File size: 17,369 Bytes
e3278e4 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 |
import json
from typing import List, Literal, Optional, Tuple, Union
from openai.types.chat.chat_completion_chunk import Choice as OpenAIStreamingChoice
from litellm.types.llms.anthropic import (
AllAnthropicToolsValues,
AnthopicMessagesAssistantMessageParam,
AnthropicFinishReason,
AnthropicMessagesRequest,
AnthropicMessagesToolChoice,
AnthropicMessagesUserMessageParam,
AnthropicResponse,
AnthropicResponseContentBlockText,
AnthropicResponseContentBlockToolUse,
AnthropicResponseUsageBlock,
ContentBlockDelta,
ContentJsonBlockDelta,
ContentTextBlockDelta,
MessageBlockDelta,
MessageDelta,
UsageDelta,
)
from litellm.types.llms.openai import (
AllMessageValues,
ChatCompletionAssistantMessage,
ChatCompletionAssistantToolCall,
ChatCompletionImageObject,
ChatCompletionImageUrlObject,
ChatCompletionRequest,
ChatCompletionSystemMessage,
ChatCompletionTextObject,
ChatCompletionToolCallFunctionChunk,
ChatCompletionToolChoiceFunctionParam,
ChatCompletionToolChoiceObjectParam,
ChatCompletionToolChoiceValues,
ChatCompletionToolMessage,
ChatCompletionToolParam,
ChatCompletionToolParamFunctionChunk,
ChatCompletionUserMessage,
)
from litellm.types.utils import Choices, ModelResponse, Usage
class AnthropicExperimentalPassThroughConfig:
def __init__(self):
pass
### FOR [BETA] `/v1/messages` endpoint support
def translatable_anthropic_params(self) -> List:
"""
Which anthropic params, we need to translate to the openai format.
"""
return ["messages", "metadata", "system", "tool_choice", "tools"]
def translate_anthropic_messages_to_openai( # noqa: PLR0915
self,
messages: List[
Union[
AnthropicMessagesUserMessageParam,
AnthopicMessagesAssistantMessageParam,
]
],
) -> List:
new_messages: List[AllMessageValues] = []
for m in messages:
user_message: Optional[ChatCompletionUserMessage] = None
tool_message_list: List[ChatCompletionToolMessage] = []
new_user_content_list: List[
Union[ChatCompletionTextObject, ChatCompletionImageObject]
] = []
## USER MESSAGE ##
if m["role"] == "user":
## translate user message
message_content = m.get("content")
if message_content and isinstance(message_content, str):
user_message = ChatCompletionUserMessage(
role="user", content=message_content
)
elif message_content and isinstance(message_content, list):
for content in message_content:
if content["type"] == "text":
text_obj = ChatCompletionTextObject(
type="text", text=content["text"]
)
new_user_content_list.append(text_obj)
elif content["type"] == "image":
image_url = ChatCompletionImageUrlObject(
url=f"data:{content['type']};base64,{content['source']}"
)
image_obj = ChatCompletionImageObject(
type="image_url", image_url=image_url
)
new_user_content_list.append(image_obj)
elif content["type"] == "tool_result":
if "content" not in content:
tool_result = ChatCompletionToolMessage(
role="tool",
tool_call_id=content["tool_use_id"],
content="",
)
tool_message_list.append(tool_result)
elif isinstance(content["content"], str):
tool_result = ChatCompletionToolMessage(
role="tool",
tool_call_id=content["tool_use_id"],
content=content["content"],
)
tool_message_list.append(tool_result)
elif isinstance(content["content"], list):
for c in content["content"]:
if c["type"] == "text":
tool_result = ChatCompletionToolMessage(
role="tool",
tool_call_id=content["tool_use_id"],
content=c["text"],
)
tool_message_list.append(tool_result)
elif c["type"] == "image":
image_str = (
f"data:{c['type']};base64,{c['source']}"
)
tool_result = ChatCompletionToolMessage(
role="tool",
tool_call_id=content["tool_use_id"],
content=image_str,
)
tool_message_list.append(tool_result)
if user_message is not None:
new_messages.append(user_message)
if len(new_user_content_list) > 0:
new_messages.append({"role": "user", "content": new_user_content_list}) # type: ignore
if len(tool_message_list) > 0:
new_messages.extend(tool_message_list)
## ASSISTANT MESSAGE ##
assistant_message_str: Optional[str] = None
