Support string format

2025-05-06 21:51:06 +08:00
parent 7a3266e4cc
commit b52997633c
3 changed files with 245 additions and 3 deletions
--- a/jambo/parser/string_type_parser.py
+++ b/jambo/parser/string_type_parser.py
@@ -1,5 +1,9 @@
 from jambo.parser._type_parser import GenericTypeParser

+from pydantic import EmailStr, HttpUrl, IPvAnyAddress
+
+from datetime import date, datetime, time
+

 class StringTypeParser(GenericTypeParser):
    mapped_type = str
@@ -10,13 +14,42 @@ class StringTypeParser(GenericTypeParser):
        "maxLength": "max_length",
        "minLength": "min_length",
        "pattern": "pattern",
+        "format": "format",
+    }
+
+    format_type_mapping = {
+        "email": EmailStr,
+        "uri": HttpUrl,
+        "ipv4": IPvAnyAddress,
+        "ipv6": IPvAnyAddress,
+        "hostname": str,
+        "date": date,
+        "time": time,
+        "date-time": datetime,
+    }
+
+    format_pattern_mapping = {
+        "hostname": r"^[a-zA-Z0-9]([a-zA-Z0-9\-]{0,61}[a-zA-Z0-9])?(\.[a-zA-Z0-9]([a-zA-Z0-9\-]{0,61}[a-zA-Z0-9])?)*$",
    }

    def from_properties(self, name, properties, required=False):
        mapped_properties = self.mappings_properties_builder(properties, required)

+        format_type = properties.get("format")
+        if format_type:
+            if format_type in self.format_type_mapping:
+                mapped_type = self.format_type_mapping[format_type]
+                if format_type in self.format_pattern_mapping:
+                    mapped_properties["pattern"] = self.format_pattern_mapping[
+                        format_type
+                    ]
+            else:
+                raise ValueError(f"Unsupported string format: {format_type}")
+        else:
+            mapped_type = str
+
        default_value = properties.get("default")
        if default_value is not None:
-            self.validate_default(str, mapped_properties, default_value)
+            self.validate_default(mapped_type, mapped_properties, default_value)

-        return str, mapped_properties
+        return mapped_type, mapped_properties
--- a/tests/parser/test_string_type_parser.py
+++ b/tests/parser/test_string_type_parser.py
@@ -1,5 +1,8 @@
 from jambo.parser import StringTypeParser

+from pydantic import EmailStr, HttpUrl, IPvAnyAddress
+
+from datetime import date, datetime, time
 from unittest import TestCase


@@ -85,3 +88,112 @@ class TestStringTypeParser(TestCase):

        with self.assertRaises(ValueError):
            parser.from_properties("placeholder", properties)
+
+    def test_string_parser_with_email_format(self):
+        parser = StringTypeParser()
+
+        properties = {
+            "type": "string",
+            "format": "email",
+        }
+
+        type_parsing, type_validator = parser.from_properties("placeholder", properties)
+
+        self.assertEqual(type_parsing, EmailStr)
+
+    def test_string_parser_with_uri_format(self):
+        parser = StringTypeParser()
+
+        properties = {
+            "type": "string",
+            "format": "uri",
+        }
+
+        type_parsing, type_validator = parser.from_properties("placeholder", properties)
+
+        self.assertEqual(type_parsing, HttpUrl)
+
+    def test_string_parser_with_ip_formats(self):
+        parser = StringTypeParser()
+
+        for ip_format in ["ipv4", "ipv6"]:
+            properties = {
+                "type": "string",
+                "format": ip_format,
+            }
+
+            type_parsing, type_validator = parser.from_properties(
+                "placeholder", properties
+            )
+
+            self.assertEqual(type_parsing, IPvAnyAddress)
+
+    def test_string_parser_with_time_format(self):
+        parser = StringTypeParser()
+
+        properties = {
+            "type": "string",
+            "format": "time",
+        }
+
+        type_parsing, type_validator = parser.from_properties("placeholder", properties)
+
+        self.assertEqual(type_parsing, time)
+
+    def test_string_parser_with_pattern_based_formats(self):
+        parser = StringTypeParser()
+
+        for format_type in ["hostname"]:
+            properties = {
+                "type": "string",
+                "format": format_type,
+            }
+
+            type_parsing, type_validator = parser.from_properties(
+                "placeholder", properties
+            )
+
+            self.assertEqual(type_parsing, str)
+            self.assertIn("pattern", type_validator)
+            self.assertEqual(
+                type_validator["pattern"], parser.format_pattern_mapping[format_type]
+            )
+
+    def test_string_parser_with_unsupported_format(self):
+        parser = StringTypeParser()
+
+        properties = {
+            "type": "string",
+            "format": "unsupported-format",
+        }
+
+        with self.assertRaises(ValueError) as context:
+            parser.from_properties("placeholder", properties)
+
+        self.assertEqual(
+            str(context.exception), "Unsupported string format: unsupported-format"
+        )
+
+    def test_string_parser_with_date_format(self):
+        parser = StringTypeParser()
+
+        properties = {
+            "type": "string",
+            "format": "date",
+        }
+
+        type_parsing, type_validator = parser.from_properties("placeholder", properties)
+
+        self.assertEqual(type_parsing, date)
+
+    def test_string_parser_with_datetime_format(self):
+        parser = StringTypeParser()
+
+        properties = {
+            "type": "string",
+            "format": "date-time",
+        }
+
+        type_parsing, type_validator = parser.from_properties("placeholder", properties)
+
+        self.assertEqual(type_parsing, datetime)
--- a/tests/test_schema_converter.py
+++ b/tests/test_schema_converter.py
@@ -1,7 +1,8 @@
 from jambo import SchemaConverter

