chore: remove Google provider entirely

fix(google): match OpenAI finish_reason format exactly

fix(google): finish_reason only in final SSE chunk

fix(google): buffer non-streaming JSON, convert to fake SSE chunks

fix(google): strip repetition_penalty and top_k from request body

fix(google): force non-streaming to bypass HF Spaces HTTP/2 edge bug

fix(google): strip extra_content/thought_signature from SSE chunks

fix(google): strip unsupported params from request body

fix(streaming): decompress upstream chunks for Google SSE

update model list by removing deprecated entries and consolidating Google models

clean error messages, add model validation, add gemini-3.1-flash-lite

fix google auth: use Authorization Bearer for OpenAI-compatible endpoint

fix streaming: async generator keeps httpx client alive internally

fix streaming: use client.send() with stream=True instead of context manager

fix httpx streaming: use client.stream() for async SSE

use shell form CMD

fix Dockerfile user for HF Spaces

fix README metadata

add app_port and base_path to README metadata

simplify Dockerfile, use python3.11

revert to /v1 prefix with APIRouter

change API prefix to /api/v1 to avoid HF routing conflict

remove APIRouter prefix, hardcode paths

add documentation for graphify and OpenAI API Proxy setup

add forwarded-allow-ips

add root endpoint for HF Spaces routing

deploy to HF Spaces: add Dockerfile, remove dashboard

Merge branch 'main' of https://huggingface.co/spaces/Wimmboo2/OpenAIProxy

initial commit

feat: add OpenAI-compatible API proxy with NVIDIA and Google integration