script/dowdload_df.sh

#!/bin/bash
#SBATCH -N 1
#SBATCH -p RM-shared
#SBATCH -t 12:00:00  # Increased time limit for full download (~150 GB total)
#SBATCH --ntasks-per-node=1
#SBATCH -J zenodo_dl
#SBATCH --output %x_%j.out
#SBATCH --error %x_%j.err
#SBATCH --mail-type=ALL
#SBATCH --mail-user=shan1
#SBATCH --signal INT@60


# =================CONFIGURATION=================
DATASETS=(
    "LibriSeVoc Dataset|15127251"
    "WaveFake|5642694"
    "Codecfake dataset - training set (part 1 of 3)|13838106"
    "Codecfake dataset - training set (part 2 of 3)|13841652"
    "Codecfake dataset - training set (part 3 of 3)|13853860"
    "Codecfake dataset - development set|13841216"
    "Codecfake dataset - test set (part 1 of 2)|13838823"
    "Codecfake dataset - test set (part 2 of 2)|11125029"
    "The train and dev sets of ADD 2022|12188127"
    "The adaption sets of ADD 2022|12188083"
    "The track 1 evaluation dataset of ADD 2022|10843991"
    "The track 2 evaluation dataset of ADD 2022|12187997"
    "The track 3.2 round 1 evaluation dataset of ADD 2022|12188035"
    "The track 3.2 round 2 evaluation dataset of ADD 2022|12188035"
    "ADD 2023 Challenge Track 1.1 Evaluation Dataset|12145773"
    "ADD 2023 Challenge Track 1.2 Training/Development Dataset|12151404"
    "ADD 2023 Challenge Track 1.2 Evaluation Dataset (Round 1)|12175884"
    "ADD 2023 Challenge Track 1.2 Evaluation Dataset (Round 2)|12176326"
    "ADD 2023 Challenge Track 2 Training/Development Dataset|12176530"
    "ADD 2023 Challenge Track 2 Evaluation Dataset|12176904"
    "ADD 2023 Challenge Track 3 Training/Development Dataset|12179632"
    "ADD 2023 Challenge Track 3 Evaluation Dataset|12179884"
)

# 1. 获取当前的任务 ID (如果是在本地运行，默认为 0)
TASK_ID=${SLURM_ARRAY_TASK_ID:-0}

# 为了避免多个任务同时写一个日志文件造成冲突，每个任务使用独立的日志
# 或者依靠 Slurm 的 .out 文件查看输出
LOG_FILE="download_log_${TASK_ID}.txt"

# 2. 检查 ID 是否越界
if [ -z "${DATASETS[$TASK_ID]}" ]; then
    echo "Error: Task ID $TASK_ID is out of bounds."
    exit 1
fi

# 3. 获取特定的一个条目
entry="${DATASETS[$TASK_ID]}"

# ===============================================

# 颜色设置
GREEN='\033[0;32m'
RED='\033[0;31m'
NC='\033[0m'

# 开始处理单个条目
RAW_NAME="${entry%%|*}"
ID="${entry##*|}"
DIR_NAME=$(echo "$RAW_NAME" | sed 's/[^a-zA-Z0-9]/_/g' | sed 's/__*/_/g' | sed 's/^_//;s/_$//')

echo -e "Job ID: $TASK_ID | Processing: $RAW_NAME (Zenodo ID: $ID)"
echo -e "Target Directory: $DIR_NAME"

mkdir -p "$DIR_NAME"

# 调用 API
echo "Fetching file list..."
API_RESPONSE=$(curl -s "https://zenodo.org/api/records/$ID")

if [[ -z "$API_RESPONSE" ]]; then
    echo -e "${RED}Error: Failed to fetch API for ID $ID${NC}"
    exit 1
fi

# 解析 JSON
FILES_TO_DOWNLOAD=$(echo "$API_RESPONSE" | python3 -c "
import sys, json
try:
    data = json.load(sys.stdin)
    if 'files' in data:
        for f in data['files']:
            url = f['links']['self'] if 'self' in f['links'] else f['links']['download']
            print(f\"{f['key']}|{url}\")
    else:
        print('ERROR_NO_FILES')
except Exception:
    print('ERROR_PARSING')
")

if [[ "$FILES_TO_DOWNLOAD" == "ERROR_NO_FILES" ]]; then
    echo -e "${RED}No public files found (Check Access Rights).${NC}"
    exit 0
fi

# 下载文件
while IFS='|' read -r FILENAME URL; do
    if [[ -z "$FILENAME" ]]; then continue; fi
    
    FILE_PATH="$DIR_NAME/$FILENAME"
    echo -e "Downloading: ${GREEN}$FILENAME${NC} ..."
    
    # wget -q 减少日志体积，只输出关键错误
    wget -c "$URL" -O "$FILE_PATH" --no-check-certificate --progress=dot:giga
    
    if [ $? -eq 0 ]; then
        echo -e "${GREEN}Success: $FILENAME${NC}"
        echo "$(date) - [OK] $FILENAME" >> "$LOG_FILE"
    else
        echo -e "${RED}Failed: $FILENAME${NC}"
        echo "$(date) - [FAIL] $FILENAME" >> "$LOG_FILE"
    fi

done <<< "$FILES_TO_DOWNLOAD"

echo "Task $TASK_ID Finished."