epicyon/schedule.py

255 lines
11 KiB
Python

__filename__ = "schedule.py"
__author__ = "Bob Mottram"
__license__ = "AGPL3+"
__version__ = "1.4.0"
__maintainer__ = "Bob Mottram"
__email__ = "bob@libreserver.org"
__status__ = "Production"
__module_group__ = "Calendar"
import os
import time
from utils import date_from_string_format
from utils import date_epoch
from utils import acct_handle_dir
from utils import has_object_dict
from utils import get_status_number
from utils import load_json
from utils import is_account_dir
from utils import acct_dir
from utils import remove_eol
from utils import date_utcnow
from outbox import post_message_to_outbox
from session import create_session
from threads import begin_thread
from siteactive import save_unavailable_sites
def _update_post_schedule(base_dir: str, handle: str, httpd,
max_scheduled_posts: int) -> None:
"""Checks if posts are due to be delivered and if so moves them to
the outbox
"""
schedule_index_filename = \
acct_handle_dir(base_dir, handle) + '/schedule.index'
if not os.path.isfile(schedule_index_filename):
return
# get the current time as an int
curr_time = date_utcnow()
days_since_epoch = (curr_time - date_epoch()).days
schedule_dir = acct_handle_dir(base_dir, handle) + '/scheduled/'
index_lines = []
delete_schedule_post = False
nickname = handle.split('@')[0]
with open(schedule_index_filename, 'r', encoding='utf-8') as sched_file:
for line in sched_file:
if ' ' not in line:
continue
date_str = line.split(' ')[0]
if 'T' not in date_str:
continue
post_id1 = line.split(' ', 1)[1]
post_id = remove_eol(post_id1)
post_filename = schedule_dir + post_id + '.json'
if delete_schedule_post:
# delete extraneous scheduled posts
if os.path.isfile(post_filename):
try:
os.remove(post_filename)
except OSError:
print('EX: _update_post_schedule unable to delete ' +
str(post_filename))
continue
# create the new index file
index_lines.append(line)
# convert string date to int
post_time = \
date_from_string_format(date_str, ["%Y-%m-%dT%H:%M:%S%z"])
post_time = post_time.replace(tzinfo=None)
post_days_since_epoch = \
(post_time - date_epoch()).days
if days_since_epoch < post_days_since_epoch:
continue
if days_since_epoch == post_days_since_epoch:
if curr_time.time().hour < post_time.time().hour:
continue
if curr_time.time().minute < post_time.time().minute:
continue
if not os.path.isfile(post_filename):
print('WARN: schedule missing post_filename=' + post_filename)
index_lines.remove(line)
continue
# load post
post_json_object = load_json(post_filename)
if not post_json_object:
print('WARN: schedule json not loaded')
index_lines.remove(line)
continue
# set the published time
# If this is not recent then http checks on the receiving side
# will reject it
_, published = get_status_number()
if post_json_object.get('published'):
post_json_object['published'] = published
if has_object_dict(post_json_object):
if post_json_object['object'].get('published'):
post_json_object['published'] = published
print('Sending scheduled post ' + post_id)
if nickname:
httpd.post_to_nickname = nickname
# create session if needed
curr_session = httpd.session
curr_proxy_type = httpd.proxy_type
if not curr_session:
curr_session = create_session(httpd.proxy_type)
httpd.session = curr_session
if not curr_session:
continue
if not post_message_to_outbox(curr_session,
httpd.translate,
post_json_object, nickname,
httpd, base_dir,
httpd.http_prefix,
httpd.domain,
httpd.domain_full,
httpd.onion_domain,
httpd.i2p_domain,
httpd.port,
httpd.recent_posts_cache,
httpd.followers_threads,
httpd.federation_list,
httpd.send_threads,
httpd.postLog,
httpd.cached_webfingers,