tool_calls: List[ChatCompletionAssistantToolCall] = []
if m["role"] == "assistant":
if isinstance(m["content"], str):
assistant_message_str = m["content"]
elif isinstance(m["content"], list):
for content in m["content"]:
if content["type"] == "text":
if assistant_message_str is None:
assistant_message_str = content["text"]
else:
assistant_message_str += content["text"]
elif content["type"] == "tool_use":
function_chunk = ChatCompletionToolCallFunctionChunk(
name=content["name"],
arguments=json.dumps(content["input"]),
)
tool_calls.append(
ChatCompletionAssistantToolCall(
id=content["id"],
type="function",
function=function_chunk,
)
)
if assistant_message_str is not None or len(tool_calls) > 0:
assistant_message = ChatCompletionAssistantMessage(
role="assistant",
content=assistant_message_str,
)
if len(tool_calls) > 0:
assistant_message["tool_calls"] = tool_calls
new_messages.append(assistant_message)
return new_messages
def translate_anthropic_tool_choice_to_openai(
self, tool_choice: AnthropicMessagesToolChoice
) -> ChatCompletionToolChoiceValues:
if tool_choice["type"] == "any":
return "required"
elif tool_choice["type"] == "auto":
return "auto"
elif tool_choice["type"] == "tool":
tc_function_param = ChatCompletionToolChoiceFunctionParam(
name=tool_choice.get("name", "")
)
return ChatCompletionToolChoiceObjectParam(
type="function", function=tc_function_param
)
else:
raise ValueError(
"Incompatible tool choice param submitted - {}".format(tool_choice)
)
def translate_anthropic_tools_to_openai(
self, tools: List[AllAnthropicToolsValues]
) -> List[ChatCompletionToolParam]:
new_tools: List[ChatCompletionToolParam] = []
mapped_tool_params = ["name", "input_schema", "description"]
for tool in tools:
function_chunk = ChatCompletionToolParamFunctionChunk(
name=tool["name"],
)
if "input_schema" in tool:
function_chunk["parameters"] = tool["input_schema"] # type: ignore
if "description" in tool:
function_chunk["description"] = tool["description"] # type: ignore
for k, v in tool.items():
if k not in mapped_tool_params: # pass additional computer kwargs
function_chunk.setdefault("parameters", {}).update({k: v})
new_tools.append(
ChatCompletionToolParam(type="function", function=function_chunk)
)
return new_tools
def translate_anthropic_to_openai(
self, anthropic_message_request: AnthropicMessagesRequest
) -> ChatCompletionRequest:
"""
This is used by the beta Anthropic Adapter, for translating anthropic `/v1/messages` requests to the openai format.
"""
new_messages: List[AllMessageValues] = []
## CONVERT ANTHROPIC MESSAGES TO OPENAI
new_messages = self.translate_anthropic_messages_to_openai(
messages=anthropic_message_request["messages"]
)
## ADD SYSTEM MESSAGE TO MESSAGES
if "system" in anthropic_message_request:
new_messages.insert(
0,
ChatCompletionSystemMessage(
role="system", content=anthropic_message_request["system"]
),
)
new_kwargs: ChatCompletionRequest = {
"model": anthropic_message_request["model"],
"messages": new_messages,
}
## CONVERT METADATA (user_id)
if "metadata" in anthropic_message_request:
if "user_id" in anthropic_message_request["metadata"]:
new_kwargs["user"] = anthropic_message_request["metadata"]["user_id"]
# Pass litellm proxy specific metadata
if "litellm_metadata" in anthropic_message_request:
# metadata will be passed to litellm.acompletion(), it's a litellm_param
new_kwargs["metadata"] = anthropic_message_request.pop("litellm_metadata")
## CONVERT TOOL CHOICE
if "tool_choice" in anthropic_message_request:
new_kwargs["tool_choice"] = self.translate_anthropic_tool_choice_to_openai(
tool_choice=anthropic_message_request["tool_choice"]
)
## CONVERT TOOLS
if "tools" in anthropic_message_request:
new_kwargs["tools"] = self.translate_anthropic_tools_to_openai(
tools=anthropic_message_request["tools"]
)
translatable_params = self.translatable_anthropic_params()
for k, v in anthropic_message_request.items():
if k not in translatable_params: # pass remaining params as is
