Added testing instructions (#31)

* renaming input_datas to input_data

* added testing instructions

README.md

@@ -61,13 +61,18 @@ Metrics from `tf.keras.metrics` and other metrics that are already implemented v
 - [ ] CategoricalCrossEntropy class
 - [ ] SparseCategoricalCrossentropy class
 
-## Local Development
+## Local Development and Testing
 
 To quickly get started with local development, run:
 ```python
 make develop
 ```
 
+To test, run:
+```python
+python3 -m pytest
+```
+
 ### Pre-commit hooks
 
 To run pre-commit against all files:

requirements.txt

@@ -3,3 +3,5 @@ black>=19.10b0
 flake8>=3.8.4
 isort>=5.7.0
 pre-commit>=2.10.1
+scikit-learn>=0.24.1
+pytest>=6.1.1

tests/test_classification_metrics.py

@@ -1,11 +1,12 @@
 import numpy as np
 import pytest
 import torch
-from input_datas import binary_prob_inputs, binary_raw_inputs
 from numpy.testing import assert_allclose
 from sklearn.metrics import accuracy_score, f1_score, precision_score
 from torch_metrics.classification import Accuracy, F1Score, Precision
 
+from input_data import binary_prob_inputs, binary_raw_inputs
+
 torch.manual_seed(42)
 
 

tests/test_regression_metrics.py

@@ -1,13 +1,16 @@
 import numpy as np
 import pytest
 import torch
-from input_datas import normal_regression_inputs, uniform_regression_inputs
 from numpy.testing import assert_allclose
-from sklearn.metrics import (mean_absolute_error, mean_squared_error,
-                             mean_squared_log_error)
-from torch_metrics.regression import (MeanAbsoluteError, MeanSquaredError,
-                                      MeanSquaredLogarithmicError,
-                                      RootMeanSquaredError)
+from sklearn.metrics import mean_absolute_error, mean_squared_error, mean_squared_log_error
+from torch_metrics.regression import (
+    MeanAbsoluteError,
+    MeanSquaredError,
+    MeanSquaredLogarithmicError,
+    RootMeanSquaredError,
+)
+
+from input_data import normal_regression_inputs, uniform_regression_inputs
 
 torch.manual_seed(42)
 
@@ -45,10 +48,7 @@ def test_mse(y_true, y_pred):
 
 
 @pytest.mark.parametrize(
-    "y_true, y_pred",
-    [
-        (uniform_regression_inputs.target, uniform_regression_inputs.preds),
-    ],
+    "y_true, y_pred", [(uniform_regression_inputs.target, uniform_regression_inputs.preds),],
 )
 def test_msle(y_true, y_pred):
     sk_preds = y_pred.numpy()