The previous method relied on the main "download link" in the list page.
But this link was broken a solid 1/4 of the time, and far more often for
some artists.
Instead, during DB build, grab and parse each actual song page too, and
grab from it all possible download links. Use a ThreadPoolExecutor to do
this in a reasonable amount of time (default of 10 workers, but user
configurable).
Then when downloading, iterate over all download links, or provide some
user options for filtering these by ID or description.
Joshua Boniface