eqbench-ja-run/setup_eqbench_run.sh

#!/usr/bin/env bash
# =============================================================================
# setup_eqbench_run.sh — EQ-Bench3 日本語版 評価実行セットアップ（ローカルvLLM）
#
# やること:
#   1. EQ-Bench3 リポジトリのクローン
#   2. 全12日本語データファイルの差し替え [Fix-JA-D1]
#   3. benchmark.py / conversation.py にコードパッチ適用 [Fix-JA-D2]
#   4. 依存関係のインストール
#   5. .env ファイルの生成（ローカルvLLMを向く）
#   6. 各種 serve スクリプトの配置
#   7. 実行手順の表示
# =============================================================================

set -euo pipefail

WORKSPACE="/workspace/eqbench-run"
HF_TOKEN="${HF_TOKEN:?HF_TOKEN が未設定です。export HF_TOKEN='hf_xxxx'}"
HF_USERNAME="${HF_USERNAME:-YUGOROU}"

PATCH_REPO="${PATCH_REPO:-${HF_USERNAME}/teememo-eq-bench-ja}"
SCRIPTS_REPO="${SCRIPTS_REPO:-${HF_USERNAME}/Test-2}"
HF_BASE="https://huggingface.co/datasets"

echo "[setup] EQ-Bench3 日本語版 評価環境セットアップ"
echo ""

# ── 依存関係 ──────────────────────────────────────────────────
echo "[setup] 依存関係のインストール..."
pip install -q requests python-dotenv tqdm numpy scipy trueskill 2>/dev/null || true
apt-get install -y tmux git 2>/dev/null || true

# ── EQ-Bench3 クローン ────────────────────────────────────────
mkdir -p "${WORKSPACE}"
if [ -d "${WORKSPACE}/eqbench3" ]; then
    echo "[setup] EQ-Bench3 既存を更新中..."
    git -C "${WORKSPACE}/eqbench3" pull --ff-only || true
else
    echo "[setup] EQ-Bench3 をクローン中..."
    git clone --depth=1 https://github.com/EQ-bench/eqbench3.git "${WORKSPACE}/eqbench3"
fi
echo "[setup] ✅ EQ-Bench3"

# ── [Fix-JA-D1] 全12日本語データファイルを差し替え ──────────
echo ""
echo "[setup] [Fix-JA-D1] 全12日本語データファイルをDL中..."
JA_COUNT=0
for pair in \
    "scenario_prompts_ja.txt|scenario_prompts.txt" \
    "scenario_notes_ja.txt|scenario_notes.txt" \
    "scenario_master_prompt_ja.txt|scenario_master_prompt.txt" \
    "scenario_master_prompt_message_drafting_ja.txt|scenario_master_prompt_message_drafting.txt" \
    "scenario_master_prompt_analysis_ja.txt|scenario_master_prompt_analysis.txt" \
    "rubric_scoring_prompt_ja.txt|rubric_scoring_prompt.txt" \
    "rubric_scoring_prompt_analysis_ja.txt|rubric_scoring_prompt_analysis.txt" \
    "rubric_scoring_criteria_ja.txt|rubric_scoring_criteria.txt" \
    "rubric_scoring_criteria_analysis_ja.txt|rubric_scoring_criteria_analysis.txt" \
    "pairwise_prompt_eqbench3_ja.txt|pairwise_prompt_eqbench3.txt" \
    "pairwise_prompt_eqbench3_analysis_ja.txt|pairwise_prompt_eqbench3_analysis.txt" \
    "debrief_prompt_ja.txt|debrief_prompt.txt"; do
    SRC="${pair%%|*}"
    DEST="${WORKSPACE}/eqbench3/data/${pair##*|}"
    [ -f "${DEST}" ] && [ ! -f "${DEST}.en.bak" ] && cp "${DEST}" "${DEST}.en.bak"
    if curl -sfL \
        -H "Authorization: Bearer ${HF_TOKEN}" \
        "${HF_BASE}/${PATCH_REPO}/resolve/main/data/${SRC}" \
        -o "${DEST}"; then
        echo "[setup]   ✅ ${pair##*|}"
        JA_COUNT=$((JA_COUNT + 1))
    else
        [ -f "${DEST}.en.bak" ] && cp "${DEST}.en.bak" "${DEST}"
        echo "[setup]   ⚠️  DL失敗 → 英語版: ${pair##*|}"
    fi
done
echo "[setup] 日本語データファイル: ${JA_COUNT}/12"

