specflow 1.0.0

A Python library for JSON Schema validation with a fluent, type-safe API

SpecFlow

A modern, type-safe Python library for JSON Schema validation with a fluent, composable API. SpecFlow provides an intuitive way to define, validate, and serialize JSON schemas programmatically.

Features

• Type-Safe Validation - Built with Python type hints for better IDE support and type checking
• Composable Schemas - Combine schemas using AnyOf, OneOf, and Not compositions
• Conditional Logic - Define conditional validation rules with if/then/else conditions
• Rich Constraints - Support for string patterns, numeric ranges, array constraints, and more
• Clear Error Messages - Descriptive validation errors with path information
• JSON Schema Compatible - Export schemas to JSON Schema format
• Extensible - Create custom constraints for domain-specific validation

Installation

pip install specflow

Quick Start

from specflow import Schema, Field

# Define a user schema
user_schema = Schema(
    title="User",
    description="A user object",
    properties=[
        Field(
            title="username",
            description="User's username",
            min_length=3,
            max_length=20,
            pattern=r"^[a-zA-Z0-9_]+$"
        ),
        Field(
            title="email",
            description="User's email address",
            pattern=r"^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$"
        ),
        Field(
            title="age",
            description="User's age",
            minimum=0,
            maximum=150,
            default=25  # int default infers Integer type
        ),
        Field(
            title="is_active",
            description="Whether the user account is active",
            default=True  # bool default infers Boolean type
        )
    ]
)

# Validate data
data = {
    "username": "john_doe",
    "email": "[e-mail]",
    "age": 25,
    "is_active": True
}

try:
    user_schema(data)
    print("✓ Validation passed!")
except ValidationError as e:
    print(f"✗ Validation failed: {e}")

Core Components

Field Function with Type Inference

SpecFlow provides a smart Field() function that automatically infers the field type based on the parameters you provide. You can also explicitly specify the type using the type_ parameter.

Automatic Type Inference

The Field() function infers the type based on:

• String-specific parameters: min_length, max_length, pattern, enum, const, or a str default
• Integer-specific parameters: numeric constraints with an int default
• Number (Float) parameters: numeric constraints with a float default
• Boolean: a bool default value
• Array-specific parameters: min_items, max_items, items, prefix_items

String Fields

from specflow import Field

# Inferred as String due to string-specific parameters
Field(
    title="username",
    description="User's username",
    min_length=3,
    max_length=20,
    pattern=r"^[a-zA-Z0-9_]+$"
)

# With enum
Field(
    title="role",
    enum=["admin", "user", "guest"]
)

# With const
Field(
    title="version",
    const="1.0.0"
)

# Explicit type
Field(
    title="name",
    type_="string",
    default="Anonymous"
)

Integer Fields

from specflow import Field

# Inferred as Integer due to int default
Field(
    title="age",
    minimum=0,
    maximum=150,
    default=25  # int default
)

# With multiple of constraint
Field(
    title="quantity",
    minimum=1,
    mult=5,  # Must be multiple of 5
    default=10
)

# Explicit type
Field(
    title="count",
    type_="integer",
    minimum=0
)

Number (Float) Fields

from specflow import Field

# Inferred as Number due to float default
Field(
    title="price",
    minimum=0.0,
    maximum=999.99,
    default=19.99  # float default
)

# With precision constraint
Field(
    title="rating",
    minimum=0.0,
    maximum=5.0,
    mult=0.5,  # Increments of 0.5
    default=4.5
)

# Explicit type
Field(
    title="temperature",
    type_="number",
    minimum=-273.15
)

Boolean Fields

from specflow import Field

# Inferred as Boolean due to bool default
Field(
    title="is_active",
    default=True
)

# Explicit type
Field(
    title="enabled",
    type_="boolean",
    default=False
)

Arrays

from specflow import Field

# Array with single item type (inferred as Array due to items parameter)
Field(
    title="tags",
    items=Field(title="tag", type_="string"),
    min_items=1,
    max_items=10
)

