fix: add docformatter to format docstrings

.pre-commit-config.yaml

@@ -31,3 +31,14 @@ repos:
     hooks:
       - id: mypy
         exclude: ^(tests/|docs/source/conf.py|datasets/)
+  - repo: local
+    hooks:
+      - id: docformatter
+        name: Format docstrings with `docformatter`
+        language: system
+        types: [python]
+        require_serial: true
+        entry: poetry run docformatter
+        args:
+          [--in-place, --wrap-summaries=88, --wrap-descriptions=87, --recursive]
+        files: ^(fedpydeseq2|tests)/

fedpydeseq2/core/deseq2_core/__init__.py

@@ -1,7 +1,7 @@
 """Module containing the core of the DESeq2 pipeline.
 
-It contains all the Mixin classes corresponding to the step of the pipeline.
-The main class defined in this module is the DESeq2FullPipe class.
+It contains all the Mixin classes corresponding to the step of the pipeline. The main
+class defined in this module is the DESeq2FullPipe class.
 """
 
 from fedpydeseq2.core.deseq2_core.deseq2_full_pipe import DESeq2FullPipe

fedpydeseq2/core/deseq2_core/build_design_matrix/build_design_matrix.py

@@ -69,7 +69,6 @@ def build_design_matrix(
 
         round_idx: int
             The updated round
-
         """
         # ---- For each design factor, get the list of each center's levels ---- #
         if len(local_states) == 0:

fedpydeseq2/core/deseq2_core/build_design_matrix/substeps.py

@@ -147,8 +147,7 @@ def set_local_design(
         shared_state,
     ):
         # pylint: disable=unused-argument
-        """
-        Set the design matrices in centers.
+        """Set the design matrices in centers.
 
         Returns their columns in order to harmonize them.
 
@@ -188,7 +187,6 @@ class LocOderDesignComputeLogMean:
     -------
     order_design_cols_compute_local_log_mean
         Order design columns and compute the local log mean.
-
     """
 
     local_adata: ad.AnnData

fedpydeseq2/core/deseq2_core/compute_cook_distance/compute_cook_distance.py

@@ -25,7 +25,6 @@ class ComputeCookDistances(
     -------
     compute_cook_distance
         The method to compute Cook's distances.
-
     """
 
     trimmed_mean_num_iter: int
@@ -39,8 +38,7 @@ def compute_cook_distance(
         round_idx,
         clean_models,
     ):
-        """
-        Compute Cook's distances.
+        """Compute Cook's distances.
 
         Parameters
         ----------
@@ -70,7 +68,6 @@ def compute_cook_distance(
 
         round_idx: int
             The updated round index.
-
         """
         local_states, agg_shared_state, round_idx = self.compute_trim_mean(
             train_data_nodes,

fedpydeseq2/core/deseq2_core/compute_cook_distance/substeps.py

@@ -24,8 +24,7 @@ def local_compute_sqerror(
         data_from_opener,
         shared_state=dict,
     ) -> None:
-        """
-        Compute the squared error between the normalized counts and the trimmed mean.
+        """Compute the squared error between the normalized counts and the trimmed mean.
 
         Parameters
         ----------
@@ -34,7 +33,6 @@ def local_compute_sqerror(
 
         shared_state : dict, optional
             Results to save in the local states.
-
         """
         cell_means = shared_state["trimmed_mean_normed_counts"]
         if isinstance(cell_means, pd.DataFrame):
@@ -61,8 +59,7 @@ def local_get_normed_count_means(
         data_from_opener,
         shared_state=dict,
     ) -> dict:
-        """
-        Send local normed counts means.
+        """Send local normed counts means.
 
         Parameters
         ----------
@@ -80,7 +77,6 @@ def local_get_normed_count_means(
             - mean_normed_counts: mean of the normalized counts
             - n_samples: number of samples
             - varEst: variance estimate
-
         """
         return {}
 
@@ -95,8 +91,7 @@ def agg_compute_dispersion_for_cook(
         self,
         shared_states: list[dict],
     ) -> dict:
-        """
-        Compute the dispersion for Cook's distance calculation.
+        """Compute the dispersion for Cook's distance calculation.
 
