refactor: improve config loader validation with Pydantic schemas

- Replace manual field type validation with Pydantic model schemas
- Add pydantic>=2.0 as core dependency
- Fix sync wrapper in decorator to properly handle rate limiting
- Update pyright settings for stricter type checking
- Fix repository URL in pyproject.toml
- Remove unused main.py
- Update test assertions for new validation error format
This commit is contained in:
2026-03-07 15:43:25 +00:00
parent 4f19c0b19e
commit 492410614f
10 changed files with 200 additions and 148 deletions

View File

@@ -20,7 +20,7 @@ from fastapi_traffic.exceptions import (
RateLimitExceeded,
)
__version__ = "0.1.0"
__version__ = "0.2.0"
__all__ = [
"Algorithm",
"Backend",

View File

@@ -89,6 +89,8 @@ class TokenBucketAlgorithm(BaseAlgorithm):
remaining=int(tokens),
reset_at=now + self.window_size,
window_size=self.window_size,
retry_after = (1 - tokens) / self.refill_rate
)
tokens = float(state.get("tokens", self.burst_size))

View File

@@ -4,7 +4,7 @@ from __future__ import annotations
from collections.abc import Callable
from dataclasses import dataclass, field
from typing import TYPE_CHECKING, Any
from typing import TYPE_CHECKING, Any, TypeAlias
from fastapi_traffic.core.algorithms import Algorithm
@@ -14,7 +14,7 @@ if TYPE_CHECKING:
from fastapi_traffic.backends.base import Backend
KeyExtractor = Callable[["Request"], str]
KeyExtractor: TypeAlias = Callable[["Request"], str]
def default_key_extractor(request: Request) -> str:
@@ -55,10 +55,10 @@ class RateLimitConfig:
if self.limit <= 0:
msg = "limit must be positive"
raise ValueError(msg)
if self.window_size <= 0:
elif self.window_size <= 0:
msg = "window_size must be positive"
raise ValueError(msg)
if self.cost <= 0:
elif self.cost <= 0:
msg = "cost must be positive"
raise ValueError(msg)

