Fix flaky NL test

Change-Id: I064c59e0c4f6d9b5ff0c888353df860dc344f74b

Author: Jon Wayne Parrott

language/snippets/ocr_nl/main_test.py

@@ -59,7 +59,7 @@ def test_text_returns_entities():
     text = "Holmes and Watson walked to the cafe."
     text_analyzer = main.TextAnalyzer()
     entities = text_analyzer.nl_detect(text)
-    assert len(entities) == 2
+    assert entities
     etype, ename, salience, wurl = text_analyzer.extract_entity_info(
         entities[0])
     assert ename == 'holmes'