🛠️ Improve email parser

harambe/parser/parser.py

@@ -4,7 +4,6 @@
 from pydantic import (
     BaseModel,
     Field,
-    NameEmail,
     ValidationError,
     create_model,
     ConfigDict,
@@ -14,6 +13,7 @@
 from harambe.errors import SchemaValidationError
 from harambe.parser.type_currency import ParserTypeCurrency
 from harambe.parser.type_date import ParserTypeDate
+from harambe.parser.type_email import ParserTypeEmail
 from harambe.parser.type_enum import ParserTypeEnum
 from harambe.parser.type_number import ParserTypeNumber
 from harambe.parser.type_phone_number import ParserTypePhoneNumber
@@ -85,7 +85,7 @@ def _get_field_types(base_url: str) -> dict[SchemaFieldType, Any]:
             "float": ParserTypeNumber,
             "double": ParserTypeNumber,
             "currency": ParserTypeCurrency(),
-            "email": NameEmail,
+            "email": ParserTypeEmail,
             "enum": ParserTypeEnum,
             "array": list,
             "object": dict[str, Any],

harambe/parser/type_email.py

@@ -0,0 +1,16 @@
+from typing import Annotated
+
+from pydantic import BeforeValidator, validate_email
+
+
+def _validate_email(value: str) -> str:
+    if isinstance(value, str):
+        value = value.strip().lower().removeprefix("mailto:")
+        return validate_email(value)[1]
+    return value
+
+
+ParserTypeEmail = Annotated[
+    str,
+    BeforeValidator(_validate_email),
+]

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "harambe-sdk"
-version = "0.28.0"
+version = "0.28.1"
 description = "Data extraction SDK for Playwright 🐒🍌"
 authors = ["awtkns <[email protected]>"]
 readme = "README.md"

test/parser/test_parser.py

@@ -565,3 +565,16 @@ def test_config_allow_extra_fields() -> None:
     output_data = validator.validate(data, base_url="http://example.com")
 
     assert output_data == data
+
+
+def test_dump_email() -> None:
+    schema = {
+        "email": {"type": "email"},
+    }
+
+    data = {"email": "[email protected]"}
+
+    validator = PydanticSchemaParser(schema)
+    output_data = validator.validate(data, base_url="http://example.com")
+    assert output_data == data
+    assert isinstance(output_data["email"], str)

test/parser/test_type_email.py

@@ -0,0 +1,37 @@
+import pytest
+from pydantic import ValidationError, BaseModel
+
+from harambe.parser.type_email import ParserTypeEmail
+
+
+class TestModel(BaseModel):
+    email: ParserTypeEmail
+
+
+@pytest.mark.parametrize(
+    "email, expected",
+    [
+        (
+            "mailto:[email protected]",
+            "[email protected]",
+        ),
+        ("John Doe <[email protected]>", "[email protected]"),
+        ("[email protected]", "[email protected]"),
+        ("[email protected]", "[email protected]"),
+        ("[email protected]", "[email protected]"),
+        ("[email protected]", "[email protected]"),
+        ("[email protected]", "[email protected]"),
+    ],
+)
+def test_parser_success(email, expected):
+    model = TestModel(email=email)
+    assert model.email == expected
+
+
+@pytest.mark.parametrize(
+    "invalid_email",
+    ["invalid", "invalid@", "invalid@.", "[email protected]", "invalid@com.", 123, None],
+)
+def test_parser_fail(invalid_email):
+    with pytest.raises(ValidationError):
+        TestModel(email=invalid_email)