View File

@@ -7,6 +7,8 @@ import os
from pathlib import Path
from typing import TYPE_CHECKING, Any, TypeVar
from pydantic import BaseModel, ConfigDict, ValidationError, field_validator
from fastapi_traffic.core.algorithms import Algorithm
from fastapi_traffic.core.config import GlobalConfig, RateLimitConfig
from fastapi_traffic.exceptions import ConfigurationError
@@ -19,35 +21,6 @@ T = TypeVar("T", RateLimitConfig, GlobalConfig)
# Environment variable prefix for config values
ENV_PREFIX = "FASTAPI_TRAFFIC_"
# Mapping of config field names to their types for validation
_RATE_LIMIT_FIELD_TYPES: dict[str, type[Any]] = {
"limit": int,
"window_size": float,
"algorithm": Algorithm,
"key_prefix": str,
"burst_size": int,
"include_headers": bool,
"error_message": str,
"status_code": int,
"skip_on_error": bool,
"cost": int,
}
_GLOBAL_FIELD_TYPES: dict[str, type[Any]] = {
"enabled": bool,
"default_limit": int,
"default_window_size": float,
"default_algorithm": Algorithm,
"key_prefix": str,
"include_headers": bool,
"error_message": str,
"status_code": int,
"skip_on_error": bool,
"exempt_ips": set,
"exempt_paths": set,
"headers_prefix": str,
}
# Fields that cannot be loaded from config files (callables, complex objects)
_NON_LOADABLE_FIELDS: frozenset[str] = frozenset(
{
@@ -59,6 +32,98 @@ _NON_LOADABLE_FIELDS: frozenset[str] = frozenset(
)
class _RateLimitSchema(BaseModel):
"""Pydantic schema for validating rate limit configuration input."""
model_config = ConfigDict(extra="forbid")
limit: int
window_size: float = 60.0
algorithm: Algorithm = Algorithm.SLIDING_WINDOW_COUNTER
key_prefix: str = "ratelimit"
burst_size: int | None = None
include_headers: bool = True
error_message: str = "Rate limit exceeded"
status_code: int = 429
skip_on_error: bool = False
cost: int = 1
@field_validator("algorithm", mode="before")
@classmethod
def _normalize_algorithm(cls, v: Any) -> Any:
if isinstance(v, str):
return v.lower()
return v
class _GlobalSchema(BaseModel):
"""Pydantic schema for validating global configuration input."""
model_config = ConfigDict(extra="forbid")
enabled: bool = True
default_limit: int = 100
default_window_size: float = 60.0
default_algorithm: Algorithm = Algorithm.SLIDING_WINDOW_COUNTER
key_prefix: str = "fastapi_traffic"
include_headers: bool = True
error_message: str = "Rate limit exceeded. Please try again later."
status_code: int = 429
skip_on_error: bool = False
exempt_ips: set[str] = set()
exempt_paths: set[str] = set()
headers_prefix: str = "X-RateLimit"
@field_validator("default_algorithm", mode="before")
@classmethod
def _normalize_algorithm(cls, v: Any) -> Any:
if isinstance(v, str):
return v.lower()
return v
# Known field names per schema (used for env-var extraction)
_RATE_LIMIT_FIELDS: frozenset[str] = frozenset(_RateLimitSchema.model_fields.keys())
_GLOBAL_FIELDS: frozenset[str] = frozenset(_GlobalSchema.model_fields.keys())
def _check_non_loadable(data: Mapping[str, Any]) -> None:
"""Raise ConfigurationError if data contains non-loadable fields."""
for key in data:
if key in _NON_LOADABLE_FIELDS:
msg = f"Field '{key}' cannot be loaded from configuration files"
raise ConfigurationError(msg)
def _format_validation_error(exc: ValidationError) -> str:
"""Convert a Pydantic ValidationError to a user-friendly message."""
errors = exc.errors()
if not errors:
return str(exc)
err = errors[0]
loc = ".".join(str(p) for p in err["loc"]) if err["loc"] else "unknown"
err_type = err["type"]
msg = err["msg"]
ctx = err.get("ctx", {})
if err_type == "extra_forbidden":
return f"Unknown configuration field: '{loc}'"
if err_type in ("int_parsing", "float_parsing"):
input_val = ctx.get("error", err.get("input", ""))
return f"Cannot parse value '{input_val}' as {loc}: {msg}"
if err_type == "bool_parsing":
return f"Cannot parse value as bool for '{loc}': {msg}"
if "enum" in err_type or err_type == "value_error":
input_val = err.get("input", "")
return f"Cannot parse value '{input_val}' as {loc}: {msg}"
return f"Invalid value for '{loc}': {msg}"
