import pandas as pd

parametres_a_filtrer = {
    '1302.0', #pH
    '1382.0', #plomb
    '1386.0', #nickel
    '1387.0', #mercure
    '1388.0', #cadmium
    '1369.0', #arsenic
    '1339.0', #nitrites
    '1340.0', #nitrates
    '1398.0', #chlore
    '6276.0' #pesticides
    } 

def filtrage(params, date):
    input_file = f"data/processed/Table{date}_no_dup.csv"
    output_file = f"data/processed/Filtre{date}.csv"

    df = pd.read_csv(input_file, encoding='utf-8-sig')

    filtered_df = df[df['cdparametre'].str.strip().isin(params)]

    filtered_df.to_csv(output_file, index=False, encoding='utf-8-sig')

    print(f"Fichier filtré créé avec {len(filtered_df)} lignes : {output_file}")

dates = ["20" + str(i).zfill(2) for i in range(18, 25)]
processed_dates = []

for date in dates:
    print(f"Filtrage des paramètres pour {date}")
    filtrage(parametres_a_filtrer, date)