validatedata

An easier way to validate data in python.

Seven validation modes – one simple syntax.

1. `validator()` – One word: speed. Ideal for high‑throughput streaming. msgspec, handwritten code, and this function will compete for first place.

2. `validate_data_fast()` – same compiled speed but with full error messages (preview of the next‑gen engine).

3. `validate_data()` – general‑purpose validation with detailed errors, nested structures, and optional mutation.

4. `@validate` – decorator for function argument validation.

5. `@validate_types` – decorator that uses Python type annotations.

6. `FastModel` – declarative, typed models with compiled validation, rich error messages, and serialization.

7. `autovalidate` / `autovalidate_package` – automatically apply @validate_types to entire modules or packages.

Validatedata gives you expressive, inline validation rules without defining model classes. It fits naturally into any Python workflow – from lightweight scripts to high‑volume data processing.

New in v0.6: - `FastModel` – declarative models with compiled validation, cross‑field checks, and zero‑overhead serialization. - `validate_data_fast` – the speed of validator() combined with rich error messages. This is an experimental fast path that will eventually replace validate_data once the API stabilises. - `autovalidate` & `autovalidate_package` – automatically apply @validate_types to whole modules or packages. - Custom type registration – add your own type checkers with register_type / unregister_type. - `check_rule` – validate rule dicts before using them. - `VALID_RULE_KEYS` – introspection of all recognised rule keys.

Benchmarks (1 million repetitions)

Test	validatedata	manual	pydantic v2	msgspec	beartype	fastjsonschema
Scalar: type (int)	0.0968s	0.0779s	0.3878s	0.0947s	0.3384s	0.1444s
Scalar: type + range	0.1160s	0.1267s	0.1249s	0.1276s	0.3628s	0.1447s
Dict (valid)	0.8388s	0.8943s	2.5218s	1.2571s		3.8539s
Dict (invalid)	0.2005s	0.1830s	2.5063s	0.7378s		3.1960s

Note

The “manual” column represents hand‑written if statements – fastest but not reusable or composable. Validatedata’s small overhead buys you maintainability and expressiveness.

Key features

• Compact pipe-syntax shorthand: 'str|strip|min:3|max:32'

• Mirror-structure rules that match the shape of your data

• Rich built-in types: email, url, ip, uuid, semver, slug, color, phone, and more

• Function and method decorators with async support

• Conditional validation, transforms, and custom error messages

• Auto‑validation of entire packages

• Register your own custom types

---

Getting started

• Quick Start
  • Installation
  • Your first validation
  • Seven ways to validate
  • The keys wrapper
  • Reading the result
  • Parameters

Guides

• Rules Reference
  • Types
  • Custom types
  • Dict rule form
  • Checking a rule dict before use
  • Shorthand rule strings
  • Nested rules
  • Transforms
  • Conditional validation
  • Custom error messages
  • Validating rule dicts
• Mirror-Structure Rules
  • The problem with explicit nested rules
  • Mirror-structure shorthand
  • Error paths
  • Multi-level nesting
  • Mixing flat and nested rules
  • Using the keys wrapper
  • Transforms on nested fields
  • Mutate and data reconstruction
  • Depth limit
  • Mixing with explicit dict rules
  • Reference: how expansion works
• Decorators
  • @validate
  • @validate_types
  • Parameters (both decorators)
  • Return values
• Auto‑validation
  • autovalidate() – single module
  • autovalidate_package() – entire package
• Fast validation with error messages – validate_data_fast
  • Basic usage
  • Differences from validate_data
  • Performance comparison
  • Current limitations (will be removed before merging)
  • When to use
  • Feedback
• FastModel – Declarative Models
  • Basic usage
  • Full example — most Rule options at once
  • Cross‑field validation
  • Serialisation / Deserialisation
  • Partial validation (no instantiation)
  • Fast boolean checks
  • Schema introspection
  • OpenAPI Schema Generation
  • Rule dict and compiled validator access
• Examples
  • Fast validation with validator()
  • User registration
  • Flask route with the decorator
  • Application config file
  • Bulk data import (high‑performance)
  • Bulk data import (with full error messages)
  • Conditional fields on a checkout form
  • Normalising data before saving
  • Auto‑validation of a module
  • Validating a whole package
  • Using the fast validator with error messages
  • Enforcing type hints with enforce_hints
  • Using a custom decorator
  • FastModel for structured data

Project

• Changelog