Skip to content

Commit

Permalink
Adding tests for cases for the edge cases for overlap with ignore_ups…
Browse files Browse the repository at this point in the history 
…tream/downstream when there is no region to match.
  • Loading branch information
@agalitsyna
agalitsyna committed Feb 7, 2024
1 parent c449aa0 commit f42af4e
Showing 1 changed file with 156 additions and 58 deletions.
214 changes: 156 additions & 58 deletions tests/test_ops.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -65,7 +65,6 @@ def mock_bioframe(num_entries=100):

############# tests #####################
def test_trim():

### trim with view_df
view_df = pd.DataFrame(
[
Expand Down Expand Up @@ -216,7 +215,6 @@ def test_trim():


def test_expand():

d = """chrom start end
0 chr1 1 5
1 chr1 50 55
Expand Down Expand Up @@ -289,6 +287,7 @@ def test_expand():
)
pd.testing.assert_frame_equal(df, fake_expanded)


def test_expand_amount_args():
d = """chrom start end
0 chr1 3 5
Expand All @@ -298,8 +297,8 @@ def test_expand_amount_args():
with pytest.raises(ValueError):
bioframe.expand(df, pad=10, scale=2.0)

def test_overlap():

def test_overlap():
### test consistency of overlap(how='inner') with pyranges.join ###
### note does not test overlap_start or overlap_end columns of bioframe.overlap
df1 = mock_bioframe()
Expand Down Expand Up @@ -467,17 +466,14 @@ def test_overlap():
columns=["chrom2", "start2", "end2", "strand", "animal"],
).astype({"start2": pd.Int64Dtype(), "end2": pd.Int64Dtype()})

assert (
bioframe.overlap(
df1,
df2,
how="outer",
cols2=["chrom2", "start2", "end2"],
return_index=True,
keep_order=False,
).shape
== (3, 12)
)
assert bioframe.overlap(
df1,
df2,
how="outer",
cols2=["chrom2", "start2", "end2"],
return_index=True,
keep_order=False,
).shape == (3, 12)

### result of overlap should still have bedframe-like properties
overlap_df = bioframe.overlap(
Expand Down Expand Up @@ -542,14 +538,14 @@ def test_overlap_preserves_coord_dtypes():
).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()})

# inner join - left keeps non-nullable numpy uint32
overlap_dtypes = bioframe.overlap(df1, df2, ensure_int=False, how="inner").dtypes
overlap_dtypes = bioframe.overlap(df1, df2, ensure_int=False, how="inner").dtypes
for col in ["start", "end"]:
assert overlap_dtypes[col] == np.uint32
for col in ["start_", "end_"]:
assert overlap_dtypes[col] == pd.Int64Dtype()

# outer join - left uint32 gets cast to numpy float64 because of NaNs on left
overlap_dtypes = bioframe.overlap(df1, df2, how="outer", ensure_int=False).dtypes
overlap_dtypes = bioframe.overlap(df1, df2, how="outer", ensure_int=False).dtypes
assert overlap_dtypes["start"] == np.float64
assert overlap_dtypes["end"] == np.float64
assert overlap_dtypes["start_"] == pd.Int64Dtype()
Expand All @@ -567,16 +563,20 @@ def test_overlap_preserves_coord_dtypes():
# convert coords to nullable *after* joining
# inner join - uint32 output becomes UInt32
# outer join - float64 output becomes Int64
overlap_dtypes = bioframe.overlap(
df1, df2, ensure_int=False, how="inner"
).convert_dtypes().dtypes
overlap_dtypes = (
bioframe.overlap(df1, df2, ensure_int=False, how="inner")
.convert_dtypes()
.dtypes
)
assert overlap_dtypes["start"] == pd.UInt32Dtype()
assert overlap_dtypes["end"] == pd.UInt32Dtype()
assert overlap_dtypes["start_"] == pd.Int64Dtype()
assert overlap_dtypes["end_"] == pd.Int64Dtype()
overlap_dtypes = bioframe.overlap(
df1, df2, ensure_int=False, how="outer"
).convert_dtypes().dtypes
overlap_dtypes = (
bioframe.overlap(df1, df2, ensure_int=False, how="outer")
.convert_dtypes()
.dtypes
)
assert overlap_dtypes["start"] == pd.Int64Dtype()
assert overlap_dtypes["end"] == pd.Int64Dtype()
assert overlap_dtypes["start_"] == pd.Int64Dtype()
Expand All @@ -593,27 +593,19 @@ def test_overlap_ensure_int():
columns=["chrom", "start", "end", "strand"],
).astype({"start": np.uint32, "end": np.uint32})
df2 = pd.DataFrame(
[
["chr1", 6, 10, "+"],
[pd.NA, pd.NA, pd.NA, "-"],
["chrX", 7, 10, "-"]
],
[["chr1", 6, 10, "+"], [pd.NA, pd.NA, pd.NA, "-"], ["chrX", 7, 10, "-"]],
columns=["chrom", "start", "end", "strand"],
).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype()})