-from pydantic import BaseModel
+from pydantic import BaseModel, HttpUrl

+from ipaddress import IPv4Address, IPv6Address
 from unittest import TestCase


@@ -397,3 +398,99 @@ class TestSchemaConverter(TestCase):

        with self.assertRaises(ValueError):
            Model(id=11)
+
+    def test_string_format_email(self):
+        schema = {
+            "title": "EmailTest",
+            "type": "object",
+            "properties": {"email": {"type": "string", "format": "email"}},
+        }
+        model = SchemaConverter.build(schema)
+        self.assertEqual(model(email="test@example.com").email, "test@example.com")
+        with self.assertRaises(ValueError):
+            model(email="invalid-email")
+
+    def test_string_format_uri(self):
+        schema = {
+            "title": "UriTest",
+            "type": "object",
+            "properties": {"website": {"type": "string", "format": "uri"}},
+        }
+        model = SchemaConverter.build(schema)
+        self.assertEqual(
+            model(website="https://example.com").website, HttpUrl("https://example.com")
+        )
+        with self.assertRaises(ValueError):
+            model(website="invalid-uri")
+
+    def test_string_format_ipv4(self):
+        schema = {
+            "title": "IPv4Test",
+            "type": "object",
+            "properties": {"ip": {"type": "string", "format": "ipv4"}},
+        }
+        model = SchemaConverter.build(schema)
+        self.assertEqual(model(ip="192.168.1.1").ip, IPv4Address("192.168.1.1"))
+        with self.assertRaises(ValueError):
+            model(ip="256.256.256.256")
+
+    def test_string_format_ipv6(self):
+        schema = {
+            "title": "IPv6Test",
+            "type": "object",
+            "properties": {"ip": {"type": "string", "format": "ipv6"}},
+        }
+        model = SchemaConverter.build(schema)
+        self.assertEqual(
+            model(ip="2001:0db8:85a3:0000:0000:8a2e:0370:7334").ip,
+            IPv6Address("2001:0db8:85a3:0000:0000:8a2e:0370:7334"),
+        )
+        with self.assertRaises(ValueError):
+            model(ip="invalid-ipv6")
+
+    def test_string_format_hostname(self):
+        schema = {
+            "title": "HostnameTest",
+            "type": "object",
+            "properties": {"hostname": {"type": "string", "format": "hostname"}},
+        }
+        model = SchemaConverter.build(schema)
+        self.assertEqual(model(hostname="example.com").hostname, "example.com")
+        with self.assertRaises(ValueError):
+            model(hostname="invalid..hostname")
+
+    def test_string_format_datetime(self):
+        schema = {
+            "title": "DateTimeTest",
+            "type": "object",
+            "properties": {"timestamp": {"type": "string", "format": "date-time"}},
+        }
+        model = SchemaConverter.build(schema)
+        self.assertEqual(
+            model(timestamp="2024-01-01T12:00:00Z").timestamp.isoformat(),
+            "2024-01-01T12:00:00+00:00",
+        )
+        with self.assertRaises(ValueError):
+            model(timestamp="invalid-datetime")
+
+    def test_string_format_time(self):
+        schema = {
+            "title": "TimeTest",
+            "type": "object",
+            "properties": {"time": {"type": "string", "format": "time"}},
+        }
+        model = SchemaConverter.build(schema)
+        self.assertEqual(
+            model(time="20:20:39+00:00").time.isoformat(), "20:20:39+00:00"
+        )
+        with self.assertRaises(ValueError):
+            model(time="25:00:00")
+
+    def test_string_format_unsupported(self):
+        schema = {
+            "title": "InvalidFormat",
+            "type": "object",
+            "properties": {"field": {"type": "string", "format": "unsupported"}},
+        }
+        with self.assertRaises(ValueError):
+            SchemaConverter.build(schema)