def main()

in utils/import_conversations_v2.py [0:0]

• 113 lines of code
• 15 McCabe index (conditional complexity)

def main():
  pargs = _ParseArgs()
  project_id = pargs.project_id
  impersonated_service_account = pargs.impersonated_service_account
  insights_endpoint = pargs.insights_endpoint
  api_version = pargs.insights_api_version
  should_redact = pargs.redact
  agent_id = pargs.agent_id
  agent_channel = int(pargs.agent_channel)
  xml_bucket = pargs.xml_gcs_bucket
  analyze_conv = pargs.analyze
  folder_name = pargs.folder_name

  if pargs.source_local_audio_path or pargs.source_audio_gcs_bucket:
    # Inputs are audio files.
    dest_bucket = pargs.dest_gcs_bucket
    if pargs.source_local_audio_path:
      source_local_audio_path = pargs.source_local_audio_path
      source_audio_base_name = os.path.basename(source_local_audio_path)
      _UploadFileToGcs(
          dest_bucket,
          source_local_audio_path,
          source_audio_base_name,
          project_id,
          impersonated_service_account,
      )
      audio_uris = [_GetGcsUri(dest_bucket, source_audio_base_name)]
    elif pargs.source_audio_gcs_bucket:
      audio_uris = _GetGcsUris(
          pargs.source_audio_gcs_bucket,
          project_id,
          impersonated_service_account,
      )
      # print(audio_uris)
    else:
      audio_uris = []

    encoding = pargs.encoding
    language_code = pargs.language_code
    sample_rate_hertz = pargs.sample_rate_hertz

    # get processed audio files
    processed_audio_files = _GetGcsUris(
        dest_bucket, project_id, impersonated_service_account, uri=False
    )
    processed_audio_files = list(
        map(lambda x: x.split('.')[0], processed_audio_files)
    )

    # remove processed files
    audio_uris = _RemoveProcessedFiles(audio_uris, processed_audio_files)

    if not audio_uris:
      print('No audio file to process')
      conversation_names = []
    else:
      conversation_names = _ImportConversationsFromAudio(
          audio_uris,
          encoding,
          language_code,
          sample_rate_hertz,
          project_id,
          dest_bucket,
          insights_endpoint,
          api_version,
          should_redact,
          agent_id,
          impersonated_service_account,
          agent_channel,
          xml_bucket,
      )
  else:
    # Inputs are transcript files.
    if pargs.source_voice_transcript_gcs_bucket:
      medium = 1
      print(pargs.source_voice_transcript_gcs_bucket)
      transcript_bucket = pargs.source_voice_transcript_gcs_bucket
    elif pargs.source_chat_transcript_gcs_bucket:
      print(pargs.source_chat_transcript_gcs_bucket)
      transcript_bucket = pargs.source_chat_transcript_gcs_bucket
      medium = 2
    else:
      print('Provide at least one bucket for (Audio/Chat)')
      return

    transcript_metadata_flag = pargs.transcript_metadata_flag
    transcript_uris = _GetGcsUris(transcript_bucket, project_id,
                                  impersonated_service_account, folder_name)

    # get processed transcripts
    processed_transcripts_uris = _GetProcessedTranscripts(project_id)

    unprocessed_transcript_uris = []

    # filter processed transcripts
    if should_redact == 'True':
      for transcript_uri in transcript_uris:
        redacted_uri = _AddRedactedFolderToGcsUri(transcript_uri)
        if redacted_uri not in processed_transcripts_uris:
          unprocessed_transcript_uris.append(transcript_uri)
    else:
      unprocessed_transcript_uris = [
          i for i in transcript_uris if i not in processed_transcripts_uris
          ]
      print('Total transcripts', len(unprocessed_transcript_uris))

    with open('transcripts_uris', 'wb') as fp:
      pickle.dump(transcript_uris, fp)

    if not transcript_uris:
      print('No transcript to ingest')
      conversation_names = []
    else:
      conversation_names = _ImportConversationsFromTranscript(
          unprocessed_transcript_uris, project_id, medium, insights_endpoint,
          api_version, should_redact, agent_id,
          impersonated_service_account, agent_channel, xml_bucket, 
          transcript_bucket, transcript_metadata_flag)

  print(
      'Created `{}` conversation IDs: {}'.format(
          len(conversation_names), conversation_names
      )
  )

  if analyze_conv == 'True':
    print('Starting analysis for conversations.')
    _AnalyzeConversations(
        conversation_names,
        insights_endpoint,
        api_version,
        impersonated_service_account,
    )