mlops_dvc/dvc.yaml

  #after performing the replace operation in csv and saving it in RAW folder now we update
# dvc.yaml
stages:
  load_data:
  # terminal command to execute
    cmd: python src/load_data.py --config=params.yaml
    # dependency for load_data.py
    deps:
    - src/get_data.py
    - src/load_data.py
    - data_given/winequality.csv
  # output -> csv file present in raw folder
    outs:
    - data/raw/winequality.csv
  # execute with  -> "dvc repro"
  # this command will generate dvc.lock file will contain dependency mentioned will be tracked as a code
  # if we execute it again it will say data and pipelines are uptodate (if no changes are made in data or load_data.py)
    #'data_given/winequality.csv.dvc' didn't change, skipping
    #Stage 'load_data' didn't change, skipping
    #Data and pipelines are up to date.

  # if changes made , if sense the changes and rerun the pipeline
    # even dvc.loc file contains the size of thee file also

###############################
  # executing the split_data.py

  split_data:
    cmd: python src/split_data.py --config=params.yaml
    deps:
    - src/split_data.py
    - data/raw/winequality.csv
    outs:
    - data/processed/train_winequality.csv
    - data/processed/test_winequality.csv

  train_and_evaluate:
    cmd: python src/train_and_evaluate.py --config=params.yaml
    deps:
    - data/processed/train_winequality.csv
    - data/processed/test_winequality.csv
    - src/train_and_evaluate.py
    params:
    - estimators.ElasticNet.params.alpha
    - estimators.ElasticNet.params.l1_ratio
    metrics:
    - report/scores.json:
        cache: false #cache memory
    - report/params.json:
        cache: false
    outs:
    - saved_models/model.joblib