fix(contract): added regex for JSON string formats date-time, date, time, and email (#930)

* Added regex for JSON string formats date-time, date, email, and time

* When generating an example resource, return ASCII for strings

* Added start string replacement for terminate_regex and tests

Author: mmaeng

src/rpdk/core/contract/resource_generator.py

@@ -28,18 +28,28 @@
 # https://github.com/aws-cloudformation/aws-cloudformation-rpdk/issues/118
 
 # Arn is just a placeholder for testing
+# format list taken from https://python-jsonschema.readthedocs.io/en/stable/validate/#jsonschema.FormatChecker.checkers
+# date-time regex from https://github.com/naimetti/rfc3339-validator
+# date is extraction from date-time
+# time is extraction from date-time
 STRING_FORMATS = {
     "arn": "^arn:aws(-(cn|gov))?:[a-z-]+:(([a-z]+-)+[0-9])?:([0-9]{12})?:[^.]+$",
-    "uri": "^(https?|ftp|file)://[0-9a-zA-Z]([-.\\w]*[0-9a-zA-Z])(:[0-9]*)*([?/#].*)?$",
+    "uri": r"^(https?|ftp|file)://[0-9a-zA-Z]([-.\w]*[0-9a-zA-Z])(:[0-9]*)*([?/#].*)?$",
+    "date-time": r"^(\d{4})-(0[1-9]|1[0-2])-(\d{2})T(?:[01]\d|2[0123]):(?:[0-5]\d):(?:[0-5]\d)(?:\.\d+)?(?:Z|[+-](?:[01]\d|2[0123]):[0-5]\d)$",
+    "date": r"^(\d{4})-(0[1-9]|1[0-2])-(\d{2})$",
+    "time": r"^(?:[01]\d|2[0123]):(?:[0-5]\d):(?:[0-5]\d)(?:\.\d+)?(?:Z|[+-](?:[01]\d|2[0123]):[0-5]\d)$",
+    "email": r"^.+@[^\.].*\.[a-z]{2,}$",
 }
 
 NEG_INF = float("-inf")
 POS_INF = float("inf")
 
 
 def terminate_regex(regex):
+    if regex.startswith("^"):
+        regex = r"\A" + regex[1:]
     if regex.endswith("$"):
-        return regex[:-1] + r"\Z"
+        regex = regex[:-1] + r"\Z"
     return regex
 
 

tests/contract/test_resource_generator.py

@@ -22,9 +22,28 @@ def test_terminate_regex_end_of_line_like_a_normal_person():
     assert re.match(modified_regex, "dfqh3eqefhq")
 
 
-def test_terminate_regex_no_termination_needed():
+def test_terminate_regex_line_start_change():
     original_regex = r"^[a-zA-Z0-9]{1,219}\Z"
-    assert terminate_regex(original_regex) == original_regex
+    terminated_regex = r"\A[a-zA-Z0-9]{1,219}\Z"
+    assert terminate_regex(original_regex) == terminated_regex
+
+
+def test_terminate_regex_line_end_change():
+    original_regex = r"\A[a-zA-Z0-9]{1,219}$"
+    terminated_regex = r"\A[a-zA-Z0-9]{1,219}\Z"
+    assert terminate_regex(original_regex) == terminated_regex
+
+
+def test_terminate_regex_line_start_and_end_change():
+    original_regex = r"^[a-zA-Z0-9]{1,219}$"
+    terminated_regex = r"\A[a-zA-Z0-9]{1,219}\Z"
+    assert terminate_regex(original_regex) == terminated_regex
+
+
+def test_terminate_regex_no_termination_needed():
+    original_regex = r"\A[a-zA-Z0-9]{1,219}\Z"
+    terminated_regex = r"\A[a-zA-Z0-9]{1,219}\Z"
+    assert terminate_regex(original_regex) == terminated_regex
 
 
 @pytest.mark.parametrize("schema_type", ["integer", "number"])
@@ -80,6 +99,22 @@ def test_generate_string_strategy_format():
     strategy = ResourceGenerator(schema).generate_schema_strategy(schema)
     assert re.fullmatch(STRING_FORMATS["arn"], strategy.example())
 
+    schema = {"type": "string", "format": "date-time"}
+    strategy = ResourceGenerator(schema).generate_schema_strategy(schema)
+    assert re.match(STRING_FORMATS["date-time"], strategy.example())
+
+    schema = {"type": "string", "format": "time"}
+    strategy = ResourceGenerator(schema).generate_schema_strategy(schema)
+    assert re.match(STRING_FORMATS["time"], strategy.example())
+
+    schema = {"type": "string", "format": "date"}
+    strategy = ResourceGenerator(schema).generate_schema_strategy(schema)
+    assert re.match(STRING_FORMATS["date"], strategy.example())
+
+    schema = {"type": "string", "format": "email"}
+    strategy = ResourceGenerator(schema).generate_schema_strategy(schema)
+    assert re.match(STRING_FORMATS["email"], strategy.example())
+
 
 def test_generate_string_strategy_length():
     schema = {"type": "string", "minLength": 5, "maxLength": 10}