def preprocessing(df, unix, diff):
df.drop(df[df['simple_journal'] == 'Refused'].index, inplace=True)
# df = df[df['shoppercountrycode'] != 'GB']
df['cvcresponsecode'] = df['cvcresponsecode'].map(lambda x: float(x))
df.loc[df['cvcresponsecode'] >= 3.0, 'cvcresponsecode'] = 3.0
df['cvcresponsecode'] = df['cvcresponsecode'].apply(lambda x: x if x <=2.0 else 3.0)
df['amount'] = df['amount'].map(lambda x: float(x))
df['card_id'] = df['card_id'].map(lambda x: x[4:])
df['ip_id'] = df['ip_id'].map(lambda x: x[2:])
df['mail_id'] = df['mail_id'].map(lambda x: x[5:])
df['country'] = df['shoppercountrycode'].apply(lambda x: x if x in ['MX', 'AU', 'GB'] else 'other')
df = df.drop(['shoppercountrycode'], axis=1)
if unix:
df['creationdate_unix'] = pd.DatetimeIndex(df['creationdate']).astype(np.int64) / 1000000000
df['creationdate_unix'].sub(df['creationdate_unix'].min(), axis=0)
df['creationdate_unix'] = df['creationdate_unix'].map(lambda x: math.log(x))
if diff:
df = time_diff(df)
df['AUD_currency'] = df[['currencycode', 'amount']].apply(lambda x: CurrencyConverter.convert_currency_from_AUD(x), axis=1)
df.dropna(axis=0, how='any', inplace=True)
df = df[~df.isin(['NaN', 'NaT', 'NA']).any(axis=1)]
return df
