Use more explicit warning regarding serialization warning for missing fields (#1415)

Author: sydney-runkle

src/errors/mod.rs

@@ -1,6 +1,3 @@
-use core::fmt;
-use std::borrow::Cow;
-
 use pyo3::prelude::*;
 
 mod line_error;
@@ -33,46 +30,3 @@ pub fn py_err_string(py: Python, err: PyErr) -> String {
         Err(_) => "Unknown Error".to_string(),
     }
 }
-
-// TODO: is_utf8_char_boundary, floor_char_boundary and ceil_char_boundary
-// with builtin methods once https://github.com/rust-lang/rust/issues/93743 is resolved
-// These are just copy pasted from the current implementation
-const fn is_utf8_char_boundary(value: u8) -> bool {
-    // This is bit magic equivalent to: b < 128 || b >= 192
-    (value as i8) >= -0x40
-}
-
-pub fn floor_char_boundary(value: &str, index: usize) -> usize {
-    if index >= value.len() {
-        value.len()
-    } else {
-        let lower_bound = index.saturating_sub(3);
-        let new_index = value.as_bytes()[lower_bound..=index]
-            .iter()
-            .rposition(|b| is_utf8_char_boundary(*b));
-
-        // SAFETY: we know that the character boundary will be within four bytes
-        unsafe { lower_bound + new_index.unwrap_unchecked() }
-    }
-}
-
-pub fn ceil_char_boundary(value: &str, index: usize) -> usize {
-    let upper_bound = Ord::min(index + 4, value.len());
-    value.as_bytes()[index..upper_bound]
-        .iter()
-        .position(|b| is_utf8_char_boundary(*b))
-        .map_or(upper_bound, |pos| pos + index)
-}
-
-pub fn write_truncated_to_50_bytes<F: fmt::Write>(f: &mut F, val: Cow<'_, str>) -> std::fmt::Result {
-    if val.len() > 50 {
-        write!(
-            f,
-            "{}...{}",
-            &val[0..floor_char_boundary(&val, 25)],
-            &val[ceil_char_boundary(&val, val.len() - 24)..]
-        )
-    } else {
-        write!(f, "{val}")
-    }
-}

src/errors/validation_exception.rs

@@ -18,7 +18,7 @@ use crate::errors::LocItem;
 use crate::get_pydantic_version;
 use crate::input::InputType;
 use crate::serializers::{DuckTypingSerMode, Extra, SerMode, SerializationState};
-use crate::tools::{safe_repr, SchemaDict};
+use crate::tools::{safe_repr, write_truncated_to_limited_bytes, SchemaDict};
 
 use super::line_error::ValLineError;
 use super::location::Location;
@@ -526,7 +526,7 @@ impl PyLineError {
             let input_value = self.input_value.bind(py);
             let input_str = safe_repr(input_value);
             write!(output, ", input_value=")?;
-            super::write_truncated_to_50_bytes(&mut output, input_str.to_cow())?;
+            write_truncated_to_limited_bytes(&mut output, &input_str.to_string(), 50)?;
 
             if let Ok(type_) = input_value.get_type().qualname() {
                 write!(output, ", input_type={type_}")?;

src/serializers/extra.rs

@@ -15,7 +15,7 @@ use crate::recursion_guard::ContainsRecursionState;
 use crate::recursion_guard::RecursionError;
 use crate::recursion_guard::RecursionGuard;
 use crate::recursion_guard::RecursionState;
-use crate::tools::safe_repr;
+use crate::tools::truncate_safe_repr;
 use crate::PydanticSerializationError;
 
 /// this is ugly, would be much better if extra could be stored in `SerializationState`
@@ -426,15 +426,10 @@ impl CollectWarnings {
                 .qualname()
                 .unwrap_or_else(|_| PyString::new_bound(value.py(), "<unknown python object>"));
 
-            let input_str = safe_repr(value);
-            let mut value_str = String::with_capacity(100);
-            value_str.push_str("with value `");
-            crate::errors::write_truncated_to_50_bytes(&mut value_str, input_str.to_cow())
-                .expect("Writing to a `String` failed");
-            value_str.push('`');
+            let value_str = truncate_safe_repr(value, None);
 
             self.add_warning(format!(
-                "Expected `{field_type}` but got `{type_name}` {value_str} - serialized value may not be as expected"
+                "Expected `{field_type}` but got `{type_name}` with value `{value_str}` - serialized value may not be as expected"
             ));
         }
     }

src/serializers/fields.rs

@@ -9,6 +9,7 @@ use smallvec::SmallVec;
 
 use crate::serializers::extra::SerCheck;
 use crate::serializers::DuckTypingSerMode;
+use crate::tools::truncate_safe_repr;
 use crate::PydanticSerializationUnexpectedValue;
 
