89 lines
3.0 KiB
Bash
Executable File
89 lines
3.0 KiB
Bash
Executable File
#!/bin/bash
|
||
# ============================================================
|
||
# 预加载 Terminal-Bench 2.1 数据集和 Docker 镜像
|
||
#
|
||
# 用途:
|
||
# terminal_bench_v2_1 运行时 Harbor 会按需拉取 Docker 镜像,
|
||
# 导致首次运行很慢。本脚本先把数据集和 89 个镜像拉到本地,
|
||
# 后续评测就只需要跑 agent,不用等镜像下载。
|
||
#
|
||
# 用法:
|
||
# bash bash/images_load/preload_terminal_bench_images.sh
|
||
#
|
||
# 注意:
|
||
# - 需要 Docker 环境
|
||
# - 89 个镜像大约需要几十 GB 磁盘空间
|
||
# - 执行时间取决于网速,可能需要 30 分钟 ~ 数小时
|
||
# ============================================================
|
||
|
||
set -euo pipefail
|
||
|
||
DATASET_NAME="terminal-bench/terminal-bench-2-1"
|
||
CACHE_DIR="${HOME}/.cache/harbor/tasks/packages/terminal-bench"
|
||
|
||
echo "============================================================"
|
||
echo "1. 预下载 Terminal-Bench 2.1 数据集"
|
||
echo "============================================================"
|
||
harbor download "${DATASET_NAME}" --cache
|
||
|
||
echo ""
|
||
echo "============================================================"
|
||
echo "2. 提取所有任务镜像列表"
|
||
echo "============================================================"
|
||
IMAGE_FILE="/tmp/terminal_bench_images.txt"
|
||
if [[ ! -d "${CACHE_DIR}" ]]; then
|
||
echo "ERROR: 数据集缓存目录不存在: ${CACHE_DIR}"
|
||
exit 1
|
||
fi
|
||
|
||
grep -hR "^docker_image" "${CACHE_DIR}" 2>/dev/null \
|
||
| sed 's/docker_image = //; s/"//g' \
|
||
| sort -u > "${IMAGE_FILE}"
|
||
|
||
IMAGE_COUNT=$(wc -l < "${IMAGE_FILE}" | tr -d ' ')
|
||
echo "找到 ${IMAGE_COUNT} 个唯一镜像,列表保存在 ${IMAGE_FILE}"
|
||
|
||
echo ""
|
||
echo "============================================================"
|
||
echo "3. 并行拉取 Docker 镜像"
|
||
echo "============================================================"
|
||
|
||
# 脚本放在 bash/images_load/,项目根目录是上一级再上一级
|
||
PROJECT_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
|
||
LOG_DIR="${PROJECT_ROOT}/logs"
|
||
LOG_FILE="${LOG_DIR}/pull_terminal_images.log"
|
||
mkdir -p "${LOG_DIR}"
|
||
> "${LOG_FILE}"
|
||
|
||
echo "开始拉取镜像(后台 4 并发),日志保存在 ${LOG_FILE}"
|
||
|
||
PULL_FAILED=0
|
||
while IFS= read -r image; do
|
||
[[ -z "$image" ]] && continue
|
||
(
|
||
echo "[$(date '+%Y-%m-%d %H:%M:%S')] Pulling ${image} ..." >> "${LOG_FILE}"
|
||
if docker pull "${image}" >> "${LOG_FILE}" 2>&1; then
|
||
echo "[$(date '+%Y-%m-%d %H:%M:%S')] OK ${image}" >> "${LOG_FILE}"
|
||
else
|
||
echo "[$(date '+%Y-%m-%d %H:%M:%S')] FAILED ${image}" >> "${LOG_FILE}"
|
||
exit 1
|
||
fi
|
||
) &
|
||
# 控制并发数 4
|
||
if (( $(jobs -r -p | wc -l) >= 4 )); then
|
||
wait -n || PULL_FAILED=1
|
||
fi
|
||
done < "${IMAGE_FILE}"
|
||
|
||
wait || PULL_FAILED=1
|
||
|
||
echo ""
|
||
echo "============================================================"
|
||
if [[ "$PULL_FAILED" -eq 0 ]]; then
|
||
echo "所有镜像拉取完成"
|
||
else
|
||
echo "部分镜像拉取失败,请查看日志: ${LOG_FILE}"
|
||
exit 1
|
||
fi
|
||
echo "============================================================"
|