test: update test cases for lambdaSpeechToScore.get_speech_to_score_tuple because of a new variable in the return tuple

tests/lambdas/test_lambdaSpeechToScore.py

@@ -68,7 +68,8 @@ def check_output_by_field(output, key, match, expected_output):
     return output
 
 
-def check_output(self, output, expected_output):
+def check_output(self, output, expected_output, check_audio_files=False):
+    from pathlib import Path
     self.maxDiff = None
     try:
         assert len(output["matched_transcripts"].strip()) > 0
@@ -92,7 +93,19 @@ def check_output(self, output, expected_output):
         output["ipa_transcript"] = expected_output["ipa_transcript"]
         output["real_transcript"] = expected_output["real_transcript"]
         output["real_transcripts_ipa"] = expected_output["real_transcripts_ipa"]
-        self.assertDictEqual(expected_output, output)
+        if check_audio_files:
+            audio_files = output["audio_files"]
+            for audio_file in audio_files:
+                path_audio_file = Path(audio_file)
+                app_logger.info(f"path_audio_file:{path_audio_file}.")
+                assert path_audio_file.is_file()
+                path_audio_file.unlink()
+        
+            output["audio_files"] = [*audio_files]
+            expected_output["audio_files"] = [*audio_files]
+            app_logger.info(f"output audio_files:{output['audio_files']}.")
+            app_logger.info(f"expected_output audio_files:{expected_output['audio_files']}.")
+        self.assertDictEqual(output, expected_output)
     except Exception as e:
         app_logger.error(f"e:{e}.")
         raise e
@@ -196,6 +209,7 @@ class TestGetAccuracyFromRecordedAudio(unittest.TestCase):
             pronunciation_accuracy,
             ipa_transcript,
             real_transcripts_ipa,
+            num_words,
             dumped,
         ) = lambdaSpeechToScore.get_speech_to_score_tuple(
             real_text=text_dict[language],
@@ -209,7 +223,8 @@ class TestGetAccuracyFromRecordedAudio(unittest.TestCase):
         assert pronunciation_accuracy <= 100
         assert len(ipa_transcript.strip()) > 0
         assert len(real_transcripts_ipa.strip()) > 0
-        check_output(self, json.loads(dumped), expected_output[language])
+        assert num_words == 6
+        check_output(self, json.loads(dumped), expected_output[language], check_audio_files=True)
 
     def test_get_speech_to_score_tuple_en_ok(self):
         from aip_trainer.lambdas import lambdaSpeechToScore
@@ -222,6 +237,7 @@ class TestGetAccuracyFromRecordedAudio(unittest.TestCase):
             pronunciation_accuracy,
             ipa_transcript,
             real_transcripts_ipa,
+            num_words,
             dumped,
         ) = lambdaSpeechToScore.get_speech_to_score_tuple(
             real_text=text_dict[language],
@@ -235,7 +251,8 @@ class TestGetAccuracyFromRecordedAudio(unittest.TestCase):
         assert pronunciation_accuracy <= 100
         assert len(ipa_transcript.strip()) > 0
         assert len(real_transcripts_ipa.strip()) > 0
-        check_output(self, json.loads(dumped), expected_output[language])
+        assert num_words == 5
+        check_output(self, json.loads(dumped), expected_output[language], check_audio_files=True)
 
     def test_get_speech_to_score_dict__de_empty_input_text(self):
         language = "de"