apply Black 2024 style in fbcode (15/17)

Summary:
Formats the covered files with pyfmt.

paintitblack

Reviewed By: zertosh

Differential Revision: D54470857

fbshipit-source-id: 6e68f9cec670e75777c696a5a47731cedab782c6

kats/data/utils.py

@@ -51,13 +51,11 @@ def load_data(file_name: str, reset_columns: bool = False) -> pd.DataFrame:
 
 
 @overload
-def load_air_passengers(return_ts: Literal[True]) -> TimeSeriesData:
-    ...
+def load_air_passengers(return_ts: Literal[True]) -> TimeSeriesData: ...
 
 
 @overload
-def load_air_passengers(return_ts: Literal[False] = ...) -> pd.DataFrame:
-    ...
+def load_air_passengers(return_ts: Literal[False] = ...) -> pd.DataFrame: ...
 
 
 def load_air_passengers(return_ts: bool = True) -> Union[pd.DataFrame, TimeSeriesData]:

kats/detectors/cusum_detection.py

@@ -1224,10 +1224,12 @@ def detector_(self, **kwargs: Any) -> List[List[CUSUMChangePoint]]:
                 if not list(change_meta_["changepoint"]):
                     continue
                 change_meta = {
-                    k: change_meta_[k][col_idx]
-                    if isinstance(change_meta_[k], np.ndarray)
-                    or isinstance(change_meta_[k], list)
-                    else change_meta_[k]
+                    k: (
+                        change_meta_[k][col_idx]
+                        if isinstance(change_meta_[k], np.ndarray)
+                        or isinstance(change_meta_[k], list)
+                        else change_meta_[k]
+                    )
                     for k in change_meta_
                 }
                 change_meta["llr"] = llr = self._get_llr(

kats/detectors/cusum_model.py

@@ -1276,9 +1276,11 @@ def _fit(
             )
 
             cps = [
-                sorted(x, key=lambda x: x.start_time)[0]
-                if x and alert_set_on_mask[i]
-                else None
+                (
+                    sorted(x, key=lambda x: x.start_time)[0]
+                    if x and alert_set_on_mask[i]
+                    else None
+                )
                 for i, x in enumerate(changepoints)
             ]
 

kats/detectors/detector_consts.py

@@ -260,10 +260,12 @@ def stat_sig(self) -> Union[bool, ArrayLike]:
         if self.num_series > 1:
             return np.array(
                 [
-                    False
-                    if cast(np.ndarray, self.upper)[i] > 1.0
-                    and cast(np.ndarray, self.lower)[i] < 1
-                    else True
+                    (
+                        False
+                        if cast(np.ndarray, self.upper)[i] > 1.0
+                        and cast(np.ndarray, self.lower)[i] < 1
+                        else True
+                    )
                     for i in range(self.current.num_series)
                 ]
             )
@@ -649,11 +651,13 @@ def get_last_n(self, N: int) -> AnomalyResponse:
 
         return AnomalyResponse(
             scores=self.scores[-N:],
-            confidence_band=None
-            if cb is None
-            else ConfidenceBand(
-                upper=cb.upper[-N:],
-                lower=cb.lower[-N:],
+            confidence_band=(
+                None
+                if cb is None
+                else ConfidenceBand(
+                    upper=cb.upper[-N:],
+                    lower=cb.lower[-N:],
+                )
             ),
             predicted_ts=None if pts is None else pts[-N:],
             anomaly_magnitude_ts=self.anomaly_magnitude_ts[-N:],

kats/detectors/distribution_distance_model.py

@@ -260,10 +260,12 @@ def fit_predict(
             window=str(self.window_size_sec) + "s",
             closed="both",
         ).agg(
-            lambda rows: rows[0]
-            if (rows.index[-1] - rows.index[0]).total_seconds()
-            > 0.9 * self.window_size_sec  # tolerance
-            else np.nan
+            lambda rows: (
+                rows[0]
+                if (rows.index[-1] - rows.index[0]).total_seconds()
+                > 0.9 * self.window_size_sec  # tolerance
+                else np.nan
+            )
         )
 
         # exclude the beginning part of NANs

kats/detectors/meta_learning/synth_metadata_reader.py

@@ -70,9 +70,11 @@ def get_metadata(self, algorithm_name: str) -> Dict[str, pd.DataFrame]:
                     .map(lambda kv: kv[a][0])
                     .map(
                         lambda kv: {
-                            k: v
-                            if k not in self.PARAMS_TO_SCALE_DOWN
-                            else v / SynthMetadataReader.NUM_SECS_IN_DAY
+                            k: (
+                                v
+                                if k not in self.PARAMS_TO_SCALE_DOWN
+                                else v / SynthMetadataReader.NUM_SECS_IN_DAY
+                            )
                             for k, v in kv.items()
                         }
                     )