         Parameters
         ----------
@@ -111,6 +106,5 @@ def agg_compute_dispersion_for_cook(
         dict
             Because it is decorated, the dictionary will have the following key:
             - cooks_dispersions: dispersion values
-
         """
         return {}

fedpydeseq2/core/deseq2_core/compute_size_factors/compute_size_factors.py

@@ -22,7 +22,6 @@ class ComputeSizeFactors(
         pipeline. It sets the size factors in the local AnnData and computes the
         Gram matrix and feature vector in order to start the next step, i.e.,
         the computation of rough dispersions.
-
     """
 
     @log_organisation_method
@@ -72,7 +71,6 @@ def compute_size_factors(
 
         round_idx: int
             The updated round index.
-
         """
         # ---- Aggregate means of log gene expressions ----#
 

fedpydeseq2/core/deseq2_core/compute_size_factors/substeps.py

@@ -49,7 +49,6 @@ class LocSetSizeFactorsComputeGramAndFeatures:
     -------
     local_set_size_factors_compute_gram_and_features
         The method to set the size factors and compute the Gram matrix and feature.
-
     """
 
     local_adata: ad.AnnData

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_MAP_dispersions/compute_MAP_dispersions.py

@@ -12,16 +12,14 @@ class ComputeMAPDispersions(
     LocFilterMAPDispersions,
     ComputeDispersionsGridSearch,
 ):
-    """
-    Mixin class to implement the computation of MAP dispersions.
+    """Mixin class to implement the computation of MAP dispersions.
 
     Methods
     -------
     fit_MAP_dispersions
         A method to fit the MAP dispersions and filter them.
         The filtering is done by removing the dispersions that are too far from the
         trend curve.
-
     """
 
     @log_organisation_method

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_dispersion_prior/compute_dispersion_prior.py

@@ -24,7 +24,6 @@ class ComputeDispersionPrior(
     -------
     compute_dispersion_prior
         The method to fit the dispersion trend.
-
     """
 
     @log_organisation_method
@@ -71,7 +70,6 @@ def compute_dispersion_prior(
 
         round_idx: int
             The updated round index.
-
         """
         # --- Return means and dispersions ---#
         # TODO : merge this step with the last steps from genewise dispersion

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_dispersion_prior/substeps.py

@@ -62,7 +62,6 @@ def get_local_mean_and_dispersion(
                 The genewise dispersions.
             - num_vars: int
                 The number of variables.
-
         """
         # Save gene-wise dispersions from the previous step.
         # Dispersions of all-zero genes should already be NaN.
@@ -89,8 +88,7 @@ class AggFitDispersionTrendAndPrior:
     @remote
     @log_remote
     def agg_fit_dispersion_trend_and_prior_dispersion(self, shared_states):
-        """
-        Fit the dispersion trend, and compute the dispersion prior.
+        """Fit the dispersion trend, and compute the dispersion prior.
 
         Parameters
         ----------
@@ -118,7 +116,6 @@ def agg_fit_dispersion_trend_and_prior_dispersion(self, shared_states):
                 The type of dispersion function (parametric or mean).
             - mean_disp: float, optional
                 The mean dispersion (if "mean" fit type).
-
         """
         genewise_dispersions = shared_states[0]["genewise_dispersions"]
         n_params = shared_states[0]["n_params"]
@@ -240,8 +237,7 @@ def loc_update_fitted_dispersions(
         data_from_opener,
         shared_state: dict,
     ) -> None:
-        """
-        Update the fitted dispersions after replacing outliers.
+        """Update the fitted dispersions after replacing outliers.
 
         Parameters
         ----------

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_genewise_dispersions/compute_MoM_dispersions/compute_MoM_dispersions.py

@@ -28,7 +28,6 @@ class ComputeMoMDispersions(
     compute_MoM_dispersions
         The method to compute the MoM dispersions, that must be used in the main
         pipeline.
-
     """
 
     @log_organisation_method
@@ -80,7 +79,6 @@ def compute_MoM_dispersions(
 
         round_idx: int
             The updated round number.
-
         """
         ###### Fit rough dispersions ######
 

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_genewise_dispersions/compute_MoM_dispersions/compute_rough_dispersions.py

@@ -26,7 +26,6 @@ class ComputeRoughDispersions(
     compute_rough_dispersions
         The method to compute the rough dispersions, that must be used in the main
         pipeline.
-
     """
 
     @log_organisation_method
@@ -78,7 +77,6 @@ def compute_rough_dispersions(
 
         round_idx: int
             The updated round number.
-
         """
         # TODO: in refit mode, we need to gather the gram matrix and the features some
         #  way

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_genewise_dispersions/compute_genewise_dispersions.py

@@ -25,8 +25,7 @@ class ComputeGenewiseDispersions(
     ComputeLFC,
     LocSetMuHat,
 ):
-    """
-    Mixin class to implement the computation of both genewise and MAP dispersions.
+    """Mixin class to implement the computation of both genewise and MAP dispersions.
 
