Add uuid validator (#772)

Co-authored-by: David Hewitt <[email protected]>

Author: JeanArhancet

Cargo.lock

@@ -43,6 +43,7 @@ idna = "0.4.0"
 base64 = "0.21.2"
 num-bigint = "0.4.3"
 python3-dll-a = "0.2.7"
+uuid = "1.3.4"
 
 [lib]
 name = "_pydantic_core"

Cargo.toml

@@ -206,6 +206,7 @@ def field_name(self) -> str:
     'url',
     'multi-host-url',
     'json',
+    'uuid',
 ]
 
 
@@ -1199,6 +1200,28 @@ def callable_schema(
     return _dict_not_none(type='callable', ref=ref, metadata=metadata, serialization=serialization)
 
 
+class UuidSchema(TypedDict, total=False):
+    type: Required[Literal['uuid']]
+    version: Literal[1, 3, 4, 5]
+    strict: bool
+    ref: str
+    metadata: Any
+    serialization: SerSchema
+
+
+def uuid_schema(
+    *,
+    version: Literal[1, 3, 4, 5] | None = None,
+    strict: bool | None = None,
+    ref: str | None = None,
+    metadata: Any = None,
+    serialization: SerSchema | None = None,
+) -> UuidSchema:
+    return _dict_not_none(
+        type='uuid', version=version, strict=strict, ref=ref, metadata=metadata, serialization=serialization
+    )
+
+
 class IncExSeqSerSchema(TypedDict, total=False):
     type: Required[Literal['include-exclude-sequence']]
     include: Set[int]
@@ -3731,6 +3754,7 @@ def definition_reference_schema(
         MultiHostUrlSchema,
         DefinitionsSchema,
         DefinitionReferenceSchema,
+        UuidSchema,
     ]
 elif False:
     CoreSchema: TypeAlias = Mapping[str, Any]
@@ -3784,6 +3808,7 @@ def definition_reference_schema(
     'multi-host-url',
     'definitions',
     'definition-ref',
+    'uuid',
 ]
 
 CoreSchemaFieldType = Literal['model-field', 'dataclass-field', 'typed-dict-field', 'computed-field']
@@ -3879,6 +3904,9 @@ def definition_reference_schema(
     'url_syntax_violation',
     'url_too_long',
     'url_scheme',
+    'uuid_type',
+    'uuid_parsing',
+    'uuid_version',
 ]
 
 

python/pydantic_core/core_schema.py

@@ -308,6 +308,14 @@ pub enum ErrorType {
     UrlScheme {
         expected_schemes: String,
     },
+    // UUID errors,
+    UuidType,
+    UuidParsing {
+        error: String,
+    },
+    UuidVersion {
+        expected_version: usize,
+    },
 }
 
 macro_rules! render {
@@ -451,6 +459,10 @@ impl ErrorType {
             Self::UrlSyntaxViolation { .. } => extract_context!(Cow::Owned, UrlSyntaxViolation, ctx, error: String),
             Self::UrlTooLong { .. } => extract_context!(UrlTooLong, ctx, max_length: usize),
             Self::UrlScheme { .. } => extract_context!(UrlScheme, ctx, expected_schemes: String),
+            Self::UuidParsing { .. } => extract_context!(UuidParsing, ctx, error: String),
+            Self::UuidVersion { .. } => {
+                extract_context!(UuidVersion, ctx, expected_version: usize)
+            }
             _ => {
                 if ctx.is_some() {
                     py_err!(PyTypeError; "'{}' errors do not require context", value)
@@ -555,6 +567,10 @@ impl ErrorType {
             Self::UrlSyntaxViolation {..} => "Input violated strict URL syntax rules, {error}",
             Self::UrlTooLong {..} => "URL should have at most {max_length} characters",
             Self::UrlScheme {..} => "URL scheme should be {expected_schemes}",
+            Self::UuidType => "UUID input should be a string, bytes or UUID object",
+            Self::UuidParsing { .. } => "Input should be a valid UUID, {error}",
+            Self::UuidVersion { .. } => "UUID version {expected_version} expected"
+
         }
     }
 
