Skip to content

modified window_ranges #117

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
marco-mariotti merged 1 commit into from
Dec 16, 2025
Merged

modified window_ranges #117

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
81 changes: 72 additions & 9 deletions pyranges/core/pyranges_main.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4718,9 +4718,11 @@ def window_ranges(
self,
window_size: int,
use_strand: VALID_USE_STRAND_TYPE = USE_STRAND_DEFAULT,
match_by: VALID_BY_TYPES = None,
group_by: VALID_BY_TYPES = None,
*,
add_window_id: bool = False,
) -> "PyRanges":
"""Return non-overlapping genomic windows.
"""Return intervals sliced in non-overlapping windows.

Every interval is split into windows of length `window_size` starting from its 5' end.

Expand All @@ -4733,11 +4735,16 @@ def window_ranges(
Whether negative strand intervals should be sliced in descending order, meaning 5' to 3'.
The default "auto" means True if PyRanges has valid strands (see .strand_valid).

match_by : str | Sequence[str] | None, default None
group_by : str | Sequence[str] | None, default None
Column name(s) used to form groups. If provided, windowing proceeds
*continuously within each group*: any leftover (partial) window at the end
of one interval is continued at the start of the next interval with the
same `match_by` value. The window “phase” resets at group boundaries.
same `group_by` value. The window “phase” resets at group boundaries.

add_window_id : bool, default False
Use only in combination with group_by. If True, adds a column "window_id" with an index-like identifier
to link the windows split in non-contigous intervals, e.g. because a certain window spans an intron.
The window_id starts at 0, and resets for each group.

Returns
-------
Expand All @@ -4748,6 +4755,7 @@ def window_ranges(
Warning
-------
The returned Pyranges may have index duplicates. Call .reset_index(drop=True) to fix it.
Moreover, the input row order may not be preserved.

See Also
--------
Expand Down Expand Up @@ -4868,7 +4876,7 @@ def window_ranges(
PyRanges with 5 rows, 5 columns, and 1 index columns.
Contains 1 chromosomes and 2 strands.

>>> gr3.window_ranges(8, match_by='ID')
>>> gr3.window_ranges(8, group_by='ID')
index | Chromosome Strand Start End ID
int64 | int64 object int64 int64 object
------- --- ------------ -------- ------- ------- --------
Expand All @@ -4884,16 +4892,60 @@ def window_ranges(
PyRanges with 10 rows, 5 columns, and 1 index columns (with 5 index duplicates).
Contains 1 chromosomes and 2 strands.

>>> gr4 = pr.PyRanges({'Chromosome':2, 'Strand':list('+++--'), 'Start':[30,10,50,90,70],
... 'End':[40,20,60,100,80], 'ID':['id1','id1','id1','id2','id2']})
>>> gr4
index | Chromosome Strand Start End ID
int64 | int64 object int64 int64 object
------- --- ------------ -------- ------- ------- --------
0 | 2 + 30 40 id1
1 | 2 + 10 20 id1
2 | 2 + 50 60 id1
3 | 2 - 90 100 id2
4 | 2 - 70 80 id2
PyRanges with 5 rows, 5 columns, and 1 index columns.
Contains 1 chromosomes and 2 strands.

>>> gr4.window_ranges(8, group_by='ID').reset_index(drop=True).head(8)
index | Chromosome Strand Start End ID
int64 | int64 object int64 int64 object
------- --- ------------ -------- ------- ------- --------
0 | 2 + 10 18 id1
1 | 2 + 18 20 id1
2 | 2 + 30 36 id1
3 | 2 + 36 40 id1
4 | 2 + 50 54 id1
5 | 2 + 54 60 id1
6 | 2 - 74 80 id2
7 | 2 - 70 74 id2
PyRanges with 8 rows, 5 columns, and 1 index columns.
Contains 1 chromosomes and 2 strands.

>>> gr4.window_ranges(8, group_by='ID', add_window_id=True).reset_index(drop=True).head(8)
index | Chromosome Strand Start End ID window_id
int64 | int64 object int64 int64 object int64
------- --- ------------ -------- ------- ------- -------- -----------
0 | 2 + 10 18 id1 1
1 | 2 + 18 20 id1 2
2 | 2 + 30 36 id1 2
3 | 2 + 36 40 id1 3
4 | 2 + 50 54 id1 3
5 | 2 + 54 60 id1 4
6 | 2 - 74 80 id2 1
7 | 2 - 70 74 id2 2
PyRanges with 8 rows, 6 columns, and 1 index columns.
Contains 1 chromosomes and 2 strands.

"""
import ruranges

use_strand = validate_and_convert_use_strand(self, use_strand)
match_by = arg_to_list(match_by)
group_ids = factorize_arange(self, match_by)
group_by = arg_to_list(group_by)
group_ids = factorize_arange(self, group_by)

negative_strand = (self[STRAND_COL] == "-").to_numpy() if use_strand else np.zeros(len(self), dtype=bool)

gr = self.sort_ranges(by=match_by, use_strand=False) if match_by else self
gr = self.sort_ranges(by=group_by, use_strand=False) if group_by else self

idx, starts, ends = ruranges.window( # type: ignore[attr-defined]
groups=group_ids,
Expand All @@ -4905,9 +4957,20 @@ def window_ranges(
df = gr.take(idx) # type: ignore[arg-type]
df.loc[:, START_COL] = starts
df.loc[:, END_COL] = ends
df = ensure_pyranges(df)

if add_window_id and group_by:
df = (df.group_cumsum(group_by, cumsum_start_column="__cstart__", cumsum_end_column="__cend__")).drop(
columns="__cstart__"
)
df = cast("pd.DataFrame", df)
df["__is_whole__"] = (df.__cend__ % window_size) == 0
df["__do_increm__"] = df.groupby(group_by)["__is_whole__"].shift(fill_value=True)
df["window_id"] = df.groupby(group_by)["__do_increm__"].cumsum()
df = ensure_pyranges(df.drop(columns=["__cend__", "__is_whole__", "__do_increm__"]))

# every interval can be processed individually. This may be optimized in the future.
return ensure_pyranges(df)
return df

def remove_nonloc_columns(self) -> "PyRanges":
"""Remove all columns that are not genome location columns (Chromosome, Start, End, Strand).
Expand Down