Added dvc.yaml with reproductible stages

.gitignore

@@ -1 +1,4 @@
 /crime_conv.csv
+/crime_test.csv
+/crime_dev.csv
+/crime_train.csv

dvc.lock

@@ -0,0 +1,34 @@
+schema: '2.0'
+stages:
+  prepare_data:
+    cmd: python3 clean_and_split_data.py
+    deps:
+    - path: crime_conv.csv
+      md5: 6074436a3903679501a4d230053f7acf
+      size: 57978034
+    outs:
+    - path: crime_dev.csv
+      md5: bf45f519bab1a91415db273e5eb99dc6
+      size: 4553128
+    - path: crime_test.csv
+      md5: c7ed98f4ee8b0ccc176e91030cc5949f
+      size: 4552734
+    - path: crime_train.csv
+      md5: 6676b883d6553adecf567c502f6c766d
+      size: 36177141
+  train:
+    cmd: python3 train_model.py
+    deps:
+    - path: crime_dev.csv
+      md5: bf45f519bab1a91415db273e5eb99dc6
+      size: 4553128
+    - path: crime_test.csv
+      md5: c7ed98f4ee8b0ccc176e91030cc5949f
+      size: 4552734
+    - path: crime_train.csv
+      md5: 6676b883d6553adecf567c502f6c766d
+      size: 36177141
+    outs:
+    - path: model/saved_model.pb
+      md5: e4e20efbab0ef424d1d18e54b65a25ad
+      size: 106610

dvc.yaml

@@ -0,0 +1,17 @@
+stages:
+  prepare_data:
+    cmd: python3 clean_and_split_data.py
+    deps:
+    - crime_conv.csv
+    outs:
+    - crime_dev.csv
+    - crime_test.csv
+    - crime_train.csv
+  train:
+    cmd: python3 train_model.py
+    deps:
+    - crime_dev.csv
+    - crime_test.csv
+    - crime_train.csv
+    outs:
+    - model/saved_model.pb