mirror of
https://github.com/BerriAI/litellm.git
synced 2025-04-26 03:04:13 +00:00
code cleanup
This commit is contained in:
parent
e9427205ef
commit
42b95c5979
3 changed files with 88 additions and 67 deletions
|
@ -35,8 +35,9 @@ from litellm.proxy._types import (
|
||||||
)
|
)
|
||||||
from litellm.proxy.auth.user_api_key_auth import user_api_key_auth
|
from litellm.proxy.auth.user_api_key_auth import user_api_key_auth
|
||||||
|
|
||||||
from .streaming_handler import ModelIteratorType, chunk_processor
|
from .streaming_handler import chunk_processor
|
||||||
from .success_handler import PassThroughEndpointLogging
|
from .success_handler import PassThroughEndpointLogging
|
||||||
|
from .types import EndpointType
|
||||||
|
|
||||||
router = APIRouter()
|
router = APIRouter()
|
||||||
|
|
||||||
|
@ -288,6 +289,12 @@ def get_response_headers(headers: httpx.Headers) -> dict:
|
||||||
return return_headers
|
return return_headers
|
||||||
|
|
||||||
|
|
||||||
|
def get_endpoint_type(url: str) -> EndpointType:
|
||||||
|
if ("generateContent") in url or ("streamGenerateContent") in url:
|
||||||
|
return EndpointType.VERTEX_AI
|
||||||
|
return EndpointType.GENERIC
|
||||||
|
|
||||||
|
|
||||||
async def pass_through_request(
|
async def pass_through_request(
|
||||||
request: Request,
|
request: Request,
|
||||||
target: str,
|
target: str,
|
||||||
|
@ -311,6 +318,8 @@ async def pass_through_request(
|
||||||
request=request, headers=headers, forward_headers=forward_headers
|
request=request, headers=headers, forward_headers=forward_headers
|
||||||
)
|
)
|
||||||
|
|
||||||
|
endpoint_type: EndpointType = get_endpoint_type(str(url))
|
||||||
|
|
||||||
_parsed_body = None
|
_parsed_body = None
|
||||||
if custom_body:
|
if custom_body:
|
||||||
_parsed_body = custom_body
|
_parsed_body = custom_body
|
||||||
|
@ -424,7 +433,7 @@ async def pass_through_request(
|
||||||
async for chunk in chunk_processor(
|
async for chunk in chunk_processor(
|
||||||
response.aiter_bytes(),
|
response.aiter_bytes(),
|
||||||
litellm_logging_obj=logging_obj,
|
litellm_logging_obj=logging_obj,
|
||||||
iterator_type=ModelIteratorType.VERTEX_AI,
|
endpoint_type=endpoint_type,
|
||||||
start_time=start_time,
|
start_time=start_time,
|
||||||
passthrough_success_handler_obj=pass_through_endpoint_logging,
|
passthrough_success_handler_obj=pass_through_endpoint_logging,
|
||||||
url_route=str(url),
|
url_route=str(url),
|
||||||
|
@ -468,7 +477,7 @@ async def pass_through_request(
|
||||||
async for chunk in chunk_processor(
|
async for chunk in chunk_processor(
|
||||||
response.aiter_bytes(),
|
response.aiter_bytes(),
|
||||||
litellm_logging_obj=logging_obj,
|
litellm_logging_obj=logging_obj,
|
||||||
iterator_type=ModelIteratorType.VERTEX_AI,
|
endpoint_type=endpoint_type,
|
||||||
start_time=start_time,
|
start_time=start_time,
|
||||||
passthrough_success_handler_obj=pass_through_endpoint_logging,
|
passthrough_success_handler_obj=pass_through_endpoint_logging,
|
||||||
url_route=str(url),
|
url_route=str(url),
|
||||||
|
|
|
@ -12,17 +12,7 @@ from litellm.llms.vertex_ai_and_google_ai_studio.gemini.vertex_and_google_ai_stu
|
||||||
from litellm.types.utils import GenericStreamingChunk
|
from litellm.types.utils import GenericStreamingChunk
