Subset an empty GenomicRangesList (#56).

setup.cfg

@@ -50,7 +50,7 @@ python_requires = >=3.8
 install_requires =
     importlib-metadata; python_version<"3.8"
     biocframe>=0.5.3,<0.6.0
-    iranges[optional]>=0.2.0,<0.3.0
+    iranges[optional]>=0.2.2,<0.3.0
     biocutils>=0.1.3,<0.2.0
     numpy
 

src/genomicranges/GenomicRangesList.py

@@ -762,8 +762,7 @@ def __getitem__(
                 If ``args`` is not a supported slice argument.
 
         Returns:
-            The genomic element as a ``GenomicRanges`` object
-            or a new ``GenomicRangesList`` of the slice.
+            A new ``GenomicRangesList`` of the slice.
         """
         if isinstance(args, int):
             return self._ranges[args]
@@ -772,49 +771,44 @@ def __getitem__(
                 _idx = self.names.map(args)
                 return self._ranges[_idx]
         else:
-            new_ranges = None
-            new_range_lengths = None
-            new_names = None
-            new_mcols = None
-            new_metadata = self.metadata
-
-            if isinstance(args, tuple):
-                # TODO: should figure out what to do with the second dimension later.
-                if len(args) >= 1:
-                    args = args[0]
-
-            if isinstance(args, list):
-                if ut.is_list_of_type(args, bool):
-                    if len(args) != len(self):
+            idx, _ = ut.normalize_subscript(args, len(self), self._names)
+
+            if isinstance(idx, list):
+                if ut.is_list_of_type(idx, bool):
+                    if len(idx) != len(self):
                         raise ValueError(
                             "`indices` is a boolean vector, length should match the size of the data."
                         )
 
-                    args = [i for i in range(len(args)) if args[i] is True]
+                    idx = [i for i in range(len(idx)) if idx[i] is True]
 
-                new_ranges = [self.ranges[i] for i in args]
-                new_range_lengths = [self._range_lengths[i] for i in args]
-                if self.names is not None:
-                    new_names = [self.names[i] for i in args]
+                new_ranges = [self.ranges[i] for i in idx]
+                new_range_lengths = [self._range_lengths[i] for i in idx]
 
-                if self.mcols is not None:
-                    new_mcols = self.mcols[args, :]
-            elif isinstance(args, slice):
-                new_ranges = self.ranges[args]
-                new_range_lengths = self._range_lengths[args]
+                new_names = None
                 if self.names is not None:
-                    new_names = self.names[args]
+                    new_names = [self.names[i] for i in idx]
 
+                new_mcols = None
                 if self.mcols is not None:
-                    new_mcols = self.mcols[args, :]
-            else:
-                raise TypeError("Arguments to slice is not a list of supported types.")
+                    new_mcols = self.mcols[idx, :]
 
-            return GenomicRangesList(
-                new_ranges, new_range_lengths, new_names, new_mcols, new_metadata
-            )
+                return GenomicRangesList(
+                    new_ranges, new_range_lengths, new_names, new_mcols, self._metadata
+                )
+            elif isinstance(idx, (slice, range)):
+                if isinstance(idx, range):
+                    idx = slice(idx.start, idx.stop, idx.step)
+
+                return GenomicRangesList(
+                    self._ranges[idx],
+                    self._range_lengths[idx],
+                    self._names[idx] if self._names is not None else self._names,
+                    self._mcols[idx, :],
+                    self._metadata,
+                )
 
-        raise TypeError("Arguments to slice is not supported.")
+            raise TypeError("Arguments to subset `GenomicRangesList` is not supported.")
 
     ##########################
     ######>> empty <<#########

tests/test_gr_methods_basic.py

@@ -74,6 +74,16 @@ def test_slices():
     assert subset_gr.seqnames == ["chr1", "chr3", "chr3"]
 
 
+def test_gr_empty_subset():
+    gre = GenomicRanges.empty()
+
+    assert gre is not None
+    assert isinstance(gre, GenomicRanges)
+    assert len(gre) == 0
+
+    subset = gre[0:10]
+
+
 def test_export():
     df = gr.to_pandas()
 

tests/test_grl_methods.py

@@ -37,12 +37,37 @@ def test_is_empty_slice():
 
     assert grl.is_empty() == False
 
-    sgrl = grl[0:1, :]
+    sgrl = grl[0:1]
     assert sgrl is not None
     assert isinstance(sgrl, GenomicRangesList)
     assert len(sgrl) == 1
 
 
+def test_slice_by_name():
+    grl = GenomicRangesList(ranges=[a, b], names=["a", "b"])
+
+    assert grl.is_empty() == False
+
+    sgrl = grl[["a"]]
+    assert sgrl is not None
+    assert isinstance(sgrl, GenomicRangesList)
+    assert len(sgrl) == 1
+
+
+def test_slice_by_bool():
+    grl = GenomicRangesList(ranges=[a, b], names=["a", "b"])
+
+    assert grl.is_empty() == False
+
+    sgrl = grl[[True, False]]
+    assert sgrl is not None
+    assert isinstance(sgrl, GenomicRangesList)
+    assert len(sgrl) == 1
+
+    with pytest.raises(Exception):
+        grl[[False]]
+
+
 def test_is_empty_True():
     grl = GenomicRangesList(GenomicRanges.empty(), range_lengths=[0])
 
@@ -95,3 +120,12 @@ def test_combine():
     cgrl = combine_sequences(grla, grlb)
 
     assert len(cgrl) == 3
+
+
+def test_empty_grl_slice():
+    grl = GenomicRangesList.empty(n=100)
+    assert isinstance(grl, GenomicRangesList)
+
+    subset = grl[0:10]
+    assert isinstance(subset, GenomicRangesList)
+    assert len(subset) == 10