class ConfigLoader:
"""Loader for rate limiting configuration from various sources.
@@ -83,88 +148,6 @@ class ConfigLoader:
"""
self._env_prefix = env_prefix
def _parse_value(self, value: str, target_type: type[Any]) -> Any:
"""Parse a string value to the target type.
Args:
value: The string value to parse.
target_type: The target type to convert to.
Returns:
The parsed value.
Raises:
ConfigurationError: If the value cannot be parsed.
"""
try:
if target_type is bool:
return value.lower() in ("true", "1", "yes", "on")
if target_type is int:
return int(value)
if target_type is float:
return float(value)
if target_type is str:
return value
if target_type is Algorithm:
return Algorithm(value.lower())
if target_type is set:
# Parse comma-separated values
if not value.strip():
return set()
return {item.strip() for item in value.split(",") if item.strip()}
except (ValueError, KeyError) as e:
msg = f"Cannot parse value '{value}' as {target_type.__name__}: {e}"
raise ConfigurationError(msg) from e
msg = f"Unsupported type: {target_type}"
raise ConfigurationError(msg)
def _validate_and_convert(
self,
data: Mapping[str, Any],
field_types: dict[str, type[Any]],
) -> dict[str, Any]:
"""Validate and convert configuration data.
Args:
data: Raw configuration data.
field_types: Mapping of field names to their expected types.
Returns:
Validated and converted configuration dictionary.
Raises:
ConfigurationError: If validation fails.
"""
result: dict[str, Any] = {}
for key, value in data.items():
if key in _NON_LOADABLE_FIELDS:
msg = f"Field '{key}' cannot be loaded from configuration files"
raise ConfigurationError(msg)
if key not in field_types:
msg = f"Unknown configuration field: '{key}'"
raise ConfigurationError(msg)
target_type = field_types[key]
if isinstance(value, str):
result[key] = self._parse_value(value, target_type)
elif target_type is set and isinstance(value, list):
result[key] = set(value)
elif target_type is Algorithm and isinstance(value, str):
result[key] = Algorithm(value.lower())
elif isinstance(value, target_type):
result[key] = value
elif target_type is float and isinstance(value, int):
result[key] = float(value)
else:
msg = f"Invalid type for '{key}': expected {target_type.__name__}, got {type(value).__name__}"
raise ConfigurationError(msg)
return result
def _load_dotenv_file(self, file_path: Path) -> dict[str, str]:
"""Load environment variables from a .env file.
@@ -248,14 +231,14 @@ class ConfigLoader:
def _extract_env_config(
self,
prefix: str,
field_types: dict[str, type[Any]],
known_fields: frozenset[str],
env_source: Mapping[str, str] | None = None,
) -> dict[str, str]:
"""Extract configuration from environment variables.
Args:
prefix: The prefix to look for (e.g., "RATE_LIMIT_" or "GLOBAL_").
field_types: Mapping of field names to their expected types.
known_fields: Set of known field names.
env_source: Optional source of environment variables. Defaults to os.environ.
Returns:
@@ -268,11 +251,29 @@ class ConfigLoader:
for key, value in source.items():
if key.startswith(full_prefix):
field_name = key[len(full_prefix) :].lower()
if field_name in field_types:
if field_name in known_fields:
result[field_name] = value
return result
def _parse_set_from_string(self, data: dict[str, Any]) -> dict[str, Any]:
"""Pre-process comma-separated string values into lists for set fields.
This handles the env-var case where sets are represented as
comma-separated strings (e.g., "127.0.0.1, 10.0.0.1").
"""
result = dict(data)
for key in ("exempt_ips", "exempt_paths"):
if key in result and isinstance(result[key], str):
value = result[key].strip()
if not value:
result[key] = []
else:
result[key] = [
item.strip() for item in value.split(",") if item.strip()
]
return result
def load_rate_limit_config_from_env(
self,
env_source: Mapping[str, str] | None = None,
@@ -294,13 +295,21 @@ class ConfigLoader:
ConfigurationError: If configuration is invalid.
"""
raw_config = self._extract_env_config(
