gpt4free/g4f/providers/retry_provider.py
hlohaus 9563f8df3a feat: Update environment variables and modify model mappings
- Added `OPENROUTER_API_KEY` and `AZURE_API_KEYS` to `example.env`.
- Updated `AZURE_DEFAULT_MODEL` to "model-router" in `example.env`.
- Added `AZURE_ROUTES` with multiple model URLs in `example.env`.
- Changed the mapping for `"phi-4-multimodal"` in `DeepInfraChat.py` to `"microsoft/Phi-4-multimodal-instruct"`.
- Added `media` parameter to `GptOss.create_completion` method and raised a `ValueError` if `media` is provided.
- Updated `model_aliases` in `any_model_map.py` to include new mappings for various models.
- Removed several model aliases from `PollinationsAI` in `any_model_map.py`.
- Added new models and updated existing models in `model_map` across various files, including `any_model_map.py` and `__init__.py`.
- Refactored `AnyModelProviderMixin` to include `model_aliases` and updated the logic for handling model aliases.
2025-08-07 01:21:22 +02:00

438 lines
No EOL
17 KiB
Python

from __future__ import annotations
import random
from ..typing import Dict, Type, List, CreateResult, Messages, AsyncResult
from .types import BaseProvider, BaseRetryProvider, ProviderType
from .response import ProviderInfo, JsonConversation, is_content
from .. import debug
from ..tools.run_tools import AuthManager
from ..errors import RetryProviderError, RetryNoProviderError, MissingAuthError, NoValidHarFileError
class RotatedProvider(BaseRetryProvider):
"""
A provider that rotates through a list of providers, attempting one provider per
request and advancing to the next one upon failure. This distributes load and
retries across multiple providers in a round-robin fashion.
"""
def __init__(
self,
providers: List[Type[BaseProvider]],
shuffle: bool = True
) -> None:
"""
Initialize the RotatedProvider.
Args:
providers (List[Type[BaseProvider]]): A non-empty list of providers to rotate through.
shuffle (bool): If True, shuffles the provider list once at initialization
to randomize the rotation order.
"""
if not isinstance(providers, list) or len(providers) == 0:
raise ValueError('RotatedProvider requires a non-empty list of providers.')
self.providers = providers
if shuffle:
random.shuffle(self.providers)
self.current_index = 0
self.last_provider: Type[BaseProvider] = None
def _get_current_provider(self) -> Type[BaseProvider]:
"""Gets the provider at the current index."""
return self.providers[self.current_index]
def _rotate_provider(self) -> None:
"""Rotates to the next provider in the list."""
self.current_index = (self.current_index + 1) % len(self.providers)
#new_provider_name = self.providers[self.current_index].__name__
#debug.log(f"Rotated to next provider: {new_provider_name}")
def create_completion(
self,
model: str,
messages: Messages,
ignored: list[str] = [], # 'ignored' is less relevant now but kept for compatibility
api_key: str = None,
**kwargs,
) -> CreateResult:
"""
Create a completion using the current provider and rotating on failure.
It will try each provider in the list once per call, rotating after each
failed attempt, until one succeeds or all have failed.
"""
exceptions: Dict[str, Exception] = {}
# Loop over the number of providers, giving each one a chance
for _ in range(len(self.providers)):
provider = self._get_current_provider()
self.last_provider = provider
self._rotate_provider()
# Skip if provider is in the ignored list
if provider.get_parent() in ignored:
continue
alias = model or getattr(provider, "default_model", None)
if hasattr(provider, "model_aliases"):
alias = provider.model_aliases.get(model, model)
if isinstance(alias, list):
alias = random.choice(alias)
debug.log(f"Attempting provider: {provider.__name__} with model: {alias}")
yield ProviderInfo(**provider.get_dict(), model=alias, alias=model)
extra_body = kwargs.copy()
current_api_key = api_key.get(provider.get_parent()) if isinstance(api_key, dict) else api_key
if not current_api_key:
current_api_key = AuthManager.load_api_key(provider)
if current_api_key:
extra_body["api_key"] = current_api_key
try:
# Attempt to get a response from the current provider
response = provider.create_function(alias, messages, **extra_body)
started = False
for chunk in response:
if chunk:
yield chunk
if is_content(chunk):
started = True
if started:
# Success, so we return and do not rotate
return
except Exception as e:
exceptions[provider.__name__] = e
debug.error(f"{provider.__name__} failed: {e}")
# If the loop completes, all providers have failed
raise_exceptions(exceptions)
async def create_async_generator(
self,
model: str,
messages: Messages,
ignored: list[str] = [],
api_key: str = None,
conversation: JsonConversation = None,
**kwargs
) -> AsyncResult:
"""
Asynchronously create a completion, rotating through providers on failure.
