rtweet 2.0.0

• This is a major release to signal the end of API v1.1 functions.

• Fix a problem with auth_sitrep() not correctly handling old tokens.

• Since httr2 > 0.2.3, rtweet refreshes OAuth 2.0 tokens automatically.
  It is no longer needed to user rtweet_oauth2(), authentication is fully handled by httr2.

• Document the deprecated functions from API v1.1 to API v2: see help("rtweet-deprecated", "rtweet").

• New function client_clean() to remove client's data (together with auth_clean() they leave the user cache empty).

rtweet 1.2.1

• Fix auth_sitrep() to work well with OAuth2 tokens.

• Added new function tweet_quoted() to search who tweets quoting a tweet.

• New tweet_post() and tweet_delete() to post and delete tweets to work with the free product.

• Fix problems with changes on is.atomic(NULL).

rtweet 1.2.0

Authentication changes

• New client_* functions to save, set and use a client app using Twitter API
  via the new authentication mechanism rtweet_oauth2() (see below).
  It has a helper function for the scopes the client is allowed to do set_scopes().
  Endpoint have different scope requirements.

• New authentication mechanism rtweet_oauth2() required by some endpoints.
  Only valid for 2 hours until it is automatically renewed, but it is left to
  the user to save it every time it is renewed.

New endpoints using API v2

New endpoints to retrieve data from twitter, no action is performed (no blocking, muting, posting new tweets...):

• New list endpoints: list_expansions(), list_fields(), list_followers(),
  list_get(), list_members(), list_membership(), list_tweets().

• New tweet endpoints: tweet_get(), tweet_liking_users(), tweet_retweeted_by().

• New user endpoints: user_blocked(),user_bookmarks(),user_by_username(),
  user_following(), user_followers(), user_liked_tweets(),
  user_list_follows(), user_lists(), user_mentions(), user_search(),
  user_self(), user_timeline(), user_tweets().

• New statistic endpoints: tweet_counts_recent(), tweet_counts_all().

Other changes

• Small breaking change: expansions and fields arguments now use NA for all and NULL for none.

• New functions to set expansions and fields for the new functions: set_expansions(), set_fields().

• Fixed a bug that prevented auth_setup_default() to work (#756 and #744)

• Fixed a bug so that the streaming functions use the current token.

rtweet 1.1.0

• Fixed a bug that prevented posting multiple media in the same tweet.
• Argument premium is correctly handled in search_fullarchive and search_30day (#720).
• ts_plot uses the first two columns of grouped data.frames for plotting (#724 and #727).
• get_timeline returns user data for multiple accounts (#723).
• network_data now handles deleted accounts (#730).
• retryonratelimit works again on paginated endpoints (#732).
• Fixed multiple bugs that prevented auth_sitrep() to work in some cases.
• Provided rbind and subsetting methods ([) for rtweet new classes.
• Fixed a bug about sending a private message (#735).
• Provided a function clean_tweets() for cleaning entities from the text of tweets (#721).
• New helper methods for extracting data from entities in tweets: entity().
• Columns of the parsed output are now always in the same order (#740).
• New methods to retrieve ids() from multiple classes and print nicer in post_tweet() (#452).
• post_tweet() now invisible returns the tweet information.
• Provides initial support for API v2 for streaming endpoints:
  filtered_stream() and sample_stream(), including adding rules
  stream_add_rules() and stream_rm_rules().
• Deprecated old stream function.
• Handle more graciously if a whole data.frame is used for lookup_tweets().

rtweet 1.0.2

rtweet 1.0.2

• Exported again tweets_with_users and users_with_tweets because Twitmo depends on them.

rtweet 1.0.1

• Fixed issue with .Rbuilignore and vignettes

• Reduced fixtures sizes by limiting the page size to the number of requests if
  it is smaller than the default page size.

rtweet 1.0.0

Breaking changes

• Data returned by rtweet is nested and uses the same names provided by
  the Twitter API. It doesn't compute or add new columns as it did previously.

• emojis, langs and stopwordslangs data are no longer provided by rtweet.

• get_friends() and get_followers() return similar formatted output with
  two columns "from_id" and "to_id" (#308, @alexpghayes).

• All paginated functions that don't return tweets now use a consistent
  pagination interface. They all store the "next cursor" in an rtweet_cursor
  attribute, which will be automatically retrieved when you use the cursor
  argument.

