changed acc to tutorial

README.md

@@ -10,6 +10,13 @@ in dvc_yaml, we will define all the stages
 scores.json stores all the prediction scores
 params.json stores all the parameters for those prediction.
 
+dvc metrics show:
+    to view the results of model.
+    stored in params.json and scores.json
+
+dvc metrics diff:
+    to view the difference in results of models with different parameters.
+
 Project Organization
 ------------
 

dvc.lock

@@ -1,34 +1,60 @@
 schema: '2.0'
 stages:
   load_data:
-    cmd: python src/data/load_data.py --config=config/params.yaml
+    cmd: python src/load_data.py --config=params.yaml
     deps:
     - path: data_given/winequality.csv
       md5: ccc8d3507eb151f53f760d36abdef888
       size: 91998
-    - path: src/data/get_data.py
-      md5: e93ff6f5120c2884787ed30217cb8afb
-      size: 730
-    - path: src/data/load_data.py
-      md5: 70a392ab799c92f1bc4dc611e2d6a476
-      size: 735
+    - path: src/get_data.py
+      md5: d32db8a35286570c0dd981f0bd7c1a25
+      size: 726
+    - path: src/load_data.py
+      md5: c07904b0e8bf3d72e3c2bf4184aebe47
+      size: 731
     outs:
     - path: data/raw/winequality.csv
       md5: d98e8f2eca228c11c4de1cc96866a54d
       size: 93598
   split_data:
-    cmd: python src/data/split_data.py --config=config/params.yaml
+    cmd: python src/split_data.py --config=params.yaml
     deps:
     - path: data/raw/winequality.csv
       md5: d98e8f2eca228c11c4de1cc96866a54d
       size: 93598
-    - path: src/data/split_data.py
-      md5: cea3d5da4e1e583a0620bac5f6512246
-      size: 1174
+    - path: src/split_data.py
+      md5: ab93250c4eb3fbb9b253fb57e93350f6
+      size: 1170
     outs:
     - path: data/processed/test_winequality.csv
       md5: 01718da3bd6e206060ed00667ca1a389
       size: 18831
     - path: data/processed/train_winequality.csv
       md5: 4bccf4d668cdc60115de42ba74d416ab
       size: 74913
+  train_and_evaluate:
+    cmd: python src/train_and_evaluate.py --config=params.yaml
+    deps:
+    - path: data/processed/test_winequality.csv
+      md5: 01718da3bd6e206060ed00667ca1a389
+      size: 18831
+    - path: data/processed/train_winequality.csv
+      md5: 4bccf4d668cdc60115de42ba74d416ab
+      size: 74913
+    - path: src/train_and_evaluate.py
+      md5: 6f4add34d08b4be577c4cf184b004cf1
+      size: 2673
+    params:
+      params.yaml:
+        estimators.ElasticNet.params.alpha: 0.9
+        estimators.ElasticNet.params.l1_ratio: 0.4
+    outs:
+    - path: report/params.json
+      md5: a4aab0167612f21a6085dd05cfafdbf6
+      size: 44
+    - path: report/scores.json
+      md5: 174868cff0fcd398bcc867bab2ac174f
+      size: 101
+    - path: saved_models/model.joblib
+      md5: 29eccfa32dedc97f770f6b245cd7aa36
+      size: 1336

dvc.yaml

@@ -1,35 +1,35 @@
 stages:
   load_data:
-    cmd: python src/data/load_data.py --config=config/params.yaml
+    cmd: python src/load_data.py --config=params.yaml
     deps:
-    - src/data/get_data.py
-    - src/data/load_data.py
+    - src/get_data.py
+    - src/load_data.py
     - data_given/winequality.csv
     outs:
     - data/raw/winequality.csv
 
   split_data:
-    cmd: python src/data/split_data.py --config=config/params.yaml
+    cmd: python src/split_data.py --config=params.yaml
     deps:
-    - src/data/split_data.py
+    - 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/models/train_model.py --config=config/params.yaml
+    cmd: python src/train_and_evaluate.py --config=params.yaml
     deps:
     - data/processed/train_winequality.csv
     - data/processed/test_winequality.csv 
-    - src/models/train_model.py
+    - src/train_and_evaluate.py
     params:
     - estimators.ElasticNet.params.alpha
     - estimators.ElasticNet.params.l1_ratio
     metrics:
-    - reports/scores.json:
+    - report/scores.json:
         cache: false
-    - reports/params.json:
+    - report/params.json:
         cache: false
     outs:
-    - models/saved_models/model.joblib
+    - saved_models/model.joblib

config/params.yaml → params.yaml

@@ -17,14 +17,14 @@ split_data:
 estimators:
   ElasticNet:
     params:
-      alpha: 0.88
-      l1_ratio: 0.89
-      # alpha: 0.9
-      # l1_ratio: 0.4
-model_dir: models/saved_models
+      # alpha: 0.88
+      # l1_ratio: 0.89
+      alpha: 0.9
+      l1_ratio: 0.4
+model_dir: saved_models
 
 reports:
-  params: reports/params.json
-  scores: reports/scores.json
+  params: report/params.json
+  scores: report/scores.json
 
 webapp_model_dir: prediction_service/model/model.joblib

report/params.json

@@ -0,0 +1,4 @@
+{
+    "alpha": 0.9,
+    "l1_ratio": 0.4
+}

report/scores.json

@@ -0,0 +1,5 @@
+{
+    "rmse": 0.8031192316982253,
+    "mae": 0.6551462834843993,
+    "r2": 0.013014969842178092
+}

saved_models/.gitignore

@@ -0,0 +1 @@
+/model.joblib

src/data/get_data.py → src/get_data.py

@@ -19,7 +19,7 @@ def get_data(config_path):
 
 if __name__=="__main__":
     args = argparse.ArgumentParser()
-    default_config_path = os.path.join("config", "params.yaml")
-    args.add_argument("--config", default = default_config_path)
+    # default_config_path = os.path.join("config", "params.yaml")
+    args.add_argument("--config", default = "params.yaml")
     parsed_args = args.parse_args()
     data = get_data(config_path=parsed_args.config)

src/data/load_data.py → src/load_data.py

@@ -14,7 +14,7 @@ def load_and_save(config_path):
 
 if __name__=="__main__":
     args = argparse.ArgumentParser()
-    default_config_path = os.path.join("config", "params.yaml")
-    args.add_argument("--config", default=default_config_path)
+    # default_config_path = os.path.join("config", "params.yaml")
+    args.add_argument("--config", default="params.yaml")
     parsed_args = args.parse_args()
     load_and_save(config_path=parsed_args.config)  

src/data/split_data.py → src/split_data.py

@@ -25,7 +25,7 @@ def split_and_saved_data(config_path):
 
 if __name__=="__main__":
     args = argparse.ArgumentParser()
-    default_config_path = os.path.join("config", "params.yaml")
-    args.add_argument("--config", default=default_config_path)
+    # default_config_path = os.path.join("config", "params.yaml")
+    args.add_argument("--config", default="params.yaml")
     parsed_args = args.parse_args()
     split_and_saved_data(config_path=parsed_args.config)

src/models/train_model.py → src/train_and_evaluate.py

@@ -87,7 +87,6 @@ def train_and_evaluate(config_path):
 
 if __name__=="__main__":
     args = argparse.ArgumentParser()
-    default_config_path = os.path.join("config", "params.yaml")
-    args.add_argument("--config", default=default_config_path)
+    args.add_argument("--config", default="params.yaml")
     parsed_args = args.parse_args()
     train_and_evaluate(config_path=parsed_args.config)