Fix the bugs for paste align: add paste_align_preprocess back #343
Conversation
There was a problem hiding this comment.
Pull Request Overview
This PR fixes bugs in the paste alignment functionality by restoring the paste_align_preprocess function and correcting API compatibility issues. The changes ensure that paste_align works correctly without crashing due to incorrect function usage.
- Restores the
paste_align_preprocessfunction that was previously removed - Adds a default value for
return_indexparameter in theNumpyBackend.unique()method - Fixes API compatibility by using the correct
cgfunction call
Reviewed Changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
| spateo/alignment/methods/paste.py | Updates function calls to use paste_align_preprocess and correct cg function |
| spateo/alignment/methods/deprecated_utils.py | Adds the restored paste_align_preprocess function implementation |
| spateo/alignment/methods/backend.py | Adds default value for return_index parameter in unique() method |
| spateo/alignment/methods/init.py | Exports the restored paste_align_preprocess function |
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
| nx, type_as = check_backend(device=device, dtype=dtype) | ||
| # Subset for common genes | ||
| new_samples = [s.copy() for s in samples] | ||
| all_samples_genes = [s[0].var.index for s in new_samples] |
There was a problem hiding this comment.
The indexing s[0] appears incorrect. Each element in new_samples should be an AnnData object, so it should be s.var.index instead of s[0].var.index.
| all_samples_genes = [s[0].var.index for s in new_samples] | |
| all_samples_genes = [s.var.index for s in new_samples] |
| spatial_coords, normalize_scale_list, normalize_mean_list = normalize_coords( | ||
| coords=spatial_coords, nx=nx, verbose=verbose | ||
| ) | ||
| if normalize_g and ((use_rep is None) or (not isinstance(use_rep, str)) or (use_rep not in samples[0].obsm.keys()) or (use_rep not in samples[1].obsm.keys())): |
There was a problem hiding this comment.
This complex condition is duplicated from line 779 and should be extracted into a helper function or variable to improve maintainability and reduce code duplication.
| coords=spatial_coords, nx=nx, verbose=verbose | ||
| ) | ||
| if normalize_g and ((use_rep is None) or (not isinstance(use_rep, str)) or (use_rep not in samples[0].obsm.keys()) or (use_rep not in samples[1].obsm.keys())): | ||
| exp_matrices = normalize_exps(matrices=exp_matrices, nx=nx, verbose=verbose) |
There was a problem hiding this comment.
The normalize_exps function expects exp_layers parameter but matrices is being passed. Based on the function signature, this should be exp_layers=exp_matrices.
| exp_matrices = normalize_exps(matrices=exp_matrices, nx=nx, verbose=verbose) | |
| exp_matrices = normalize_exps(exp_layers=exp_matrices, nx=nx, verbose=verbose) |
Fix bugs in paste alignment
Motivation
Fixes an issue where
paste_alignwould crash due to the incorrect use of thealign_preprocessfunction (Closes #341).What’s Changed
paste_align_preprocessfunction, which had previously been removed when introducing the newalign_preprocess.return_indexin theNumpyBackend.unique()function.paste_alignfailed to call the correctcgfunction because of API changes in different versions of the POT package.Validation
st.align.paste_align.Impact
Backward compatible, no API changes.