@@ -674,6 +690,8 @@ impl ErrorType {
             Self::UrlSyntaxViolation { error } => render!(tmpl, error),
             Self::UrlTooLong { max_length } => to_string_render!(tmpl, max_length),
             Self::UrlScheme { expected_schemes } => render!(tmpl, expected_schemes),
+            Self::UuidParsing { error } => render!(tmpl, error),
+            Self::UuidVersion { expected_version } => to_string_render!(tmpl, expected_version),
             _ => Ok(tmpl.to_string()),
         }
     }
@@ -734,6 +752,9 @@ impl ErrorType {
             Self::UrlSyntaxViolation { error } => py_dict!(py, error),
             Self::UrlTooLong { max_length } => py_dict!(py, max_length),
             Self::UrlScheme { expected_schemes } => py_dict!(py, expected_schemes),
+
+            Self::UuidParsing { error } => py_dict!(py, error),
+            Self::UuidVersion { expected_version } => py_dict!(py, expected_version),
             _ => Ok(None),
         }
     }

src/errors/types.rs

@@ -809,6 +809,13 @@ impl<'a> From<&'a PyBytes> for EitherBytes<'a> {
 }
 
 impl<'a> EitherBytes<'a> {
+    pub fn as_slice(&'a self) -> &[u8] {
+        match self {
+            EitherBytes::Cow(bytes) => bytes,
+            EitherBytes::Py(py_bytes) => py_bytes.as_bytes(),
+        }
+    }
+
     pub fn len(&'a self) -> PyResult<usize> {
         match self {
             EitherBytes::Cow(bytes) => Ok(bytes.len()),

src/input/return_enums.rs

@@ -187,6 +187,11 @@ pub(crate) fn infer_to_python_known(
                 let py_url: PyMultiHostUrl = value.extract()?;
                 py_url.__str__().into_py(py)
             }
+            ObType::Uuid => {
+                let py_uuid: &PyAny = value.downcast()?;
+                let uuid = super::type_serializers::uuid::uuid_to_string(py_uuid)?;
+                uuid.into_py(py)
+            }
             ObType::PydanticSerializable => serialize_with_serializer()?,
             ObType::Dataclass => serialize_dict(dataclass_to_dict(value)?)?,
             ObType::Enum => {
@@ -483,6 +488,11 @@ pub(crate) fn infer_serialize_known<S: Serializer>(
             pydantic_serializer.serialize(serializer)
         }
         ObType::Dataclass => serialize_dict!(dataclass_to_dict(value).map_err(py_err_se_err)?),
+        ObType::Uuid => {
+            let py_uuid: &PyAny = value.downcast().map_err(py_err_se_err)?;
+            let uuid = super::type_serializers::uuid::uuid_to_string(py_uuid).map_err(py_err_se_err)?;
+            serializer.serialize_str(&uuid)
+        }
         ObType::Enum => {
             let v = value.getattr(intern!(value.py(), "value")).map_err(py_err_se_err)?;
             infer_serialize(v, serializer, include, exclude, extra)
@@ -584,6 +594,11 @@ pub(crate) fn infer_json_key_known<'py>(ob_type: &ObType, key: &'py PyAny, extra
             let iso_time = super::type_serializers::datetime_etc::time_to_string(py_time)?;
             Ok(Cow::Owned(iso_time))
         }
+        ObType::Uuid => {
+            let py_uuid: &PyAny = key.downcast()?;
+            let uuid = super::type_serializers::uuid::uuid_to_string(py_uuid)?;
+            Ok(Cow::Owned(uuid))
+        }
         ObType::Timedelta => {
             let py_timedelta: &PyDelta = key.downcast()?;
             extra.config.timedelta_mode.json_key(py_timedelta)

