#346 Prevent double nesting of features field

backend/models.py

@@ -971,17 +971,20 @@ def get_doc_annotation_dict(self, json_format="raw", anonymize=True):
         if json_format == "raw" or json_format == "csv":
             doc_dict = self.data
         elif json_format == "gate":
+            # GATE json format are expected to have an existing "features" field
+            features_dict = self.data["features"] if "features" in self.data and isinstance(self.data["features"], dict) else {}
 
-            ignore_keys = {"text", self.project.document_id_field}
-            features_dict = {key: value for key, value in self.data.items() if key not in ignore_keys}
+            # Add any non-compliant top-level fields into the "features" field instead
+            ignore_keys = {"text", "features", self.project.document_id_field}
+            features_dict.update({key: value for key, value in self.data.items() if key not in ignore_keys})
 
             doc_dict = {
                 "text": self.data["text"],
                 "features": features_dict,
                 "offset_type": "p",
                 "name": get_value_from_key_path(self.data, self.project.document_id_field)
             }
-            pass
+
 
         # Insert annotation sets into the doc dict
         annotations = self.annotations.filter(status=Annotation.COMPLETED)

backend/tests/test_models.py

@@ -1111,6 +1111,11 @@ def setUp(self):
                     "feature1": "Testvalue 1",
                     "feature2": "Testvalue 1",
                     "feature3": "Testvalue 1",
+                    "features": {
+                        "gate_format_feature1": "Gate feature test value",
+                        "gate_format_feature2": "Gate feature test value",
+                        "gate_format_feature3": "Gate feature test value",
+                    }
 
                 }
             )
@@ -1146,19 +1151,28 @@ def setUp(self):
     def test_export_raw(self):
 
         for document in self.project.documents.all():
+            # Fields should remain exactly the same as what's been uploaded
+            # aside from  annotation_sets
             doc_dict = document.get_doc_annotation_dict("raw")
             print(doc_dict)
             self.assertTrue("id" in doc_dict)
             self.assertTrue("text" in doc_dict)
             self.assertTrue("feature1" in doc_dict)
             self.assertTrue("feature2" in doc_dict)
             self.assertTrue("feature3" in doc_dict)
+            self.assertTrue("features" in doc_dict)
+            doc_features = doc_dict["features"]
+            self.assertTrue("gate_format_feature1" in doc_features)
+            self.assertTrue("gate_format_feature2" in doc_features)
+            self.assertTrue("gate_format_feature3" in doc_features)
 
             self.check_raw_gate_annotation_formatting(doc_dict)
 
     def test_export_gate(self):
 
         for document in self.project.documents.all():
+            # All top-level fields apart from name, text, features and annotation_sets should be
+            # nested inside the features field
             doc_dict = document.get_doc_annotation_dict("gate")
             print(doc_dict)
 
@@ -1169,6 +1183,10 @@ def test_export_gate(self):
             self.assertTrue("feature1" in doc_features)
             self.assertTrue("feature2" in doc_features)
             self.assertTrue("feature3" in doc_features)
+            self.assertFalse("features" in doc_features, "Double nesting of features field")
+            self.assertTrue("gate_format_feature1" in doc_features)
+            self.assertTrue("gate_format_feature2" in doc_features)
+            self.assertTrue("gate_format_feature3" in doc_features)
 
             self.check_raw_gate_annotation_formatting(doc_dict)