Task 42, creating tests for SampleData GoodFile and BadFile (#56)

Closes #52 

I misunderstood the requirements first. I thought we had to create tests
to make sure the GoodFile had the correct data for each field. I
understood the requirements later.

---------

Co-authored-by: Nargis Sultani <[email protected]>

README.md

@@ -24,23 +24,24 @@ There are few files in `src/validator` that will be of interest.
 - The repo also includes 2 test datasets for manual testing, one with all valid data, and one where each line
 represents a different failed validation, or different permutation of of the same
 failed validation.
-  - [`SBL_Validations_SampleData_GoodFile_03312023.csv`](SBL_Validations_SampleData_GoodFile_03312023.csv)
-  - [`SBL_Validations_SampleData_BadFile_03312023.csv`](SBL_Validations_SampleData_BadFile_03312023.csv)
+- [`sbl-validations-pass.csv`](src/tests/data/sbl-validations-pass.csv)
+- [`sbl-validations-fail.csv`](src/tests/data/sbl-validations-fail.csv)
 
 ### Manual Test
 ```sh
 # Test validating the "good" file
 # If passing lei value, pass lei as first arg and csv_path as second arg
-python src/validator/main.py 000TESTFIUIDDONOTUSE SBL_Validations_SampleData_GoodFile_03312023.csv
+python src/validator/main.py 000TESTFIUIDDONOTUSE src/tests/data/sbl-validations-pass.csv
 # else just pass the csv_path as arg
-python src/validator/main.py SBL_Validations_SampleData_GoodFile_03312023.csv
+python src/validator/main.py src/tests/data/sbl-validations-pass.csv
 
 # Test validating the "bad" file
-python src/validator/main.py 000TESTFIUIDDONOTUSE SBL_Validations_SampleData_BadFile_03312023.csv
+python src/validator/main.py 000TESTFIUIDDONOTUSE src/tests/data/sbl-validations-fail.csv
 # or
-python src/validator/main.py SBL_Validations_SampleData_BadFile_03312023.csv
+python src/validator/main.py src/tests/data/sbl-validations-fail.csv
 ```
 
+
 ## Development
 
 Development Process

src/tests/test_sample_data.py

@@ -0,0 +1,54 @@
+import os
+import sys
+
+import pandas as pd
+import pytest
+
+from validator.create_schemas import validate_phases
+
+ROOT_DIR = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))  # noqa: E402
+sys.path.append(ROOT_DIR)  # noqa: E402
+
+GOOD_FILE_PATH = "./src/tests/data/sbl-validations-pass.csv"
+BAD_FILE_PATH = "./src/tests/data/sbl-validations-fail.csv"
+
+
+class TestValidatingSampleData:
+    valid_response = {"response": "No validations errors or warnings"}
+
+    good_file_df = pd.read_csv(GOOD_FILE_PATH, dtype=str, na_filter=False)
+    bad_file_df = pd.read_csv(BAD_FILE_PATH, dtype=str, na_filter=False)
+
+    def test_invalid_data_file(self):
+        failed_fpath = "./file-does-not-exist.csv"
+        with pytest.raises(Exception) as exc:
+            pd.read_csv(failed_fpath, dtype=str, na_filter=False)
+        assert exc.type == FileNotFoundError
+
+    def test_run_validation_on_good_data_invalid_lei(self):
+        lei = "000TESTFIUIDDONOTUS1"
+        validation_result = validate_phases(self.good_file_df, lei)
+
+        assert len(validation_result) == 1
+        assert validation_result[0] != self.valid_response
+
+    def test_run_validation_on_good_data_valid_lei(self):
+        lei = "000TESTFIUIDDONOTUSE"
+        validation_result = validate_phases(self.good_file_df, lei)
+
+        assert len(validation_result) == 1
+        assert validation_result[0] == self.valid_response
+
+    def test_run_validation_on_bad_data_invalid_lei(self):
+        lei = "000TESTFIUIDDONOTUS1"
+        validation_result = validate_phases(self.bad_file_df, lei)
+
+        assert len(validation_result) >= 1
+        assert validation_result[0] != self.valid_response
+
+    def test_run_validation_on_bad_data_valid_lei(self):
+        lei = "000TESTFIUIDDONOTUSE"
+        validation_result = validate_phases(self.bad_file_df, lei)
+
+        assert len(validation_result) >= 1
+        assert validation_result[0] != self.valid_response