kats/detectors/outlier.py

@@ -62,7 +62,7 @@ def __init__(
 
     def __clean_ts__(
         self,
-        original: Union[pd.Series, pd.DataFrame]
+        original: Union[pd.Series, pd.DataFrame],
         # pyre-fixme[11]: Annotation `Timestamp` is not defined as a type.
     ) -> Tuple[List[int], List[float], List[pd.Timestamp]]:
         """

kats/detectors/prophet_detector.py

@@ -41,6 +41,8 @@
 import sys
 
 NOT_SUPPRESS_PROPHET_FIT_LOGS_VAR_NAME = "NOT_SUPPRESS_PROPHET_FIT_LOGS"
+
+
 # this is a bug in prophet which was discussed in open source thread
 # issues was also suggested
 # details https://github.com/facebook/prophet/issues/223#issuecomment-326455744

kats/detectors/stat_sig_detector.py

@@ -922,7 +922,6 @@ def predict(
 
 
 class MultiStatSigDetectorModel(StatSigDetectorModel):
-
     """
     MultiStatSigDetectorModel is a multivariate version of the StatSigDetector.  It applies a univariate
     t-test to each of the components of the multivariate time series to see if the means between the control

kats/metrics/metrics.py

@@ -24,6 +24,7 @@
 # from numpy.typing import ArrayLike
 ArrayLike = Union[np.ndarray, Sequence[float]]
 
+
 # Type aliases
 #
 # Most metrics have the shape:
@@ -47,13 +48,13 @@ def __call__(
         y_true: ArrayLike,
         y_pred: ArrayLike,
         sample_weight: Optional[ArrayLike] = ...,
-    ) -> np.ndarray:
-        ...  # pragma: no cover
+    ) -> np.ndarray: ...  # pragma: no cover
 
 
 class Metric(Protocol):
-    def __call__(self, y_true: ArrayLike, y_pred: ArrayLike) -> float:
-        ...  # pragma: no cover
+    def __call__(
+        self, y_true: ArrayLike, y_pred: ArrayLike
+    ) -> float: ...  # pragma: no cover
 
 
 class WeightedMetric(Protocol):
@@ -62,8 +63,7 @@ def __call__(
         y_true: ArrayLike,
         y_pred: ArrayLike,
         sample_weight: Optional[ArrayLike] = ...,
-    ) -> float:
-        ...  # pragma: no cover
+    ) -> float: ...  # pragma: no cover
 
 
 class MultiOutputMetric(Protocol):
@@ -73,8 +73,7 @@ def __call__(
         y_pred: ArrayLike,
         sample_weight: Optional[ArrayLike] = ...,
         multioutput: Union[str, ArrayLike] = ...,
-    ) -> float:
-        ...  # pragma: no cover
+    ) -> float: ...  # pragma: no cover
 
 
 class ThresholdMetric(Protocol):
@@ -83,8 +82,7 @@ def __call__(
         y_true: ArrayLike,
         y_pred: ArrayLike,
         threshold: float,
-    ) -> float:
-        ...  # pragma: no cover
+    ) -> float: ...  # pragma: no cover
 
 
 class MultiThresholdMetric(Protocol):
@@ -93,8 +91,7 @@ def __call__(
         y_true: ArrayLike,
         y_pred: ArrayLike,
         threshold: ArrayLike,
-    ) -> np.ndarray:
-        ...  # pragma: no cover
+    ) -> np.ndarray: ...  # pragma: no cover
 
 
 KatsMetric = Union[

kats/models/ensemble/kats_ensemble.py

@@ -747,7 +747,7 @@ def _fit_single(
         data: TimeSeriesData,
         # pyre-fixme[24]: Generic type `Callable` expects 2 type parameters.
         model_func: Callable,
-        model_param: Params
+        model_param: Params,
         # pyre-fixme[24]: Generic type `Model` expects 1 type parameter.
     ) -> Model:
         """Private method to fit individual model

kats/models/globalmodel/data_processor.py

@@ -263,12 +263,12 @@ def __init__(
         if params.model_type == "rnn" and params.seasonality > 1:
             init_seasonality = self._get_seasonality(train_x, params.seasonality)
             # bound initial seasonalities
-            init_seasonality[
-                init_seasonality < params.init_seasonality[0]
-            ] = params.init_seasonality[0]
-            init_seasonality[
-                init_seasonality > params.init_seasonality[1]
-            ] = params.init_seasonality[1]
+            init_seasonality[init_seasonality < params.init_seasonality[0]] = (
+                params.init_seasonality[0]
+            )
+            init_seasonality[init_seasonality > params.init_seasonality[1]] = (
+                params.init_seasonality[1]
+            )
             # pyre-fixme[4]: Attribute must be annotated.
             self.init_seasonality = torch.tensor(init_seasonality, dtype=tdtype)
         else:
@@ -451,7 +451,6 @@ def _get_array(
         Optional[np.ndarray],
         Optional[np.ndarray],
     ]:
-
         """
 
