prmdr-242 added more validation and test

lambdas/services/lloyd_george_validator.py

@@ -9,12 +9,26 @@ class LGFileNameException(ValueError):
     """One or more of the files do not match naming convention"""
     pass
 
-def validate_lg_file_type(file_type):
+class LGInvalidFilesException(Exception):
+    """One or more of the files are not valid"""
+    pass
+
+def validate_lg_file_type(file_type: str):
     if file_type != 'application/pdf':
         raise LGFileTypeException
 
-def validate_file_name(name):
+def validate_file_name(name: str):
     lg_regex = '[0-9]+of[0-9]+_Lloyd_George_Record_\\[[A-Za-z]+\\s[A-Za-z]+]_\\[[0-9]{10}]_\\[\\d\\d-\\d\\d-\\d\\d\\d\\d].pdf'
     if not re.fullmatch(lg_regex, name):
         raise LGFileNameException
 
+def check_for_duplicate_files(file_list: list[str]):
+    if len(file_list) > len(set(file_list)):
+        raise LGInvalidFilesException
+
+def check_for_number_of_files_match_expected(file_name: str, total_files_number: str):
+    lg_number_regex = 'of[0-9]+'
+    expected_number_of_files = re.search(lg_number_regex, file_name)
+    if expected_number_of_files and not expected_number_of_files.group()[2:] == total_files_number:
+        raise LGInvalidFilesException
+

lambdas/tests/unit/services/test_lloyd_george_validator.py

@@ -1,6 +1,7 @@
 import pytest
 
-from services.lloyd_george_validator import validate_lg_file_type, LGFileTypeException, LGFileNameException, validate_file_name
+from services.lloyd_george_validator import validate_lg_file_type, LGFileTypeException, LGFileNameException, LGInvalidFilesException, validate_file_name, check_for_duplicate_files, check_for_number_of_files_match_expected
+
 
 def test_catching_error_when_file_type_not_pdf():
     with pytest.raises(LGFileTypeException):
@@ -25,3 +26,39 @@ def test_valid_file_name():
         validate_file_name(file_name)
     except LGFileNameException:
         assert False, 'One or more of the files do not match naming convention'
+
+def test_files_with_duplication():
+    with pytest.raises(LGInvalidFilesException):
+        lg_file_list = [
+                '1of1_Lloyd_George_Record_[Joe Bloggs]_[1111111111]_[25-12-2019].pdf',
+                '1of1_Lloyd_George_Record_[Joe Bloggs]_[1111111111]_[25-12-2019].pdf'
+            ]
+        check_for_duplicate_files(lg_file_list)
+
+def test_files_without_duplication():
+    try:
+        lg_file_list = [
+            '1of2_Lloyd_George_Record_[Joe Bloggs]_[1111111111]_[25-12-2019].pdf',
+            '2of2_Lloyd_George_Record_[Joe Bloggs]_[1111111111]_[25-12-2019].pdf'
+        ]
+        check_for_duplicate_files(lg_file_list)
+    except LGFileNameException:
+        assert False, 'One or more of the files are not valid'
+
+def test_files_list_with_missing_files():
+    with pytest.raises(LGInvalidFilesException):
+        lg_file_list = [
+            '1of3_Lloyd_George_Record_[Joe Bloggs]_[1111111111]_[25-12-2019].pdf',
+            '2of3_Lloyd_George_Record_[Joe Bloggs]_[1111111111]_[25-12-2019].pdf'
+        ]
+        check_for_number_of_files_match_expected(lg_file_list[0], str(len(lg_file_list)))
+
+def test_files_without_missing_files():
+    try:
+        lg_file_list = [
+            '1of2_Lloyd_George_Record_[Joe Bloggs]_[1111111111]_[25-12-2019].pdf',
+            '2of2_Lloyd_George_Record_[Joe Bloggs]_[1111111111]_[25-12-2019].pdf'
+        ]
+        check_for_number_of_files_match_expected(lg_file_list[0], str(len(lg_file_list)))
+    except LGFileNameException:
+        assert False, 'One or more of the files are not valid'