Feat/backed adata

[noamteyssier]

No description provided.

[gemini-code-assist]

Summary of Changes

Hello @noamteyssier, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request introduces robust support for backed AnnData objects, enabling the pdex function to process data efficiently whether it's in-memory or stored on disk. The changes include updating core dependencies to facilitate broader data manipulation, refining internal matrix isolation logic to guarantee in-memory processing of backed data, and enhancing sparse matrix detection for improved compatibility. New tests have been added to rigorously confirm that pdex yields identical results for both in-memory and backed AnnData, thereby ensuring reliability and consistency across different data storage paradigms.

Highlights

• Dependency Updates: Added numpy, pandas, and scipy as core dependencies, and refined development dependencies by removing pyright and pytest-benchmark while upgrading ruff.
• Backed AnnData Support: Enhanced the _isolate_matrix function to explicitly handle backed AnnData objects, ensuring that data slices are always returned in-memory as np.ndarray or csr_matrix.
• Sparse Matrix Detection: Improved sparse matrix detection by replacing isinstance(..., csr_matrix) with the more general issparse function from scipy.sparse.
• Comprehensive Testing: Introduced new test fixtures and a dedicated test suite to validate pdex functionality and ensure consistent results when working with backed AnnData objects.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Changelog

• pyproject.toml
  • Added numpy, pandas, and scipy to the main project dependencies.
  • Removed pyright and pytest-benchmark from development dependencies.
  • Updated the version of the ruff dependency.
• src/pdex/init.py
  • Imported issparse from scipy.sparse.
  • Modified _isolate_matrix to ensure backed sparse or dense data is converted to in-memory csr_matrix or np.ndarray respectively.
• src/pdex/_utils.py
  • Imported issparse from scipy.sparse.
  • Replaced isinstance(chunk, csr_matrix) with issparse(chunk) in the _detect_is_log1p function.
• tests/conftest.py
  • Added a new small_adata_backed pytest fixture to create and load AnnData objects in backed mode for testing.
• tests/test_pdex.py
  • Introduced the TestPdexBacked class with new test cases to compare pdex output for backed and in-memory AnnData in both 'ref' and 'all' modes.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request introduces support for backed AnnData objects, which is an excellent enhancement for scalability when working with large datasets. The core logic in _isolate_matrix to handle data loading from backed storage into memory is well-implemented. The switch to scipy.sparse.issparse is also a good generalization. My feedback is focused on improving the test suite to ensure this new capability is comprehensively validated across all relevant modes.

[gemini-code-assist]

This test class can be improved for better maintainability and test coverage.

1. The tests test_ref_mode_backed_matches_inmemory and test_all_mode_backed_matches_inmemory contain significant code duplication. They can be combined into a single, parameterized test using pytest.mark.parametrize to improve maintainability.
2. The on_target mode also uses the modified _isolate_matrix function, but it is not currently tested with backed AnnData objects. Adding a test for this mode is crucial for ensuring complete coverage of the new feature.

Here is a suggested implementation that addresses both points:

class TestPdexBacked:
    """Backed AnnData should produce the same results as in-memory."""

    @pytest.mark.parametrize("mode", ["ref", "all"])
    def test_modes_backed_matches_inmemory(self, small_adata, small_adata_backed, mode):
        inmem = pdex(small_adata, groupby="guide", mode=mode, is_log1p=False)
        backed = pdex(small_adata_backed, groupby="guide", mode=mode, is_log1p=False)
        assert inmem.shape == backed.shape
        for col in ["p_value", "statistic", "fold_change", "percent_change"]:
            np.testing.assert_allclose(
                inmem[col].to_numpy(),
                backed[col].to_numpy(),
                rtol=1e-6,
                err_msg=f"Mismatch in column {col}",
            )

    def test_on_target_mode_backed_matches_inmemory(self, on_target_adata, tmp_path):
        import anndata as ad

        path = tmp_path / "test.h5ad"
        on_target_adata.write_h5ad(path)
        on_target_adata_backed = ad.read_h5ad(path, backed="r")

        inmem = pdex(
            on_target_adata,
            groupby="guide",
            mode="on_target",
            gene_col="target_gene",
            is_log1p=False,
        )
        backed = pdex(
            on_target_adata_backed,
            groupby="guide",
            mode="on_target",
            gene_col="target_gene",
            is_log1p=False,
        )
        assert inmem.shape == backed.shape
        for col in ["p_value", "statistic", "fold_change", "percent_change"]:
            np.testing.assert_allclose(
                inmem[col].to_numpy(),
                backed[col].to_numpy(),
                rtol=1e-6,
                err_msg=f"Mismatch in column {col}",
            )