"RATE_LIMIT_", _RATE_LIMIT_FIELD_TYPES, env_source
"RATE_LIMIT_", _RATE_LIMIT_FIELDS, env_source
)
config_dict = self._validate_and_convert(raw_config, _RATE_LIMIT_FIELD_TYPES)
_check_non_loadable(raw_config)
try:
schema = _RateLimitSchema(**raw_config) # type: ignore[arg-type] # Pydantic coerces str→typed values at runtime
except ValidationError as e:
raise ConfigurationError(_format_validation_error(e)) from e
config_dict = schema.model_dump(exclude_defaults=True)
# Apply overrides
for key, value in overrides.items():
if key in _NON_LOADABLE_FIELDS or key in _RATE_LIMIT_FIELD_TYPES:
if key in _NON_LOADABLE_FIELDS or key in _RATE_LIMIT_FIELDS:
config_dict[key] = value
# Ensure required field 'limit' is present
@@ -353,11 +362,19 @@ class ConfigLoader:
if not isinstance(raw_config, dict):
msg = "JSON root must be an object"
raise ConfigurationError(msg)
config_dict = self._validate_and_convert(raw_config, _RATE_LIMIT_FIELD_TYPES)
_check_non_loadable(raw_config)
try:
schema = _RateLimitSchema(**raw_config) # type: ignore[arg-type] # Pydantic coerces str→typed values at runtime
except ValidationError as e:
raise ConfigurationError(_format_validation_error(e)) from e
config_dict = schema.model_dump(exclude_defaults=True)
# Apply overrides
for key, value in overrides.items():
if key in _NON_LOADABLE_FIELDS or key in _RATE_LIMIT_FIELD_TYPES:
if key in _NON_LOADABLE_FIELDS or key in _RATE_LIMIT_FIELDS:
config_dict[key] = value
# Ensure required field 'limit' is present
@@ -388,13 +405,24 @@ class ConfigLoader:
ConfigurationError: If configuration is invalid.
"""
raw_config = self._extract_env_config(
"GLOBAL_", _GLOBAL_FIELD_TYPES, env_source
"GLOBAL_", _GLOBAL_FIELDS, env_source
)
config_dict = self._validate_and_convert(raw_config, _GLOBAL_FIELD_TYPES)
_check_non_loadable(raw_config)
# Pre-process comma-separated strings into lists for set fields
processed = self._parse_set_from_string(raw_config)
try:
schema = _GlobalSchema(**processed) # type: ignore[arg-type] # Pydantic coerces str→typed values at runtime
except ValidationError as e:
raise ConfigurationError(_format_validation_error(e)) from e
config_dict = schema.model_dump(exclude_defaults=True)
# Apply overrides
for key, value in overrides.items():
if key in _NON_LOADABLE_FIELDS or key in _GLOBAL_FIELD_TYPES:
if key in _NON_LOADABLE_FIELDS or key in _GLOBAL_FIELDS:
config_dict[key] = value
return GlobalConfig(**config_dict)
@@ -439,11 +467,23 @@ class ConfigLoader:
"""
path = Path(file_path)
raw_config = self._load_json_file(path)
config_dict = self._validate_and_convert(raw_config, _GLOBAL_FIELD_TYPES)
if not isinstance(raw_config, dict):
msg = "JSON root must be an object"
raise ConfigurationError(msg)
_check_non_loadable(raw_config)
try:
schema = _GlobalSchema(**raw_config)
except ValidationError as e:
raise ConfigurationError(_format_validation_error(e)) from e
config_dict = schema.model_dump(exclude_defaults=True)
# Apply overrides
for key, value in overrides.items():
if key in _NON_LOADABLE_FIELDS or key in _GLOBAL_FIELD_TYPES:
if key in _NON_LOADABLE_FIELDS or key in _GLOBAL_FIELDS:
config_dict[key] = value
return GlobalConfig(**config_dict)

View File

@@ -51,7 +51,6 @@ def rate_limit(
/,
) -> Callable[[F], F]: ...
def rate_limit(
limit: int,
window_size: float = 60.0,
@@ -139,9 +138,23 @@ def rate_limit(
def sync_wrapper(*args: Any, **kwargs: Any) -> Any:
import asyncio
return asyncio.get_event_loop().run_until_complete(
async_wrapper(*args, **kwargs)
)
async def _sync_rate_limit() -> Any:
request = _extract_request(args, kwargs)
if request is None:
return func(*args, **kwargs)
limiter = get_limiter()
result = await limiter.hit(request, config)
response = func(*args, **kwargs)
if config.include_headers and hasattr(response, "headers"):
for key, value in result.info.to_headers().items():
response.headers[key] = value
return response
return asyncio.get_event_loop().run_until_complete(_sync_rate_limit())
if _is_coroutine_function(func):
return async_wrapper # type: ignore[return-value]