#!/bin/bash
#SBATCH -J qwen3omni_check
#SBATCH -p gpu
#SBATCH -N 1
#SBATCH --gres=gpu:1
#SBATCH --cpus-per-task=4
#SBATCH --qos normal

#SBATCH -o /workspace/echoloc/Dataset/VStyle/sft/logs/qwen3omni_check_%j.out
#SBATCH -e /workspace/echoloc/Dataset/VStyle/sft/logs/qwen3omni_check_%j.err

set -euo pipefail

# ===== Conda =====
export CONDA_PREFIX=/workspace/miniconda3/envs/qwen3omni_thinker
export PATH=$CONDA_PREFIX/bin:$PATH
export LD_LIBRARY_PATH="${CONDA_PREFIX}/lib:${LD_LIBRARY_PATH:-}"

echo "===== ENV ====="
which python
python --version

echo $LD_LIBRARY_PATH

python << 'PYCODE'
import json

MASTER_JSONL = "/workspace/echoloc/datas/ds3_vstyle_controls_3_17/ds3_vstyle_controls_3_17.jsonl"

count = 0
with open(MASTER_JSONL, "r", encoding="utf-8") as f:
    for line in f:
        entry = json.loads(line)
        controls = entry.get("generated_control", {}).get("Control", [])

        if len(controls) > 1:
            count += 1
            print(f"\n--- Multi-Control Case #{count} ---")
            print(f"ID: {entry.get('instruct_id')}")
            print(f"Control Count: {len(controls)}")
            for i, c in enumerate(controls):
                print(f"  Segment {i+1} Text: {c.get('sample_text')}")

            if count >= 3: # 只看前3个例子
                break

PYCODE