# Array with tuple validation (prefix items)
Field(
    title="coordinates",
    prefix_items=[
        Field(title="latitude", type_="number"),
        Field(title="longitude", type_="number")
    ]
)

# Mixed array with prefix items and additional items
Field(
    title="mixed",
    prefix_items=[
        Field(title="name", type_="string"),
        Field(title="age", default=0)  # Integer inferred
    ],
    items=Field(title="flags", default=False)  # Boolean inferred
)

# Explicit type
Field(
    title="numbers",
    type_="array",
    items=Field(title="num", default=0)
)

Schemas

Schemas are composite objects that group multiple properties:

from specflow import Schema, Field

address_schema = Schema(
    title="Address",
    properties=[
        Field(title="street", type_="string"),
        Field(title="city", type_="string"),
        Field(title="zipcode", pattern=r"^\d{5}$")
    ]
)

# Nested schemas
user_schema = Schema(
    title="User",
    properties=[
        Field(title="name", type_="string"),
        address_schema  # Nested schema
    ]
)

Compositions

AnyOf

Validates if the data matches at least one of the specified schemas:

from specflow import AnyOf, Schema, Field

contact_schema = Schema(
    title="Contact",
    properties=[
        Field(title="name", type_="string"),
        AnyOf(
            Field(title="email", type_="string"),
            Field(title="phone", type_="string")
        )
    ]
)

# Valid: has name and email
data1 = {"name": "John", "email": "[e-mail]"}

# Valid: has name and phone
data2 = {"name": "Jane", "phone": "[phone]"}

# Valid: has all three
data3 = {"name": "Bob", "email": "[e-mail]", "phone": "[phone]"}

OneOf

Validates if the data matches exactly one of the specified schemas:

from specflow import OneOf, Field

payment_method = OneOf(
    Field(title="credit_card", type_="string"),
    Field(title="paypal_email", type_="string"),
    Field(title="bank_account", type_="string")
)

# Valid: exactly one payment method
data = {"credit_card": "4111-1111-1111-1111"}

# Invalid: multiple payment methods
invalid_data = {
    "credit_card": "4111-1111-1111-1111",
    "paypal_email": "[e-mail]"
}

Not

Validates if the data does not match the specified schema:

from specflow import Not, Schema, Field

schema = Schema(
    title="Example",
    properties=[
        Field(title="username", type_="string"),
        Not(
            Field(title="banned_word", const="admin")
        )
    ]
)

Conditions

Define conditional validation rules with if/then/else logic:

from specflow import Schema, Condition, Field

# If country is "US", then require state; otherwise require province
address_schema = Schema(
    title="Address",
    properties=[
        Field(title="country", type_="string"),
        Field(title="state", type_="string", nullable=True),
        Field(title="province", type_="string", nullable=True)
    ],
    conditions=[
        Condition(
            if_=Field(title="country", const="US"),
            then_=Field(title="state", min_length=2),
            else_=Field(title="province", min_length=1)
        )
    ]
)

Validation

Basic Validation

try:
    schema(data)
    print("Validation passed!")
except ValidationError as e:
    print(f"Validation error: {e}")

Strict vs Non-Strict Mode

# Strict mode (default): extra fields not allowed
schema(data, strict=True)

# Non-strict mode: extra fields allowed
schema(data, strict=False)

Error Paths

SpecFlow provides detailed error paths for nested validation failures:

from specflow import Schema, Field

schema = Schema(
    title="User",
    properties=[
        Field(title="name", type_="string"),
        Field(
            title="addresses",
            items=Schema(
                title="Address",
                properties=[
                    Field(title="street", type_="string"),
                    Field(title="zipcode", pattern=r"^\d{5}$")
                ]
            )
        )
    ]
)

data = {
    "name": "John",
    "addresses": [
        {"street": "123 Main St", "zipcode": "12345"},
        {"street": "456 Oak Ave", "zipcode": "INVALID"}
    ]
}

try:
    schema(data)
except ValidationError as e:
    print(e)
    # Output: Validation failed at addresses[1].zipcode: Must match pattern: ^\d{5}$, got INVALID

