Improve values handling in index and deindex methods (fixes #794)

addok/helpers/index.py

@@ -125,10 +125,12 @@ def index(pipe, key, doc, tokens, **kwargs):
                 if callable(boost):
                     boost = boost(doc)
                 boost = boost + importance
-                if not isinstance(values, (list, tuple)):
+                if isinstance(values, str):
                     values = [values]
+                if not all(isinstance(item, str) for item in values):
+                    raise ValueError("{} must not be a string or an array of strings".format(name))
                 for value in values:
-                    extract_tokens(tokens, str(value), boost=boost)
+                    extract_tokens(tokens, value, boost=boost)
         index_tokens(pipe, tokens, key, **kwargs)
 
     @staticmethod
@@ -139,8 +141,10 @@ def deindex(db, key, doc, tokens, **kwargs):
                 continue
             values = doc.get(name)
             if values:
-                if not isinstance(values, (list, tuple)):
+                if isinstance(values, str):
                     values = [values]
+                if not all(isinstance(item, str) for item in values):
+                    raise ValueError("{} must not be a string or an array of strings".format(name))
                 for value in values:
                     tokens.extend(deindex_field(key, value))
 

tests/test_index_utils.py

@@ -279,6 +279,24 @@ def test_doc_with_null_value_should_not_be_index_if_not_allowed(config):
     index_document(doc)
     assert not DB.exists("w|cergy")
 
+
+def test_doc_with_array_of_array_value_should_not_be_index_if_not_allowed(config):
+    config.FIELDS = [
+        {"key": "name", "null": False},
+        {"key": "city"},
+    ]
+    doc = {
+        "id": "xxxx",
+        "_id": "yyyy",
+        "lat": "49.32545",
+        "lon": "4.2565",
+        "name": [["Lilas"]],
+        "city": "Cergy",
+    }
+    index_document(doc)
+    assert not DB.exists("w|cergy")
+
+
 def test_create_edge_ngrams(config):
     config.MIN_EDGE_NGRAMS = 2
     config.INDEX_EDGE_NGRAMS = False