import requests
import zipfile
import io
import pandas as pd

def run_fao_test():
    from scripts.bigquery_config import get_bigquery_client
    from scripts.bigquery_helpers import load_to_bigquery, read_from_bigquery

    print("--- MEMULAI TEST LOAD FAO KE BIGQUERY ---")

    # 1. Extract
    url = "https://bulks-faostat.fao.org/production/Food_Security_Data_E_All_Data_(Normalized).zip"
    response = requests.get(url, timeout=120)
    with zipfile.ZipFile(io.BytesIO(response.content)) as z:
        csv_name = [f for f in z.namelist() if f.endswith('.csv')][0]
        df = pd.read_csv(z.open(csv_name), encoding='latin-1')

    # Ambil 5 baris teratas
    df_top5 = df.head(5)

    print("HASIL 5 DATA TERATAS:")
    print("====================================================")
    print(df_top5.to_string(index=False))
    print("====================================================")

    # 2. Load ke BigQuery
    client = get_bigquery_client()
    load_to_bigquery(client, df_top5, "raw_fao_test", layer="bronze", write_disposition="WRITE_TRUNCATE")
    print(f"Total data yang berhasil di-load: {len(df_top5)} baris.")

    # 3. Verify
    df_check = read_from_bigquery(client, "raw_fao_test", layer="bronze")
    print(f"Verify dari BigQuery: {len(df_check)} baris tersimpan.")

if __name__ == "__main__":
    run_fao_test()