Schema Export

Export schemas to JSON Schema format:

schema_dict = schema.to_dict()
print(schema_dict)

Advanced Examples

E-commerce Product Schema

from specflow import Schema, OneOf, Field

product_schema = Schema(
    title="Product",
    description="E-commerce product",
    properties=[
        Field(
            title="id",
            pattern=r"^PRD-\d{6}$"
        ),
        Field(
            title="name",
            min_length=3,
            max_length=100
        ),
        Field(
            title="description",
            max_length=1000
        ),
        Field(
            title="price",
            minimum=0.01,
            mult=0.01,
            default=0.0  # Float default infers Number
        ),
        Field(
            title="stock",
            minimum=0,
            default=0  # Int default infers Integer
        ),
        Field(
            title="categories",
            items=Field(title="category", type_="string"),
            min_items=1,
            max_items=5
        ),
        Field(
            title="tags",
            items=Field(title="tag", type_="string"),
            max_items=10
        ),
        Field(
            title="in_stock",
            default=True
        ),
        OneOf(
            Field(title="color", type_="string"),
            Field(title="size", type_="string"),
            Field(title="material", type_="string")
        )
    ]
)

API Response Schema with Conditions

from specflow import Schema, Condition, Field

api_response = Schema(
    title="APIResponse",
    properties=[
        Field(title="status_code", default=200),
        Field(title="success", default=True),
        Field(title="message", type_="string", nullable=True),
        Field(title="data", type_="string", nullable=True),
        Field(title="error", type_="string", nullable=True)
    ],
    conditions=[
        Condition(
            if_=Field(title="success", default=True),
            then_=Field(title="data", min_length=1),
            else_=Field(title="error", min_length=1)
        )
    ]
)

Error Handling

SpecFlow raises ValidationError exceptions with detailed information:

from specflow.core.exceptions import ValidationError

try:
    schema(data)
except ValidationError as e:
    print(f"Message: {e.message}")
    print(f"Path: {e.path}")
    print(f"Full error: {e}")

Custom Constraints

You can create your own custom constraints by extending the Constraint base class. This allows you to implement domain-specific validation rules that go beyond the built-in constraints.

The Constraint Interface

To create a custom constraint, you need to:

1. Import Constraint from specflow
2. Extend Constraint[T] where T is the type you're validating (str, int, float, bool)
3. Implement three required properties/methods:
   • _name: Returns the constraint name (for serialization)
   • _value: Returns the constraint value (for serialization)
   • __call__: Performs the actual validation logic

Basic Custom Constraint Example

from specflow import Constraint, ValidationError, Schema, Field

class EmailDomain(Constraint[str]):
    """Validates that an email address belongs to a specific domain."""
    
    def __init__(self, domain: str) -> None:
        self._domain = domain
    
    @property
    def _name(self) -> str:
        return "emailDomain"
    
    @property
    def _value(self) -> str:
        return self._domain
    
    def __call__(self, to_validate: str) -> None:
        if not to_validate.endswith(f"@{self._domain}"):
            raise ValidationError(
                f"Email must be from domain '{self._domain}', got '{to_validate}'"
            )

# Usage
user_schema = Schema(
    title="User",
    properties=[
        Field(
            title="email",
            type_="string",
            constraints=[EmailDomain("company.com")]
        )
    ]
)

# Valid
user_schema({"email": "[e-mail]"})

# Invalid - raises ValidationError
try:
    user_schema({"email": "[e-mail]"})
except ValidationError as e:
    print(e)  # Validation failed at email: Email must be from domain 'company.com', got '[e-mail]'

Advanced Custom Constraint Examples

Password Strength Validator

import re
from specflow import Constraint, ValidationError, Schema, Field

class PasswordStrength(Constraint[str]):
    """Validates password contains uppercase, lowercase, digit, and special char."""
    
