3 Commits

Author	SHA1	Message	Date
ed	ae9cc1d494	feat(twitter_threads): full-thread extraction (conversations) + front-matter from target + corpus dedupe ... fetch_thread_from_url now passes -o conversations=true (full threads, not single tweets), sorts posts chronologically, and sets root_post_id to the URL's target tweet. render_markdown builds front-matter from the target (root_post_id) post, not the earliest conversation tweet (fixes wrong @handle). Added dedupe_corpus.py: keeps deepest thread per conversation, deletes duplicates, keeps+marks unique branches (branch_of) + writes threads_index.json. 33 tests pass.	2026-07-05 18:38:43 -04:00
ed	53721a7796	fix(twitter_threads): align gallery-dl parser to live schema + add track extraction scripts ... Live Task 5.3 validation against real gallery-dl output: (1) pass -o text-tweets=true so text-only tweets are returned (media-only default yielded []); (2) fix author name/nick swap (gallery-dl name=handle, nick=display). Added convert_cookies.py (JSON cookie export -> Netscape) + run_corpus.py (8-URL pipeline -> docs/twitter/) to the track dir. Extracted 8/8 threads. 33 tests still pass.	2026-07-05 18:28:40 -04:00
ed	06ff9299e2	feat(twitter_threads): fetch_thread.py — html.parser (local HTML) + gallery-dl subprocess (URL) + URL query-strip + CLI ... fetch_thread_from_html parses the data-attr article contract via stdlib html.parser; _normalize_url strips ?s=20 quote-share suffix (spec FR2); fetch_thread_from_url wraps gallery-dl --dump-json (best-effort, manually smoke-tested); dual-import (absolute/relative) makes it runnable via -m AND as a bare script (VC6/G5); CLI writes thread_data.json. TDD: red(no module) -> green(7 fetch + 22 regression = 29 passed).	2026-07-05 17:36:03 -04:00