Merge pull request #41 from flatironinstitute/dev

v0.5.5

docs/changelog.rst

@@ -1,6 +1,17 @@
 Change Log
 ==========
 
+Inferelator v0.5.5 `April 29, 2021`
+-----------------------------------
+
+New Functionality:
+
+- Added ``.set_regression_parameters(tol=None)`` to parameterize tolerances in AMuSR regression
+
+Code Refactoring:
+
+- Profiled and optimized AMuSR code
+
 Inferelator v0.5.4 `April 23, 2021`
 -----------------------------------
 

docs/conf.py

@@ -23,7 +23,7 @@
 author = 'Chris Jackson'
 
 # The full version, including alpha/beta/rc tags
-release = 'v0.5.4'
+release = 'v0.5.5'
 
 
 # -- General configuration ---------------------------------------------------

inferelator/crossvalidation_workflow.py

@@ -58,6 +58,7 @@ class CrossValidationManager(object):
     size_sample_stratified_column = None
     size_sample_with_replacement = False
     size_sample_seed = None
+    size_sample_only = False
 
     # Workflow storage
     _baseline_workflow = None
@@ -180,7 +181,8 @@ def add_grouping_dropin(self, metadata_column_name, group_size=None, seed=42):
         self.dropin_max_size = group_size
         self.dropin_seed = seed
 
-    def add_size_subsampling(self, size_vector, stratified_column_name=None, with_replacement=False, seed=42):
+    def add_size_subsampling(self, size_vector, stratified_column_name=None, with_replacement=False, seed=42,
+                             size_sample_only=None):
         """
         Resample expression data to a ratio of the original data.
 
@@ -206,6 +208,7 @@ def add_size_subsampling(self, size_vector, stratified_column_name=None, with_re
         self.size_sample_stratified_column = stratified_column_name
         self.size_sample_with_replacement = with_replacement
         self.size_sample_seed = seed
+        self.size_sample_only = size_sample_only if size_sample_only is not None else self.size_sample_only
 
     def run(self):
         """
@@ -222,7 +225,10 @@ def run(self):
         self._check_grid_search_params_exist()
 
         # Run base grid search
-        results = self._grid_search()
+        if self.size_sample_only:
+            results = []
+        else:
+            results = self._grid_search()
 
         # Run size sampling
         if self.size_sample_vector is not None:
@@ -382,8 +388,7 @@ def _grid_search(self, test=None, value=None, mask_function=None):
             # Drop any observations which are False in the mask (if set)
             if mask_function is not None:
                 mask = mask_function()
-                cv_workflow.expression_matrix.drop(cv_workflow.expression_matrix.columns[~mask], axis=1, inplace=True)
-                cv_workflow.meta_data.drop(cv_workflow.meta_data.index[~mask], axis=0, inplace=True)
+                cv_workflow.data = cv_workflow.data.get_sample_data(mask.index[mask])
                 n_obs = mask.sum()
             else:
                 n_obs = cv_workflow._num_obs
@@ -448,7 +453,7 @@ def _check_metadata_column_exists(self, col_name):
         :param col_name: str
         """
 
-        if col_name in self.workflow.meta_data.columns:
+        if col_name in self.workflow.data.meta_data.columns:
             return True
         else:
             raise ValueError("Column {col} is not present in the loaded metadata".format(col=col_name))
@@ -458,7 +463,7 @@ def _dropout_cv(self):
         Run grid search on all data minus one group at a time
         """
 
-        meta_data = self.workflow.meta_data.copy()
+        meta_data = self.workflow.data.meta_data.copy()
         col = self.dropout_column
         max_size = self.dropout_max_size
 
@@ -511,7 +516,7 @@ def _dropin_cv(self):
         Run grid search on one group from the data at a time
         """
 
-        meta_data = self.workflow.meta_data.copy()
+        meta_data = self.workflow.data.meta_data.copy()
         col = self.dropin_column
         max_size = self.dropin_max_size
 
@@ -551,7 +556,7 @@ def _size_cv(self):
 
         for i, size_ratio in enumerate(self.size_sample_vector):
             rgen = np.random.RandomState(self.size_sample_seed + i)
-            meta_data = self.workflow.meta_data.copy()
+            meta_data = self.workflow.data.meta_data.copy()
 
             if self.size_sample_stratified_column is not None:
                 strat_col = self.size_sample_stratified_column

inferelator/distributed/dask_functions.py

@@ -16,7 +16,8 @@
 
 
 def amusr_regress_dask(X, Y, priors, prior_weight, n_tasks, genes, tfs, G, remove_autoregulation=True,
-                       lambda_Bs=None, lambda_Ss=None, Cs=None, Ss=None, regression_function=None):
+                       lambda_Bs=None, lambda_Ss=None, Cs=None, Ss=None, regression_function=None, 
+                       tol=None, rel_tol=None):
     """
     Execute multitask (AMUSR)
 
@@ -54,7 +55,7 @@ def regression_maker(j, x_df, y_list, prior, tf):
 
         prior = format_prior(prior, gene, tasks, prior_weight, tfs=tf)
         return j, regression_function(x, y, tf, tasks, gene, prior,
-                                      lambda_Bs=lambda_Bs, lambda_Ss=lambda_Ss, Cs=Cs, Ss=Ss)
+                                      lambda_Bs=lambda_Bs, lambda_Ss=lambda_Ss, Cs=Cs, Ss=Ss, tol=tol, rel_tol=rel_tol)
 
     def response_maker(y_df, i):
         y = []