reformatting using black / isort

batchglm/models/glm_beta/model.py

@@ -1,5 +1,6 @@
 import abc
 from typing import Any, Callable, Dict, Optional, Tuple, Union
+
 import dask
 
 try:

batchglm/models/glm_norm/model.py

@@ -1,5 +1,6 @@
 import abc
 from typing import Any, Callable, Dict, Optional, Tuple, Union
+
 import dask
 
 try:

batchglm/train/numpy/base_glm/estimator.py

@@ -165,9 +165,9 @@ def train(
                     # Reverse update by feature if update leads to worse loss:
                     ll_proposal = -self.model_container.ll_byfeature_j(j=idx_update)
                     idx_bad_step = idx_update[np.where(ll_proposal > ll_current[idx_update])[0]]
-                    
+
                     theta_scale_new = self.model_container.theta_scale.compute()
-                   
+
                     theta_scale_new[:, idx_bad_step] = theta_scale_new[:, idx_bad_step] - b_step[:, idx_bad_step]
                     self.model_container.theta_scale = theta_scale_new
                 else:
@@ -190,9 +190,9 @@ def train(
                     # Reverse update by feature if update leads to worse loss:
                     ll_proposal = -self.model_container.ll_byfeature_j(j=idx_update)
                     idx_bad_step = idx_update[np.where(ll_proposal > ll_current[idx_update])[0]]
-                   
+
                     theta_location_new = self.model_container.theta_location.compute()
-                   
+
                     theta_location_new[:, idx_bad_step] = theta_location_new[:, idx_bad_step] - a_step[:, idx_bad_step]
                     self.model_container.theta_location = theta_location_new
                 else:

batchglm/train/numpy/base_glm/modelContainer.py

@@ -1,5 +1,5 @@
 import abc
-from typing import Callable, Union, Any
+from typing import Any, Callable, Union
 
 import dask.array
 import numpy as np
@@ -257,10 +257,7 @@ def hessian_location_scale(self) -> Union[np.ndarray, dask.array.core.Array]:
         :return: (features x inferred param x inferred param)
         """
         w = self.hessian_weight_location_scale
-        return np.einsum(
-            "fob,oc->fbc",
-            np.einsum("ob,of->fob", self.xh_loc, w), self.xh_scale
-        )
+        return np.einsum("fob,oc->fbc", np.einsum("ob,of->fob", self.xh_loc, w), self.xh_scale)
 
     @property
     @abc.abstractmethod
@@ -358,8 +355,7 @@ def ll(self) -> Union[np.ndarray, dask.array.core.Array]:
     def ll_j(self, j) -> Union[np.ndarray, dask.array.core.Array]:
         pass
 
-
-    @property # type: ignore
+    @property  # type: ignore
     @dask_compute
     def ll_byfeature(self) -> np.ndarray:
         return np.sum(self.ll, axis=0)

batchglm/train/numpy/glm_nb/modelContainer.py

@@ -1,4 +1,4 @@
-from typing import Union, Callable
+from typing import Callable, Union
 
 import dask
 import numpy as np
@@ -8,7 +8,6 @@
 
 
 class ModelContainer(BaseModelContainer):
-
     @property
     def fim_weight(self):
         raise NotImplementedError("This method is currently unimplemented as it isn't used by any built-in procedures.")
@@ -54,7 +53,7 @@ def jac_weight(self):
     @property
     def jac_weight_j(self):
         raise NotImplementedError("This method is currently unimplemented as it isn't used by any built-in procedures.")
-    
+
     @property
     def jac_weight_scale(self) -> Union[np.ndarray, dask.array.core.Array]:
         """

tests/numpy/test_accuracy_extreme_values.py

@@ -15,9 +15,7 @@ class _TestAccuracyXtremeAll(TestAccuracy):
     Test whether numerical extremes throw error in initialisation or during first training steps.
     """
 
-    def _test_accuracy_extreme_values(
-        self, idx: Union[List[int], int, np.ndarray], val: float, noise_model: str
-    ):
+    def _test_accuracy_extreme_values(self, idx: Union[List[int], int, np.ndarray], val: float, noise_model: str):
         model = get_generated_model(noise_model=noise_model, num_conditions=2, num_batches=4, sparse=False, mode=None)
         model._x[:, idx] = val
         estimator = get_estimator(noise_model=noise_model, model=model, init_location="standard", init_scale="standard")
@@ -74,5 +72,6 @@ def test_beta(self) -> bool:
         # self._test_zero_variance(noise_model="beta")
         return True
 
+
 if __name__ == "__main__":
     unittest.main()

tests/numpy/utils.py

@@ -2,25 +2,27 @@
 
 import numpy as np
 
+from batchglm.models.base_glm import _ModelGLM
 from batchglm.models.glm_beta import Model as BetaModel
 from batchglm.models.glm_nb import Model as NBModel
 from batchglm.models.glm_norm import Model as NormModel
-from batchglm.models.base_glm import _ModelGLM
-# from batchglm.train.numpy.glm_beta import Estimator as BetaEstimator
-from batchglm.train.numpy.glm_nb import Estimator as NBEstimator
+
 # from batchglm.train.numpy.glm_norm import Estimator as NormEstimator
 from batchglm.train.numpy.base_glm import EstimatorGlm
 
+# from batchglm.train.numpy.glm_beta import Estimator as BetaEstimator
+from batchglm.train.numpy.glm_nb import Estimator as NBEstimator
+
 
 def get_estimator(noise_model: str, **kwargs) -> EstimatorGlm:
     if noise_model == "nb":
         return NBEstimator(**kwargs)
     elif noise_model == "norm":
         raise NotImplementedError("Norm Estimator is not yet implemented.")
-        #estimator = NormEstimator(**kwargs)
+        # estimator = NormEstimator(**kwargs)
     elif noise_model == "beta":
         raise NotImplementedError("Beta Estimator is not yet implemented.")
-        #estimator = BetaEstimator(**kwargs)
+        # estimator = BetaEstimator(**kwargs)
     raise ValueError(f"Noise model {noise_model} not recognized.")
 
 
@@ -36,7 +38,9 @@ def get_model(noise_model: str) -> _ModelGLM:
     raise ValueError(f"Noise model {noise_model} not recognized.")
 
 
-def get_generated_model(noise_model: str, num_conditions: int, num_batches: int, sparse: bool, mode: Optional[str] = None) -> _ModelGLM:
+def get_generated_model(
+    noise_model: str, num_conditions: int, num_batches: int, sparse: bool, mode: Optional[str] = None
+) -> _ModelGLM:
     model = get_model(noise_model=noise_model)
 
     def random_uniform(low: float, high: float):