gistfile1.py

dep_codes = {"ZA" : "971",
             "ZB" : "972",
             "ZC" : "973",
             "ZD" : "974",
             "ZM" : "976",
             "ZN" : "988",
             "ZP" : "987",
             "ZS" : "975",
             "ZW" : "986",
             "ZX" : "977_978",
             "ZZ" : "fr_etranger",
             "ZT" : "978",
             "ZY" : "977"
            }



def clean_elec(df):
    cols = df.columns
    keywords = ["code","libellé","nuance","abstention", "blanc", "nuls","nom","voix","commune","département","circonscription"]
    new_cols = []
    for col in cols:
        if any(x in col.lower() for x in keywords) and "%" not in col:
            new_cols.append(col)
    df = df[new_cols]
    df.columns  = [i.lower().replace(" du ","_").replace(" de la ","_").replace("é","e").replace(" et ","_").strip().replace(" ","_") for i in df.columns]


    if "code_departement" in df.columns:
        df["code_departement"] = df["code_departement"].replace(dep_codes)
        df["code_departement"] = df["code_departement"].astype(str).str.pad(2,"left","0")

    if "libelle_commune" in df.columns:

        df.loc[(df["code_departement"].astype(str) == "977_978") & (df["code_commune"].astype(str) == "701"), ["code_departement","libelle_departement"]] = ["977","Saint-Barthélémy"]
        df.loc[(df["code_departement"].astype(str) == "977_978") & (df["code_commune"].astype(str) == "801"), ["code_departement","libelle_departement"]] = ["978","Saint-Martin"]

        df["code_commune"] = df["code_commune"].astype(str).str.pad(3,"left","0")

    if "nom" in df.columns:
        name_present = True
    else:
        name_present = False

        if "code_b.vote" in df.columns:
            df = df.drop(columns=["libelle_commune"],axis=1)


    categories_candidats = list(pd.Series([i.split(".")[0] for i in df.columns if ".1" in i]).unique())

    keywords = ["blanc","nul","abstention"]

    max_cand = [int(re.search('.(\d+)',i)[1]) for i in df.columns if "." in i]
    max_cand = max(max_cand) + 1

    n = max_cand
    for col in df.columns:
        if any(x in col.lower() for x in keywords):
            for cat in categories_candidats:
                
                is_name = len([i for i in categories_candidats if "nom" in i]) > 0
                
                if "voix" in cat:
                    df["voix." + str(n)] = df[col]
                
                elif ("nuance" in cat and is_name) or "prenom" in cat:
                    df[cat + "." + str(n)] = np.nan
                else:
                    df[cat + "." + str(n)] = col


            n +=1

    for cat in categories_candidats:
        for col in df.columns:
            if cat == col:
                df = df.rename(columns={col : col + ".0"})

    max_cand = [int(re.search('.(\d+)',i)[1]) for i in df.columns if "." in i]
    max_cand = max(max_cand) + 1


    temp_list = []
    for cand in range(max_cand):
        cols = [i for i in df.columns if "." in i and int(re.search('.(\d+)',i)[1]) == cand]

        df_temp = df[cols].reset_index(drop=True)
        df_temp.columns = [i.split(".")[0] for i in df_temp.columns]
        temp_list.append(df_temp)

    geo_codes= [i for i in df.columns if (any(x in i.lower() for x in ["code","libelle",'departement','commune','circonscription']) and "nuance" not in i and "liste" not in i)]
    len_cand = len(temp_list)
    codgeo = pd.concat([df[geo_codes] for i in range(len_cand)])
    temp_df = pd.concat(temp_list)
    df = pd.concat([codgeo,temp_df],axis=1)

    if "libelle_circonscription" in df.columns:
        df = df.drop(["libelle_circonscription"],axis=1)


    df.columns = [i.replace(" ","_").replace("libelle_","").replace("_candidat","") for i in df.columns]
    df = df[df["voix"].notna()]
    
    
    df = df.drop(columns={"etendu_liste"},errors="ignore").rename(columns={ "nom_tête_de_liste" : "nom"})
    

    return df