camel-ai 0.2.45__py3-none-any.whl → 0.2.46__py3-none-any.whl

camel/models/qwen_model.py

@@ -13,29 +13,19 @@
 # ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
 
 import os
-from typing import Any, Dict, List, Optional, Type, Union
-
-from openai import AsyncOpenAI, AsyncStream, OpenAI, Stream
-from pydantic import BaseModel
+from typing import Any, Dict, Optional, Union
 
 from camel.configs import QWEN_API_PARAMS, QwenConfig
-from camel.messages import OpenAIMessage
-from camel.models import BaseModelBackend
-from camel.models._utils import try_modify_message_with_format
-from camel.types import (
-    ChatCompletion,
-    ChatCompletionChunk,
-    ModelType,
-)
+from camel.models.openai_compatible_model import OpenAICompatibleModel
+from camel.types import ModelType
 from camel.utils import (
     BaseTokenCounter,
-    OpenAITokenCounter,
     api_keys_required,
 )
 
 
-class QwenModel(BaseModelBackend):
-    r"""Qwen API in a unified BaseModelBackend interface.
+class QwenModel(OpenAICompatibleModel):
+    r"""Qwen API in a unified OpenAICompatibleModel interface.
 
     Args:
         model_type (Union[ModelType, str]): Model for which a backend is
@@ -81,104 +71,13 @@ class QwenModel(BaseModelBackend):
         )
         timeout = timeout or float(os.environ.get("MODEL_TIMEOUT", 180))
         super().__init__(
-            model_type, model_config_dict, api_key, url, token_counter, timeout
-        )
-        self._client = OpenAI(
-            timeout=self._timeout,
-            max_retries=3,
-            api_key=self._api_key,
-            base_url=self._url,
-        )
-        self._async_client = AsyncOpenAI(
-            timeout=self._timeout,
-            max_retries=3,
-            api_key=self._api_key,
-            base_url=self._url,
-        )
-
-    async def _arun(
-        self,
-        messages: List[OpenAIMessage],
-        response_format: Optional[Type[BaseModel]] = None,
-        tools: Optional[List[Dict[str, Any]]] = None,
-    ) -> Union[ChatCompletion, AsyncStream[ChatCompletionChunk]]:
-        r"""Runs inference of Qwen chat completion.
-
-        Args:
-            messages (List[OpenAIMessage]): Message list with the chat history
-                in OpenAI API format.
-
-        Returns:
-            Union[ChatCompletion, AsyncStream[ChatCompletionChunk]]:
-                `ChatCompletion` in the non-stream mode, or
-                `AsyncStream[ChatCompletionChunk]` in the stream mode.
-        """
-        request_config = self._prepare_request(
-            messages, response_format, tools
-        )
-
-        response = await self._async_client.chat.completions.create(
-            messages=messages,
-            model=self.model_type,
-            **request_config,
-        )
-        return response
-
-    def _run(
-        self,
-        messages: List[OpenAIMessage],
-        response_format: Optional[Type[BaseModel]] = None,
-        tools: Optional[List[Dict[str, Any]]] = None,
-    ) -> Union[ChatCompletion, Stream[ChatCompletionChunk]]:
-        r"""Runs inference of Qwen chat completion.
-
-        Args:
-            messages (List[OpenAIMessage]): Message list with the chat history
-                in OpenAI API format.
-
-        Returns:
-            Union[ChatCompletion, Stream[ChatCompletionChunk]]:
-                `ChatCompletion` in the non-stream mode, or
-                `Stream[ChatCompletionChunk]` in the stream mode.
-        """
-        request_config = self._prepare_request(
-            messages, response_format, tools
-        )
-
-        response = self._client.chat.completions.create(
-            messages=messages,
-            model=self.model_type,
-            **request_config,
+            model_type=model_type,
+            model_config_dict=model_config_dict,
+            api_key=api_key,
+            url=url,
+            token_counter=token_counter,
+            timeout=timeout,
         )
-        return response
-
-    def _prepare_request(
-        self,
-        messages: List[OpenAIMessage],
-        response_format: Optional[Type[BaseModel]] = None,
-        tools: Optional[List[Dict[str, Any]]] = None,
-    ) -> Dict[str, Any]:
-        request_config = self.model_config_dict.copy()
-        if tools:
-            request_config["tools"] = tools
-        elif response_format:
-            try_modify_message_with_format(messages[-1], response_format)
-            request_config["response_format"] = {"type": "json_object"}
-
-        return request_config
-
-    @property
-    def token_counter(self) -> BaseTokenCounter:
-        r"""Initialize the token counter for the model backend.
-
-        Returns:
-            OpenAITokenCounter: The token counter following the model's
-                tokenization style.
-        """
-
-        if not self._token_counter:
-            self._token_counter = OpenAITokenCounter(ModelType.GPT_4O_MINI)
-        return self._token_counter
 
