model-skill

Pydantic v2 validates data models through field-level and model-level validators. Use field_validator decorators to enforce constraints on individual fields, and model_validator for cross-field validation rules. Type hints provide automatic type checking, while model_validate() and model_dump() handle serialization and deserialization with full type safety.

Yes. Pydantic models work well for API schemas across backend services. Define models with typed fields, validation rules, and serialization options using json_encoders and example schemas. model_validate() deserializes incoming JSON into typed objects, and model_dump() serializes responses back to JSON with strict type enforcement.

Pydantic enforces field constraints through type hints, field validators, and default values. Combine type safety (int, str, datetime) with field_validator for custom logic, and set defaults or required status on each field. Enums provide strict value control, reducing data errors and boilerplate across data pipelines.

Use model_validator with mode='after' to validate relationships between fields after individual field validation completes. Access all model fields in the validator function to enforce integrity rules that depend on multiple fields, such as date ranges or conditional requirements.

No. Pydantic supports schema inheritance and composition. Create base models with shared validation rules, then extend them for specific use cases. This reduces duplication and keeps validation logic centralized across related data structures in APIs and data pipelines.

Pydantic raises a ValidationError with detailed field-level error messages. Errors include the field name, the invalid value, and the constraint that failed. This immediate feedback helps catch data integrity issues early in backend services and prevents invalid data from propagating through pipelines.