# ── [Fix-JA-D2] コードパッチ (benchmark.py / conversation.py) ─
echo ""
echo "[setup] [Fix-JA-D2] コードパッチを適用中..."
PATCH_DIR="${WORKSPACE}/eqbench-patch"
mkdir -p "${PATCH_DIR}"
PATCH_OK=0
for pf in patch_benchmark.py patch_conversation.py; do
    if curl -sfL \
        -H "Authorization: Bearer ${HF_TOKEN}" \
        "${HF_BASE}/${PATCH_REPO}/resolve/main/patch/${pf}" \
        -o "${PATCH_DIR}/${pf}"; then
        PATCH_OK=$((PATCH_OK + 1))
    else
        echo "[setup]   ⚠️  DL失敗: ${pf}"
    fi
done

if [ "${PATCH_OK}" -eq 2 ]; then
    cd "${WORKSPACE}/eqbench3"
    python3 "${PATCH_DIR}/patch_benchmark.py"
    python3 "${PATCH_DIR}/patch_conversation.py"
    echo "[setup]   ✅ コードパッチ完了"
else
    echo "[setup]   ❌ パッチ取得失敗。PATCH_REPO=${PATCH_REPO} を確認してください。"
fi
cd "${WORKSPACE}"

# ── serve スクリプトのDL ──────────────────────────────────────
echo ""
echo "[setup] serve スクリプトのDL..."
for script in serve_test.sh serve_judge.sh; do
    if curl -fL \
        -H "Authorization: Bearer ${HF_TOKEN}" \
        "${HF_BASE}/${SCRIPTS_REPO}/resolve/main/eqbench-ja-run/${script}" \
        -o "${WORKSPACE}/${script}"; then
        chmod +x "${WORKSPACE}/${script}"
        echo "[setup]   ✅ ${script}"
    else
        echo "[setup]   ❌ DL失敗: ${script}"
    fi
done

# ── .env ファイルの生成 ────────────────────────────────────────
echo "[setup] .env ファイルを生成中..."
cat > "${WORKSPACE}/eqbench3/.env" << ENVEOF
# EQ-Bench3 ローカルvLLM設定（自動生成）
# 受験者: TeenEmo LoRA (port 8000) / 採点者: Qwen3.5-35B-A3B (port 8001)

TEST_API_URL=http://localhost:8000/v1/chat/completions
TEST_API_KEY=dummy

JUDGE_API_URL=http://localhost:8001/v1/chat/completions
JUDGE_API_KEY=dummy

MAX_RETRIES=3
RETRY_DELAY=5
REQUEST_TIMEOUT=300
ENVEOF
echo "[setup]   ✅ .env 生成完了"

echo ""
echo "[setup] =============================="
echo "[setup] ✅ セットアップ完了"
echo "[setup] =============================="
echo ""
echo "【実行手順】"
echo ""
echo "# Step 1: TeenEmo サーバー起動（port 8000）"
echo "tmux new-session -d -s eq_run"
echo "tmux new-window -t eq_run -n test"
echo "tmux send-keys -t eq_run:test \"cd ${WORKSPACE} && export HF_TOKEN='\$HF_TOKEN' && ./serve_test.sh\" Enter"
echo ""
echo "# Step 2: 採点者サーバー起動（port 8001）"
echo "tmux new-window -t eq_run -n judge"
echo "tmux send-keys -t eq_run:judge \"cd ${WORKSPACE} && ./serve_judge.sh\" Enter"
echo ""
echo "# Step 3: 起動確認（Application startup complete. が出るまで待つ）"
echo "tmux capture-pane -t eq_run:test -p | tail -3"
echo "tmux capture-pane -t eq_run:judge -p | tail -3"
echo ""
echo "# Step 4: EQ-Bench3 実行"
echo "cd ${WORKSPACE}/eqbench3"
echo "python eqbench3.py \\"
echo "  --test-model teenemo-dpo \\"
echo "  --model-name TeenEmo-DPO \\"
echo "  --judge-model Qwen/Qwen3.5-35B-A3B \\"
echo "  --no-elo \\"
echo "  --save-interval 1 \\"
echo "  --iterations 1"
echo ""
echo "⚠️  OOMが発生した場合は順次実行モードを使用してください（README参照）"