httpd.person_cache,
httpd.allow_deletion,
curr_proxy_type,
httpd.project_version,
httpd.debug,
httpd.yt_replace_domain,
httpd.twitter_replacement_domain,
httpd.show_published_date_only,
httpd.allow_local_network_access,
httpd.city, httpd.system_language,
httpd.shared_items_federated_domains,
httpd.shared_item_federation_tokens,
httpd.low_bandwidth,
httpd.signing_priv_key_pem,
httpd.peertube_instances,
httpd.theme_name,
httpd.max_like_count,
httpd.max_recent_posts,
httpd.cw_lists,
httpd.lists_enabled,
httpd.content_license_url,
httpd.dogwhistles,
httpd.min_images_for_accounts,
httpd.buy_sites,
httpd.sites_unavailable):
index_lines.remove(line)
try:
os.remove(post_filename)
except OSError:
print('EX: _update_post_schedule unable to delete ' +
str(post_filename))
continue
# move to the outbox
outbox_post_filename = \
post_filename.replace('/scheduled/', '/outbox/')
os.rename(post_filename, outbox_post_filename)
print('Scheduled post sent ' + post_id)
index_lines.remove(line)
if len(index_lines) > max_scheduled_posts:
delete_schedule_post = True
# write the new schedule index file
schedule_index_file = \
acct_handle_dir(base_dir, handle) + '/schedule.index'
with open(schedule_index_file, 'w+', encoding='utf-8') as schedule_file:
for line in index_lines:
schedule_file.write(line)
def run_post_schedule(base_dir: str, httpd, max_scheduled_posts: int):
"""Dispatches scheduled posts
"""
while True:
time.sleep(60)
# for each account
for _, dirs, _ in os.walk(base_dir + '/accounts'):
for account in dirs:
if '@' not in account:
continue
if not is_account_dir(account):
continue
# scheduled posts index for this account
schedule_index_filename = \
base_dir + '/accounts/' + account + '/schedule.index'
if not os.path.isfile(schedule_index_filename):
continue
_update_post_schedule(base_dir, account,
httpd, max_scheduled_posts)
break
def run_post_schedule_watchdog(project_version: str, httpd) -> None:
"""This tries to keep the scheduled post thread running even if it dies
"""
print('THREAD: Starting scheduled post watchdog ' + project_version)
post_schedule_original = \
httpd.thrPostSchedule.clone(run_post_schedule)
begin_thread(httpd.thrPostSchedule, 'run_post_schedule_watchdog')
curr_sites_unavailable = httpd.sites_unavailable.copy()
while True:
time.sleep(20)
# save the list of unavailable sites
if str(curr_sites_unavailable) != str(httpd.sites_unavailable):
save_unavailable_sites(httpd.base_dir, httpd.sites_unavailable)
curr_sites_unavailable = httpd.sites_unavailable.copy()
if httpd.thrPostSchedule.is_alive():
continue
httpd.thrPostSchedule.kill()
print('THREAD: restarting scheduled post watchdog')
httpd.thrPostSchedule = \
post_schedule_original.clone(run_post_schedule)
begin_thread(httpd.thrPostSchedule, 'run_post_schedule_watchdog')
print('Restarting scheduled posts...')
def remove_scheduled_posts(base_dir: str, nickname: str, domain: str) -> None:
"""Removes any scheduled posts
"""
# remove the index
schedule_index_filename = \
acct_dir(base_dir, nickname, domain) + '/schedule.index'
if os.path.isfile(schedule_index_filename):
try:
os.remove(schedule_index_filename)
except OSError:
print('EX: remove_scheduled_posts unable to delete ' +
schedule_index_filename)
# remove the scheduled posts
scheduled_dir = acct_dir(base_dir, nickname, domain) + '/scheduled'
if not os.path.isdir(scheduled_dir):
return
for scheduled_post_filename in os.listdir(scheduled_dir):
file_path = os.path.join(scheduled_dir, scheduled_post_filename)
if os.path.isfile(file_path):
try:
os.remove(file_path)
except OSError:
print('EX: remove_scheduled_posts unable to delete ' +
file_path)