Add extra parameter to the validate functions (#1722)

Author: anvilpete

benches/main.rs

@@ -6,7 +6,7 @@ from _typeshed import SupportsAllComparisons
 from typing_extensions import LiteralString, Self, TypeAlias
 
 from pydantic_core import ErrorDetails, ErrorTypeInfo, InitErrorDetails, MultiHostHost
-from pydantic_core.core_schema import CoreConfig, CoreSchema, ErrorType
+from pydantic_core.core_schema import CoreConfig, CoreSchema, ErrorType, ExtraBehavior
 
 __all__ = [
     '__version__',
@@ -92,6 +92,7 @@ class SchemaValidator:
         input: Any,
         *,
         strict: bool | None = None,
+        extra: ExtraBehavior | None = None,
         from_attributes: bool | None = None,
         context: Any | None = None,
         self_instance: Any | None = None,
@@ -106,6 +107,8 @@ class SchemaValidator:
             input: The Python object to validate.
             strict: Whether to validate the object in strict mode.
                 If `None`, the value of [`CoreConfig.strict`][pydantic_core.core_schema.CoreConfig] is used.
+            extra: Whether to ignore, allow, or forbid extra data during model validation.
+                If `None`, the value of [`CoreConfig.extra_fields_behavior`][pydantic_core.core_schema.CoreConfig] is used.
             from_attributes: Whether to validate objects as inputs to models by extracting attributes.
                 If `None`, the value of [`CoreConfig.from_attributes`][pydantic_core.core_schema.CoreConfig] is used.
             context: The context to use for validation, this is passed to functional validators as
@@ -130,6 +133,7 @@ class SchemaValidator:
         input: Any,
         *,
         strict: bool | None = None,
+        extra: ExtraBehavior | None = None,
         from_attributes: bool | None = None,
         context: Any | None = None,
         self_instance: Any | None = None,
@@ -150,6 +154,7 @@ class SchemaValidator:
         input: str | bytes | bytearray,
         *,
         strict: bool | None = None,
+        extra: ExtraBehavior | None = None,
         context: Any | None = None,
         self_instance: Any | None = None,
         allow_partial: bool | Literal['off', 'on', 'trailing-strings'] = False,
@@ -169,6 +174,8 @@ class SchemaValidator:
             input: The JSON data to validate.
             strict: Whether to validate the object in strict mode.
                 If `None`, the value of [`CoreConfig.strict`][pydantic_core.core_schema.CoreConfig] is used.
+            extra: Whether to ignore, allow, or forbid extra data during model validation.
+                If `None`, the value of [`CoreConfig.extra_fields_behavior`][pydantic_core.core_schema.CoreConfig] is used.
             context: The context to use for validation, this is passed to functional validators as
                 [`info.context`][pydantic_core.core_schema.ValidationInfo.context].
             self_instance: An instance of a model set attributes on from validation.
@@ -190,6 +197,7 @@ class SchemaValidator:
         input: _StringInput,
         *,
         strict: bool | None = None,
+        extra: ExtraBehavior | None = None,
         context: Any | None = None,
         allow_partial: bool | Literal['off', 'on', 'trailing-strings'] = False,
         by_alias: bool | None = None,
@@ -205,6 +213,8 @@ class SchemaValidator:
             input: The input as a string, or bytes/bytearray if `strict=False`.
             strict: Whether to validate the object in strict mode.
                 If `None`, the value of [`CoreConfig.strict`][pydantic_core.core_schema.CoreConfig] is used.
+            extra: Whether to ignore, allow, or forbid extra data during model validation.
+                If `None`, the value of [`CoreConfig.extra_fields_behavior`][pydantic_core.core_schema.CoreConfig] is used.
             context: The context to use for validation, this is passed to functional validators as
                 [`info.context`][pydantic_core.core_schema.ValidationInfo.context].
             allow_partial: Whether to allow partial validation; if `True` errors in the last element of sequences
@@ -227,6 +237,7 @@ class SchemaValidator:
         field_value: Any,
         *,
         strict: bool | None = None,
+        extra: ExtraBehavior | None = None,
         from_attributes: bool | None = None,
         context: Any | None = None,
         by_alias: bool | None = None,
@@ -241,6 +252,8 @@ class SchemaValidator:
             field_value: The value to assign to the field.
             strict: Whether to validate the object in strict mode.
                 If `None`, the value of [`CoreConfig.strict`][pydantic_core.core_schema.CoreConfig] is used.
+            extra: Whether to ignore, allow, or forbid extra data during model validation.
+                If `None`, the value of [`CoreConfig.extra_fields_behavior`][pydantic_core.core_schema.CoreConfig] is used.
             from_attributes: Whether to validate objects as inputs to models by extracting attributes.
                 If `None`, the value of [`CoreConfig.from_attributes`][pydantic_core.core_schema.CoreConfig] is used.
             context: The context to use for validation, this is passed to functional validators as