src/serializers/infer.rs

@@ -39,6 +39,8 @@ pub struct ObTypeLookup {
     // types from this package
     url: usize,
     multi_host_url: usize,
+    // uuid type
+    uuid: usize,
     // enum type
     enum_type: usize,
     // generator
@@ -83,6 +85,7 @@ impl ObTypeLookup {
             enum_type: py.import("enum").unwrap().getattr("Enum").unwrap().get_type_ptr() as usize,
             generator: py.import("types").unwrap().getattr("GeneratorType").unwrap().as_ptr() as usize,
             path: py.import("pathlib").unwrap().getattr("Path").unwrap().as_ptr() as usize,
+            uuid: py.import("uuid").unwrap().getattr("UUID").unwrap().as_ptr() as usize,
         }
     }
 
@@ -130,6 +133,7 @@ impl ObTypeLookup {
             ObType::Enum => self.enum_type == ob_type,
             ObType::Generator => self.generator == ob_type,
             ObType::Path => self.path == ob_type,
+            ObType::Uuid => is_uuid(op_value),
             ObType::Unknown => false,
         };
 
@@ -218,6 +222,8 @@ impl ObTypeLookup {
             ObType::Generator
         } else if ob_type == self.path {
             ObType::Path
+        } else if ob_type == self.uuid || is_uuid(op_value) {
+            ObType::Uuid
         } else {
             // this allows for subtypes of the supported class types,
             // if `ob_type` didn't match any member of self, we try again with the next base type pointer
@@ -256,6 +262,13 @@ fn is_dataclass(op_value: Option<&PyAny>) -> bool {
     }
 }
 
+fn is_uuid(op_value: Option<&PyAny>) -> bool {
+    if let Some(value) = op_value {
+        value.hasattr(intern!(value.py(), "int")).unwrap_or(false)
+    } else {
+        false
+    }
+}
 fn is_pydantic_serializable(op_value: Option<&PyAny>) -> bool {
     if let Some(value) = op_value {
         value
@@ -315,6 +328,8 @@ pub enum ObType {
     Generator,
     // Path
     Path,
+    // Uuid
+    Uuid,
     // unknown type
     Unknown,
 }

src/serializers/ob_type.rs

@@ -129,6 +129,7 @@ combined_serializer! {
         Dataclass: super::type_serializers::dataclass::DataclassSerializer;
         Url: super::type_serializers::url::UrlSerializer;
         MultiHostUrl: super::type_serializers::url::MultiHostUrlSerializer;
+        Uuid: super::type_serializers::uuid::UuidSerializer;
         Any: super::type_serializers::any::AnySerializer;
         Format: super::type_serializers::format::FormatSerializer;
         ToString: super::type_serializers::format::ToStringSerializer;
@@ -254,6 +255,7 @@ impl PyGcTraverse for CombinedSerializer {
             CombinedSerializer::Recursive(inner) => inner.py_gc_traverse(visit),
             CombinedSerializer::TuplePositional(inner) => inner.py_gc_traverse(visit),
             CombinedSerializer::TupleVariable(inner) => inner.py_gc_traverse(visit),
+            CombinedSerializer::Uuid(inner) => inner.py_gc_traverse(visit),
         }
     }
 }

src/serializers/shared.rs

@@ -22,6 +22,7 @@ pub mod tuple;
 pub mod typed_dict;
 pub mod union;
 pub mod url;
+pub mod uuid;
 pub mod with_default;
 
 use super::computed_fields::ComputedFields;