|
||||||
|
|
||||||
from .success_handler import PassThroughEndpointLogging
|
from .success_handler import PassThroughEndpointLogging
|
||||||
|
from .types import EndpointType
|
||||||
|
|
||||||
class ModelIteratorType(Enum):
|
|
||||||
VERTEX_AI = "vertexAI"
|
|
||||||
# Add more iterator types here as needed
|
|
||||||
|
|
||||||
|
|
||||||
MODEL_ITERATORS: Dict[ModelIteratorType, type] = {
|
|
||||||
ModelIteratorType.VERTEX_AI: VertexAIIterator,
|
|
||||||
# Add more mappings here as needed
|
|
||||||
}
|
|
||||||
|
|
||||||
|
|
||||||
def get_litellm_chunk(
|
def get_litellm_chunk(
|
||||||
|
@ -37,73 +27,89 @@ def get_litellm_chunk(
|
||||||
return None
|
return None
|
||||||
|
|
||||||
|
|
||||||
|
def get_iterator_class_from_endpoint_type(
|
||||||
|
endpoint_type: EndpointType,
|
||||||
|
) -> Optional[type]:
|
||||||
|
if endpoint_type == EndpointType.VERTEX_AI:
|
||||||
|
return VertexAIIterator
|
||||||
|
return None
|
||||||
|
|
||||||
|
|
||||||
async def chunk_processor(
|
async def chunk_processor(
|
||||||
aiter_bytes: AsyncIterable[bytes],
|
aiter_bytes: AsyncIterable[bytes],
|
||||||
litellm_logging_obj: LiteLLMLoggingObj,
|
litellm_logging_obj: LiteLLMLoggingObj,
|
||||||
iterator_type: ModelIteratorType,
|
endpoint_type: EndpointType,
|
||||||
start_time: datetime,
|
start_time: datetime,
|
||||||
passthrough_success_handler_obj: PassThroughEndpointLogging,
|
passthrough_success_handler_obj: PassThroughEndpointLogging,
|
||||||
url_route: str,
|
url_route: str,
|
||||||
) -> AsyncIterable[bytes]:
|
) -> AsyncIterable[bytes]:
|
||||||
|
|
||||||
IteratorClass = MODEL_ITERATORS[iterator_type]
|
iteratorClass = get_iterator_class_from_endpoint_type(endpoint_type)
|
||||||
model_iterator = IteratorClass(sync_stream=False, streaming_response=aiter_bytes)
|
if iteratorClass is None:
|
||||||
custom_stream_wrapper = litellm.utils.CustomStreamWrapper(
|
# Generic endpoint - litellm does not do any tracking / logging for this
|
||||||
completion_stream=aiter_bytes, model=None, logging_obj=litellm_logging_obj
|
async for chunk in aiter_bytes:
|
||||||
)
|
yield chunk
|
||||||
buffer = b""
|
else:
|
||||||
all_chunks = []
|
# known streaming endpoint - litellm will do tracking / logging for this
|
||||||
async for chunk in aiter_bytes:
|
model_iterator = iteratorClass(
|
||||||
buffer += chunk
|
sync_stream=False, streaming_response=aiter_bytes
|
||||||
try:
|
)
|
||||||
_decoded_chunk = chunk.decode("utf-8")
|
custom_stream_wrapper = litellm.utils.CustomStreamWrapper(
|
||||||
_chunk_dict = json.loads(_decoded_chunk)
|
completion_stream=aiter_bytes, model=None, logging_obj=litellm_logging_obj
|
||||||
litellm_chunk = get_litellm_chunk(
|
)
|
||||||
model_iterator, custom_stream_wrapper, _chunk_dict
|
buffer = b""
|
||||||
)
|
all_chunks = []
|
||||||
if litellm_chunk:
|
async for chunk in aiter_bytes:
|
||||||
all_chunks.append(litellm_chunk)
|
buffer += chunk
|
||||||
except json.JSONDecodeError:
|
try:
|
||||||
pass
|
_decoded_chunk = chunk.decode("utf-8")
|
||||||
finally:
|
_chunk_dict = json.loads(_decoded_chunk)
|
||||||
yield chunk # Yield the original bytes
|
|
||||||
|
|
||||||
# Process any remaining data in the buffer
|
|
||||||
if buffer:
|
|
||||||
try:
|
|
||||||
