epicyon/newsdaemon.py

144 lines
4.7 KiB
Python
Raw Normal View History

2020-10-07 12:05:49 +00:00
__filename__ = "newsdaemon.py"
__author__ = "Bob Mottram"
__license__ = "AGPL3+"
__version__ = "1.1.0"
__maintainer__ = "Bob Mottram"
__email__ = "bob@freedombone.net"
__status__ = "Production"
import os
2020-10-07 12:05:49 +00:00
import time
from newswire import getDictFromNewswire
from posts import createNewsPost
from utils import saveJson
2020-10-07 12:05:49 +00:00
def updateFeedsIndex(baseDir: str, filename: str) -> None:
"""Updates the index used for imported RSS feeds
"""
indexFilename = baseDir + '/accounts/feeds.index'
if os.path.isfile(indexFilename):
if filename not in open(indexFilename).read():
try:
with open(indexFilename, 'r+') as feedsFile:
content = feedsFile.read()
feedsFile.seek(0, 0)
feedsFile.write(filename + '\n' + content)
print('DEBUG: feeds post added to index')
except Exception as e:
print('WARN: Failed to write entry to feeds posts index ' +
indexFilename + ' ' + str(e))
else:
feedsFile = open(indexFilename, 'w+')
if feedsFile:
feedsFile.write(filename + '\n')
feedsFile.close()
def convertRSStoActivityPub(baseDir: str, httpPrefix: str,
domain: str, port: int,
newswire: {},
translate: {}) -> None:
"""Converts rss items in a newswire into posts
"""
basePath = baseDir + '/accounts/feeds'
if not os.path.isdir(basePath):
os.mkdir(basePath)
nickname = 'feeds'
for dateStr, item in newswire.items():
# convert the date to the format used by ActivityPub
dateStr = dateStr.replace(' ', 'T')
dateStr = dateStr.replace('+00:00', 'Z')
# file where the post is stored
filename = basePath + '/' + dateStr + '.json'
if os.path.isfile(filename):
# if a local post exists as html then change the link
# to the local one
htmlFilename = basePath + '/' + dateStr + '.html'
if os.path.isfile(htmlFilename):
item[1] = '/feeds/' + dateStr + '.html'
# don't create the post if it already exists
continue
rssTitle = item[0]
url = item[1]
2020-10-07 13:55:27 +00:00
rssDescription = ''
# get the rss description if it exists
2020-10-07 13:58:39 +00:00
if len(item) >= 5:
2020-10-07 13:55:27 +00:00
rssDescription = item[4]
# add the off-site link to the description
if rssDescription:
rssDescription += \
'\n\n' + translate['Read more...'] + '\n' + url
else:
rssDescription = url
# create the activitypub post
blog = createNewsPost(baseDir,
nickname, domain, port,
httpPrefix, dateStr,
rssTitle, rssDescription,
None, None, None, False)
# save the post and update the index
if saveJson(blog, filename):
updateFeedsIndex(baseDir, filename)
def runNewswireDaemon(baseDir: str, httpd,
httpPrefix: str, domain: str, port: int,
translate: {}) -> None:
2020-10-07 12:05:49 +00:00
"""Periodically updates RSS feeds
"""
# initial sleep to allow the system to start up
time.sleep(50)
while True:
# has the session been created yet?
if not httpd.session:
print('Newswire daemon waiting for session')
time.sleep(60)
continue
# try to update the feeds
newNewswire = None
try:
newNewswire = getDictFromNewswire(httpd.session, baseDir)
except Exception as e:
print('WARN: unable to update newswire ' + str(e))
time.sleep(120)
continue
httpd.newswire = newNewswire
print('Newswire updated')
convertRSStoActivityPub(baseDir,
httpPrefix, domain, port,
newNewswire, translate)
print('Newswire feed converted to ActivityPub')
2020-10-07 12:05:49 +00:00
# wait a while before the next feeds update
time.sleep(1200)
def runNewswireWatchdog(projectVersion: str, httpd) -> None:
"""This tries to keep the newswire update thread running even if it dies
"""
print('Starting newswire watchdog')
newswireOriginal = \
httpd.thrPostSchedule.clone(runNewswireDaemon)
httpd.thrNewswireDaemon.start()
while True:
time.sleep(50)
if not httpd.thrNewswireDaemon.isAlive():
httpd.thrNewswireDaemon.kill()
httpd.thrNewswireDaemon = \
newswireOriginal.clone(runNewswireDaemon)
httpd.thrNewswireDaemon.start()
print('Restarting newswire daemon...')