Compare commits

..

No commits in common. "97267c5f294893b3d72b44ea4f450ab21da6fee2" and "d27dd89a1fcd0ca158d7045f875c6365dc59a227" have entirely different histories.

108
check.py
View file

@ -6,41 +6,12 @@ import configparser
import os import os
import re import re
import smtplib import smtplib
import typing
import warnings import warnings
from datetime import date
from email.mime.text import MIMEText from email.mime.text import MIMEText
from email.utils import formatdate, make_msgid from email.utils import formatdate, make_msgid
import requests import requests
import yaml
from requests.adapters import HTTPAdapter
from urllib3.exceptions import InsecureRequestWarning # type: ignore from urllib3.exceptions import InsecureRequestWarning # type: ignore
from urllib3.util.url import parse_url # type: ignore
class AbsoluteDNSAdapter(HTTPAdapter):
"""A custom adapter for requests to ensure hostnames are treated as absolute."""
def add_dot_to_hostname(self, url: str) -> str:
"""Append a dot to the hostname to treat it as an absolute domain name."""
parsed_url = parse_url(url)
# Append a dot to the hostname if it's not already there.
hostname = parsed_url.host
if not hostname.endswith("."):
hostname += "."
# Reconstruct the URL with the modified hostname.
new_url: str = parsed_url._replace(host=hostname).url
return new_url
def send(self, request, **kwargs): # type: ignore
"""Override the send method to modify the request URL before sending."""
# Modify the request URL to ensure the hostname is treated as absolute.
request.url = self.add_dot_to_hostname(request.url)
return super().send(request, **kwargs)
config_file_path = os.path.expanduser( config_file_path = os.path.expanduser(
os.path.join( os.path.join(
@ -64,11 +35,10 @@ headers = {"User-Agent": AGENT, "Accept": "text/html"}
s = requests.Session() s = requests.Session()
s.headers.update(headers) s.headers.update(headers)
# Create a session and mount the custom adapter for both HTTP and HTTPS requests. MAIL_FROM = "edward@4angle.com"
adapter = AbsoluteDNSAdapter() MAIL_TO_NAME = "Edward Betts"
s.mount("http://", adapter) MAIL_TO_ADDRESS = "edward@4angle.com"
s.mount("https://", adapter) SMTP_HOST = "4angle.com"
not_here_list = [ not_here_list = [
"The specified URL was not found.", "The specified URL was not found.",
@ -82,7 +52,31 @@ not_here_list = [
"This page doesn't exist (404)", "This page doesn't exist (404)",
"Coming soon", "Coming soon",
"NOT_FOUND", "NOT_FOUND",
"Resource Not Found", ]
conferences = [
# ("FOSDEM", "https://fosdem.org/2024"),
# ("PyCascades", "https://2024.pycascades.com/"),
# ("foss-north", "https://foss-north.se/2024"),
# ("Wikimedia Hackathon", "https://www.mediawiki.org/wiki/Wikimedia_Hackathon_2024"),
("FOSS4G", "https://2024.foss4g.org/"),
# ("FOSS4G Europe", "https://2024.europe.foss4g.org/"),
("FOSSY", "https://2024.fossy.us/"),
# ("North Bay Python", "https://2024.northbaypython.org/"),
# ("DebConf", "https://wiki.debian.org/DebConf/24"),
# ("State of the Map US", "https://2024.stateofthemap.us/"),
# ("WikiConference North America", "https://wikiconference.org/wiki/2024/Main_Page"),
# ("PyCon DE", "https://2024.pycon.de/"),
# ("PyData London", "https://pydata.org/london2024"),
# ("Pass the SALT", "https://2024.pass-the-salt.org/"),
("SotM Baltics", "https://2024.sotm-baltics.org/"),
("EuroSciPy", "https://www.euroscipy.org/2024/"),
# ("EuroPython", "https://ep2024.europython.eu/"),
("Semantic Web in Libraries", "https://swib.org/swib24/"),
("SotM Africa", "https://2024.stateofthemap.africa/"),
("FOSS4G Oceania", "https://2024.foss4g-oceania.org/"),
# ("All Things Open", "https://2024.allthingsopen.org/"),
("GLAMhack24", "https://opendata.ch/events/glamhack2024/"),
] ]
@ -111,15 +105,12 @@ def check_conference(name: str, url: str) -> tuple[bool, str]:
def send_mail(subject: str, body: str) -> None: def send_mail(subject: str, body: str) -> None:
"""Send an e-mail.""" """Send an e-mail."""
mail_from_address = config["mail"]["from_address"] mail_from = MAIL_FROM
mail_from_name = config["mail"]["from_name"]
mail_to_address = config["mail"]["to_address"]
mail_to_name = config["mail"]["to_name"]
msg = MIMEText(body, "plain", "UTF-8") msg = MIMEText(body, "plain", "UTF-8")
msg["Subject"] = subject msg["Subject"] = subject
msg["To"] = f"{mail_to_name} <{mail_to_address}>" msg["To"] = f"{MAIL_TO_NAME} <{MAIL_TO_ADDRESS}>"
msg["From"] = f"{mail_from_name} <{mail_from_address}>" msg["From"] = f"Edward Betts <{mail_from}>"
msg["Date"] = formatdate() msg["Date"] = formatdate()
msg["Message-ID"] = make_msgid() msg["Message-ID"] = make_msgid()
@ -127,49 +118,20 @@ def send_mail(subject: str, body: str) -> None:
for header_name, value in config["mail_headers"].items(): for header_name, value in config["mail_headers"].items():
msg[header_name] = value msg[header_name] = value
s = smtplib.SMTP(config["mail"]["smtp_host"]) s = smtplib.SMTP(SMTP_HOST)
s.sendmail(mail_from_address, [mail_to_address], msg.as_string()) s.sendmail(mail_from, [MAIL_TO_ADDRESS], msg.as_string())
s.quit() s.quit()
def load_yaml(name: str) -> typing.Any:
"""Load YAML."""
filename = os.path.expanduser(config["data"][name])
assert os.path.exists(filename)
return yaml.safe_load(open(filename))
def main(show_not_live: bool = False) -> None: def main(show_not_live: bool = False) -> None:
"""Check each conference.""" """Check each conference."""
today = date.today() for name, url in conferences:
this_year = today.year
conferences = load_yaml("conferences")
live_conferences = load_yaml("live")
live_set = {(c["conference"], c["year"]) for c in live_conferences}
new_live = False
for name, src_url in conferences.items():
for year in this_year, this_year + 1:
if (name, year) in live_set:
continue
assert "{year}" in src_url
url = src_url.format(year=year)
live, msg = check_conference(name, url) live, msg = check_conference(name, url)
if not live: if not live:
continue continue
body = f"{name}\n{url}\nWeb page title: {msg}" body = f"{name}\n{url}\nWeb page title: {msg}"
send_mail(f"Conference site live: {name}", body) send_mail(f"Conference site live: {name}", body)
new_live = True
live_conferences.append({"conference": name, "year": year, "live": today})
if new_live:
live_filename = os.path.expanduser(config["data"]["live"])
with open(live_filename, "w") as out:
yaml.dump(live_conferences, stream=out, sort_keys=False)
if __name__ == "__main__": if __name__ == "__main__":
main() main()