Overlap point segment patch #183

bioframe/core/arrops.py

@@ -264,6 +264,22 @@ def _overlap_intervals_legacy(starts1, ends1, starts2, ends2, closed=False, sort
 
     return overlap_ids
 
+def get_pseudo_segment(starts, ends):
+    """
+    Get pseudo-segment for overlapping intervals.
+
+    Parameters
+    ----------
+    starts, ends : numpy.ndarray
+
+    Returns
+    -------
+    pseudo_ends : numpy.ndarray
+    An array of pseudo-ends for overlapping intervals.
+    """
+    pseudo_ends = ends.copy()
+    pseudo_ends[ends == starts] += 1
+    return [starts, pseudo_ends]
 
 def overlap_intervals(starts1, ends1, starts2, ends2, closed=False, sort=False):
     """
@@ -296,8 +312,11 @@ def overlap_intervals(starts1, ends1, starts2, ends2, closed=False, sort=False):
 
     starts1 = np.asarray(starts1)
     ends1 = np.asarray(ends1)
+    starts1, ends1 = get_pseudo_segment(starts1, ends1)
+
     starts2 = np.asarray(starts2)
     ends2 = np.asarray(ends2)
+    starts2, ends2 = get_pseudo_segment(starts2, ends2)
 
     # Concatenate intervals lists
     n1 = len(starts1)

tests/test_ops.py

@@ -427,6 +427,94 @@ def test_overlap():
     )
     assert len(b) == 3
 
+    ### test overlap with point and segment data
+    df1 = pd.DataFrame(
+        [
+            ['chr1', 1, 1]
+        ],
+        columns=['chrom','start','end']
+    ).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()})
+
+    df2 = pd.DataFrame(
+        [
+            ['chr1', 1, 2]
+        ],
+        columns=['chrom','start','end']
+    ).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()})
+
+    b = bioframe.overlap(
+        df1,
+        df2,
+        on=None,
+        how="left",
+        return_index=True,
+        return_input=False,
+    )
+    assert np.sum(pd.isna(b["index_"].values)) == 0
+    b = bioframe.overlap(
+        df2,
+        df1,
+        on=None,
+        how="left",
+        return_index=True,
+        return_input=False,
+    )
+    assert np.sum(pd.isna(b["index_"].values)) == 0
+
+    b = bioframe.overlap(
+        df1,
+        df2,
+        on=None,
+        how="right",
+        return_index=True,
+        return_input=False,
+    )
+    assert np.sum(pd.isna(b["index"].values)) == 0
+    b = bioframe.overlap(
+        df2,
+        df1,
+        on=None,
+        how="right",
+        return_index=True,
+        return_input=False,
+    )
+    assert np.sum(pd.isna(b["index"].values)) == 0
+
+    ### Two adjacent point should not overlap with each other
+    df1 = pd.DataFrame(
+        [
+            ['chr1', 1, 1]
+        ],
+        columns=['chrom','start','end']
+    ).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()})
+
+    df2 = pd.DataFrame(
+        [
+            ['chr1', 2, 2]
+        ],
+        columns=['chrom','start','end']
+    ).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()})
+
+    b = bioframe.overlap(
+        df1,
+        df2,
+        on=None,
+        how="left",
+        return_index=True,
+        return_input=False,
+    )
+    assert np.sum(pd.isna(b["index_"].values)) == 1
+    b = bioframe.overlap(
+        df2,
+        df1,
+        on=None,
+        how="left",
+        return_index=True,
+        return_input=False,
+    )
+    assert np.sum(pd.isna(b["index_"].values)) == 1
+
+
     ### test keep_order and NA handling
     df1 = pd.DataFrame(
         [
@@ -1403,14 +1491,80 @@ def test_subtract():
         .sort_values(["chrom", "start", "end"])
         .reset_index(drop=True)
     )
-
     pd.testing.assert_frame_equal(
         df_result.astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()}),
         bioframe.subtract(df1, df2)
         .sort_values(["chrom", "start", "end"])
         .reset_index(drop=True),
     )
 
+    # Test the case when substraction from point bioframe
+    df1 = pd.DataFrame(
+        [
+            ['chr1', 1, 1]
+        ],
+        columns=['chrom','start','end']
+    ).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()})
+
+    df2 = pd.DataFrame(
+        [
+            ['chr1', 0, 2]
+        ],
+        columns=['chrom','start','end']
+    ).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()})
+
+    df_result = (
+        pd.DataFrame(
+            [
+                ["chr1", 0, 1],
+                ["chr1", 1, 2],
+            ],
+            columns=["chrom", "start", "end"],
+        )
+        .sort_values(["chrom", "start", "end"])
+        .reset_index(drop=True)
+    )
+    pd.testing.assert_frame_equal(
+        df_result.astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()}),
+        bioframe.subtract(df2, df1)
+        .sort_values(["chrom", "start", "end"])
+        .reset_index(drop=True),
+    )
+
+    # Test the case when substraction from point is at the beginning bioframe
+
+    df1 = pd.DataFrame(
+        [
+            ['chr1', 1, 1]
+        ],
+        columns=['chrom','start','end']
+    ).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()})
+
+    df2 = pd.DataFrame(
+        [
+            ['chr1', 1, 2]
+        ],
+        columns=['chrom','start','end']
+    ).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()})
+
+    df_result = (
+        pd.DataFrame(
+            [
+                ["chr1", 1, 2],
+            ],
+            columns=["chrom", "start", "end"],
+        )
+        .sort_values(["chrom", "start", "end"])
+        .reset_index(drop=True)
+    )
+
+    pd.testing.assert_frame_equal(
+        df_result.astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()}),
+        bioframe.subtract(df2, df1)
+        .sort_values(["chrom", "start", "end"])
+        .reset_index(drop=True),
+    )
+
     # subtract should ignore null rows
     df1 = pd.DataFrame(
         [[pd.NA, pd.NA, pd.NA], ["chr1", 1, 5]],