Disable archive.org queries after timeout

[HebaruSan]

Problem

#346 tried to address the download counter's log spam and downtime by setting an instance variable when a timeout occurred and gating further queries with that variable. This was somewhat effective in that last night's count pass completed, but it still wasted time querying archive.org and spammed Discord.

Cause

After performing a batched query to archive.org, DownloadCounter replaces its InternetArchiveBatchedQuery object with a fresh one. Before #346, the timeout exception was preventing this from happening, so the same query object got stuffed with all mods over time, .full() was always True, and we tried to query archive.org for every mod (rather than once per 30), over and over, with an increasingly gigantic query.

The main thing from #346 that was helpful was catching the exception, which allowed the main loop to replace the query object, reducing the queries back to once per 30. But setting self.connect_timed_out then had no effect because the query object was discarded immediately afterwards.

Changes

• Now the changes from Exclude internetarchive's internal logging from Discord #346 are removed, and instead the main DownloadCounter class itself catches the timeout exception from archive.org, discards its ia_query reference, and stops querying archive.org once this has happened.
• GraphQLQuery is renamed GitHubBatchedQuery for consistency with the other classes
• DownloadCounter.GITHUB_PATH_PATTERN is moved to GitHubBatchedQuery.PATH_PATTERN, and so on for the other path pattern variables