src/serializers/type_serializers/mod.rs

@@ -0,0 +1,91 @@
+use std::borrow::Cow;
+
+use pyo3::prelude::*;
+use pyo3::types::PyDict;
+
+use crate::definitions::DefinitionsBuilder;
+
+use super::{
+    infer_json_key, infer_serialize, infer_to_python, py_err_se_err, BuildSerializer, CombinedSerializer, Extra,
+    IsType, ObType, SerMode, TypeSerializer,
+};
+
+pub(crate) fn uuid_to_string(py_uuid: &PyAny) -> PyResult<String> {
+    Ok(py_uuid.str()?.to_string())
+}
+
+#[derive(Debug, Clone)]
+pub struct UuidSerializer;
+
+impl_py_gc_traverse!(UuidSerializer {});
+
+impl BuildSerializer for UuidSerializer {
+    const EXPECTED_TYPE: &'static str = "uuid";
+
+    fn build(
+        _schema: &PyDict,
+        _config: Option<&PyDict>,
+        _definitions: &mut DefinitionsBuilder<CombinedSerializer>,
+    ) -> PyResult<CombinedSerializer> {
+        Ok(Self {}.into())
+    }
+}
+
+impl TypeSerializer for UuidSerializer {
+    fn to_python(
+        &self,
+        value: &PyAny,
+        include: Option<&PyAny>,
+        exclude: Option<&PyAny>,
+        extra: &Extra,
+    ) -> PyResult<PyObject> {
+        let py = value.py();
+        match extra.ob_type_lookup.is_type(value, ObType::Uuid) {
+            IsType::Exact | IsType::Subclass => match extra.mode {
+                SerMode::Json => Ok(uuid_to_string(value)?.into_py(py)),
+                _ => Ok(value.into_py(py)),
+            },
+            IsType::False => {
+                extra.warnings.on_fallback_py(self.get_name(), value, extra)?;
+                infer_to_python(value, include, exclude, extra)
+            }
+        }
+    }
+
+    fn json_key<'py>(&self, key: &'py PyAny, extra: &Extra) -> PyResult<Cow<'py, str>> {
+        match extra.ob_type_lookup.is_type(key, ObType::Uuid) {
+            IsType::Exact | IsType::Subclass => {
+                let str = uuid_to_string(key)?;
+                Ok(Cow::Owned(str))
+            }
+            IsType::False => {
+                extra.warnings.on_fallback_py(self.get_name(), key, extra)?;
+                infer_json_key(key, extra)
+            }
+        }
+    }
+
+    fn serde_serialize<S: serde::ser::Serializer>(
+        &self,
+        value: &PyAny,
+        serializer: S,
+        include: Option<&PyAny>,
+        exclude: Option<&PyAny>,
+        extra: &Extra,
+    ) -> Result<S::Ok, S::Error> {
+        match extra.ob_type_lookup.is_type(value, ObType::Uuid) {
+            IsType::Exact | IsType::Subclass => {
+                let s = uuid_to_string(value).map_err(py_err_se_err)?;
+                serializer.serialize_str(&s)
+            }
+            IsType::False => {
+                extra.warnings.on_fallback_ser::<S>(self.get_name(), value, extra)?;
+                infer_serialize(value, serializer, include, exclude, extra)
+            }
+        }
+    }
+
+    fn get_name(&self) -> &str {
+        Self::EXPECTED_TYPE
+    }
+}

src/serializers/type_serializers/uuid.rs

@@ -53,6 +53,7 @@ mod tuple;
 mod typed_dict;
 mod union;
 mod url;
+mod uuid;
 mod with_default;
 
 pub use with_default::DefaultType;
@@ -507,6 +508,8 @@ pub fn build_validator<'a>(
         // url types
         url::UrlValidator,
         url::MultiHostUrlValidator,
+        // uuid types
+        uuid::UuidValidator,
         // recursive (self-referencing) models
         definitions::DefinitionRefValidator,
         definitions::DefinitionsValidatorBuilder,
@@ -650,6 +653,8 @@ pub enum CombinedValidator {
     // url types
     Url(url::UrlValidator),
     MultiHostUrl(url::MultiHostUrlValidator),
+    // uuid types
+    Uuid(uuid::UuidValidator),
     // reference to definition, useful for recursive (self-referencing) models
     DefinitionRef(definitions::DefinitionRefValidator),
     // input dependent