feat(plot_qc_metrics): plot single cell quality control metrics #354
Conversation
…cs quality control metrics
There was a problem hiding this comment.
Pull Request Overview
This PR adds a new function plot_qc_metrics to generate QC metric violin plots for single-cell or spatial transcriptomics data and includes tests to validate its output.
- Implements
plot_qc_metricsinsrc/spac/visualization.pywith optional grouping, log transformation, and customization. - Adds unit tests in
tests/test_visualization/test_plot_qc_metrics.pyto check return types and basic functionality.
Reviewed Changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
| tests/test_visualization/test_plot_qc_metrics.py | New tests for plot_qc_metrics, verifying output types for default, log, and annotation modes |
| src/spac/visualization.py | Introduces plot_qc_metrics, including parameter handling and Scanpy violin plotting logic |
Comments suppressed due to low confidence (2)
src/spac/visualization.py:3864
- The docstring refers to 'group_column' but the function parameter is named 'annotation'. Update the text to use 'annotation' for consistency.
If group_column is None, returns a dictionary with keys
tests/test_visualization/test_plot_qc_metrics.py:41
- [nitpick] Consider adding an assertion to verify the 'axes' element in the returned dictionary for the annotation case, ensuring both 'figure' and 'axes' are returned as expected.
self.assertIsInstance(result["A"]["figure"], Figure)
| should_exist=True) | ||
|
|
||
| if annotation is not None: | ||
| check_annotation(adata, annotations=annotation) |
There was a problem hiding this comment.
Hi @abombin , you can combine that check with the check on line 3875 (just create one list with all the annotations)
| Column name in adata.obs to group the data by (default: None). | ||
| log : bool, optional | ||
| Whether to log-transform the data (default: False). | ||
| size : float, optional |
There was a problem hiding this comment.
@abombin , would you like to check that the size is within the correct range?
| results = {} | ||
| for group in adata.obs[annotation].unique(): | ||
| adata_subset = adata[adata.obs[annotation] == group] | ||
| violin_plot = sc.pl.violin( |
There was a problem hiding this comment.
@abombin , I suggest to use "partial" to define the call once, and call it in various part of the code (e.g. here and in line 3903). This way, if you ever change it, you change it in one place.
| adata = self.create_test_adata() | ||
| result = plot_qc_metrics(adata, log=True) | ||
| self.assertIsInstance(result["figure"], Figure) | ||
| self.assertTrue(isinstance(result["axes"], (np.ndarray, Axes))) |
There was a problem hiding this comment.
@abombin , would you like to check other aspects of the figure other than type of what is being returned? Something to verify that the figure actually has a valid plot in it?
Add plot_qc_metrics function to visualize sc/spatial transcriptomics quality control metrics.
Add tests for plot_qc_metrics function.