Minecraft Bot LLM Backend
System Design

# ── Stage 1: builder (compile llama-cpp-python)
FROM python:3.11-slim AS builder
RUN apt-get install -y build-essential cmake wget ...

ENV CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS"
RUN pip install -r requirements.txt --target /build/deps

# ── Stage 2: runtime (slim image)
FROM python:3.11-slim
RUN useradd -m -u 1000 user   # HF Spaces yêu cầu non-root
USER user

# Download GGUF model tại BUILD time (~4.4 GB)
RUN python -c "from huggingface_hub import hf_hub_download; \
    hf_hub_download(repo_id='Qwen/Qwen2.5-Coder-7B-Instruct-GGUF', \
    filename='qwen2.5-coder-7b-instruct-q4_k_m.gguf', \
    local_dir='/app/models')"

ENV MODEL_PATH=/app/models/qwen2.5-coder-7b-instruct-q4_k_m.gguf
# BEARER_TOKEN được inject từ HF Secret — không hard-code ở đây!

EXPOSE 7860
CMD ["python", "-m", "uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

from fastapi import FastAPI, HTTPException, Depends
from fastapi.security import HTTPBearer, HTTPAuthorizationCredentials
import os

BEARER_TOKEN = os.environ.get("BEARER_TOKEN", "")  # ← từ HF Secret

# Auth middleware
def verify_token(creds: HTTPAuthorizationCredentials):
    if creds.credentials != BEARER_TOKEN:
        raise HTTPException(401, "Invalid Bearer Token")

# OpenAI-compatible endpoint
@app.post("/v1/chat/completions", dependencies=[Depends(verify_token)])
async def chat_completions(request: ChatCompletionRequest):
    result = llm.create_chat_completion(
        messages=[{"role": m.role, "content": m.content} for m in request.messages],
        max_tokens=request.max_tokens,
        temperature=request.temperature,
    )
    return ChatCompletionResponse(...)   # wrapped in OpenAI schema

llama-cpp-python==0.3.4     # GGUF inference engine
fastapi==0.115.6          # API framework
uvicorn[standard]==0.32.1  # ASGI server
pydantic==2.10.3          # Schema validation
huggingface-hub==0.27.0   # Model download
httpx==0.28.1             # HTTP client

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://<username>-<space>.hf.space/v1",
  apiKey:  process.env.BEARER_TOKEN,
});

const response = await client.chat.completions.create({
  model:    "qwen2.5-coder-7b-instruct",
  messages: [
    { role: "system",  content: "You are a Minecraft bot brain..." },
    { role: "user",    content: "Bot at x=120. Nearest: oak_log. Chop it." },
  ],
  max_tokens:  512,
  temperature: 0.2,
});

from openai import OpenAI
import os

client = OpenAI(
    base_url="https://<username>-<space>.hf.space/v1",
    api_key=os.environ.get("BEARER_TOKEN"),
)

response = client.chat.completions.create(
    model="qwen2.5-coder-7b-instruct",
    messages=[
        {"role": "system", "content": "You are a Minecraft bot brain..."},
        {"role": "user",   "content": "Bot at x=120. Nearest: oak_log. Chop it."},
    ],
    max_tokens=512,
    temperature=0.2,
)