mend

tdc/model_server/tokenizers/geneformer.py

@@ -102,11 +102,11 @@ def tokenize_cell_vectors(self,
             X_norm = (X_view / n_counts * target_sum / norm_factor_vector)
             X_norm = sp.csr_matrix(X_norm)
 
-            tokenized_cells += [
+            tokenized_cells.append([
                 self.rank_genes(X_norm[i].data,
                                 coding_miRNA_tokens[X_norm[i].indices])
                 for i in range(X_norm.shape[0])
-            ]
+            ])
 
             # add custom attributes for subview to dict
             if self.custom_attr_name_dict is not None:

tdc/test/test_model_server.py

@@ -103,7 +103,8 @@ def testGeneformerTokenizer(self):
         geneformer = tdc_hf_interface("Geneformer")
         model = geneformer.load()
         input_tensor = torch.tensor(cells)
-        raise Exception("shape is", input_tensor.shape, "values are\n", input_tensor)
+        input_tensor = torch.squeeze(input_tensor)
+        # raise Exception("shape is", input_tensor.shape, "values are\n", input_tensor)
         out = model(input_tensor)
         assert out, "FAILURE: Geneformer output is false-like. Value = {}".format(out)
         assert len(out) == len(cells), "FAILURE: Geneformer output and cells input don't have the same length. {} vs {}".format(len(out), len(cells))