added error checking to return smoothed anndatas

gssnng/score_cells.py

@@ -81,7 +81,7 @@ def run_gssnng(
 
     samp_neighbors = None
     error_checking(mat, samp_neighbors, recompute_neighbors,
-                   gs_obj, score_method, ranked, method_params)
+                   gs_obj, score_method, ranked, method_params, 0)
 
     if method_params == None:
         method_params = dict()
@@ -154,7 +154,7 @@ def with_gene_sets(
 
     samp_neighbors = None
     error_checking(adata, samp_neighbors, recompute_neighbors,
-                   gs_obj, score_method, ranked, method_params)
+                   gs_obj, score_method, ranked, method_params, 0)
 
     if method_params == None:
         method_params = dict()

gssnng/smooth_anndatas.py

@@ -1,6 +1,6 @@
 import anndata
 from gssnng.score_cells import _proc_data
-#from gssnng.util import error_checking
+from gssnng.util import error_checking
 from typing import Union
 
 def smooth_anndata(
@@ -34,16 +34,18 @@ def smooth_anndata(
 
     return_data = 1
     noise_trials = 0  ### not used currently
-    samp_neighbors = None
+    samp_neighbors = None ### also not used
+    just_smoothing=1
 
-    #error_checking2(adata, recompute_neighbors, method_params)  # UPDATE
+    error_checking(adata, samp_neighbors, recompute_neighbors,
+                   None, None, None, method_params, just_smoothing)
 
     if method_params == None:
         method_params = dict()
 
     # score each cell with the list of gene sets
     data_list = _proc_data(adata, None, groupby, smooth_mode, recompute_neighbors,
-                                  None, method_params, None,
+                                  None, method_params, samp_neighbors,
                                   noise_trials, None, cores, return_data)
 
     print("**done**")

gssnng/util.py

@@ -14,7 +14,8 @@ def error_checking(
         gs_obj,
         score_method,
         ranked,
-        method_params
+        method_params,
+        just_smoothing
 ):
     """
     QC on the adata. Need to make sure there's enough neighbors available given the sampling size.
@@ -23,30 +24,33 @@ def error_checking(
     :param samp_neighbors: integer, number of neighbors to sample
     """
 
-    if type(method_params) != type(dict()):
-        raise Exception('ERROR: please use a dictionary to pass method params')
-
-    if any([xi in adata.obs.columns for xi in gs_obj.get_gs_names()]):
-        #raise Exception('ERROR: gene set names in columns of adata.obs, please drop.')
-        print("Warning! Dropping gene set names from obs!")
-        genesetlist = [x.name for x in gs_obj.set_list]
-        for gsi in genesetlist:
-            print('dropping: ' + gsi)
-            adata.obs.drop(columns=[gsi], inplace=True)
-
     if 'gssnng_groupby' in adata.obs.columns:
         adata.obs.drop(columns='gssnng_groupby', inplace=True)
         #raise Exception("Error: please drop 'gssnng_groupby' as a column name.")
         print('... and dropping gssnng_groupby column...')
 
-    if ranked == False and score_method == 'singscore':
-        raise Exception('ERROR: singscore requires ranked data, set ranked parameter to True')
-
     if (recompute_neighbors == None) or (recompute_neighbors == 0):
         n_neighbors = adata.uns['neighbors']['params']['n_neighbors'] #[0]# in older AnnData versions need this??
     else:
         n_neighbors = recompute_neighbors
 
+    if just_smoothing == 0:
+        # then do all other checks
+        if type(method_params) != type(dict()):
+            raise Exception('ERROR: please use a dictionary to pass method params')
+
+        if any([xi in adata.obs.columns for xi in gs_obj.get_gs_names()]):
+            #raise Exception('ERROR: gene set names in columns of adata.obs, please drop.')
+            print("Warning! Dropping gene set names from obs!")
+            genesetlist = [x.name for x in gs_obj.set_list]
+            for gsi in genesetlist:
+                print('dropping: ' + gsi)
+                adata.obs.drop(columns=[gsi], inplace=True)
+
+        if ranked == False and score_method == 'singscore':
+            raise Exception('ERROR: singscore requires ranked data, set ranked parameter to True')
+
+
     #if n_neighbors < samp_neighbors:
     #    print('*******')
     #    print('WARNING: Number of neighbors too low for sampling parameter!')