!/usr/bin/env python3

"""Generate Gemini feed by Atom.

Documentation

Instance

"""

import xml.etree.ElementTree as ET

from datetime import datetime

from email.message import Message

from socket import create_connection

from ssl import SSLContext, CERT_NONE, PROTOCOL_TLS_CLIENT

from urllib.parse import urlsplit, urljoin, uses_relative, uses_netloc, unquote

for urljoin:

uses_relative.append("gemini")

uses_netloc.append("gemini")

def _is_gemini_link(value):

return value and value.lower().startswith("gemini://")

def _main(raw_url):

splitted_url = urlsplit(raw_url)

if splitted_url.scheme != "gemini":

print("59 Only Gemini links are allowed\r")

return

for _ in range(6):

with create_connection((splitted_url.hostname, splitted_url.port or 1965)) as raw_conn:

context = SSLContext(PROTOCOL_TLS_CLIENT)

context.check_hostname = False

context.verify_mode = CERT_NONE

with context.wrap_socket(raw_conn, server_hostname=splitted_url.hostname) as conn:

conn.sendall((raw_url + '\r\n').encode("UTF-8"))

fp = conn.makefile("rb")

splitted = fp.readline().decode("UTF-8").strip().split(maxsplit=1)

status = splitted[0]

if status.startswith("3") and len(splitted) == 2:

# redirect

raw_url = urljoin(raw_url, splitted[1])

continue

if not status.startswith("2"):

print(f"43 Remote server error: {' '.join(splitted)}\r")

return

mime = splitted[1].lower() if len(splitted) == 2 else "text/gemini"

mime_matched = (

mime.startswith("text/")

or mime.startswith("application/xml")

or mime.startswith("application/atom")

)

if not mime_matched:

print(f"43 Only links to `text/*` are allowed: {mime}\r")

return

m = Message()

m['content-type'] = mime

root = {}

try:

for child in ET.fromstring(fp.read().decode(m.get_param('charset') or "UTF-8")):

_, _, tag = child.tag.rpartition('}')

root.setdefault(tag, []).append(child)

except ET.ParseError:

print("43 Parse Atom error\r")

return

print("20 text/gemini\r")

title = root.get("title")

print(f"# {title[0].text if title else raw_url}\r")

print("\r")

for entry in root.get("entry") or []:

gemini_link = None

entry_dict = {}

for child in entry:

_, _, tag = child.tag.rpartition('}')

entry_dict[tag] = child.text

if tag == "link" and not _is_gemini_link(gemini_link):

gemini_link = urljoin(raw_url, child.attrib.get("href", ""))

if not _is_gemini_link(gemini_link):

continue

try:

entry_date = (

datetime.fromisoformat(entry_dict.get("updated", ""))

.date().strftime("%Y-%m-%d - ")

)

except ValueError:

entry_date = ""

print(f"=> {gemini_link} {entry_date}{entry_dict.get('title') or gemini_link}\r")

return

if __name__ == '__main__':

_main(unquote(urlsplit(input().strip()).query))