File size: 8,178 Bytes
62da328 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 |
# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ========= Copyright 2023-2024 @ CAMEL-AI.org. All Rights Reserved. =========
import os
from typing import Any, Dict, List, Optional, Union
from openai import OpenAI, Stream
from camel.configs import DEEPSEEK_API_PARAMS, DeepSeekConfig
from camel.logger import get_logger
from camel.messages import OpenAIMessage
from camel.models.base_model import BaseModelBackend
from camel.types import (
ChatCompletion,
ChatCompletionChunk,
ModelType,
)
from camel.utils import BaseTokenCounter, OpenAITokenCounter, api_keys_required
from retry import retry
import json
logger = get_logger(__name__)
class DeepSeekModel(BaseModelBackend):
r"""DeepSeek API in a unified BaseModelBackend interface.
Args:
model_type (Union[ModelType, str]): Model for which a backend is
created.
model_config_dict (Optional[Dict[str, Any]], optional): A dictionary
that will be fed into:obj:`openai.ChatCompletion.create()`. If
:obj:`None`, :obj:`DeepSeekConfig().as_dict()` will be used.
(default: :obj:`None`)
api_key (Optional[str], optional): The API key for authenticating with
the DeepSeek service. (default: :obj:`None`)
url (Optional[str], optional): The url to the DeepSeek service.
(default: :obj:`https://api.deepseek.com`)
token_counter (Optional[BaseTokenCounter], optional): Token counter to
use for the model. If not provided, :obj:`OpenAITokenCounter`
will be used. (default: :obj:`None`)
References:
https://api-docs.deepseek.com/
"""
def __init__(
self,
model_type: Union[ModelType, str],
model_config_dict: Optional[Dict[str, Any]] = None,
api_key: Optional[str] = None,
url: Optional[str] = None,
token_counter: Optional[BaseTokenCounter] = None,
) -> None:
if model_config_dict is None:
model_config_dict = DeepSeekConfig().as_dict()
api_key = api_key or os.environ.get("DEEPSEEK_API_KEY")
url = url or os.environ.get(
"DEEPSEEK_API_BASE_URL",
"https://api.deepseek.com",
)
super().__init__(
model_type, model_config_dict, api_key, url, token_counter
)
self._client = OpenAI(
timeout=180,
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(
model=ModelType.GPT_4O_MINI
)
return self._token_counter
@retry((ValueError, TypeError, json.decoder.JSONDecodeError), delay=10, logger=logger)
def run(
self,
messages: List[OpenAIMessage],
) -> Union[ChatCompletion, Stream[ChatCompletionChunk]]:
r"""Runs inference of DeepSeek 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.
"""
# deepseek reasoner has limitations
# reference: https://api-docs.deepseek.com/guides/reasoning_model#api-parameters
if self.model_type in [
ModelType.DEEPSEEK_REASONER,
]:
import re
logger.warning(
"You are using a DeepSeek Reasoner model, "
"which has certain limitations, reference: "
"`https://api-docs.deepseek.com/guides/reasoning_model#api-parameters`"
)
# Check and remove unsupported parameters and reset the fixed
# parameters
unsupported_keys = [
"temperature",
"top_p",
"presence_penalty",
"frequency_penalty",
"logprobs",
"top_logprobs",
"tools",
]
for key in unsupported_keys:
if key in self.model_config_dict:
del self.model_config_dict[key]
# Remove thinking content from messages before sending to API
# This ensures only the final response is sent, excluding
# intermediate thought processes
messages = [
{ # type: ignore[misc]
**msg,
'content': re.sub(
r'<think>.*?</think>',
'',
msg['content'], # type: ignore[arg-type]
flags=re.DOTALL,
).strip(),
}
for msg in messages
]
response = self._client.chat.completions.create(
messages=messages,
model=self.model_type,
**self.model_config_dict,
)
# Handle reasoning content with <think> tags at the beginning
if (
self.model_type
in [
ModelType.DEEPSEEK_REASONER,
]
and os.environ.get("GET_REASONING_CONTENT", "false").lower()
== "true"
):
reasoning_content = response.choices[0].message.reasoning_content
combined_content = (
f"<think>\n{reasoning_content}\n</think>\n"
if reasoning_content
else ""
) + response.choices[0].message.content
response = ChatCompletion.construct(
id=response.id,
choices=[
dict(
index=response.choices[0].index,
message={
"role": response.choices[0].message.role,
"content": combined_content,
"tool_calls": None,
},
finish_reason=response.choices[0].finish_reason
if response.choices[0].finish_reason
else None,
)
],
created=response.created,
model=response.model,
object="chat.completion",
usage=response.usage,
)
return response
def check_model_config(self):
r"""Check whether the model configuration contains any
unexpected arguments to DeepSeek API.
Raises:
ValueError: If the model configuration dictionary contains any
unexpected arguments to DeepSeek API.
"""
for param in self.model_config_dict:
if param not in DEEPSEEK_API_PARAMS:
raise ValueError(
f"Unexpected argument `{param}` is "
"input into DeepSeek 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) |