     def check_model_config(self):
         r"""Check whether the model configuration contains any
@@ -194,13 +93,3 @@ class QwenModel(BaseModelBackend):
                     f"Unexpected argument `{param}` is "
                     "input into Qwen model backend."
                 )
-
-    @property
-    def stream(self) -> bool:
-        r"""Returns whether the model is in stream mode, which sends partial
-        results each time.
-
-        Returns:
-            bool: Whether the model is in stream mode.
-        """
-        return self.model_config_dict.get('stream', False)

camel/models/reka_model.py

@@ -41,7 +41,7 @@ except (ImportError, AttributeError):
 
 
 class RekaModel(BaseModelBackend):
-    r"""Reka API in a unified BaseModelBackend interface.
+    r"""Reka API in a unified OpenAICompatibleModel interface.
 
     Args:
         model_type (Union[ModelType, str]): Model for which a backend is

camel/models/sglang_model.py

@@ -119,7 +119,9 @@ class SGLangModel(BaseModelBackend):
                 )
 
                 server_process = _execute_shell_command(cmd)
-                _wait_for_server("http://localhost:30000")
+                _wait_for_server(
+                    base_url="http://localhost:30000", timeout=self._timeout
+                )
                 self._url = "http://127.0.0.1:30000/v1"
                 self.server_process = server_process  # type: ignore[assignment]
                 # Start the inactivity monitor in a background thread
@@ -356,12 +358,12 @@ def _execute_shell_command(command: str) -> subprocess.Popen:
     return subprocess.Popen(parts, text=True, stderr=subprocess.STDOUT)
 
 
-def _wait_for_server(base_url: str, timeout: Optional[int] = 30) -> None:
+def _wait_for_server(base_url: str, timeout: Optional[float] = 30) -> None:
     r"""Wait for the server to be ready by polling the /v1/models endpoint.
 
     Args:
         base_url (str): The base URL of the server
-        timeout (Optional[int]): Maximum time to wait in seconds.
+        timeout (Optional[float]): Maximum time to wait in seconds.
             (default: :obj:`30`)
     """
     import requests

camel/models/siliconflow_model.py

@@ -14,12 +14,12 @@
 import os
 from typing import Any, Dict, List, Optional, Type, Union
 
-from openai import AsyncStream, OpenAI, Stream
+from openai import AsyncStream
 from pydantic import BaseModel
 
 from camel.configs import SILICONFLOW_API_PARAMS, SiliconFlowConfig
 from camel.messages import OpenAIMessage
-from camel.models.base_model import BaseModelBackend
+from camel.models.openai_compatible_model import OpenAICompatibleModel
 from camel.types import (
     ChatCompletion,
     ChatCompletionChunk,
@@ -27,13 +27,12 @@ from camel.types import (
 )
 from camel.utils import (
     BaseTokenCounter,
-    OpenAITokenCounter,
     api_keys_required,
 )
 
 
-class SiliconFlowModel(BaseModelBackend):
-    r"""SiliconFlow API in a unified BaseModelBackend interface.
+class SiliconFlowModel(OpenAICompatibleModel):
+    r"""SiliconFlow API in a unified OpenAICompatibleModel interface.
 
     Args:
         model_type (Union[ModelType, str]): Model for which a backend is
@@ -80,38 +79,13 @@ class SiliconFlowModel(BaseModelBackend):
         )
         timeout = timeout or float(os.environ.get("MODEL_TIMEOUT", 180))
         super().__init__(
-            model_type, model_config_dict, api_key, url, token_counter, timeout
+            model_type=model_type,
+            model_config_dict=model_config_dict,
+            api_key=api_key,
+            url=url,
+            token_counter=token_counter,
+            timeout=timeout,
         )
-        self._client = OpenAI(
-            timeout=self._timeout,
-            max_retries=3,
-            api_key=self._api_key,
-            base_url=self._url,
-        )
-
-    def _run(
-        self,
-        messages: List[OpenAIMessage],
-        response_format: Optional[Type[BaseModel]] = None,
-        tools: Optional[List[Dict[str, Any]]] = None,
-    ) -> Union[ChatCompletion, Stream[ChatCompletionChunk]]:
-        r"""Runs inference of SiliconFlow chat completion.
-
-        Args:
-            messages (List[OpenAIMessage]): Message list with the chat history
-                in OpenAI API format.
-
-        Returns:
-            Union[ChatCompletion, Stream[ChatCompletionChunk]]:
-                `ChatCompletion` in the non-stream mode, or
-                `Stream[ChatCompletionChunk]` in the stream mode.
-        """
-        response = self._client.chat.completions.create(
-            messages=messages,
-            model=self.model_type,
-            **self.model_config_dict,
-        )
-        return response
 
