rio-tiler 6.4.1__py3-none-any.whl → 6.4.3__py3-none-any.whl

rio_tiler/__init__.py

@@ -1,6 +1,6 @@
 """rio-tiler."""
 
-__version__ = "6.4.1"
+__version__ = "6.4.3"
 
 from . import (  # noqa
     colormap,

rio_tiler/io/rasterio.py

@@ -266,7 +266,7 @@ class Reader(BaseReader):
             "height": self.dataset.height,
             "overviews": self.dataset.overviews(1),
         }
-        if self.dataset.scales[0] and self.dataset.offsets[0]:
+        if self.dataset.scales[0] != 1.0 or self.dataset.offsets[0] != 0.0:
             meta.update(
                 {"scale": self.dataset.scales[0], "offset": self.dataset.offsets[0]}
             )

rio_tiler/utils.py

@@ -33,6 +33,27 @@ def _chunks(my_list: Sequence, chuck_size: int) -> Generator[Sequence, None, Non
         yield my_list[i : i + chuck_size]
 
 
+# Ref: https://stackoverflow.com/posts/73905572
+def _weighted_quantiles(
+    values: NDArray[numpy.floating],
+    weights: NDArray[numpy.floating],
+    quantiles: float = 0.5,
+) -> float:
+    i = numpy.argsort(values)
+    c = numpy.cumsum(weights[i])
+    return float(values[i[numpy.searchsorted(c, numpy.array(quantiles) * c[-1])]])
+
+
+# Ref: https://stackoverflow.com/questions/2413522
+def _weighted_stdev(
+    values: NDArray[numpy.floating],
+    weights: NDArray[numpy.floating],
+) -> float:
+    average = numpy.average(values, weights=weights)
+    variance = numpy.average((values - average) ** 2, weights=weights)
+    return float(math.sqrt(variance))
+
+
 def get_array_statistics(
     data: numpy.ma.MaskedArray,
     categorical: bool = False,
@@ -85,25 +106,31 @@ def get_array_statistics(
     percentiles = percentiles or [2, 98]
 
     if len(data.shape) < 3:
-        data = numpy.expand_dims(data, axis=0)
+        data = numpy.ma.expand_dims(data, axis=0)
 
     output: List[Dict[Any, Any]] = []
     percentiles_names = [f"percentile_{int(p)}" for p in percentiles]
 
+    if coverage is not None:
+        assert coverage.shape == (
+            data.shape[1],
+            data.shape[2],
+        ), f"Invalid shape ({coverage.shape}) for Coverage, expected {(data.shape[1], data.shape[2])}"
+
+    else:
+        coverage = numpy.ones((data.shape[1], data.shape[2]))
+
     # Avoid non masked nan/inf values
     numpy.ma.fix_invalid(data, copy=False)
 
     for b in range(data.shape[0]):
-        keys, counts = numpy.unique(data[b].compressed(), return_counts=True)
+        data_comp = data[b].compressed()
+
+        keys, counts = numpy.unique(data_comp, return_counts=True)
 
         valid_pixels = float(numpy.ma.count(data[b]))
         masked_pixels = float(numpy.ma.count_masked(data[b]))
         valid_percent = round((valid_pixels / data[b].size) * 100, 2)
-        info_px = {
-            "valid_pixels": valid_pixels,
-            "masked_pixels": masked_pixels,
-            "valid_percent": valid_percent,
-        }
 
         if categorical:
             out_dict = dict(zip(keys.tolist(), counts.tolist()))
@@ -115,28 +142,26 @@ def get_array_statistics(
                 h_keys,
             ]
         else:
-            h_counts, h_keys = numpy.histogram(data[b].compressed(), **kwargs)
+            h_counts, h_keys = numpy.histogram(data_comp, **kwargs)
             histogram = [h_counts.tolist(), h_keys.tolist()]
 
-        if valid_pixels:
-            percentiles_values = numpy.percentile(
-                data[b].compressed(), percentiles
-            ).tolist()
-        else:
-            percentiles_values = (numpy.nan,) * len(percentiles_names)
-
-        if coverage is not None:
-            assert coverage.shape == (
-                data.shape[1],
-                data.shape[2],
-            ), f"Invalid shape ({coverage.shape}) for Coverage, expected {(data.shape[1], data.shape[2])}"
-
-            array = data[b] * coverage
+        # Data coverage fractions
+        data_cov = data[b] * coverage
+        # Coverage Array + data mask
+        masked_coverage = numpy.ma.MaskedArray(coverage, mask=data_cov.mask)
 
