Fix labels to validate exported as floats (#1636)

asreview · Jan 5, 2024 · bc8558d · bc8558d
1 parent 07ec269
commit bc8558d
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 8 deletions.
diff --git a/asreview/data/base.py b/asreview/data/base.py
@@ -476,16 +476,19 @@ def to_dataframe(self, labels=None, ranking=None, keep_old_labels=False):
 
         # if there are labels, add them to the frame
         if labels is not None:
-            # unnest the nested (record_id, label) tuples
+            # unnest list of nested (record_id, label) tuples
             labeled_record_ids = [x[0] for x in labels]
             labeled_values = [x[1] for x in labels]
 
             if keep_old_labels:
-                result_df["asreview_label_to_validate"] = result_df[col_label]
+                result_df["asreview_label_to_validate"] = \
+                    result_df[col_label].replace(LABEL_NA, None).astype("Int64")
 
             # remove the old results and write the values
             result_df[col_label] = LABEL_NA
             result_df.loc[labeled_record_ids, col_label] = labeled_values
+            result_df[col_label] = result_df[col_label] \
+                .replace(LABEL_NA, None).astype("Int64")
 
         # if there is a ranking, apply this ranking as order
         if ranking is not None:
@@ -494,11 +497,6 @@ def to_dataframe(self, labels=None, ranking=None, keep_old_labels=False):
             # append a column with 1 to n
             result_df["asreview_ranking"] = np.arange(1, len(result_df) + 1)
 
-        # replace labeled NA values by np.nan
-        if col_label in list(result_df):
-            result_df[col_label] = result_df[col_label].astype(object)
-            result_df.loc[result_df[col_label] == LABEL_NA, col_label] = np.nan
-
         return result_df
 
     def duplicated(self, pid="doi"):

diff --git a/asreview/webapp/api/projects.py b/asreview/webapp/api/projects.py
@@ -1179,7 +1179,7 @@ def api_export_dataset(project):
         if "asreview_prior" in as_data.df:
             as_data.df.drop("asreview_prior", axis=1, inplace=True)
 
-        state_df["asreview_prior"] = state_df.query_strategy.eq("prior").astype(int)
+        state_df["asreview_prior"] = state_df.query_strategy.eq("prior").astype("Int64")
         as_data.df = as_data.df.join(state_df["asreview_prior"], on="record_id")
 
         # Adding Notes from State file to the exported dataset