NEWS

twscrapeR 0.1.3.9000

Python calls now go through a single bridge module (inst/python/twscraper_bridge.py) imported once per session, with real arguments instead of generated Python source. Tweet and user fields are mapped in one place.
get_retweeters_batch() fetches all tweets with one twscrape API instance and one event loop instead of one Python run per tweet.
to_dataframe() builds columns directly instead of one tibble per row (about 7x faster for 5,000 tweets, same output). Without 'tibble' it now returns a data.frame with the same column names as the tibble (id instead of tweet_id/user_id). 'purrr' and 'dplyr' are no longer suggested.
.onLoad() no longer overrides a user-set cli.ignore_unknown_rstudio_theme option.

options(twscrapeR.db_file = ...) and a db_file argument in every scraping function, so searches use the same account database as add_account().
install_agent_skill() installs an Agent Skills skill (inst/skills/twscraper) that teaches AI agents such as Posit Assistant or Claude Code to use twscrapeR conversationally: setup, choosing functions, returned fields, recipes and common errors.

Usernames were inserted unescaped into generated Python code in user_tweets(), get_user(), get_followers(), get_following(), user_media(), user_tweets_and_replies() and verified_followers(); a quote or backslash broke the call. Inputs are now validated and passed as arguments.
Scraping functions ignored a custom db_file used in add_account() and always read accounts.db.
Tweet dates are parsed as UTC; before, the UTC timestamp from twscrape was read as local time.
Large numeric tweet IDs are no longer sent in scientific notation; IDs must contain only digits.
An unknown username now gives a clear error instead of a Python AttributeError.
filter_by_lang(), filter_by_date() and sort_tweets() no longer misalign results when some tweets lack the field.
print() of a user no longer fails when verified is missing.

Compatibility release for twscrape 0.18.0 after X/Twitter SearchTimeline changes.

Added product to search_tweets() so callers can choose "Latest" or "Top" results.
Added dry tests for search_tweets() product validation, generated Python search calls, and query escaping.

Updated README setup guidance for Python >= 3.10 and twscrape >= 0.18.0.
Documented cookie setup via either TWS_COOKIES or TWS_AUTH_TOKEN + TWS_CT0.

Removed the local X Client Transaction ID monkeypatch now that the upstream fix ships in twscrape 0.18.0.
Restored live search compatibility when the Python environment is upgraded from twscrape 0.17.0 to 0.18.0.

Quality and maintenance release after validating v0.1.1 live scraping.

Added dry testthat coverage for environment credential helpers, Python string escaping, tweet ID extraction, and batch retweeters conversion.
Added GitHub Actions CI to run dry R tests on push and pull request without requiring secrets or live X/Twitter access.
Added NEWS.md and docs/RELEASE.md so release history and the .9000 development workflow are explicit.

Startup messages now show the installed package version instead of a hardcoded version.
Python string escaping no longer over-escapes newline and carriage-return characters.

Kept the XClientTransaction monkeypatch tracker open because upstream twscrape fixes are still unreleased.

Maintenance release focused on making live scraping work reliably after changes in X/Twitter's web bundles.

Restored real tweet retrieval for twscrape 0.17.0 by applying a runtime compatibility patch for the x-client-transaction-id parser used by twscrape.
Replaced the hardcoded startup version with the installed package version.
Corrected Python string escaping for newline and carriage-return characters.

Added add_account_from_env() to load credentials from .Renviron or environment variables using TWS_* names.
Added support for separate TWS_AUTH_TOKEN and TWS_CT0 variables, with automatic cookie-string construction.
Added get_retweeters_batch() for retrieving retweeters across multiple tweets while preserving source_tweet_id.
Added dry testthat coverage for credential helpers, tweet ID extraction, retweeters batch conversion, and string escaping.
Added GitHub Actions CI for dry R tests.

Documented rate limiting, multiple accounts, the No account available for queue message, and the IndexError: list index out of range upstream twscrape failure mode.

Initial public release.

Guided Python and twscrape setup via setup_twscraper().
Account management with add_account(), list_accounts(), and delete_account().
Tweet search, hashtag search, mention search, user timelines, replies, media, followers, following, retweeters, and verified followers.
Conversion helpers such as to_dataframe() and export helpers for CSV/JSON.
README and quickstart examples for first-time users.