Take a string and a pandas series. Return which element in the series is the string most similar to, and the percentage of similarity

gistfile1.txt

from difflib import SequenceMatcher


def similitud(a, serie):
    """
      Calculates greatest similarity between a string (a) and a pandas series containing strings
      :return: number between 0 and 1 indicating similarity percentage, and element from the 
        pandas series it is most similar to
    """
    
    df_resultados = pd.DataFrame(columns=serie, index=['similitud'])
    A = str(a).strip()
    A = ''.join(filter(lambda x: x in printable, A))
    for col in df_resultados.columns:
        b = str(col).strip()
        b = ''.join(filter(lambda x: x in printable, b))
        df_resultados.at['similitud', col] = SequenceMatcher(None, A.upper(), b.upper()).ratio()
        
    df_resultados = df_resultados.astype('float32')
    col_max = df_resultados.idxmax(axis=1)[0]
    similitud_max = df_resultados.max(axis=1)[0]
    return similitud_max, col_max

I've found myself using this snippet often when standardizing fields imputed by humans, for instance in survey tables.