allow user to save an empty label dataframe (#1012)

server/app/scanpy_engine/scanpy_engine.py

@@ -350,34 +350,37 @@ class ScanpyEngine(CXGDriver):
                         )
 
     @requires_data
-    def _validate_label_data(self):
+    def _validate_label_data(self, labels=None):
         """
         labels is None if disabled, empty if enabled by no data
         """
-        if self.labels is None or self.labels.empty:
+        if labels is None:
+            labels = self.labels
+
+        if labels is None or labels.empty:
             return
 
         # all lables must have a name, which must be unique and not used in obs column names
-        if not self.labels.columns.is_unique:
+        if not labels.columns.is_unique:
             raise KeyError(f"All column names specified in {self.config['label_file']} must be unique.")
 
         # the label index must be unique, and must have same values the anndata obs index
-        if not self.labels.index.is_unique:
+        if not labels.index.is_unique:
             raise KeyError(f"All row index values specified in the label file "
                            f"`{self.config['label_file']}` must be unique.")
 
-        if not self.labels.index.equals(self.original_obs_index):
+        if not labels.index.equals(self.original_obs_index):
             raise KeyError("Label file row index does not match H5AD file index. "
                            "Please ensure that column zero (0) in the label file contain the same "
                            "index values as the H5AD file.")
 
-        duplicate_columns = list(set(self.labels.columns) & set(self.data.obs.columns))
+        duplicate_columns = list(set(labels.columns) & set(self.data.obs.columns))
         if len(duplicate_columns) > 0:
             raise KeyError(f"Labels file may not contain column names which overlap "
                            f"with h5ad obs columns {duplicate_columns}")
 
         # labels must have same count as obs annotations
-        if self.labels.shape[0] != self.data.obs.shape[0]:
+        if labels.shape[0] != self.data.obs.shape[0]:
             raise ValueError("Labels file must have same number of rows as h5ad file.")
 
     @staticmethod
@@ -467,8 +470,9 @@ class ScanpyEngine(CXGDriver):
             raise ValueError("Only OBS dimension access is supported")
 
         new_label_df = decode_matrix_fbs(fbs)
-        new_label_df.index = self.original_obs_index
-        self._validate_label_data()  # paranoia
+        if not new_label_df.empty:
+            new_label_df.index = self.original_obs_index
+        self._validate_label_data(labels=new_label_df)  # paranoia
 
         # if any of the new column labels overlap with our existing labels, raise error
         duplicate_columns = list(set(new_label_df.columns) & set(self.data.obs.columns))
@@ -477,7 +481,7 @@ class ScanpyEngine(CXGDriver):
                            f"with h5ad obs columns {duplicate_columns}")
 
         # update our internal state and save it.  Multi-threading often enabled,
-        # so treat this as a critical section critical section.
+        # so treat this as a critical section.
         with self.label_lock:
             self.labels = new_label_df
             write_labels(fname, self.labels)