dvc.lock

schema: '2.0'
stages:
  prepare:
    cmd: python src/prepare.py data/data.xml
    deps:
    - path: data/data.xml
      md5: 22a1a2931c8370d3aeedd7183606fd7f
      size: 14445097
    - path: src/prepare.py
      md5: f09ea0c15980b43010257ccb9f0055e2
      size: 1576
    params:
      params.yaml:
        prepare.seed: 20170428
        prepare.split: 0.2
    outs:
    - path: data/prepared
      md5: 153aad06d376b6595932470e459ef42a.dir
      size: 8437363
      nfiles: 2
  featurize:
    cmd: python src/featurization.py data/prepared data/features
    deps:
    - path: data/prepared
      md5: 153aad06d376b6595932470e459ef42a.dir
      size: 8437363
      nfiles: 2
    - path: src/featurization.py
      md5: e0265fc22f056a4b86d85c3056bc2894
      size: 2490
    params:
      params.yaml:
        featurize.max_features: 200
        featurize.ngrams: 2
    outs:
    - path: data/features
      md5: f35d4cc2c552ac959ae602162b8543f3.dir
      size: 2232588
      nfiles: 2
  train:
    cmd: python src/train.py data/features model.pkl
    deps:
    - path: data/features
      md5: f35d4cc2c552ac959ae602162b8543f3.dir
      size: 2232588
      nfiles: 2
    - path: src/train.py
      md5: c3961d777cfbd7727f9fde4851896006
      size: 967
    params:
      params.yaml:
        train.min_split: 0.01
        train.n_est: 50
        train.seed: 20170428
    outs:
    - path: model.pkl
      md5: 46865edbf3d62fc5c039dd9d2b0567a4
      size: 1763725
  evaluate:
    cmd: python src/evaluate.py model.pkl data/features
    deps:
    - path: data/features
      md5: f35d4cc2c552ac959ae602162b8543f3.dir
      size: 2232588
      nfiles: 2
    - path: model.pkl
      md5: 46865edbf3d62fc5c039dd9d2b0567a4
      size: 1763725
    - path: src/evaluate.py
      md5: 44e714021a65edf881b1716e791d7f59
      size: 2346
    outs:
    - path: evaluation.json
      md5: 099ba05938dcee126403b109f3ce793f
      size: 73
    - path: evaluation/importance.png
      md5: 3c00de7be61ab7ebd141e68510f2757d
      size: 34065
    - path: evaluation/plots/confusion_matrix.json
      md5: 86c0c21030f85476b76de983d70effb1
      size: 122246
    - path: evaluation/plots/precision_recall.json
      md5: 819d0ad9bba6dd0e5ca76d4706f45da3
      size: 150773
    - path: evaluation/plots/roc.json
      md5: 6ff55945578272fc388e662680855e59
      size: 40653