from langchain_openai import ChatOpenAI
from langchain_core.messages import AIMessage, AIMessageChunk
from lwe.core.provider import Provider, PresetValue
from lwe.core import constants
[docs]
class CustomChatOpenAI(ChatOpenAI):
@property
def _llm_type(self):
"""Return type of llm."""
return "chat_openai"
[docs]
class ProviderChatOpenai(Provider):
"""
Access to OpenAI chat models via the OpenAI API
"""
[docs]
def default_config(self):
return {
"validate_models": True,
}
[docs]
def setup(self):
self.log.info(f"Setting up Provider Chat OpenAI plugin, running with backend: {self.backend.name}")
super().setup()
self.validate_models = self.config.get("plugins.provider_chat_openai.validate_models")
@property
def capabilities(self):
return {
"chat": True,
"validate_models": self.validate_models,
}
@property
def default_model(self):
return constants.API_BACKEND_DEFAULT_MODEL
@property
def static_models(self):
return {
"gpt-3.5-turbo": {
"max_tokens": 16384,
},
"gpt-3.5-turbo-16k": {
"max_tokens": 16384,
},
"gpt-3.5-turbo-0613": {
"max_tokens": 4096,
},
"gpt-3.5-turbo-16k-0613": {
"max_tokens": 16384,
},
"gpt-3.5-turbo-1106": {
"max_tokens": 16384,
},
"gpt-3.5-turbo-0125": {
"max_tokens": 16384,
},
"gpt-4": {
"max_tokens": 8192,
},
"gpt-4-32k": {
"max_tokens": 32768,
},
"gpt-4-0613": {
"max_tokens": 8192,
},
"gpt-4-32k-0613": {
"max_tokens": 32768,
},
"gpt-4-turbo": {
"max_tokens": 131072,
},
"gpt-4-turbo-2024-04-09": {
"max_tokens": 131072,
},
"gpt-4-turbo-preview": {
"max_tokens": 131072,
},
"gpt-4-1106-preview": {
"max_tokens": 131072,
},
"gpt-4-0125-preview": {
"max_tokens": 131072,
},
"gpt-4o": {
"max_tokens": 131072,
},
"chatgpt-4o-latest": {
"max_tokens": 131072,
},
"gpt-4o-2024-05-13": {
"max_tokens": 131072,
},
"gpt-4o-2024-08-06": {
"max_tokens": 131072,
},
"gpt-4o-2024-11-20": {
"max_tokens": 131072,
},
"gpt-4o-mini": {
"max_tokens": 131072,
},
"gpt-4o-mini-2024-07-18": {
"max_tokens": 131072,
},
"gpt-4.1": {
"max_tokens": 1047576,
},
"gpt-4.1-2025-04-14": {
"max_tokens": 1047576,
},
"gpt-4.1-mini": {
"max_tokens": 1047576,
},
"gpt-4.1-mini-2025-04-14": {
"max_tokens": 1047576,
},
"gpt-4.1-nano": {
"max_tokens": 1047576,
},
"gpt-4.1-nano-2025-04-14": {
"max_tokens": 1047576,
},
"gpt-4.5-preview": {
"max_tokens": 131072,
},
"gpt-4.5-preview-2025-02-27": {
"max_tokens": 131072,
},
"o1-preview": {
"max_tokens": 131072,
},
"o1-preview-2024-09-12": {
"max_tokens": 131072,
},
"o1-mini": {
"max_tokens": 131072,
},
"o1-mini-2024-09-12": {
"max_tokens": 131072,
},
"o1": {
"max_tokens": 204800,
},
"o1-2024-12-17": {
"max_tokens": 204800,
},
"o1-pro": {
"max_tokens": 204800,
},
"o1-pro-2025-03-19": {
"max_tokens": 204800,
},
"o3-mini": {
"max_tokens": 204800,
},
"o3-mini-2025-01-31": {
"max_tokens": 204800,
},
"o3": {
"max_tokens": 204800,
},
"o3-2025-04-16": {
"max_tokens": 204800,
},
"o3-pro": {
"max_tokens": 204800,
},
"o3-pro-2025-06-10": {
"max_tokens": 204800,
},
"o4-mini": {
"max_tokens": 204800,
},
"o4-mini-2025-04-16": {
"max_tokens": 204800,
},
"gpt-5": {
"max_tokens": 409600,
},
"gpt-5-2025-08-07": {
"max_tokens": 409600,
},
"gpt-5-mini": {
"max_tokens": 409600,
},
"gpt-5-mini-2025-08-07": {
"max_tokens": 409600,
},
"gpt-5-nano": {
"max_tokens": 409600,
},
"gpt-5-nano-2025-08-07": {
"max_tokens": 409600,
},
"gpt-5-pro": {
"max_tokens": 409600,
},
"gpt-5-pro-2025-10-06": {
"max_tokens": 409600,
},
"gpt-5.1": {
"max_tokens": 409600,
},
"gpt-5.1-2025-11-13": {
"max_tokens": 409600,