python/pydantic_core/_pydantic_core.pyi

@@ -1,5 +1,6 @@
 use std::error::Error;
 use std::fmt;
+use std::str::FromStr;
 
 use pyo3::exceptions::PyException;
 use pyo3::prelude::*;
@@ -176,7 +177,7 @@ macro_rules! py_schema_err {
 pub(crate) use py_schema_err;
 
 #[derive(Debug, Clone, Copy, Eq, PartialEq)]
-pub(crate) enum ExtraBehavior {
+pub enum ExtraBehavior {
     Allow,
     Forbid,
     Ignore,
@@ -197,12 +198,22 @@ impl ExtraBehavior {
         )?
         .flatten();
         let res = match extra_behavior.as_ref().map(|s| s.to_str()).transpose()? {
-            Some("allow") => Self::Allow,
-            Some("ignore") => Self::Ignore,
-            Some("forbid") => Self::Forbid,
-            Some(v) => return py_schema_err!("Invalid extra_behavior: `{}`", v),
+            Some(s) => Self::from_str(s)?,
             None => default,
         };
         Ok(res)
     }
 }
+
+impl FromStr for ExtraBehavior {
+    type Err = PyErr;
+
+    fn from_str(s: &str) -> Result<Self, Self::Err> {
+        match s {
+            "allow" => Ok(Self::Allow),
+            "forbid" => Ok(Self::Forbid),
+            "ignore" => Ok(Self::Ignore),
+            s => py_schema_err!("Invalid extra_behavior: `{}`", s),
+        }
+    }
+}

src/build_tools.rs

@@ -45,7 +45,7 @@ impl PyUrl {
     pub fn py_new(py: Python, url: &Bound<'_, PyAny>) -> PyResult<Self> {
         let schema_obj = SCHEMA_DEFINITION_URL
             .get_or_init(py, || build_schema_validator(py, "url"))
-            .validate_python(py, url, None, None, None, None, false.into(), None, None)?;
+            .validate_python(py, url, None, None, None, None, None, false.into(), None, None)?;
         schema_obj.extract(py)
     }
 
@@ -225,7 +225,7 @@ impl PyMultiHostUrl {
     pub fn py_new(py: Python, url: &Bound<'_, PyAny>) -> PyResult<Self> {
         let schema_obj = SCHEMA_DEFINITION_MULTI_HOST_URL
             .get_or_init(py, || build_schema_validator(py, "multi-host-url"))
-            .validate_python(py, url, None, None, None, None, false.into(), None, None)?;
+            .validate_python(py, url, None, None, None, None, None, false.into(), None, None)?;
         schema_obj.extract(py)
     }
 

src/url.rs

@@ -241,13 +241,15 @@ impl ArgumentsV3Validator {
 
         let validate_by_alias = state.validate_by_alias_or(self.validate_by_alias);
         let validate_by_name = state.validate_by_name_or(self.validate_by_name);
+        let extra_behavior = state.extra_behavior_or(self.extra);
 
         // Keep track of used keys for extra behavior:
-        let mut used_keys: Option<AHashSet<&str>> = if self.extra == ExtraBehavior::Ignore || mapping.is_py_get_attr() {
-            None
-        } else {
-            Some(AHashSet::with_capacity(self.parameters.len()))
-        };
+        let mut used_keys: Option<AHashSet<&str>> =
+            if extra_behavior == ExtraBehavior::Ignore || mapping.is_py_get_attr() {
+                None
+            } else {
+                Some(AHashSet::with_capacity(self.parameters.len()))
+            };
 
         for parameter in &self.parameters {
             let lookup_key = parameter
@@ -492,7 +494,7 @@ impl ArgumentsV3Validator {
             mapping.iterate(ValidateExtra {
                 used_keys,
                 errors: &mut errors,
-                extra_behavior: self.extra,
+                extra_behavior,
             })??;
         }
 
@@ -524,6 +526,7 @@ impl ArgumentsV3Validator {
 
         let validate_by_alias = state.validate_by_alias_or(self.validate_by_alias);
         let validate_by_name = state.validate_by_name_or(self.validate_by_name);
+        let extra_behavior = state.extra_behavior_or(self.extra);
 
