Inmoose

None

- fix dispersion estimation and MAD computation in `deseq2` module
- improve GLM fit in `deseq2` module
- create `cohort_qc` module with `CohortMetric` and `QCReport` classes to
evaluate dataset quality after batch effect correction
- improve `cohort_qc` module when the number of covariates is 0 or 1
- improve documentation for `diffexp` module
- improve documentation for `consensus_clustering` module
- batch effect correction functions now accept `AnnData` as input
- fix p-value adjustment in `deseq2` module
- add logos and badges in the README

- opt out of `numpy` 2.1.0 until `statsmodels` supports it
- fix an array indexing bug in function `meta_de`
- function `meta_de` is now robust to NaN *p*-values
- allow arbitrary types for `batch` and `group` arguments in function
`sim_rnaseq`

- fix a bug when splicing a `DEResults`
- fix coverage report upload in CI/CD pipeline
- improve documentation landing page and references
- use a package-specific logger

- harmonize differential expression analysis results across `deseq2`, `edgepy`
and `limma` modules
- add a meta-analysis feature for differential expression

- move most of C++ code to Python and Cython
- compatibility with `numpy` 2.0.0