    def __init__(self, min_length: int = 8) -> None:
        self._min_length = min_length
    
    @property
    def _name(self) -> str:
        return "passwordStrength"
    
    @property
    def _value(self) -> int:
        return self._min_length
    
    def __call__(self, to_validate: str) -> None:
        if len(to_validate) < self._min_length:
            raise ValidationError(
                f"Password must be at least {self._min_length} characters"
            )
        
        if not re.search(r"[A-Z]", to_validate):
            raise ValidationError("Password must contain an uppercase letter")
        
        if not re.search(r"[a-z]", to_validate):
            raise ValidationError("Password must contain a lowercase letter")
        
        if not re.search(r"\d", to_validate):
            raise ValidationError("Password must contain a digit")
        
        if not re.search(r"[!@#$%^&*(),.?\":{}|<>]", to_validate):
            raise ValidationError("Password must contain a special character")

# Usage
registration_schema = Schema(
    title="Registration",
    properties=[
        Field(title="username", min_length=3),
        Field(
            title="password",
            type_="string",
            constraints=[PasswordStrength(min_length=12)]
        )
    ]
)

Age Range Validator (Date-based)

from datetime import datetime, date
from specflow import Constraint, ValidationError, Schema, Field

class AgeRange(Constraint[str]):
    """Validates age based on birthdate is within a range."""
    
    def __init__(self, min_age: int, max_age: int) -> None:
        self._min_age = min_age
        self._max_age = max_age
    
    @property
    def _name(self) -> str:
        return "ageRange"
    
    @property
    def _value(self) -> list[int]:
        return [self._min_age, self._max_age]
    
    def __call__(self, to_validate: str) -> None:
        try:
            birthdate = datetime.strptime(to_validate, "%Y-%m-%d").date()
        except ValueError:
            raise ValidationError(
                f"Invalid date format. Expected YYYY-MM-DD, got '{to_validate}'"
            )
        
        today = date.today()
        age = today.year - birthdate.year - (
            (today.month, today.day) < (birthdate.month, birthdate.day)
        )
        
        if age < self._min_age:
            raise ValidationError(
                f"Age must be at least {self._min_age}, person is {age}"
            )
        
        if age > self._max_age:
            raise ValidationError(
                f"Age must be at most {self._max_age}, person is {age}"
            )

# Usage
user_schema = Schema(
    title="User",
    properties=[
        Field(title="name", type_="string"),
        Field(
            title="birthdate",
            type_="string",
            constraints=[AgeRange(min_age=18, max_age=65)]
        )
    ]
)

Credit Card Validator (Luhn Algorithm)

from specflow import Constraint, ValidationError, Schema, Field

class CreditCardNumber(Constraint[str]):
    """Validates credit card number using the Luhn algorithm."""
    
    @property
    def _name(self) -> str:
        return "creditCard"
    
    @property
    def _value(self) -> bool:
        return True
    
    def __call__(self, to_validate: str) -> None:
        # Remove spaces and dashes
        card_number = to_validate.replace(" ", "").replace("-", "")
        
        # Check if it's all digits
        if not card_number.isdigit():
            raise ValidationError("Credit card must contain only digits")
        
        # Check length (most cards are 13-19 digits)
        if not 13 <= len(card_number) <= 19:
            raise ValidationError(
                f"Credit card must be 13-19 digits, got {len(card_number)}"
            )
        
        # Luhn algorithm
        def luhn_check(card: str) -> bool:
            digits = [int(d) for d in card]
            checksum = 0
            
            for i, digit in enumerate(reversed(digits)):
                if i % 2 == 1:
                    digit *= 2
                    if digit > 9:
                        digit -= 9
                checksum += digit
            
            return checksum % 10 == 0
        
        if not luhn_check(card_number):
            raise ValidationError("Invalid credit card number (failed Luhn check)")

