[speech-to-text] API generates invalid JSON transcription

[nevinchanyi]

[REQUIRED] Step 2: Describe your configuration

• Extension name: _ speech-to-text
• Extension version: _ googlecloud/speech-to-text@0.1.6 (slightly modified)
• Configuration values (redact info where appropriate):
  • _ language code is en-US
  • _ The modified index.js:
    https://gitlab.com/nevinchanyik/RED/-/blob/main/RED/GoogleCloudAPI/index.js?ref_type=heads
  • _ The modified transcribe-audio.js:
    https://gitlab.com/nevinchanyik/RED/-/blob/main/RED/GoogleCloudAPI/transcribe-audio.js?ref_type=heads
  • _ The link to the .m4a audio file that needs to be transcribed:
    https://firebasestorage.googleapis.com/v0/b/red-speech-to-text.firebasestorage.app/o/en-US%2FWt4XTR1eEodQYRBDkokFEBXJzNr1%2FWt4XTR1eEodQYRBDkokFEBXJzNr1-769399928.814137.m4a?alt=media&token=862e7332-1e7e-41ce-96e1-6cb507d7b677
  • _ The link to the .txt transcription file:
    https://firebasestorage.googleapis.com/v0/b/red-speech-to-text.firebasestorage.app/o/en-US%2FWt4XTR1eEodQYRBDkokFEBXJzNr1%2FWt4XTR1eEodQYRBDkokFEBXJzNr1-769399928.814137.m4a.wav_transcription.txt?alt=media&token=92dd3535-886c-4559-81d4-7637ce25e5c3

[REQUIRED] Step 3: Describe the problem

Steps to reproduce:

1. Upload the .m4a example audio file to a storage in the path such as: {lang_code}/{user_id}/audio.m4a
2. Wait for the transcription and check the result – it produces invalid JSON, which I cannot parse on the iOS app not with JSONDecoder, not with JSONSerialization.

Expected result

API generates a valid JSON transcription.

Actual result

iOS (Swift) cannot decode the invalid JSON file. I think this {} in the empty array might be the reason:

, {
    "alternatives": [ {
    
    } ],
    "languageCode": "en-us",
    "resultEndTime": "221.340s"
  },

[cabljac]

Hi there, appreciate this has been open for a while. I've marked this for reproduction and further investigation