         Helper function for transforming TS to arrays, including truncating/padding values,

kats/models/globalmodel/model.py

@@ -545,9 +545,7 @@ def _format_fcst(
         for i, idx in enumerate(ids):
 
             df = pd.DataFrame(
-                fcst[i].transpose()[
-                    :steps,
-                ],
+                fcst[i].transpose()[:steps,],
                 columns=cols,
             )
             df["time"] = pd.date_range(
@@ -651,12 +649,12 @@ def save_model(self, file_name: str) -> None:
         info = {
             "gmparam_string": self.params.to_string(),
             "state_dict": self.rnn.state_dict() if self.rnn is not None else None,
-            "encoder_state_dict": self.encoder.state_dict()
-            if self.encoder is not None
-            else None,
-            "decoder_state_dict": self.decoder.state_dict()
-            if self.decoder is not None
-            else None,
+            "encoder_state_dict": (
+                self.encoder.state_dict() if self.encoder is not None else None
+            ),
+            "decoder_state_dict": (
+                self.decoder.state_dict() if self.decoder is not None else None
+            ),
         }
         with open(file_name, "wb") as f:
             joblib.dump(info, f)
@@ -1099,7 +1097,11 @@ def _single_pass_s2s(
                 cur_step + 1
             )
 
-            (x_t, anchor_level, x_lt,) = self._process_s2s(
+            (
+                x_t,
+                anchor_level,
+                x_lt,
+            ) = self._process_s2s(
                 prev_idx, cur_idx, batch.x, x_lt, period, params.input_window
             )
 

kats/models/globalmodel/stdmodel.py

@@ -195,7 +195,6 @@ def _deseasonal(
     def _predict_seasonality(
         self, steps: int, tsd_model: Union[ProphetModel, np.ndarray]
     ) -> np.ndarray:
-
         """Predict the future seasonality.
 
         Args:

kats/models/globalmodel/utils.py

@@ -73,7 +73,7 @@ def get_filters(isna_idx, seasonality) -> np.ndarray:
         else:
             i += 1
     filters = np.array([True] * n)
-    for (i, j) in flips:
+    for i, j in flips:
         filters[i:j] = False
     return filters
 
@@ -188,19 +188,23 @@ def split(
         split_data = [
             (
                 {t: train_TSs[t] for t in keys[~index[i]]},
-                {t: valid_TSs[t] for t in keys[~index[i]]}
-                if valid_TSs is not None
-                else None,
+                (
+                    {t: valid_TSs[t] for t in keys[~index[i]]}
+                    if valid_TSs is not None
+                    else None
+                ),
             )
             for i in range(splits)
         ]
     else:
         split_data = [
             (
                 {t: train_TSs[t] for t in keys[index[i]]},
-                {t: valid_TSs[t] for t in keys[index[i]]}
-                if valid_TSs is not None
-                else None,
+                (
+                    {t: valid_TSs[t] for t in keys[index[i]]}
+                    if valid_TSs is not None
+                    else None
+                ),
             )
             for i in range(splits)
         ]

kats/models/metalearner/get_metadata.py

@@ -220,7 +220,7 @@ def _tune_single(
         # pyre-fixme[24]: Generic type `Callable` expects 2 type parameters.
         single_model: Callable,
         # pyre-fixme[24]: Generic type `Callable` expects 2 type parameters.
-        single_params: Callable
+        single_params: Callable,
         # pyre-fixme[24]: Generic type `dict` expects 2 type parameters, use
         #  `typing.Dict` to avoid runtime subscripting errors.
     ) -> Tuple[Dict, float]:

kats/models/metalearner/metalearner_hpt.py

@@ -456,9 +456,7 @@ def build_network(
         print("Multi-task neural network structure:")
         print(self.model)
 
-    def _prepare_data(
-        self, val_size: float
-    ) -> Tuple[
+    def _prepare_data(self, val_size: float) -> Tuple[
         torch.FloatTensor,
         Optional[torch.LongTensor],
         Optional[torch.FloatTensor],

kats/models/ml_ar.py

@@ -934,9 +934,9 @@ def _merge_past_and_future_reg(
                 num_rows_dat = norm_in_data[target_var].shape[0]
                 num_cols_dat = norm_in_data[target_var].shape[1]
 
-                full_mat[
-                    tv_idx : (tv_idx + num_rows_dat), 0:num_cols_dat
-                ] = norm_in_data[target_var]
+                full_mat[tv_idx : (tv_idx + num_rows_dat), 0:num_cols_dat] = (
+                    norm_in_data[target_var]
+                )
 
                 full_mat[
                     tv_idx : (tv_idx + num_rows_dat),