},
"gpt-5.2-chat-latest": {
"max_tokens": 131072,
},
"gpt-5.2-pro": {
"max_tokens": 409600,
},
"gpt-5.2-pro-2025-12-11": {
"max_tokens": 409600,
},
"gpt-5.2": {
"max_tokens": 409600,
},
"gpt-5.2-2025-12-11": {
"max_tokens": 409600,
},
"gpt-5.3-chat-latest": {
"max_tokens": 131072,
},
"gpt-5.4-2026-03-05": {
"max_tokens": 1047576,
},
"gpt-5.4": {
"max_tokens": 1047576,
},
"gpt-5.4-pro": {
"max_tokens": 1047576,
},
"gpt-5.5-2026-04-23": {
"max_tokens": 1047576,
},
"gpt-5.5": {
"max_tokens": 1047576,
},
}
[docs]
def prepare_messages_method(self):
return self.prepare_messages_for_llm_chat
[docs]
def llm_factory(self):
return CustomChatOpenAI
[docs]
def llm_pre_init(self, customizations):
model_name = customizations.get(self.model_property_name, "")
if self.is_reasoning_model(model_name):
customizations["temperature"] = 1
if not self.is_legacy_reasoning_model(model_name):
customizations["use_responses_api"] = True
return customizations
[docs]
def llm_pre_call(self, llm, messages):
model_name = getattr(llm, self.model_property_name)
if self.is_legacy_reasoning_model(model_name):
messages = [{**m, "role": "user"} if m["role"] == "system" else m for m in messages]
return messages
[docs]
def is_reasoning_model(self, model_name):
if model_name.startswith("o1") or model_name.startswith("o3") or model_name.startswith("o4") or model_name.startswith("gpt-5"):
return True
return False
[docs]
def is_legacy_reasoning_model(self, model_name):
if model_name.startswith("o1-mini") or model_name.startswith("o1-preview"):
return True
return False
[docs]
def format_responses_content(self, content_list):
return "".join([content['text'] for content in content_list if 'text' in content])
[docs]
def handle_non_streaming_response(self, response: AIMessage):
if isinstance(response.content, list):
response.content = self.format_responses_content(response.content)
return response
[docs]
def handle_streaming_chunk(self, chunk: AIMessageChunk | str, _previous_chunks: list[AIMessageChunk | str]) -> str:
if isinstance(chunk, str):
return chunk
elif isinstance(chunk.content, str):
return chunk.content
return self.format_responses_content(chunk.content)
[docs]
def customization_config(self):
return {
"verbose": PresetValue(bool),
"model_name": PresetValue(str, options=self.available_models),
"temperature": PresetValue(float, min_value=0.0, max_value=2.0),
"reasoning_effort": PresetValue(str, options=["low", "medium", "high", "xhigh"], include_none=True),
"verbosity": PresetValue(str, options=["low", "medium", "high"], include_none=True),
"openai_api_base": PresetValue(str, include_none=True),
"openai_api_key": PresetValue(str, include_none=True, private=True),
"openai_organization": PresetValue(str, include_none=True, private=True),
"request_timeout": PresetValue(int),
"max_retries": PresetValue(int, 1, 10),
"max_tokens": PresetValue(int, include_none=True),
"top_p": PresetValue(float, min_value=0.0, max_value=1.0),
"presence_penalty": PresetValue(float, min_value=-2.0, max_value=2.0),
"frequency_penalty": PresetValue(float, min_value=-2.0, max_value=2.0),
"seed": PresetValue(int, include_none=True),
"logprobs": PresetValue(bool, include_none=True),
"top_logprobs": PresetValue(int, min_value=0, max_value=20, include_none=True),
"logit_bias": dict,
"stop": PresetValue(str, include_none=True),
"tools": None,
"tool_choice": None,
"model_kwargs": {
"response_format": dict,
"user": PresetValue(str),
},
}