"""
exceptions: Dict[str, Exception] = {}
for _ in range(len(self.providers)):
provider = self._get_current_provider()
self._rotate_provider()
self.last_provider = provider
if provider.get_parent() in ignored:
continue
alias = model or getattr(provider, "default_model", None)
if hasattr(provider, "model_aliases"):
alias = provider.model_aliases.get(model, model)
if isinstance(alias, list):
alias = random.choice(alias)
debug.log(f"Attempting provider: {provider.__name__} with model: {alias}")
yield ProviderInfo(**provider.get_dict(), model=alias)
extra_body = kwargs.copy()
current_api_key = api_key.get(provider.get_parent()) if isinstance(api_key, dict) else api_key
if not current_api_key:
current_api_key = AuthManager.load_api_key(provider)
if current_api_key:
extra_body["api_key"] = current_api_key
if conversation and hasattr(conversation, provider.__name__):
extra_body["conversation"] = JsonConversation(**getattr(conversation, provider.__name__))
try:
response = provider.async_create_function(alias, messages, **extra_body)
started = False
async for chunk in response:
if isinstance(chunk, JsonConversation):
if conversation is None: conversation = JsonConversation()
setattr(conversation, provider.__name__, chunk.get_dict())
yield conversation
elif chunk:
yield chunk
if is_content(chunk):
started = True
if started:
return # Success
except Exception as e:
exceptions[provider.__name__] = e
debug.error(f"{provider.__name__} failed: {e}")
raise_exceptions(exceptions)
# Maintain API compatibility
create_function = create_completion
async_create_function = create_async_generator
class IterListProvider(BaseRetryProvider):
def __init__(
self,
providers: List[Type[BaseProvider]] = [],
shuffle: bool = True
) -> None:
"""
Initialize the BaseRetryProvider.
Args:
providers (List[Type[BaseProvider]]): List of providers to use.
shuffle (bool): Whether to shuffle the providers list.
single_provider_retry (bool): Whether to retry a single provider if it fails.
max_retries (int): Maximum number of retries for a single provider.
"""
self.providers = providers
self.shuffle = shuffle
self.working = True
self.last_provider: Type[BaseProvider] = None
def create_completion(
self,
model: str,
messages: Messages,
ignored: list[str] = [],
api_key: str = None,
**kwargs,
) -> CreateResult:
"""
Create a completion using available providers.
Args:
model (str): The model to be used for completion.
messages (Messages): The messages to be used for generating completion.
Yields:
CreateResult: Tokens or results from the completion.
Raises:
Exception: Any exception encountered during the completion process.
"""
exceptions = {}
started: bool = False
for provider in self.get_providers(ignored):
self.last_provider = provider
alias = model
if not model:
alias = getattr(provider, "default_model", None)
if hasattr(provider, "model_aliases"):
alias = provider.model_aliases.get(model, model)
if isinstance(alias, list):
alias = random.choice(alias)
debug.log(f"Using provider: {provider.__name__} with model: {alias}")
yield ProviderInfo(**provider.get_dict(), model=alias)
extra_body = kwargs.copy()
if isinstance(api_key, dict):
api_key = api_key.get(provider.get_parent())
if not api_key:
api_key = AuthManager.load_api_key(provider)
if api_key:
extra_body["api_key"] = api_key
try:
response = provider.create_function(alias, messages, **extra_body)
for chunk in response:
if chunk:
yield chunk
if is_content(chunk):
started = True
if started:
return
except Exception as e:
exceptions[provider.__name__] = e
debug.error(f"{provider.__name__}:", e)
if started:
raise e
yield e
raise_exceptions(exceptions)
async def create_async_generator(
self,
model: str,
messages: Messages,
ignored: list[str] = [],
api_key: str = None,
conversation: JsonConversation = None,
**kwargs
) -> AsyncResult:
exceptions = {}
started: bool = False
for provider in self.get_providers(ignored):
self.last_provider = provider
alias = model
if not model:
alias = getattr(provider, "default_model", None)
if hasattr(provider, "model_aliases"):
alias = provider.model_aliases.get(model, model)
if isinstance(alias, list):
alias = random.choice(alias)
debug.log(f"Using {provider.__name__} provider with model {alias}")
yield ProviderInfo(**provider.get_dict(), model=alias)
extra_body = kwargs.copy()
current_provider_api_key = None
if isinstance(api_key, dict):
current_provider_api_key = api_key.get(provider.get_parent())
if not api_key:
current_provider_api_key = AuthManager.load_api_key(provider)
if current_provider_api_key:
extra_body["api_key"] = current_provider_api_key
if conversation is not None and hasattr(conversation, provider.__name__):
extra_body["conversation"] = JsonConversation(**getattr(conversation, provider.__name__))
try:
response = provider.async_create_function(model, messages, **extra_body)
if hasattr(response, "__aiter__"):
async for chunk in response:
if isinstance(chunk, JsonConversation):
if conversation is None:
conversation = JsonConversation()
setattr(conversation, provider.__name__, chunk.get_dict())
yield conversation
elif chunk:
yield chunk
if is_content(chunk):
started = True
elif response:
response = await response
if response:
yield response
started = True
if started:
return
except Exception as e:
exceptions[provider.__name__] = e
debug.error(f"{provider.__name__}:", e)
if started:
raise e
yield e
raise_exceptions(exceptions)
create_function = create_completion
async_create_function = create_async_generator
def get_providers(self, ignored: list[str]) -> list[ProviderType]:
providers = [p for p in self.providers if p.__name__ not in ignored]
if self.shuffle:
random.shuffle(providers)
return providers
class RetryProvider(IterListProvider):
def __init__(
self,
providers: List[Type[BaseProvider]],
shuffle: bool = True,
single_provider_retry: bool = False,
max_retries: int = 3,
) -> None:
"""
Initialize the BaseRetryProvider.
Args:
providers (List[Type[BaseProvider]]): List of providers to use.
shuffle (bool): Whether to shuffle the providers list.
single_provider_retry (bool): Whether to retry a single provider if it fails.
max_retries (int): Maximum number of retries for a single provider.
"""
super().__init__(providers, shuffle)
self.single_provider_retry = single_provider_retry
self.max_retries = max_retries
def create_completion(
self,
model: str,
messages: Messages,
**kwargs,
) -> CreateResult:
"""
Create a completion using available providers.
Args:
model (str): The model to be used for completion.
messages (Messages): The messages to be used for generating completion.
Yields:
CreateResult: Tokens or results from the completion.
Raises:
Exception: Any exception encountered during the completion process.
"""
if self.single_provider_retry:
exceptions = {}
started: bool = False
provider = self.providers[0]
self.last_provider = provider
for attempt in range(self.max_retries):
try:
if debug.logging:
print(f"Using {provider.__name__} provider (attempt {attempt + 1})")
response = provider.create_function(model, messages, **kwargs)
for chunk in response:
yield chunk
if is_content(chunk):
started = True
if started:
return
except Exception as e:
exceptions[provider.__name__] = e
if debug.logging:
print(f"{provider.__name__}: {e.__class__.__name__}: {e}")
if started:
raise e
raise_exceptions(exceptions)
else:
yield from super().create_completion(model, messages, **kwargs)
async def create_async_generator(
self,
model: str,
messages: Messages,
**kwargs
) -> AsyncResult:
exceptions = {}
started = False
if self.single_provider_retry:
provider = self.providers[0]
self.last_provider = provider
for attempt in range(self.max_retries):
try:
debug.log(f"Using {provider.__name__} provider (attempt {attempt + 1})")
response = provider.async_create_function(model, messages, **kwargs)
if hasattr(response, "__aiter__"):
async for chunk in response:
yield chunk
if is_content(chunk):
started = True
else:
response = await response
if response:
yield response
started = True
if started:
return
except Exception as e:
exceptions[provider.__name__] = e
if debug.logging:
print(f"{provider.__name__}: {e.__class__.__name__}: {e}")
raise_exceptions(exceptions)
else:
async for chunk in super().create_async_generator(model, messages, **kwargs):
yield chunk
def raise_exceptions(exceptions: dict) -> None:
"""
Raise a combined exception if any occurred during retries.
Raises:
RetryProviderError: If any provider encountered an exception.
RetryNoProviderError: If no provider is found.
"""
if exceptions:
for provider_name, e in exceptions.items():
if isinstance(e, (MissingAuthError, NoValidHarFileError)):
raise e
if len(exceptions) == 1:
raise list(exceptions.values())[0]
raise RetryProviderError("RetryProvider failed:\n" + "\n".join([
f"{p}: {type(exception).__name__}: {exception}" for p, exception in exceptions.items()
])) from list(exceptions.values())[0]
raise RetryNoProviderError("No content response from any provider. ")