# Usage
payment_schema = Schema(
    title="Payment",
    properties=[
        Field(
            title="card_number",
            type_="string",
            constraints=[CreditCardNumber()]
        ),
        Field(title="cvv", pattern=r"^\d{3,4}$")
    ]
)

IP Address Validator

import ipaddress
from specflow import Constraint, ValidationError, Schema, Field

class IPAddress(Constraint[str]):
    """Validates IPv4 or IPv6 addresses."""
    
    def __init__(self, version: int | None = None) -> None:
        """
        Args:
            version: IP version (4 or 6). If None, accepts both.
        """
        if version not in (None, 4, 6):
            raise ValueError("version must be 4, 6, or None")
        self._version = version
    
    @property
    def _name(self) -> str:
        return "ipAddress"
    
    @property
    def _value(self) -> int | None:
        return self._version
    
    def __call__(self, to_validate: str) -> None:
        try:
            ip = ipaddress.ip_address(to_validate)
            
            if self._version == 4 and ip.version != 4:
                raise ValidationError(f"Must be IPv4 address, got IPv{ip.version}")
            
            if self._version == 6 and ip.version != 6:
                raise ValidationError(f"Must be IPv6 address, got IPv{ip.version}")
                
        except ValueError:
            version_str = f"IPv{self._version}" if self._version else "IP"
            raise ValidationError(f"Invalid {version_str} address: '{to_validate}'")

# Usage
network_schema = Schema(
    title="NetworkConfig",
    properties=[
        Field(
            title="ipv4",
            type_="string",
            constraints=[IPAddress(version=4)]
        ),
        Field(
            title="ipv6",
            type_="string",
            constraints=[IPAddress(version=6)]
        ),
        Field(
            title="any_ip",
            type_="string",
            constraints=[IPAddress()]
        )
    ]
)

Numeric Range with Exclusions

from specflow import Constraint, ValidationError, Schema, Field

class RangeWithExclusions(Constraint[int]):
    """Validates integer is in range but not in excluded values."""
    
    def __init__(self, minimum: int, maximum: int, exclude: list[int]) -> None:
        self._minimum = minimum
        self._maximum = maximum
        self._exclude = set(exclude)
    
    @property
    def _name(self) -> str:
        return "rangeWithExclusions"
    
    @property
    def _value(self) -> dict[str, int | list[int]]:
        return {
            "minimum": self._minimum,
            "maximum": self._maximum,
            "exclude": list(self._exclude)
        }
    
    def __call__(self, to_validate: int) -> None:
        if to_validate < self._minimum or to_validate > self._maximum:
            raise ValidationError(
                f"Must be between {self._minimum} and {self._maximum}, got {to_validate}"
            )
        
        if to_validate in self._exclude:
            raise ValidationError(
                f"Value {to_validate} is not allowed (excluded values: {sorted(self._exclude)})"
            )

# Usage
config_schema = Schema(
    title="ServerConfig",
    properties=[
        Field(
            title="port",
            type_="integer",
            constraints=[RangeWithExclusions(1000, 9999, [3000, 5000, 8080])]
        )
    ]
)

Tips for Creating Custom Constraints

1. Keep them focused: Each constraint should validate one specific rule
2. Provide clear error messages: Users should understand what went wrong
3. Handle edge cases: Consider None values and invalid types
4. Make them reusable: Design constraints that can be used across different schemas
5. Use proper type hints: Specify Constraint[str], Constraint[int], etc.
6. Return meaningful values: The _value property should represent the constraint's configuration

Combining Multiple Constraints

You can apply multiple constraints to a single field:

from specflow import Schema, Field

user_schema = Schema(
    title="User",
    properties=[
        Field(
            title="email",
            type_="string",
            constraints=[
                EmailDomain("company.com"),
                # Add other constraints here
            ]
        ),
        Field(
            title="password",
            type_="string",
            constraints=[
                PasswordStrength(min_length=12),
                # Constraints are evaluated in order
            ]
        )
    ]
)

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Support

For issues, questions, or contributions, please visit the GitHub repository.