def encode(self, xs: pd.Series):
if get_config('keep_original'):
self.mask = xs.isna()
if self.by is not None:
return xs.fillna(getattr(xs, self.by)())
else:
return xs.fillna(self.value)
def encode(self, xs:pd.Series)->pd.Series:
if self.date_range is None:
self.date_range = pd.date_range(
self.start, self.end, self.periods, self.freq, self.tz,
self.normalize, closed=self.closed
)
result = pd.Series(
np.zeros_like(xs, dtype=np.int64),
index=xs.index, name=xs.name
)
for i, (start, end) in enumerate(zip(self.date_range, self.date_range[1:])):
result[xs.between(start, end)] = i
if get_config('keep_original'):
self.original_xs = xs
return result
def encode(self, xs: pd.Series) -> pd.Series:
if self.upper is None:
upper = xs.quantile(self.percentile)
lower = xs.quantile(1.0 - self.percentile)
self.upper = upper
self.lower = lower
else:
upper = self.upper
lower = self.lower
upper_mask = xs > upper
lower_mask = xs < lower
if get_config('keep_original'):
self.lower_mask = lower_mask
self.lower_original_values = xs[self.lower_mask]
self.upper_mask = upper_mask
self.upper_original_values = xs[self.upper_mask]
xs = xs.mask(upper_mask, upper)
xs = xs.mask(lower_mask, lower)
return xs
def encode(self, df:pd.DataFrame):
mask = df.isna().any(axis=1)
self.original_dropped_index = df.index[mask]
if get_config('keep_original'):
self.original = df.loc[mask]
return df.dropna()
def decode(self, xs:pd.Series)->pd.Series:
if get_config('keep_original'):
return self.original_xs
else:
return pd.Series(self.date_range.take(xs))