# inner join
overlap_dtypes = bioframe.overlap(
df1, df2, how="inner", ensure_int=True
).dtypes
overlap_dtypes = bioframe.overlap(df1, df2, how="inner", ensure_int=True).dtypes
for col in ["start", "end"]:
assert overlap_dtypes[col] == np.uint32
for col in ["start_", "end_"]:
assert overlap_dtypes[col] == pd.Int64Dtype()

# outer join - left uint32 gets cast to UInt32 before the join
overlap_dtypes = bioframe.overlap(
df1, df2, how="outer", ensure_int=True
).dtypes
overlap_dtypes = bioframe.overlap(df1, df2, how="outer", ensure_int=True).dtypes
for col in ["start", "end"]:
assert overlap_dtypes[col] == pd.UInt32Dtype()
for col in ["start_", "end_"]:
Expand Down Expand Up @@ -1078,10 +1070,8 @@ def test_closest():
)

df2 = pd.DataFrame(
[["chr1", 1, 2],
["chr1", 2, 8],
["chr1", 10, 11]],
columns=["chrom", "start", "end"]
[["chr1", 1, 2], ["chr1", 2, 8], ["chr1", 10, 11]],
columns=["chrom", "start", "end"],
)

### closest(df1, df2, k=1, direction_col="strand") ###
Expand Down Expand Up @@ -1113,13 +1103,18 @@ def test_closest():
"distance": pd.Int64Dtype(),
}
)
pd.testing.assert_frame_equal(df,
bioframe.closest(df1, df2,
k=1,
ignore_upstream=False,
ignore_downstream=True,
ignore_overlaps=True,
direction_col="strand"))
pd.testing.assert_frame_equal(
df,
bioframe.closest(
df1,
df2,
k=1,
ignore_upstream=False,
ignore_downstream=True,
ignore_overlaps=True,
direction_col="strand",
),
)

### closest(df1, df2, k=1, ignore_upstream=True, ignore_downstream=False,
### ignore_overlaps=True, direction_col="strand") ###
Expand All @@ -1134,17 +1129,119 @@ def test_closest():
"distance": pd.Int64Dtype(),
}
)
pd.testing.assert_frame_equal(df,
bioframe.closest(df1, df2,
k=1,
ignore_upstream=True,
ignore_downstream=False,
ignore_overlaps=True,
direction_col="strand"))
pd.testing.assert_frame_equal(
df,
bioframe.closest(
df1,
df2,
k=1,
ignore_upstream=True,
ignore_downstream=False,
ignore_overlaps=True,
direction_col="strand",
),
)

### closest(df1, df2, k=1, ignore_upstream=False, ignore_downstream=True,
### ignore_overlaps=True) when upstream region is present ###
df1 = pd.DataFrame(
[
["chr1", 3, 5],
],
columns=["chrom", "start", "end"],
)

df2 = pd.DataFrame(
[["chr1", 1, 2], ["chr1", 10, 11]], columns=["chrom", "start", "end"]
)

def test_coverage():
d = """chrom start end chrom_ start_ end_ distance
0 chr1 3 5 chr1 1 2 1
"""
df = pd.read_csv(StringIO(d), sep=r"\s+").astype(
{
"start_": pd.Int64Dtype(),
"end_": pd.Int64Dtype(),
"distance": pd.Int64Dtype(),
}
)
pd.testing.assert_frame_equal(
df,
bioframe.closest(
df1,
df2,
k=1,
ignore_upstream=False,
ignore_downstream=True,
ignore_overlaps=True,
),
)

