Remove retweeted original tweet from the list of tweets, fixes #310

[Flaburgan]

get_typed_objectinutils.pyis a recursive deep-walker that collects **all** objects with a__typenamefield anywhere in the API response. When Twitter returns a retweet, the retweeted tweet is nested insideretweeted_status_result.resultand also carries__typename: "Tweet". It therefore ends up in the flat tweetsdict built byto_old_rep, and _parse_items` then yields every entry in that dict without distinguishing top-level timeline items from nested lookup objects.

The nested tweet is legitimately needed in res["tweets"] as a lookup so Tweet.parse() can populate retweetedTweet. The bug is that _parse_items has no way to know which entries are top-level vs. which were collected as nested references.

Expected behaviour

user_tweets should yield the RT wrapper once. The original tweet should only be accessible via tweet.retweetedTweet, not as a separate item in the iteration.

Note

The bug existed conceptually since 9238398 (July 2023, the commit that introduced retweetedTweet support), but likely only became reliably reproducible after the May 2026 schema change: the old Twitter schema may not have included a legacy key on the nested retweeted object, which was the filter gate ([x for x in tmp["Tweet"] if "legacy" in x]). The schema migration explicitly preserves that filter condition, causing nested tweets to now consistently pass through.