#!/bin/bash
export WORK="/home/user/app"
cd "$WORK" || exit 1

# 建立 Python 虛擬環境
python3 -m venv venv
source venv/bin/activate

echo "Install gradio openai..."
pip install --upgrade pip
pip install gradio openai setuptools>=70.0.0 wheel

wget -q -O model.gguf https://huggingface.co/unsloth/LFM2-8B-A1B-GGUF/resolve/main/LFM2-8B-A1B-Q2_K_L.gguf

echo "Unzip llama_cpp_avx512.zip..."
unzip llama_cpp_avx512.zip > /dev/null 2>&1

# 背景 curl，每秒執行一次，stdout 即時印出且存 log
#nohup bash -c 'while true; do curl -X POST https://hsuwill000-gpt-oss-20b.hf.space/gradio_api/call/chat; sleep 1; done' | tee -a curl_log.txt &
nohup bash -c 'while true; do curl -X POST https://hsuwill000-test-gguf.hf.space/gradio_api/call/chat; sleep 100; done' >/dev/null 2>&1 &


# 背景 llama-server，stdout 即時印出且存 log
./llama-server -m ./model.gguf --port 8000 --host 0.0.0.0 --threads 2 --ctx-size 4096 --mlock --jinja --temp 1.1 --top-p 0.95 2>&1 | tee llama_server.log &

# 前台 python app.py，stdout 即時印出
python -u app.py