         // go through non variadic parameters, getting the value from args or kwargs and validating it
         for (index, parameter) in self.parameters.iter().filter(|p| !p.is_variadic()).enumerate() {
@@ -687,7 +690,7 @@ impl ArgumentsV3Validator {
 
                         match maybe_var_kwargs_parameter {
                             None => {
-                                if self.extra == ExtraBehavior::Forbid {
+                                if extra_behavior == ExtraBehavior::Forbid {
                                     errors.push(ValLineError::new_with_loc(
                                         ErrorTypeDefaults::UnexpectedKeywordArgument,
                                         value,

src/validators/arguments_v3.rs

@@ -155,6 +155,7 @@ impl Validator for DataclassArgsValidator {
         let mut used_keys: AHashSet<&str> = AHashSet::with_capacity(self.fields.len());
 
         let state = &mut state.rebind_extra(|extra| extra.data = Some(output_dict.clone()));
+        let extra_behavior = state.extra_behavior_or(self.extra_behavior);
 
         let validate_by_alias = state.validate_by_alias_or(self.validate_by_alias);
         let validate_by_name = state.validate_by_name_or(self.validate_by_name);
@@ -308,7 +309,7 @@ impl Validator for DataclassArgsValidator {
                         Ok(either_str) => {
                             if !used_keys.contains(either_str.as_cow()?.as_ref()) {
                                 // Unknown / extra field
-                                match self.extra_behavior {
+                                match extra_behavior {
                                     ExtraBehavior::Forbid => {
                                         errors.push(ValLineError::new_with_loc(
                                             ErrorTypeDefaults::UnexpectedKeywordArgument,
@@ -379,6 +380,7 @@ impl Validator for DataclassArgsValidator {
         state: &mut ValidationState<'_, 'py>,
     ) -> ValResult<PyObject> {
         let dict = obj.downcast::<PyDict>()?;
+        let extra_behavior = state.extra_behavior_or(self.extra_behavior);
 
         let ok = |output: PyObject| {
             dict.set_item(field_name, output)?;
@@ -426,7 +428,7 @@ impl Validator for DataclassArgsValidator {
             // Handle extra (unknown) field
             // We partially use the extra_behavior for initialization / validation
             // to determine how to handle assignment
-            match self.extra_behavior {
+            match extra_behavior {
                 // For dataclasses we allow assigning unknown fields
                 // to match stdlib dataclass behavior
                 ExtraBehavior::Allow => ok(field_value.clone().unbind()),

src/validators/dataclass.rs

@@ -4,6 +4,7 @@ use std::sync::Arc;
 use pyo3::types::{PyDict, PyString};
 use pyo3::{prelude::*, IntoPyObjectExt, PyTraverseError, PyVisit};
 
+use crate::build_tools::ExtraBehavior;
 use crate::errors::{ErrorType, LocItem, ValError, ValResult};
 use crate::input::{BorrowInput, GenericIterator, Input};
 use crate::py_gc::PyGcTraverse;
@@ -220,6 +221,7 @@ pub struct InternalValidator {
     // TODO, do we need data?
     data: Option<Py<PyDict>>,
     strict: Option<bool>,
+    extra_behavior: Option<ExtraBehavior>,
     from_attributes: Option<bool>,
     context: Option<PyObject>,
     self_instance: Option<PyObject>,
@@ -252,6 +254,7 @@ impl InternalValidator {
             validator,
             data: extra.data.as_ref().map(|d| d.clone().into()),
             strict: extra.strict,
+            extra_behavior: extra.extra_behavior,
             from_attributes: extra.from_attributes,
             context: extra.context.map(|d| d.clone().unbind()),
             self_instance: extra.self_instance.map(|d| d.clone().unbind()),
@@ -277,6 +280,7 @@ impl InternalValidator {
             input_type: self.validation_mode,
             data: self.data.as_ref().map(|data| data.bind(py).clone()),
             strict: self.strict,
+            extra_behavior: self.extra_behavior,
             from_attributes: self.from_attributes,
             field_name: Some(PyString::new(py, field_name)),
             context: self.context.as_ref().map(|data| data.bind(py)),
@@ -315,6 +319,7 @@ impl InternalValidator {
             input_type: self.validation_mode,
             data: self.data.as_ref().map(|data| data.bind(py).clone()),
             strict: self.strict,
+            extra_behavior: self.extra_behavior,
             from_attributes: self.from_attributes,
             field_name: None,
             context: self.context.as_ref().map(|data| data.bind(py)),