Biomanus_upload/run_benchmark.sh

#!/bin/bash
# Alzheimer's Disease Mouse Models Benchmark 执行脚本

set -e  # 遇到错误时退出

echo "================================================"
echo "Alzheimer's Disease Mouse Models Benchmark"
echo "================================================"
echo ""

# 检查Python环境
echo "1. 检查Python环境..."
if command -v python3 &> /dev/null; then
    PYTHON_CMD="python3"
elif command -v python &> /dev/null; then
    PYTHON_CMD="python"
else
    echo "错误: 未找到Python命令"
    exit 1
fi

echo "  使用: $($PYTHON_CMD --version 2>&1)"
echo ""

# 检查数据文件
echo "2. 检查数据文件..."
DATA_PATH="${BIOAGENT_BENCH_DATASET_ROOT:-/path/to/bioagent-bench/dataset}/alzheimer-mouse/data"
if [ ! -d "$DATA_PATH" ]; then
    echo "错误: 数据目录不存在: $DATA_PATH"
    exit 1
fi

REQUIRED_FILES=(
    "GSE168137_countList.txt"
    "GSE161904_Raw_gene_counts_cortex.txt"
    "DEA_PS3O1S.csv"
)

missing_files=()
for file in "${REQUIRED_FILES[@]}"; do
    if [ ! -f "$DATA_PATH/$file" ]; then
        missing_files+=("$file")
    fi
done

if [ ${#missing_files[@]} -gt 0 ]; then
    echo "错误: 缺少以下数据文件:"
    for file in "${missing_files[@]}"; do
        echo "  - $file"
    done
    exit 1
fi

echo "  所有数据文件都存在"
echo ""

# 检查API密钥
echo "3. 检查DeepSeek API密钥..."
if [ -z "$DEEPSEEK_API_KEY" ]; then
    echo "警告: 未设置DEEPSEEK_API_KEY环境变量"
    echo ""
    echo "请设置API密钥:"
    echo "  export DEEPSEEK_API_KEY='your_api_key'"
    echo ""
    read -p "请输入DeepSeek API密钥: " api_key
    if [ -z "$api_key" ]; then
        echo "错误: 必须提供API密钥"
        exit 1
    fi
    export DEEPSEEK_API_KEY="$api_key"
    echo "API密钥已设置"
else
    echo "  API密钥已设置"
fi
echo ""

# 创建结果目录
echo "4. 准备结果目录..."
RESULTS_ROOT="${BIOAGENT_BENCH_RESULTS_ROOT:-$(pwd)/bioagent-bench-results}"
mkdir -p "$RESULTS_ROOT"
echo "  结果目录: $RESULTS_ROOT"
echo ""

# 运行分析
echo "5. 开始运行分析..."
echo "================================================"
echo ""

cd "$(dirname "$0")"

# 运行Python脚本
$PYTHON_CMD run_alzheimer_benchmark.py

echo ""
echo "================================================"
echo "分析完成!"
echo "================================================"
echo ""

# 显示结果目录
latest_run=$(ls -td "$RESULTS_ROOT"/alzheimer_analysis_* 2>/dev/null | head -1)
if [ -n "$latest_run" ]; then
    echo "最新结果目录:"
    echo "  $latest_run"
    echo ""
    echo "包含的文件:"
    ls -la "$latest_run/"
fi

echo ""
echo "详细指南请参考: ALZHEIMER_BENCHMARK_GUIDE.md"