website-title

#!/usr/bin/env python3

"""Retrieve a website and extract the title

Default to OpenGraph Title property, and fall back to <title>.
"""


import os
import sys
import requests
from bs4 import BeautifulSoup


# fallback url
fallback_url = 'https://github.com/malte70/scripts'


def usage(output=sys.stdout) -> None:
	argv0 = os.path.basename(sys.argv[0])
	print(f"Usage: {argv0} <url>", file=output)


def get_title(url: str) -> str | None:
	"""Get a web site's title.

	Uses og:title by default, or the classic <title> tag. Returns None if neither are found.

	Args:
		url (str): Website URL

	Returns:
		str | None: The title, or None if no title could be detected.
	"""
	# making requests instance
	reqs = requests.get(url)

	# print("get_title(): Got " + str(len(reqs.text)) + " Bytes...", file=sys.stderr)

	# using the BeautifulSoup module
	soup = BeautifulSoup(reqs.text, 'html.parser')

	# Try to get <title>
	# for title in soup.find_all('title'):
	# 	if len(title) > 1:
	# 		return title.get_text()
	title = soup.find("title")
	if title is not None and len(title.get_text()) > 1:
		return title.get_text()

	# Try to get og:title meta property
	title = soup.find("meta", property="og:title")
	if title is not None and len(title["content"]) > 1:
		return title["content"]

	# No title found :(
	return None


if __name__ == "__main__":
	if len(sys.argv) == 1:
		print("Using fallback URL...")
		url = fallback_url
		title = get_title(url)
		print(f"Website-URL: {url}")
		print(f"Title:       {title}")

	else:
		for arg in sys.argv[1:]:
			if arg == "--help" or arg == "-h":
				usage()

			elif arg[0] == "-":
				print(f"Unknown option: {arg}", file=sys.stderr)
				usage(sys.stderr)
				sys.exit(1)

			else:
				url = arg
				title = get_title(url)
				# print(f"Website-URL: {url}")
				# print(f"Title:       {title}")
				print(title)