 use super::computed_fields::ComputedFields;
@@ -210,7 +211,24 @@ impl GeneralFieldsSerializer {
             // Check for missing fields, we can't have extra fields here
             && self.required_fields > used_req_fields
         {
-            Err(PydanticSerializationUnexpectedValue::new_err(None))
+            let required_fields = self.required_fields;
+            let type_name = match extra.model {
+                Some(model) => model
+                    .get_type()
+                    .qualname()
+                    .ok()
+                    .unwrap_or_else(|| PyString::new_bound(py, "<unknown python object>"))
+                    .to_string(),
+                None => "<unknown python object>".to_string(),
+            };
+            let field_value = match extra.model {
+                Some(model) => truncate_safe_repr(model, Some(100)),
+                None => "<unknown python object>".to_string(),
+            };
+
+            Err(PydanticSerializationUnexpectedValue::new_err(Some(format!(
+                "Expected {required_fields} fields but got {used_req_fields} for type `{type_name}` with value `{field_value}` - serialized value may not be as expected."
+            ))))
         } else {
             Ok(output_dict)
         }

src/serializers/type_serializers/model.rs

@@ -167,9 +167,9 @@ impl TypeSerializer for ModelSerializer {
     ) -> PyResult<PyObject> {
         let model = Some(value);
         let duck_typing_ser_mode = extra.duck_typing_ser_mode.next_mode();
+
         let model_extra = Extra {
             model,
-            field_name: None,
             duck_typing_ser_mode,
             ..*extra
         };
@@ -221,7 +221,6 @@ impl TypeSerializer for ModelSerializer {
         let duck_typing_ser_mode = extra.duck_typing_ser_mode.next_mode();
         let model_extra = Extra {
             model,
-            field_name: None,
             duck_typing_ser_mode,
             ..*extra
         };

src/serializers/type_serializers/union.rs

@@ -8,10 +8,9 @@ use std::borrow::Cow;
 use crate::build_tools::py_schema_err;
 use crate::common::union::{Discriminator, SMALL_UNION_THRESHOLD};
 use crate::definitions::DefinitionsBuilder;
-use crate::errors::write_truncated_to_50_bytes;
 use crate::lookup_key::LookupKey;
 use crate::serializers::type_serializers::py_err_se_err;
-use crate::tools::{safe_repr, SchemaDict};
+use crate::tools::{truncate_safe_repr, SchemaDict};
 use crate::PydanticSerializationUnexpectedValue;
 
 use super::{
@@ -446,15 +445,10 @@ impl TaggedUnionSerializer {
             Discriminator::Function(func) => func.call1(py, (value,)).ok(),
         };
         if discriminator_value.is_none() {
-            let input_str = safe_repr(value);
-            let mut value_str = String::with_capacity(100);
-            value_str.push_str("with value `");
-            write_truncated_to_50_bytes(&mut value_str, input_str.to_cow()).expect("Writing to a `String` failed");
-            value_str.push('`');
-
+            let value_str = truncate_safe_repr(value, None);
             extra.warnings.custom_warning(
                 format!(
-                    "Failed to get discriminator value for tagged union serialization {value_str} - defaulting to left to right union serialization."
+                    "Failed to get discriminator value for tagged union serialization with value `{value_str}` - defaulting to left to right union serialization."
                 )
             );
         }

src/tools.rs

@@ -1,4 +1,4 @@
-use std::borrow::Cow;
+use core::fmt;
 
 use pyo3::exceptions::PyKeyError;
 use pyo3::prelude::*;
@@ -96,15 +96,6 @@ pub enum ReprOutput<'py> {
     Fallback(String),
 }
 
-impl ReprOutput<'_> {
-    pub fn to_cow(&self) -> Cow<'_, str> {
-        match self {
-            ReprOutput::Python(s) => s.to_string_lossy(),
-            ReprOutput::Fallback(s) => s.into(),
-        }
-    }
-}
-
 impl std::fmt::Display for ReprOutput<'_> {
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
         match self {
@@ -124,6 +115,15 @@ pub fn safe_repr<'py>(v: &Bound<'py, PyAny>) -> ReprOutput<'py> {
     }
 }
 
