Fetch recently played episodes from Overcast.fm. Then, publish history to my website.

overcast-recently-played.py

'''
You'll need to pip install some dependencies:

* python-dateutil
* requests

Also, populate your EMAIL and PASSWORD below.
'''

from xml.etree import ElementTree
from datetime import datetime
from dateutil.tz import UTC
from dateutil.parser import parse as parse_dt

import conf
import re
import sys
import requests
import pickle
import os.path
import json



# load stored session, or re-authenticate
if os.path.exists(conf.SESSION_PATH):
    print('Found saved session. Restoring!')
    session = pickle.loads(open(conf.SESSION_PATH, 'rb').read())
else:
    print('No saved session. Authenticating!')
    session = requests.Session()
    response = session.post('https://overcast.fm/login', data={
        'email': conf.EMAIL,
        'password': conf.PASSWORD
    })

    if response.status_code != 200:
        print('Authentication failed')
        sys.exit(0)

    print('Authenticated successfully. Saving session.')

    with open(conf.SESSION_PATH, 'wb') as saved_session:
        saved_session.write(pickle.dumps(session))

# fetch the latest detailed OPML export from Overcast
print('Fetching latest OPML export from Overcast')
response = session.get('https://overcast.fm/account/export_opml/extended')
if response.status_code != 200:
    print('Failed to fetch OPML. Exiting.')
    print(response.text)
    print(response.headers)
    sys.exit(0)

# cache the last OPML file
try:
    with open('/tmp/overcast.opml', 'w') as f:
        f.write(response.text)
except:
    print('Unable to cache OPML file.')

# parse the OPML
tree = ElementTree.fromstring(response.text)

# find all podcasts and their episodes
podcasts = tree.findall(".//*[@type='rss']")

# look for recently played episodes
now = datetime.utcnow().astimezone(UTC)
for podcast in podcasts:
    pod_title = podcast.attrib['title']
    for episode in list(podcast):
        # skip unplayed episodes
        played = episode.attrib.get('played', '0') == '1'
        if not played:
            continue

        # skip episodes played over 5 days ago
        user_activity_date_raw = episode.attrib.get('userUpdatedDate')
        user_activity_date = parse_dt(user_activity_date_raw)
        recency = now - user_activity_date
        if recency.days > 5:
            continue

        # parse out the remaining details we care about
        title = episode.attrib['title']
        published = parse_dt(episode.attrib['pubDate'])
        url = episode.attrib['url']
        overcast_url = episode.attrib['overcastUrl']
        overcast_id = episode.attrib['overcastId']
        progress = episode.attrib.get('progress')

        # fetch the epside artwork
        response = session.get(overcast_url)
        results = re.findall('img class="art fullart" src="(.*)"', response.text)
        artwork_url = ''
        if len(results) == 1:
            artwork_url = results[0]
        else:
            print('Cannot find podcast artwork for this show... skipping...')

        # fetch the episode summary
        results = re.findall('meta name="og:description" content="(.*)"', response.text)
        summary = title
        if len(results) == 1 and len(results[0]):
            summary = results[0]

        # publish if needed
        footprint = conf.HISTORY_PATH + '/' + overcast_id
        if os.path.exists(footprint):
            print('Skipping already published ep ->', title)
            continue

        print('Played episode of ', pod_title)
        print('    ->', title)
        print('    ->', summary)
        print('    ->', artwork_url)
        print('    ->', url)
        print('    ->', overcast_url)
        print('    ->', user_activity_date_raw)

        # build payload
        data = {
            'title': title,
            'summary': summary,
            'type': 'podcast',
            'author': pod_title,
            'link': overcast_url,
            'listenDateTime': user_activity_date_raw
        }

        poster_response = requests.get(artwork_url)
        response = requests.post(conf.ENDPOINT_URL, data={
            'payload': json.dumps(data)
        }, files={
            'photo': (
                artwork_url.rsplit('/', 1)[1],
                poster_response.content,
                poster_response.headers['Content-Type'],
                {'Expires': '0'}
            )
        })

        if response.status_code in (200, 201, 202):
            open(footprint, 'w').write(json.dumps(data))
            print('Successfully published!')
            print(response.headers)
        else:
            print('Failed to publish!')
            print(response.status_code)
            print(response.text)

Hello,

is extended opml download broken somehow? i can do it via browser, but with this python script, i only get a file which is not the xml i was looking for. It seems that the authentication is not valid any longer after the session is saved ... any hints?

@StrikeZW everything seems to be working fine for me. Have you created a conf.py file with SESSION_PATH, EMAIL, and PASSWORD, HISTORY_PATH, and ENDPOINT_URL defined?