_chunk_dict = json.loads(buffer.decode("utf-8"))
|
|
||||||
|
|
||||||
if isinstance(_chunk_dict, list):
|
|
||||||
for _chunk in _chunk_dict:
|
|
||||||
litellm_chunk = get_litellm_chunk(
|
|
||||||
model_iterator, custom_stream_wrapper, _chunk
|
|
||||||
)
|
|
||||||
if litellm_chunk:
|
|
||||||
all_chunks.append(litellm_chunk)
|
|
||||||
elif isinstance(_chunk_dict, dict):
|
|
||||||
litellm_chunk = get_litellm_chunk(
|
litellm_chunk = get_litellm_chunk(
|
||||||
model_iterator, custom_stream_wrapper, _chunk_dict
|
model_iterator, custom_stream_wrapper, _chunk_dict
|
||||||
)
|
)
|
||||||
if litellm_chunk:
|
if litellm_chunk:
|
||||||
all_chunks.append(litellm_chunk)
|
all_chunks.append(litellm_chunk)
|
||||||
except json.JSONDecodeError:
|
except json.JSONDecodeError:
|
||||||
pass
|
pass
|
||||||
|
finally:
|
||||||
|
yield chunk # Yield the original bytes
|
||||||
|
|
||||||
complete_streaming_response: litellm.ModelResponse = litellm.stream_chunk_builder(
|
# Process any remaining data in the buffer
|
||||||
chunks=all_chunks
|
if buffer:
|
||||||
)
|
try:
|
||||||
end_time = datetime.now()
|
_chunk_dict = json.loads(buffer.decode("utf-8"))
|
||||||
|
|
||||||
if passthrough_success_handler_obj.is_vertex_route(url_route):
|
if isinstance(_chunk_dict, list):
|
||||||
_model = passthrough_success_handler_obj.extract_model_from_url(url_route)
|
for _chunk in _chunk_dict:
|
||||||
complete_streaming_response.model = _model
|
litellm_chunk = get_litellm_chunk(
|
||||||
litellm_logging_obj.model = _model
|
model_iterator, custom_stream_wrapper, _chunk
|
||||||
litellm_logging_obj.model_call_details["model"] = _model
|
)
|
||||||
|
if litellm_chunk:
|
||||||
|
all_chunks.append(litellm_chunk)
|
||||||
|
elif isinstance(_chunk_dict, dict):
|
||||||
|
litellm_chunk = get_litellm_chunk(
|
||||||
|
model_iterator, custom_stream_wrapper, _chunk_dict
|
||||||
|
)
|
||||||
|
if litellm_chunk:
|
||||||
|
all_chunks.append(litellm_chunk)
|
||||||
|
except json.JSONDecodeError:
|
||||||
|
pass
|
||||||
|
|
||||||
asyncio.create_task(
|
complete_streaming_response: litellm.ModelResponse = (
|
||||||
litellm_logging_obj.async_success_handler(
|
litellm.stream_chunk_builder(chunks=all_chunks)
|
||||||
result=complete_streaming_response,
|
)
|
||||||
start_time=start_time,
|
end_time = datetime.now()
|
||||||
end_time=end_time,
|
|
||||||
|
if passthrough_success_handler_obj.is_vertex_route(url_route):
|
||||||
|
_model = passthrough_success_handler_obj.extract_model_from_url(url_route)
|
||||||
|
complete_streaming_response.model = _model
|
||||||
|
litellm_logging_obj.model = _model
|
||||||
|
litellm_logging_obj.model_call_details["model"] = _model
|
||||||
|
|
||||||
|
asyncio.create_task(
|
||||||
|
litellm_logging_obj.async_success_handler(
|
||||||
|
result=complete_streaming_response,
|
||||||
|
start_time=start_time,
|
||||||
|
end_time=end_time,
|
||||||
|
)
|
||||||
)
|
)
|
||||||
)
|
|
||||||
|
|
6
litellm/proxy/pass_through_endpoints/types.py
Normal file
6
litellm/proxy/pass_through_endpoints/types.py
Normal file
|
@ -0,0 +1,6 @@
|
||||||
|
from enum import Enum
|
||||||
|
|
||||||
|
|
||||||
|
class EndpointType(str, Enum):
|
||||||
|
VERTEX_AI = "vertex-ai"
|
||||||
|
GENERIC = "generic"
|
Loading…
Add table
Add a link
Reference in a new issue