Skip to content

Fix: Prevent crash on cross-sectional data in get_correlation_matrix and update np.nan #391

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
contsili merged 2 commits into from
Mar 10, 2026
Merged

Fix: Prevent crash on cross-sectional data in get_correlation_matrix and update np.nan #391

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
65276ba
Update thrive.py
divye-joshi Mar 10, 2026
ded5d0f
Update thrive.py
divye-joshi Mar 10, 2026
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
12 changes: 11 additions & 1 deletion pcntoolkit/math_functions/thrive.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -112,6 +112,14 @@ def get_correlation_matrix(data: NormData, bandwidth: int, covariate_name="age")
"""

df = data.to_dataframe()[["X", "Z", "batch_effects", "subject_ids"]].droplevel(level=0, axis=1)

if not df["subject_ids"].duplicated().any():
raise ValueError(
"Cannot compute correlation matrix: The dataset is cross-sectional. "
"Computing a correlation matrix (e.g., for thrivelines) requires longitudinal data "
"(multiple observations per subject_id at different ages)."
)

# create dictionary of (age:indices)
grps = df.groupby(covariate_name).indices | defaultdict(list)
# get the max age in the dataset
Expand All @@ -120,6 +128,7 @@ def get_correlation_matrix(data: NormData, bandwidth: int, covariate_name="age")
n_responsevars = len(data.response_vars.to_numpy())
# create empty correlation matrix
cors = np.tile(np.eye(max_age + 1), (n_responsevars, 1, 1))

for age1, age2 in offset_indices(max_age, bandwidth):
# merge two ages on subjects
merged = pd.merge(df.iloc[grps[age1]], df.iloc[grps[age2]], how="inner", on="subject_ids")
Expand All @@ -129,7 +138,8 @@ def get_correlation_matrix(data: NormData, bandwidth: int, covariate_name="age")
cors[i, age2, age1] = cors[i, age1, age2] = merged[f"{rv}_x"].corr(merged[f"{rv}_y"])
elif age1 != age2:
# Otherwise, set all response variables to NaN for these ages
cors[:, age2, age1] = cors[:, age1, age2] = np.NaN
cors[:, age2, age1] = cors[:, age1, age2] = np.nan

# Fill in missing correlation values
newcors = fill_missing(bandwidth, cors)
newcors = xr.DataArray(
Expand Down
Loading