Fix unittest errors and Update encoder naming convention (#1617)

* fix unittest
* change query encoder naming convention to python

pyserini/search/lucene/_impact_searcher.py

@@ -110,9 +110,11 @@ def from_prebuilt_index(cls, prebuilt_index_name: str, query_encoder: Union[Quer
         return cls(index_dir, query_encoder, min_idf, encoder_type, prebuilt_index_name=prebuilt_index_name)
 
     def encode(self, query):
-        if self.encoder_type == 'pytorch':
+        if self.encoder_type == 'onnx':
+            encoded_query = self.object.encode_with_onnx(query)
+        elif self.encoder_type == 'pytorch':
             encoded_query = self.query_encoder.encode(query)
-        else: raise ValueError(f'Invalid query encoder type: {type(query_encoder)} for encode')
+        else: raise ValueError(f'Invalid query encoder type: {type(self.query_encoder)} for encode')
         return encoded_query
 
     @staticmethod

tests/test_encoder.py

@@ -226,10 +226,10 @@ def test_onnx_encode_unicoil(self):
         temp_object = LuceneImpactSearcher(f'{self.index_dir}lucene9-index.cacm', 'SpladePlusPlusEnsembleDistil', encoder_type='onnx')
 
         # this function will never be called in _impact_searcher, here to check quantization correctness
-        results = temp_object.object.encodeWithOnnx("here is a test")
-        self.assertAlmostEqual(results.get("here"), 156, delta=2e-4)
-        self.assertAlmostEqual(results.get("a"), 31, delta=2e-4)
-        self.assertAlmostEqual(results.get("test"), 149, delta=2e-4)
+        results = temp_object.encode("here is a test")
+        self.assertEqual(results.get("here"), 156)
+        self.assertEqual(results.get("a"), 31)
+        self.assertEqual(results.get("test"), 149)
 
         temp_object.close()
         del temp_object