+        if valid_pixels:
+            # TODO: when switching to numpy~=2.0
+            # percentiles_values = numpy.percentile(
+            #     data_comp, percentiles, weights=coverage.flatten()
+            # ).tolist()
+            percentiles_values = [
+                _weighted_quantiles(data_comp, masked_coverage.compressed(), pp / 100.0)
+                for pp in percentiles
+            ]
         else:
-            array = data[b]
+            percentiles_values = [numpy.nan] * len(percentiles_names)
 
-        count = array.count() if coverage is None else coverage.sum()
         if valid_pixels:
             majority = float(keys[counts.tolist().index(counts.max())].tolist())
             minority = float(keys[counts.tolist().index(counts.min())].tolist())
@@ -144,21 +169,40 @@ def get_array_statistics(
             majority = numpy.nan
             minority = numpy.nan
 
+        _count = masked_coverage.sum()
+        _sum = data_cov.sum()
+
         output.append(
             {
+                # Minimum value, not taking coverage fractions into account.
                 "min": float(data[b].min()),
+                # Maximum value, not taking coverage fractions into account.
                 "max": float(data[b].max()),
-                "mean": float(array.mean()),
-                "count": float(count),
-                "sum": float(array.sum()),
-                "std": float(array.std()),
-                "median": float(numpy.ma.median(array)),
+                # Mean value, weighted by the percent of each cell that is covered.
+                "mean": float(_sum / _count),
+                # Sum of all non-masked cell coverage fractions.
+                "count": float(_count),
+                # Sum of values, weighted by their coverage fractions.
+                "sum": float(_sum),
+                # Population standard deviation of cell values, taking into account coverage fraction.
+                "std": _weighted_stdev(data_comp, masked_coverage.compressed()),
+                # Median value of cells, weighted by the percent of each cell that is covered.
+                "median": _weighted_quantiles(data_comp, masked_coverage.compressed()),
+                # The value occupying the greatest number of cells.
                 "majority": majority,
+                # The value occupying the least number of cells.
                 "minority": minority,
+                # Unique values.
                 "unique": float(counts.size),
+                # quantiles
                 **dict(zip(percentiles_names, percentiles_values)),
                 "histogram": histogram,
-                **info_px,
+                # Number of non-masked cells, not taking coverage fractions into account.
+                "valid_pixels": valid_pixels,
+                # Number of masked cells, not taking coverage fractions into account.
+                "masked_pixels": masked_pixels,
+                # Percent of valid cells
+                "valid_percent": valid_percent,
             }
         )
 

{rio_tiler-6.4.1.dist-info → rio_tiler-6.4.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.3
 Name: rio-tiler
-Version: 6.4.1
+Version: 6.4.3
 Summary: User friendly Rasterio plugin to read raster datasets.
 Project-URL: Homepage, https://cogeotiff.github.io/rio-tiler/
 Project-URL: Documentation, https://cogeotiff.github.io/rio-tiler/