     async def _arun(
         self,
@@ -123,18 +97,6 @@ class SiliconFlowModel(BaseModelBackend):
             "SiliconFlow does not support async inference."
         )
 
-    @property
-    def token_counter(self) -> BaseTokenCounter:
-        r"""Initialize the token counter for the model backend.
-
-        Returns:
-            BaseTokenCounter: The token counter following the model's
-                tokenization style.
-        """
-        if not self._token_counter:
-            self._token_counter = OpenAITokenCounter(ModelType.GPT_4O_MINI)
-        return self._token_counter
-
     def check_model_config(self):
         r"""Check whether the model configuration contains any
         unexpected arguments to SiliconFlow API.
@@ -149,13 +111,3 @@ class SiliconFlowModel(BaseModelBackend):
                     f"Unexpected argument `{param}` is "
                     "input into SiliconFlow model backend."
                 )
-
-    @property
-    def stream(self) -> bool:
-        """Returns whether the model is in stream mode, which sends partial
-        results each time.
-
-        Returns:
-            bool: Whether the model is in stream mode.
-        """
-        return self.model_config_dict.get('stream', False)

camel/models/togetherai_model.py

@@ -13,27 +13,18 @@
 # ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
 
 import os
-from typing import Any, Dict, List, Optional, Type, Union
-
-from openai import AsyncOpenAI, AsyncStream, OpenAI, Stream
-from pydantic import BaseModel
+from typing import Any, Dict, Optional, Union
 
 from camel.configs import TOGETHERAI_API_PARAMS, TogetherAIConfig
-from camel.messages import OpenAIMessage
-from camel.models import BaseModelBackend
-from camel.types import (
-    ChatCompletion,
-    ChatCompletionChunk,
-    ModelType,
-)
+from camel.models.openai_compatible_model import OpenAICompatibleModel
+from camel.types import ModelType
 from camel.utils import (
     BaseTokenCounter,
-    OpenAITokenCounter,
     api_keys_required,
 )
 
 
-class TogetherAIModel(BaseModelBackend):
+class TogetherAIModel(OpenAICompatibleModel):
     r"""Constructor for Together AI backend with OpenAI compatibility.
 
