lambda_function.py 文件源码

def lambda_handler(event, context):
    rss = event['rss']
    bucket_name = event['bucket']
    logging.info("Processing url: %s" % rss)
    logging.info("Using bucket: %s" % bucket_name)

    session = Session(region_name="us-west-2")
    polly = session.client("polly")
    s3 = resource('s3')
    bucket = s3.Bucket(bucket_name)

    logging.info("getting list of existing objects in the given bucket")
    files = set(o.key for o in bucket.objects.all())

    feed = feedparser.parse(rss)

    title = feed['feed']['title']
    fg = FeedGenerator()
    fg.load_extension('podcast')
    fg.title('Audio podcast based on: %s' % title)
    fg.link(href=feed.feed.link, rel='alternate')
    fg.subtitle(feed.feed.description)

    ENTRY_URL = "http://{bucket}.s3-website.{region}.amazonaws.com/{filename}"

    for entry in get_entries(feed):
        filename = "%s.mp3" % entry['id']
        fe = fg.add_entry()
        fe.id(entry['id'])
        fe.title(entry['title'])
        fe.published(entry['published'])
        entry_url = ENTRY_URL.format(bucket=bucket_name, filename=filename, region=os.environ["AWS_REGION"])
        fe.enclosure(entry_url, 0, 'audio/mpeg')
        if filename in files:
            logging.info('Article "%s" with id %s already exist, skipping.'
                         % (entry['title'], entry['id']))
            continue
        try:
            logging.info("Next entry, size: %d" % len(entry['content']))
            logging.debug("Content: %s" % entry['content'])
            response = polly.synthesize_speech(
                    Text=entry['content'],
                    OutputFormat="mp3",
                    VoiceId="Joanna")
            with closing(response["AudioStream"]) as stream:
                bucket.put_object(Key=filename, Body=stream.read())
        except BotoCoreError as error:
            logging.error(error)
    bucket.put_object(Key='podcast.xml', Body=fg.rss_str(pretty=True))