style: pre-commit fixes

Author: pre-commit-ci[bot]

docs/paper.md

@@ -91,9 +91,9 @@ training, snow-covered area map generation, AOI simplification, canopy-aware
 post-processing of snow-covered area (SCA) time series, and validation against
 airborne reference snow surveys. Sample data and a pre-trained model are
 provided to demonstrate the library's functions. PlanetSCA requires users to
-have an account with Planet Data Explorer (https://www.planet.com/explorer/)
-and an API key to utilize its search and download capabilities. Limited access
-to free PlanetScope images is available at
+have an account with Planet Data Explorer (https://www.planet.com/explorer/) and
+an API key to utilize its search and download capabilities. Limited access to
+free PlanetScope images is available at
 https://www.planet.com/industries/education-and-research/.
 
 # Statement of Need
@@ -135,8 +135,8 @@ A persistent challenge in satellite-based snow detection is the obstruction of
 the sensor's view by forest canopy, which can cause snow-covered pixels beneath
 dense tree cover to be systematically misclassified as snow-free. PlanetSCA
 addresses this through a `post_process` module that provides canopy-aware
-spatial infilling of SCA predictions using a canopy height model (CHM).
-Forested pixels classified as snow-free are reclassified as snow-covered when a
+spatial infilling of SCA predictions using a canopy height model (CHM). Forested
+pixels classified as snow-free are reclassified as snow-covered when a
 sufficient fraction of their open-area neighbors are snow-covered, with
 configurable kernel radius and threshold. The module further provides temporal
 median filtering applied independently to forested and open pixels, nodata
@@ -145,9 +145,9 @@ day-since-disappearance (DSD) dates for characterizing snowmelt timing. A
 complementary `validate` module supports quantitative evaluation of PlanetSCA
 predictions against Airborne Snow Observatory (ASO) lidar-derived reference
 data, computing standard binary classification metrics (precision, recall,
-F1-score, accuracy, and Cohen's kappa) and producing a four-class
-snow-by-canopy validation raster for stratified assessment of model performance
-in forested versus open terrain.
+F1-score, accuracy, and Cohen's kappa) and producing a four-class snow-by-canopy
+validation raster for stratified assessment of model performance in forested
+versus open terrain.
 
 # Acknowledgements
 

src/planetsca/validate.py

@@ -508,15 +508,15 @@ def validation_tif_4class(
     )
     chm_binary = chm_tif.squeeze()
 
-    assert (
-        aso_binary.shape == chm_binary.shape
-    ), "ASO and CHM rasters must have the same shape"
-    assert (
-        aso_binary.rio.crs == chm_binary.rio.crs
-    ), "ASO and CHM rasters must have the same CRS"
-    assert (
-        aso_binary.rio.transform() == chm_binary.rio.transform()
-    ), "ASO and CHM rasters must have the same transform"
+    assert aso_binary.shape == chm_binary.shape, (
+        "ASO and CHM rasters must have the same shape"
+    )
+    assert aso_binary.rio.crs == chm_binary.rio.crs, (
+        "ASO and CHM rasters must have the same CRS"
+    )
+    assert aso_binary.rio.transform() == chm_binary.rio.transform(), (
+        "ASO and CHM rasters must have the same transform"
+    )
 
     combined = np.full(aso_binary.shape, np.nan, dtype=np.float32)
     combined[(aso_binary == 0) & (chm_binary == 1)] = 0