### closest(df1, df2, k=1, ignore_upstream=False, ignore_downstream=True,
### ignore_overlaps=True) when upstream region is absent ###

df2 = pd.DataFrame(
[["chr1", 5, 6], ["chr1", 10, 11]], columns=["chrom", "start", "end"]
)

d = """chrom start end chrom_ start_ end_ distance
0 chr1 3 5 NaN NaN NaN NaN
"""
df = pd.read_csv(StringIO(d), sep=r"\s+").astype(
{
"chrom_": "O",
"start_": pd.Int64Dtype(),
"end_": pd.Int64Dtype(),
"distance": pd.Int64Dtype(),
}
)
pd.testing.assert_frame_equal(
df,
bioframe.closest(
df1,
df2,
k=1,
ignore_upstream=False,
ignore_downstream=True,
ignore_overlaps=True,
),
)

### closest(df1, df2, k=1, ignore_upstream=True, ignore_downstream=False,
### ignore_overlaps=True) when upstream region is absent ###

df2 = pd.DataFrame(
[
["chr1", 1, 2],
],
columns=["chrom", "start", "end"],
)

d = """chrom start end chrom_ start_ end_ distance
0 chr1 3 5 NaN NaN NaN NaN
"""
df = pd.read_csv(StringIO(d), sep=r"\s+").astype(
{
"chrom_": "O",
"start_": pd.Int64Dtype(),
"end_": pd.Int64Dtype(),
"distance": pd.Int64Dtype(),
}
)
pd.testing.assert_frame_equal(
df,
bioframe.closest(
df1,
df2,
k=1,
ignore_upstream=True,
ignore_downstream=False,
ignore_overlaps=True,
),
)


def test_coverage():
#### coverage does not exceed length of original interval
df1 = pd.DataFrame([["chr1", 3, 8]], columns=["chrom", "start", "end"])
df2 = pd.DataFrame([["chr1", 2, 10]], columns=["chrom", "start", "end"])
Expand Down Expand Up @@ -1180,8 +1277,7 @@ def test_coverage():
0 chr1 3 8 5"""
df = pd.read_csv(StringIO(d), sep=r"\s+")
pd.testing.assert_frame_equal(
df,
bioframe.coverage(df1, df2, cols1=cols1, cols2=cols2)
df, bioframe.coverage(df1, df2, cols1=cols1, cols2=cols2)
)

### coverage of NA interval returns zero for coverage
Expand Down Expand Up @@ -1461,7 +1557,6 @@ def test_subtract():


def test_setdiff():

cols1 = ["chrom1", "start", "end"]
cols2 = ["chrom2", "start", "end"]
df1 = pd.DataFrame(
Expand Down Expand Up @@ -1624,7 +1719,12 @@ def test_count_overlaps():

counts_nans_inserted_after = (
pd.concat([pd.DataFrame([pd.NA]), counts_no_nans, pd.DataFrame([pd.NA])])
).astype({"start": pd.Int64Dtype(), "end": pd.Int64Dtype(),})[
).astype(
{
"start": pd.Int64Dtype(),
"end": pd.Int64Dtype(),
}
)[
["chrom1", "start", "end", "strand", "animal", "count"]
]

Expand Down Expand Up @@ -1666,7 +1766,6 @@ def test_count_overlaps():


def test_assign_view():

## default assignment case
view_df = pd.DataFrame(
[
Expand Down Expand Up @@ -1785,7 +1884,6 @@ def test_assign_view():


def test_sort_bedframe():

view_df = pd.DataFrame(
[
["chrX", 1, 8, "oranges"],
Expand Down

0 comments on commit f42af4e

Please sign in to comment.