{rio_tiler-6.4.1.dist-info → rio_tiler-6.4.3.dist-info}/RECORD

@@ -1,4 +1,4 @@
-rio_tiler/__init__.py,sha256=HaxwIqy3HP3MuUeSf0jq4_QjmwhQTXmgjQLxxd9VXVE,192
+rio_tiler/__init__.py,sha256=rx1oh4GVoIrzSnsvVZYRSRTIPaB0IUT6P9mDoAbWRW8,192
 rio_tiler/colormap.py,sha256=hspVEbnvbRPGPEykWxV_PoqQpltbfSR7_g35mzd1d6s,9755
 rio_tiler/constants.py,sha256=55i-7JZDupTXZdLgxL03KsgM4lAzuGuIVP1zZKktzp0,426
 rio_tiler/errors.py,sha256=ag-wNr13JaQ4YcQzWWPFzjegeGdzJIMMrOrbb3coQ40,1733
@@ -10,7 +10,7 @@ rio_tiler/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 rio_tiler/reader.py,sha256=W08HPOfE48BrbsDZgHKBPhJrrZsOLz-r0wq9r27VnAs,21681
 rio_tiler/tasks.py,sha256=O4laTCr66XHJ-4REWg3SAaA1pre0OMG0NklbKZ2g9Fc,3181
 rio_tiler/types.py,sha256=508_E5R-nKrAo8TWB-6h7pF6SYwNM8kJPiHjQw-k-tA,1443
-rio_tiler/utils.py,sha256=zhfrhjAdaa2eJYVQwjzuy68m6Uj0bIYSaOVWK-dc4ek,24426
+rio_tiler/utils.py,sha256=WHAjsO1C-oW-cMb8Q50D6fCDvLji-eT2oxN9SUjzZxQ,26514
 rio_tiler/cmap_data/__init__.py,sha256=8FtVmfpTjXlvhxQ5QesN0UC1m_B3MuF3LbGbhMC5Rw4,1039
 rio_tiler/cmap_data/accent.npy,sha256=Qde1ldOoXghe4L05v1QbVvnMA1ldwNjKWPf5xCBbmI4,1152
 rio_tiler/cmap_data/accent_r.npy,sha256=ba-GWSMSPRAcm9CGzlXJeNG4ABbBHDCV602hAWV2idc,1152
@@ -225,7 +225,7 @@ rio_tiler/cmap_data/ylorrd.npy,sha256=9ImXljw40oe60w8uV4EMDPY4aFFVkGbyCBi6SlTX83
 rio_tiler/cmap_data/ylorrd_r.npy,sha256=K5uiHNHbLxV5SizyT09cSVAxldE-BW5GpOXxUp7UsTE,1152
 rio_tiler/io/__init__.py,sha256=_L4iILm6vSiJ14GEDDOvkuUHRtbWC9oqx6Bu8PxHhvA,270
 rio_tiler/io/base.py,sha256=r1p9yEnQc3dI9Ey-sg8q0jfD4K1l0T70kKVWUceoGG4,45260
-rio_tiler/io/rasterio.py,sha256=FzAMkP2xUNOvHfAcmOwilUFr6j3ao08l1Xm7ZpKX3kk,32171
+rio_tiler/io/rasterio.py,sha256=-JRm-KLUGXB-NQxYCvZPUXH2SvTfZ-lFietzNcvW4pM,32184
 rio_tiler/io/stac.py,sha256=soHDrWBKp7bgepwVe2dDBgVPp3iEAbTsx7BBHhtLDiE,10182
 rio_tiler/io/xarray.py,sha256=d8JO45Wv3Jvk0aSbUSUbfOEUzTTefdN8JkrUmDwZFo8,17785
 rio_tiler/mosaic/__init__.py,sha256=Yj6CKpnFl8PJhLSp-a55wo33hKZ8-6OOBJtWA1HZVy8,118
@@ -233,8 +233,8 @@ rio_tiler/mosaic/reader.py,sha256=_YBwTJwHvXOzKwpNpOmmh0F4yicyxgWo9vHyof3w_Do,96
 rio_tiler/mosaic/methods/__init__.py,sha256=tgkXM9skaTLXIm5QFoheOEznQXM97KGflcAWHfkrt1g,612
 rio_tiler/mosaic/methods/base.py,sha256=9YZJWVRwH5Fk9KO9q5CW52Q8Mf60tAJ21oM4ixEDXBo,1424
 rio_tiler/mosaic/methods/defaults.py,sha256=z34lna2wGXnAPwculjk_6hDrloqS8wzer68FFoIo7pg,6744
-rio_tiler-6.4.1.dist-info/METADATA,sha256=sYF6Ga3VhAdIM8-XbJGP986aVknaJggIRJ5sYjhRIWU,11869
-rio_tiler-6.4.1.dist-info/WHEEL,sha256=TJPnKdtrSue7xZ_AVGkp9YXcvDrobsjBds1du3Nx6dc,87
-rio_tiler-6.4.1.dist-info/licenses/AUTHORS.txt,sha256=FCVd4Tjg-8syl0ZugCunpXER8X2-XonW2ZfllyTnRvE,158
-rio_tiler-6.4.1.dist-info/licenses/LICENSE,sha256=vq8Tt4KoYQT9JxAjQ4yXMmmhFYRTsBRgrOj-ao-bC5o,1517
-rio_tiler-6.4.1.dist-info/RECORD,,
+rio_tiler-6.4.3.dist-info/METADATA,sha256=fjeMD-0DJ4uGXv0wKufK6f2C9TFMVbEJzoVDivgpwbU,11869
+rio_tiler-6.4.3.dist-info/WHEEL,sha256=bq9SyP5NxIRA9EpQgMCd-9RmPHWvbH-4lTDGwxgIR64,87
+rio_tiler-6.4.3.dist-info/licenses/AUTHORS.txt,sha256=FCVd4Tjg-8syl0ZugCunpXER8X2-XonW2ZfllyTnRvE,158
+rio_tiler-6.4.3.dist-info/licenses/LICENSE,sha256=vq8Tt4KoYQT9JxAjQ4yXMmmhFYRTsBRgrOj-ao-bC5o,1517
+rio_tiler-6.4.3.dist-info/RECORD,,

{rio_tiler-6.4.1.dist-info → rio_tiler-6.4.3.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: hatchling 1.21.1
+Generator: hatchling 1.22.3
 Root-Is-Purelib: true
 Tag: py3-none-any