add pipeline dvc

.gitignore

@@ -1 +1,10 @@
-/housing_price_dataset.csv
+/data/housing_price_dataset.csv
+/hp_model.h5
+/hp_test_predictions.csv
+/hp_test_metrics.csv
+/plot_rmse.png
+/plot_mae.png
+/plot_r2.png
+/hp_train.csv
+/hp_dev.csv
+/hp_test.csv

data_processing_dvc.py

@@ -0,0 +1,14 @@
+from sklearn.model_selection import train_test_split
+import pandas as pd
+import sys
+
+housing_price_dataset = pd.read_csv(sys.argv[1])
+
+housing_price_dataset = pd.get_dummies(housing_price_dataset, columns=['Neighborhood'])
+
+hp_train_test, hp_dev = train_test_split(housing_price_dataset, test_size=0.1)
+hp_train, hp_test = train_test_split(hp_train_test, test_size=1000)
+
+hp_train.to_csv('hp_train.csv', index=False)
+hp_dev.to_csv('hp_dev.csv', index=False)
+hp_test.to_csv('hp_test.csv', index=False)

dvc.yaml

@@ -0,0 +1,38 @@
+stages:
+  data_processing:
+    cmd: python data_processing_dvc.py data/housing_price_dataset.csv
+    deps:
+      - data_processing_dvc.py
+      - data/housing_price_dataset.csv
+    outs:
+      - hp_train.csv
+      - hp_dev.csv
+      - hp_test.csv
+  create_model:
+     cmd: python create_model.py ${create_model.epochs} ${create_model.learning_rate} ${create_model.batch_size}
+     deps:
+       - create_model.py
+       - helper.py
+       - hp_train.csv
+       - hp_dev.csv
+     outs:
+       - hp_model.h5
+     params:
+       - create_model.epochs
+       - create_model.learning_rate
+       - create_model.batch_size
+  evaluate:
+    cmd: python evaluate.py ${evaluate.build_number}
+    deps:
+      - evaluate.py
+      - helper.py
+      - hp_model.h5
+      - hp_test.csv
+    outs:
+      - hp_test_predictions.csv
+      - hp_test_metrics.csv
+      - plot_rmse.png
+      - plot_mae.png
+      - plot_r2.png
+    params:
+      - evaluate.build_number

params.yaml

@@ -0,0 +1,6 @@
+create_model:
+  epochs: 20
+  learning_rate: 0.001
+  batch_size: 32
+evaluate:
+  build_number: 0