Crawling error rises as invalid URLs are extracted

[MrOrz]

In google webmaster tool, we are receiving crawl error in weird URLs:

These URLs are extracted from pages like this:
https://cofacts.g0v.tw/article/2yje6no2cqv2v

「www.sushiexpress.com.tw，爭鮮也已經出面回應這是假訊息。」 is being extracted as URL, and "http://" is not prepended, thus the crawler goes to https://cofacts.g0v.tw/article/www.sushiexpress.com.tw，⋯⋯ instead.

We should fix the URL scrapping logic, prepend http:// or https:// when needed, and rewrite all articles & replies' hyperlinks field to remove the wrongly extracted URLs.