new_kwargs[k] = v # type: ignore
return new_kwargs
def _translate_openai_content_to_anthropic(
self, choices: List[Choices]
) -> List[
Union[AnthropicResponseContentBlockText, AnthropicResponseContentBlockToolUse]
]:
new_content: List[
Union[
AnthropicResponseContentBlockText, AnthropicResponseContentBlockToolUse
]
] = []
for choice in choices:
if (
choice.message.tool_calls is not None
and len(choice.message.tool_calls) > 0
):
for tool_call in choice.message.tool_calls:
new_content.append(
AnthropicResponseContentBlockToolUse(
type="tool_use",
id=tool_call.id,
name=tool_call.function.name or "",
input=json.loads(tool_call.function.arguments),
)
)
elif choice.message.content is not None:
new_content.append(
AnthropicResponseContentBlockText(
type="text", text=choice.message.content
)
)
return new_content
def _translate_openai_finish_reason_to_anthropic(
self, openai_finish_reason: str
) -> AnthropicFinishReason:
if openai_finish_reason == "stop":
return "end_turn"
elif openai_finish_reason == "length":
return "max_tokens"
elif openai_finish_reason == "tool_calls":
return "tool_use"
return "end_turn"
def translate_openai_response_to_anthropic(
self, response: ModelResponse
) -> AnthropicResponse:
## translate content block
anthropic_content = self._translate_openai_content_to_anthropic(choices=response.choices) # type: ignore
## extract finish reason
anthropic_finish_reason = self._translate_openai_finish_reason_to_anthropic(
openai_finish_reason=response.choices[0].finish_reason # type: ignore
)
# extract usage
usage: Usage = getattr(response, "usage")
anthropic_usage = AnthropicResponseUsageBlock(
input_tokens=usage.prompt_tokens or 0,
output_tokens=usage.completion_tokens or 0,
)
translated_obj = AnthropicResponse(
id=response.id,
type="message",
role="assistant",
model=response.model or "unknown-model",
stop_sequence=None,
usage=anthropic_usage,
content=anthropic_content,
stop_reason=anthropic_finish_reason,
)
return translated_obj
def _translate_streaming_openai_chunk_to_anthropic(
self, choices: List[OpenAIStreamingChoice]
) -> Tuple[
Literal["text_delta", "input_json_delta"],
Union[ContentTextBlockDelta, ContentJsonBlockDelta],
]:
text: str = ""
partial_json: Optional[str] = None
for choice in choices:
if choice.delta.content is not None:
text += choice.delta.content
elif choice.delta.tool_calls is not None:
partial_json = ""
for tool in choice.delta.tool_calls:
if (
tool.function is not None
and tool.function.arguments is not None
):
partial_json += tool.function.arguments
if partial_json is not None:
return "input_json_delta", ContentJsonBlockDelta(
type="input_json_delta", partial_json=partial_json
)
else:
return "text_delta", ContentTextBlockDelta(type="text_delta", text=text)
def translate_streaming_openai_response_to_anthropic(
self, response: ModelResponse
) -> Union[ContentBlockDelta, MessageBlockDelta]:
## base case - final chunk w/ finish reason
if response.choices[0].finish_reason is not None:
delta = MessageDelta(
stop_reason=self._translate_openai_finish_reason_to_anthropic(
response.choices[0].finish_reason
),
)
if getattr(response, "usage", None) is not None:
litellm_usage_chunk: Optional[Usage] = response.usage # type: ignore
elif (
hasattr(response, "_hidden_params")
and "usage" in response._hidden_params
):
litellm_usage_chunk = response._hidden_params["usage"]
else:
litellm_usage_chunk = None
if litellm_usage_chunk is not None:
usage_delta = UsageDelta(
input_tokens=litellm_usage_chunk.prompt_tokens or 0,
output_tokens=litellm_usage_chunk.completion_tokens or 0,
)
else:
usage_delta = UsageDelta(input_tokens=0, output_tokens=0)
return MessageBlockDelta(
type="message_delta", delta=delta, usage=usage_delta
)
(
type_of_content,
content_block_delta,
) = self._translate_streaming_openai_chunk_to_anthropic(
choices=response.choices # type: ignore
)
return ContentBlockDelta(
type="content_block_delta",
index=response.choices[0].index,
delta=content_block_delta,
)
|