119 lines
3.8 KiB
Python
119 lines
3.8 KiB
Python
# SPDX-License-Identifier: Apache-2.0
|
|
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
|
|
|
|
|
from http import HTTPStatus
|
|
|
|
from fastapi import APIRouter, Depends, FastAPI, HTTPException, Request
|
|
from fastapi.exceptions import RequestValidationError
|
|
from fastapi.responses import JSONResponse
|
|
from typing_extensions import assert_never
|
|
|
|
from vllm.entrypoints.openai.api_server import validate_json_request
|
|
from vllm.entrypoints.openai.protocol import (
|
|
DetokenizeRequest,
|
|
DetokenizeResponse,
|
|
ErrorResponse,
|
|
TokenizeRequest,
|
|
TokenizeResponse,
|
|
)
|
|
from vllm.entrypoints.serve.tokenize.serving import OpenAIServingTokenization
|
|
from vllm.entrypoints.utils import (
|
|
with_cancellation,
|
|
)
|
|
from vllm.logger import init_logger
|
|
|
|
logger = init_logger(__name__)
|
|
|
|
|
|
def tokenization(request: Request) -> OpenAIServingTokenization:
|
|
return request.app.state.openai_serving_tokenization
|
|
|
|
|
|
router = APIRouter()
|
|
|
|
|
|
@router.post(
|
|
"/tokenize",
|
|
dependencies=[Depends(validate_json_request)],
|
|
responses={
|
|
HTTPStatus.BAD_REQUEST.value: {"model": ErrorResponse},
|
|
HTTPStatus.NOT_FOUND.value: {"model": ErrorResponse},
|
|
HTTPStatus.INTERNAL_SERVER_ERROR.value: {"model": ErrorResponse},
|
|
HTTPStatus.NOT_IMPLEMENTED.value: {"model": ErrorResponse},
|
|
},
|
|
)
|
|
@with_cancellation
|
|
async def tokenize(request: TokenizeRequest, raw_request: Request):
|
|
handler = tokenization(raw_request)
|
|
|
|
try:
|
|
generator = await handler.create_tokenize(request, raw_request)
|
|
except NotImplementedError as e:
|
|
raise HTTPException(
|
|
status_code=HTTPStatus.NOT_IMPLEMENTED.value, detail=str(e)
|
|
) from e
|
|
except Exception as e:
|
|
raise HTTPException(
|
|
status_code=HTTPStatus.INTERNAL_SERVER_ERROR.value, detail=str(e)
|
|
) from e
|
|
|
|
if isinstance(generator, ErrorResponse):
|
|
return JSONResponse(
|
|
content=generator.model_dump(), status_code=generator.error.code
|
|
)
|
|
elif isinstance(generator, TokenizeResponse):
|
|
return JSONResponse(content=generator.model_dump())
|
|
|
|
assert_never(generator)
|
|
|
|
|
|
@router.post(
|
|
"/detokenize",
|
|
dependencies=[Depends(validate_json_request)],
|
|
responses={
|
|
HTTPStatus.BAD_REQUEST.value: {"model": ErrorResponse},
|
|
HTTPStatus.NOT_FOUND.value: {"model": ErrorResponse},
|
|
HTTPStatus.INTERNAL_SERVER_ERROR.value: {"model": ErrorResponse},
|
|
},
|
|
)
|
|
@with_cancellation
|
|
async def detokenize(request: DetokenizeRequest, raw_request: Request):
|
|
handler = tokenization(raw_request)
|
|
|
|
try:
|
|
generator = await handler.create_detokenize(request, raw_request)
|
|
except OverflowError as e:
|
|
raise RequestValidationError(errors=[str(e)]) from e
|
|
except Exception as e:
|
|
raise HTTPException(
|
|
status_code=HTTPStatus.INTERNAL_SERVER_ERROR.value, detail=str(e)
|
|
) from e
|
|
|
|
if isinstance(generator, ErrorResponse):
|
|
return JSONResponse(
|
|
content=generator.model_dump(), status_code=generator.error.code
|
|
)
|
|
elif isinstance(generator, DetokenizeResponse):
|
|
return JSONResponse(content=generator.model_dump())
|
|
|
|
assert_never(generator)
|
|
|
|
|
|
def attach_router(app: FastAPI):
|
|
if getattr(app.state.args, "enable_tokenizer_info_endpoint", False):
|
|
"""Conditionally register the tokenizer info endpoint if enabled."""
|
|
|
|
@router.get("/tokenizer_info")
|
|
async def get_tokenizer_info(raw_request: Request):
|
|
"""Get comprehensive tokenizer information."""
|
|
result = await tokenization(raw_request).get_tokenizer_info()
|
|
return JSONResponse(
|
|
content=result.model_dump(),
|
|
status_code=result.error.code
|
|
if isinstance(result, ErrorResponse)
|
|
else 200,
|
|
)
|
|
|
|
app.include_router(router)
|