     The switch between genewise and MAP dispersions is done by setting the `fit_mode`
     argument in the `fit_dispersions` to either "MLE" or "MAP".
@@ -43,8 +42,6 @@ class ComputeGenewiseDispersions(
         in downstream steps (cooks distance, etc).
         3. Compute an estimate of the mean from these dispersions.
         4. Fit the dispersions using a grid search.
-
-
     """
 
     @log_organisation_method

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_genewise_dispersions/get_num_replicates/get_num_replicates.py

@@ -21,8 +21,7 @@ class GetNumReplicates(
     def get_num_replicates(
         self, train_data_nodes, aggregation_node, local_states, round_idx, clean_models
     ):
-        """
-        Compute the number of replicates for each combination of factors.
+        """Compute the number of replicates for each combination of factors.
 
         Parameters
         ----------

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_genewise_dispersions/get_num_replicates/substeps.py

@@ -18,8 +18,7 @@ class LocGetDesignMatrixLevels:
     @log_remote_data
     @reconstruct_adatas
     def loc_get_design_matrix_levels(self, data_from_opener, shared_state=dict) -> dict:
-        """
-        Get the values of the local design matrix.
+        """Get the values of the local design matrix.
 
         Parameters
         ----------
@@ -33,7 +32,6 @@ def loc_get_design_matrix_levels(self, data_from_opener, shared_state=dict) -> d
         dict
             Dictionary with the following key:
             - unique_counts: unique values and counts of the local design matrix
-
         """
         unique_counts = self.local_adata.obsm["design_matrix"].value_counts()
 
@@ -46,8 +44,7 @@ class AggGetCountsLvlForCells:
     @remote
     @log_remote
     def agg_get_counts_lvl_for_cells(self, shared_states: list[dict]) -> dict:
-        """
-        Aggregate the counts of the design matrix values.
+        """Aggregate the counts of the design matrix values.
 
         Parameters
         ----------
@@ -79,8 +76,7 @@ class LocFinalizeCellCounts:
     @log_remote_data
     @reconstruct_adatas
     def loc_finalize_cell_counts(self, data_from_opener, shared_state=dict) -> None:
-        """
-        Finalize the cell counts.
+        """Finalize the cell counts.
 
         Parameters
         ----------
@@ -91,7 +87,6 @@ def loc_finalize_cell_counts(self, data_from_opener, shared_state=dict) -> None:
             Dictionary with keys labeling the different values taken by the
             overall design matrix. Each values of the dictionary contains the
             sum of the counts of the corresponding design matrix value and the level.
-
         """
         counts_by_lvl = shared_state["counts_by_lvl"]
 

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_genewise_dispersions/substeps.py

@@ -41,7 +41,6 @@ def fit_lin_mu(
         refit_mode : bool
             Whether to run the pipeline in refit mode. If True, the pipeline will be run
             on `refit_adata`s instead of `local_adata`s. (default: ``False``).
-
         """
         if refit_mode:
             adata = self.refit_adata

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_lfc/compute_lfc.py

@@ -40,8 +40,6 @@ class ComputeLFC(
         The main method to compute the log fold changes by
         running the IRLS algorithm and catching it with the
         FedProxQuasiNewton algorithm.
-
-
     """
 
     @log_organisation_method
@@ -90,7 +88,6 @@ def compute_lfc(
 
         round_idx: int
             The updated round index.
-
         """
         #### ---- Initialization ---- ####
 

fedpydeseq2/core/deseq2_core/deseq2_lfc_dispersions/compute_lfc/substeps.py

@@ -34,7 +34,6 @@ class LocGetGramMatrixAndLogFeatures:
         If the gram matrix is full rank, it shares the features vector
         and the gram matrix. If the gram matrix is not full rank, it shares
         the normed log means and the number of observations.
-
     """
 
     local_adata: AnnData
@@ -106,7 +105,6 @@ def get_gram_matrix_and_log_features(
                     The normed log means.
                 - n_obs: int
                     The number of observations.
-
         """
         global_gram_matrix = self.local_adata.uns["_global_gram_matrix"]
 
@@ -236,8 +234,6 @@ def create_beta_init(self, shared_states: list[dict]) -> dict[str, Any]:
                 (n_non_zero_genes,).
             - round_number_irls: int
                 The current round number of the IRLS algorithm.
-
-
         """
         # Get the global quantities
         gram_full_rank = shared_states[0]["gram_full_rank"]
@@ -310,7 +306,6 @@ class LocSaveLFC:
         is expected to be applied after catching the IRLS method
         with the fed prox quasi newton method, and takes as an input a
         shared state from the last iteration of that method.
-
     """
 
     local_adata: AnnData
@@ -358,7 +353,6 @@ def save_lfc_to_local(
         refit_mode : bool
             Whether to run the pipeline on `refit_adata` instead of `local_adata`.
             (default: False).
-
         """
         beta = shared_state["beta"]