Skip to content

Speed up sparse tensor export/import using numpy #465

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

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Speed up sparse tensor export/import using numpy #465

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b0401e8
use numpy to export/import sparse tensors
gqcollins Nov 7, 2025
d974d5c
check imported tensor dimensions
gqcollins Nov 14, 2025
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
13 changes: 4 additions & 9 deletions pyttb/export_data.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,12 +109,7 @@ def export_sparse_array(
"""Export sparse array data in coordinate format."""
if not fmt_data:
fmt_data = "%.16e"
# TODO: looping through all values may take a long time, can this be more efficient?
for i in range(A.nnz):
# 0-based indexing in package, 1-based indexing in file
subs = A.subs[i, :] + index_base
subs.tofile(fp, sep=" ", format="%d")
print(end=" ", file=fp)
val = A.vals[i][0]
val.tofile(fp, sep=" ", format=fmt_data)
print(file=fp)
# 0-based indexing in package, 1-based indexing in file
subs = A.subs + index_base
vals = A.vals[:, 0].reshape(-1, 1)
np.savetxt(fp, np.hstack((subs, vals)), fmt="%d " * subs.shape[1] + fmt_data)
13 changes: 7 additions & 6 deletions pyttb/import_data.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -99,12 +99,13 @@ def import_sparse_array(
fp: TextIO, n: int, nz: int, index_base: int = 1
) -> tuple[np.ndarray, np.ndarray]:
"""Extract sparse data subs and vals from coordinate format data."""
subs = np.zeros((nz, n), dtype="int64")
vals = np.zeros((nz, 1))
for k in range(nz):
line = fp.readline().strip().split(" ")
subs[k, :] = [np.int64(i) - index_base for i in line[:-1]]
vals[k, 0] = line[-1]
data = np.loadtxt(fp)
subs = data[:, :-1].astype("int64") - index_base
vals = data[:, -1].reshape(-1, 1)
if subs.shape[0] != nz:
raise ValueError("Imported nonzeros are not of expected size")
if subs.shape[1] != n:
raise ValueError("Imported tensor is not of expected shape")
return subs, vals


Expand Down
Loading