+pub fn truncate_safe_repr(v: &Bound<'_, PyAny>, max_len: Option<usize>) -> String {
+    let max_len = max_len.unwrap_or(50); // default to 100 bytes
+    let input_str = safe_repr(v);
+    let mut limited_str = String::with_capacity(max_len);
+    write_truncated_to_limited_bytes(&mut limited_str, &input_str.to_string(), max_len)
+        .expect("Writing to a `String` failed");
+    limited_str
+}
+
 pub fn extract_i64(v: &Bound<'_, PyAny>) -> Option<i64> {
     #[cfg(PyPy)]
     if !v.is_instance_of::<pyo3::types::PyInt>() {
@@ -146,3 +146,47 @@ pub(crate) fn new_py_string<'py>(py: Python<'py>, s: &str, cache_str: StringCach
         pystring_fast_new(py, s, ascii_only)
     }
 }
+
+// TODO: is_utf8_char_boundary, floor_char_boundary and ceil_char_boundary
+// with builtin methods once https://github.com/rust-lang/rust/issues/93743 is resolved
+// These are just copy pasted from the current implementation
+const fn is_utf8_char_boundary(value: u8) -> bool {
+    // This is bit magic equivalent to: b < 128 || b >= 192
+    (value as i8) >= -0x40
+}
+
+pub fn floor_char_boundary(value: &str, index: usize) -> usize {
+    if index >= value.len() {
+        value.len()
+    } else {
+        let lower_bound = index.saturating_sub(3);
+        let new_index = value.as_bytes()[lower_bound..=index]
+            .iter()
+            .rposition(|b| is_utf8_char_boundary(*b));
+
+        // SAFETY: we know that the character boundary will be within four bytes
+        unsafe { lower_bound + new_index.unwrap_unchecked() }
+    }
+}
+
+pub fn ceil_char_boundary(value: &str, index: usize) -> usize {
+    let upper_bound = Ord::min(index + 4, value.len());
+    value.as_bytes()[index..upper_bound]
+        .iter()
+        .position(|b| is_utf8_char_boundary(*b))
+        .map_or(upper_bound, |pos| pos + index)
+}
+
+pub fn write_truncated_to_limited_bytes<F: fmt::Write>(f: &mut F, val: &str, max_len: usize) -> std::fmt::Result {
+    if val.len() > max_len {
+        let mid_point = max_len.div_ceil(2);
+        write!(
+            f,
+            "{}...{}",
+            &val[0..floor_char_boundary(val, mid_point)],
+            &val[ceil_char_boundary(val, val.len() - (mid_point - 1))..]
+        )
+    } else {
+        write!(f, "{val}")
+    }
+}

tests/serializers/test_model.py

@@ -13,7 +13,13 @@
 import pytest
 from dirty_equals import IsJson
 
-from pydantic_core import PydanticSerializationError, SchemaSerializer, SchemaValidator, core_schema
+from pydantic_core import (
+    PydanticSerializationError,
+    PydanticSerializationUnexpectedValue,
+    SchemaSerializer,
+    SchemaValidator,
+    core_schema,
+)
 
 from ..conftest import plain_repr
 
@@ -1084,20 +1090,68 @@ class Model:
 
 
 def test_no_warn_on_exclude() -> None:
-    warnings.simplefilter('error')
+    with warnings.catch_warnings():
+        warnings.simplefilter('error')
+
+        s = SchemaSerializer(
+            core_schema.model_schema(
+                BasicModel,
+                core_schema.model_fields_schema(
+                    {
+                        'a': core_schema.model_field(core_schema.int_schema()),
+                        'b': core_schema.model_field(core_schema.int_schema()),
+                    }
+                ),
+            )
+        )
+
+        value = BasicModel(a=0, b=1)
+        assert s.to_python(value, exclude={'b'}) == {'a': 0}
+        assert s.to_python(value, mode='json', exclude={'b'}) == {'a': 0}
+
+
+def test_warn_on_missing_field() -> None:
+    class AModel(BasicModel): ...
+
+    class BModel(BasicModel): ...
 
     s = SchemaSerializer(
         core_schema.model_schema(
             BasicModel,
             core_schema.model_fields_schema(
                 {
-                    'a': core_schema.model_field(core_schema.int_schema()),
-                    'b': core_schema.model_field(core_schema.int_schema()),
+                    'root': core_schema.model_field(
+                        core_schema.tagged_union_schema(
+                            choices={
+                                'a': core_schema.model_schema(
+                                    AModel,
+                                    core_schema.model_fields_schema(
+                                        {
+                                            'type': core_schema.model_field(core_schema.literal_schema(['a'])),
+                                            'a': core_schema.model_field(core_schema.int_schema()),
+                                        }
+                                    ),
+                                ),
+                                'b': core_schema.model_schema(
+                                    BModel,
+                                    core_schema.model_fields_schema(
+                                        {
+                                            'type': core_schema.model_field(core_schema.literal_schema(['b'])),
+                                            'b': core_schema.model_field(core_schema.int_schema()),
+                                        }
+                                    ),
+                                ),
+                            },
+                            discriminator='type',
+                        )
+                    ),
                 }
             ),
         )
     )
 
-    value = BasicModel(a=0, b=1)
-    assert s.to_python(value, exclude={'b'}) == {'a': 0}
-    assert s.to_python(value, mode='json', exclude={'b'}) == {'a': 0}
+    with pytest.raises(
+        PydanticSerializationUnexpectedValue, match='Expected 2 fields but got 1 for type `.*AModel` with value `.*`.+'
+    ):
+        value = BasicModel(root=AModel(type='a'))
+        s.to_python(value)