     Args:
@@ -80,160 +71,12 @@ class TogetherAIModel(BaseModelBackend):
         )
         timeout = timeout or float(os.environ.get("MODEL_TIMEOUT", 180))
         super().__init__(
-            model_type, model_config_dict, api_key, url, token_counter, timeout
-        )
-
-        self._client = OpenAI(
-            timeout=self._timeout,
-            max_retries=3,
-            api_key=self._api_key,
-            base_url=self._url,
-        )
-        self._async_client = AsyncOpenAI(
-            timeout=self._timeout,
-            max_retries=3,
-            api_key=self._api_key,
-            base_url=self._url,
-        )
-
-    @property
-    def token_counter(self) -> BaseTokenCounter:
-        r"""Initialize the token counter for the model backend.
-
-        Returns:
-            BaseTokenCounter: The token counter following the model's
-                tokenization style.
-        """
-        if not self._token_counter:
-            self._token_counter = OpenAITokenCounter(ModelType.GPT_4O_MINI)
-        return self._token_counter
-
-    def _run(
-        self,
-        messages: List[OpenAIMessage],
-        response_format: Optional[Type[BaseModel]] = None,
-        tools: Optional[List[Dict[str, Any]]] = None,
-    ) -> Union[ChatCompletion, Stream[ChatCompletionChunk]]:
-        r"""Runs inference of OpenAI chat completion.
-
-        Args:
-            messages (List[OpenAIMessage]): Message list with the chat history
-                in OpenAI API format.
-            response_format (Optional[Type[BaseModel]]): The format of the
-                response.
-            tools (Optional[List[Dict[str, Any]]]): The schema of the tools to
-                use for the request.
-
-        Returns:
-            Union[ChatCompletion, Stream[ChatCompletionChunk]]:
-                `ChatCompletion` in the non-stream mode, or
-                `Stream[ChatCompletionChunk]` in the stream mode.
-        """
-        response_format = response_format or self.model_config_dict.get(
-            "response_format", None
-        )
-        if response_format:
-            return self._request_parse(messages, response_format, tools)
-        else:
-            return self._request_chat_completion(messages, tools)
-
-    async def _arun(
-        self,
-        messages: List[OpenAIMessage],
-        response_format: Optional[Type[BaseModel]] = None,
-        tools: Optional[List[Dict[str, Any]]] = None,
-    ) -> Union[ChatCompletion, AsyncStream[ChatCompletionChunk]]:
-        r"""Runs inference of OpenAI chat completion in async mode.
-
-        Args:
-            messages (List[OpenAIMessage]): Message list with the chat history
-                in OpenAI API format.
-            response_format (Optional[Type[BaseModel]]): The format of the
-                response.
-            tools (Optional[List[Dict[str, Any]]]): The schema of the tools to
-                use for the request.
-
-        Returns:
-            Union[ChatCompletion, AsyncStream[ChatCompletionChunk]]:
-                `ChatCompletion` in the non-stream mode, or
-                `AsyncStream[ChatCompletionChunk]` in the stream mode.
-        """
-        response_format = response_format or self.model_config_dict.get(
-            "response_format", None
-        )
-        if response_format:
-            return await self._arequest_parse(messages, response_format, tools)
-        else:
-            return await self._arequest_chat_completion(messages, tools)
-
-    def _request_chat_completion(
-        self,
-        messages: List[OpenAIMessage],
-        tools: Optional[List[Dict[str, Any]]] = None,
-    ) -> Union[ChatCompletion, Stream[ChatCompletionChunk]]:
-        request_config = self.model_config_dict.copy()
-
-        if tools:
-            request_config["tools"] = tools
-
-        return self._client.chat.completions.create(
-            messages=messages,
-            model=self.model_type,
-            **request_config,
-        )
-
-    async def _arequest_chat_completion(
-        self,
-        messages: List[OpenAIMessage],
-        tools: Optional[List[Dict[str, Any]]] = None,
-    ) -> Union[ChatCompletion, AsyncStream[ChatCompletionChunk]]:
-        request_config = self.model_config_dict.copy()
-
-        if tools:
-            request_config["tools"] = tools
-
-        return await self._async_client.chat.completions.create(
-            messages=messages,
-            model=self.model_type,
-            **request_config,
-        )
-
-    def _request_parse(
-        self,
-        messages: List[OpenAIMessage],
-        response_format: Type[BaseModel],
-        tools: Optional[List[Dict[str, Any]]] = None,
-    ) -> ChatCompletion:
-        request_config = self.model_config_dict.copy()
-
-        request_config["response_format"] = response_format
-
-        if tools is not None:
-            request_config["tools"] = tools
-
-        return self._client.beta.chat.completions.parse(
-            messages=messages,
-            model=self.model_type,
-            **request_config,
-        )
-
-    async def _arequest_parse(
-        self,
-        messages: List[OpenAIMessage],
-        response_format: Type[BaseModel],
-        tools: Optional[List[Dict[str, Any]]] = None,
-    ) -> ChatCompletion:
-        request_config = self.model_config_dict.copy()
-
-        request_config["response_format"] = response_format
-
-        if tools is not None:
-            request_config["tools"] = tools
-
-        return await self._async_client.beta.chat.completions.parse(
-            messages=messages,
-            model=self.model_type,
-            **request_config,
+            model_type=model_type,
+            model_config_dict=model_config_dict,
+            api_key=api_key,
+            url=url,
+            token_counter=token_counter,
+            timeout=timeout,
         )
 
     def check_model_config(self):
@@ -250,13 +93,3 @@ class TogetherAIModel(BaseModelBackend):
                     f"Unexpected argument `{param}` is "
                     "input into TogetherAI model backend."
                 )
-
-    @property
-    def stream(self) -> bool:
-        r"""Returns whether the model is in stream mode, which sends partial
-        results each time.
-
-        Returns:
-            bool: Whether the model is in stream mode.
-        """
-        return self.model_config_dict.get('stream', False)