"""
GitHub Actions 用: 从 CFTC 下载 COT 持仓报告
"""
import pandas as pd
import os, io, zipfile, requests

OUTPUT = 'data/cloud'
os.makedirs(OUTPUT, exist_ok=True)

print("Fetching CFTC COT reports...")

# 当年 + 去年数据
years = [2025, 2026]
all_data = []

for year in years:
    url = f"https://www.cftc.gov/files/dea/history/fut_disagg_xls_{year}.zip"
    try:
        resp = requests.get(url, timeout=60)
        if resp.status_code == 200:
            z = zipfile.ZipFile(io.BytesIO(resp.content))
            for fname in z.namelist():
                if fname.endswith('.xls') or fname.endswith('.xlsx'):
                    df = pd.read_excel(z.open(fname))
                    # 筛选原油相关
                    oil = df[df['Market_and_Exchange_Names'].str.contains(
                        'CRUDE OIL|PETROLEUM|BRENT|WTI', case=False, na=False)]
                    if len(oil) > 0:
                        all_data.append(oil)
                        print(f"  ✓ {year}: {len(oil)} rows from {fname}")
    except Exception as e:
        print(f"  ✗ {year}: {str(e)[:50]}")

if all_data:
    df_all = pd.concat(all_data, ignore_index=True)
    outpath = os.path.join(OUTPUT, 'cftc_positioning.csv')
    df_all.to_csv(outpath, index=False)
    print(f"✓ Saved {len(df_all)} rows to {outpath}")
else:
    print("✗ No CFTC data fetched")