[bunkr] Broken extractor

[Yakabuff]

bunkr now inserts different TLDs into every URL which breaks the extractor.

Currently, we assume that URLs in a bunkr album either

1. Use the same TLD (starts with /) and we need to extract CDN URL from HTML
2. Directly links the CDN and the file can be downloaded directly

We will need to add a case where the URL in album does not start with / and is not a CDN URL but also matches base pattern

[Yakabuff]

                     else:
                         domain = domain.replace("cdn", "media-files", 1)
                     url = urlunsplit((scheme, domain, path, query, fragment))
+                else:
+                    scheme, domain, path, query, fragment = urlsplit(url)
+                    try:
+                        url = self._extract_file(text.unescape(path))
+                    except Exception as exc:
+                        self.log.error("%s: %s", exc.__class__.__name__, exc)
+                        continue