saezlab/liana-py

AI Summary: Users are unable to run a specific `liana-py` tutorial on Python 3.13 because the required `corneto==0.9.1-alpha.6` fails to install due to a build error with its `PySCIPOpt` dependency. While newer `corneto` versions (>=1.0) are available for Python 3.13, they are incompatible with `liana`, creating a dependency conflict that blocks the tutorial's execution.

AI Summary: The user is encountering a `DatabaseError: file is not a database` error when using the LIANA+ framework's `get_metalinks` function. This indicates a problem with the database connection, likely due to an incorrect or missing database file path. The task involves debugging the code to identify the correct database path and ensure the database file exists and is accessible.

AI Summary: Debug a KeyError: 'cellphone_pvals' occurring in the LIANA+ Python library's `li.mt.rank_aggregate.by_sample` function when permutation testing (n_perms > 0) is enabled. The error only appears when using this function and not when using `li.mt.cellphonedb.by_sample`, suggesting the issue lies in how `rank_aggregate` handles p-values during aggregation with permutations. The task involves analyzing the provided error traceback, understanding the LIANA+ codebase (specifically the `liana_pipe`, `_aggregate`, and `_rank_aggregate` functions), and identifying the source of the missing 'cellphone_pvals' key. Reproducible with the `li.testing.datasets.kang_2018()` dataset.

AI Summary: The task is to debug a `tqdm` progress bar error occurring within the LIANA+ framework when used with Python 3.11 on macOS. The error, stemming from multiprocessing issues, prevents the `tqdm` progress bar from functioning correctly, even when disabled. The solution likely involves investigating the interaction between `tqdm`, multiprocessing, and Python 3.11 on macOS, potentially requiring adjustments to the LIANA+ code or its dependencies.

AI Summary: The LIANA+ framework, used for cell-cell communication analysis, has a dependency (PySCIPOpt) that does not support the linux/arm64 architecture, preventing the creation of Docker images for this platform. The task is to either find a workaround to build the Docker image despite the PySCIPOpt limitation or to find a suitable replacement for PySCIPOpt within LIANA+.

AI Summary: The task is to standardize the docstring format across the LIANA+ Python codebase. Currently, a mix of NumPy and Google style docstrings is used, and the project should consistently use either one. The issue suggests prioritizing NumPy style, as it's the default, but this needs verification.

AI Summary: The task is to investigate and resolve a potential configuration duplication within the `bump_version` settings of the LIANA+ project. This likely involves reviewing the project's configuration files to identify and correct or remove the redundant entry.

AI Summary: The user is encountering an AttributeError when using the `liana` Python package. The error indicates that the `resource` and `rs` attributes are missing, preventing access to the `get_metalinks` function for retrieving metabolite-mediated cell-cell communication data. The task involves debugging this issue, potentially by examining the `liana` package's documentation and code, checking for version compatibility issues, or identifying a missing dependency.

AI Summary: The task is to fix a bug in the LIANA+ Python package's circle plot function. Currently, the node size in the plot only reflects the number of outgoing interactions from source cell types, neglecting incoming interactions to target cell types. The fix involves modifying a single line of code in the `_circle_plot.py` file to calculate node size based on the total number of incoming and outgoing interactions for each cell type.