model_training.SCALE_POS_WEIGHT
      - model_training.N_ESTIMATORS
      - model_training.TREE_METHOD
      - model_training.DEVICE
      - model_training.SEED
      - model_training.N_TRIALS


      - model_training.TRIALS
      - model_training.NO_LAYERS
      - model_training.HIDDEN_DIMS
      - model_training.DROPOUT_RATE
      - model_training.LEARNING_RATE
      - model_training.WEIGHT_DECAY
      - model_training.BATCH_SIZE
      - model_training.N_EPOCHS
      - model_training.OPTIMIZER

  prediction:
    cmd: python src/prediction.py
    deps:
      - src/prediction.py
      - data/processed/val.csv
    params:
      - experiment.EXPERIMENT_NAME
      - prediction.DATA
      - prediction.MODEL
      - prediction.SUBMISSION_FILE_PATH
      - prediction.XGB_MODEL_PATH
      - prediction.MLP_MODEL_PATH
    outs:
      - prediction.csv