MLproject

name: HDB Resale Prices Pipeline

python_env: python_env.yaml

entry_points:
  load_raw_data:
    command: "python load_raw_data.py"

  data_validate:
    parameters:
      filepath: path
    command: "python scripts/data_validate.py --filepath {filepath}"

  preprocess:
    parameters:
      filepath: path
      train_ratio: {type: float, default: 0.7}
      val_ratio: {type: float, default: 0.2}
      test_ratio: {type: float, default: 0.1}
    command: "python scripts/preprocess.py --filepath {filepath} --train-ratio {train_ratio} --val-ratio {val_ratio} --test-ratio {test_ratio}"

  train:
    parameters:
      datadir: path
      n_estimators: {type: int, default: 10}
      max_features: {type: str, default: "sqrt"}
      max_depth: {type: int, default: 1}
      min_samples_split: {type: int, default: 2}
      min_samples_leaf: {type: int, default: 1}
    command: "python scripts/train.py --datadir {datadir} --n-estimators {n_estimators} --max-features {max_features} --max-depth {max_depth} --min-samples-split {min_samples_split} --min-samples-leaf {min_samples_leaf}"
      
  evaluate:
    parameters:
      datadir: path
      modeldir: path
    command: "python scripts/evaluate.py --datadir {datadir} --modeldir {modeldir}"

  model_validate:
    parameters:
      datadir: path
      modeldir: path
      test_score: {type: float}
      eval_threshold: {type: float, default: 150000}
    command: "python scripts/model_validate.py --datadir {datadir} --modeldir {modeldir} --test-score {test_score} --eval-threshold {eval_threshold}"
  
  main:
    parameters:
      eval_mae_threshold: {type: int, default: 150000}
      keras_hidden_units: {type: int, default: 20}
      max_row_limit: {type: int, default: 100000}
    command: "python scripts/main.py --eval-mae-threshold {eval_mae_threshold} --keras-hidden-units {keras_hidden_units}
                             --max-row-limit {max_row_limit}"