• Functions that return tweets (e.g. get_favorites(), get_my_timeline(),
  get_timeline(), get_mentions(), lists_statuses() and search_tweets())
  now expose a consistent pagination interface. They all support max_id and
  since_id to find earlier and later tweets respectively, as well as
  retryonratelimit to wait as long as needed when rate limited (#510).

• suggested_slugs(), suggested_users(), suggested_users_all() have been
  removed as they stopped working when Twitter remove the suggested users
  endpoint in June 2019 (https://twittercommunity.com/t/124732).

• parse = FALSE always means return the raw "JSON". Previously some functions
  (e.g. my_friendships()) would return the raw HTTP response instead (#504).

• rtweet no longer re-exports the magrittr pipe %>%; if you want to continue
  using it, you'll need to library(magrittr) or library(dplyr) (#522).

Deprecations

• The authentication system has been rewritten, check the following section.

• lookup_collections() and get_collections() has been hard deprecated
  because the underlying Twitter API has been deprecated.

• previous_cursor() has been hard deprecated. It could only be used with
  lists_memberships() and it has been dropped in favour of making regular
  pagination better.

• tweet_shot() has been hard deprecated as the screenshots do not have the
  tweet. It might come back with webshot2 (#458).

• The home argument to get_timeline() has been deprecated. You can only
  retrieve the home timeline for the logged in user, and that's the job of
  get_my_timeline() (#550).

• Due to the changes on rtweet data format, all the functions related to
  flattening the data (write_as_csv(), save_as_csv() flatten(),
  unflatten(), read_twitter_csv()) are deprecated.
  Users should decide how to flatten the nested structure of the data.

• lookup_statuses() has been deprecated in favour of lookup_tweets().

• as_userid() has been deprecated since in case of ambiguity the default is
  to assume a numeric string is a user id (#520). All functions now use a
  single user_type() function so behaviour is identical for all rtweet
  functions.

• get_timelines() has been deprecated since it does that same thing as
  get_timeline() (#509).

• stream_tweets2() has been deprecated in favour of stream_tweets().

Authentication

rtweet's authentication system has been completely written. It is now based
around three authentication options: rtweet_user(), rtweet_app(), and
rtweet_bot(). Authentication no longer touches ~/.Renviron file; instead
auth_save() and auth_as() allow you to explicitly save and load
authentication mechanisms from a system config directory. See vignette("auth")
for more details.

• The httpuv package is now only suggested, since it's only needed for
  interactive auth, and you'll be prompted to install it when needed.

• bearer_token() has been deprecated in favour of rtweet_app(), which takes
  the bearer token found in your Twitter developer portal. invalidate_bearer()
  has been deprecated since this is something you should do yourself in the
  Twitter developer portal.

• create_token() has been deprecated in favour of the combination of
  rtweet_user()/rtweet_bot()/rtweet_app() + auth_as() + auth_save().

• get_token() and get_tokens() have been deprecated in favour of
  auth_get() and auth_list().

• auth_as() accepts path to an authentication to make it easier to use
  authentications outside a user account (#602, @maelle)

• auth_setup_default() will not only authenticate and save but use the
  default token.

• The new auth_sitrep() helps reports the different authentications of the user

Other changes

• Update to new rOpenSci Code of Conduct: https://ropensci.org/code-of-conduct/

• lookup_users() and search_users() now returns a data frame containing
  all information about each user (not their latest tweet). If you want to get
  that data you can use tweets_data().

• rtweet 1.0.0 implements a consistent strategy for handling rate limits.
  By default, if a paginated function (i.e. a rtweet function that performs
  multiple calls to the twitter API) is rate-limited it will return all results
  received up to that point, along with a warning telling you how to get more
  results. Alternatively, if you want to automatically wait until the
  rate-limit is reset, you can set retryratelimit = TRUE.

• The default value of retryonratelimit comes from the option
  rtweet.retryonratelimit so you can globally set it to TRUE if desired
  (#173).

• All functions that perform multiple requests on your behalf now display
  a progress bar so you know what's happening. If you don't want it, you can
  turn it off with verbose = FALSE (#518).

• Banned or protected accounts now trigger a warning instead of an error,
  but if data from other users is requested it is not served by the API and
  returned as NA (#590, @simonheb).

• Added support for posting alt-text metadata with images tweeted with status
  updated via post_tweet(). (#425, @hrbrmstr)

• stream_tweets() has been overhauled to only write valid data. This obsoletes
  all previous strategy to clean up bad data after the fact (#350, #356).

• The maintainer changed.

• New user_block() and user_unblock() to block and unblock users (#593,
  @simonheb).

• The new tweet_threading function is now faster and more reliable (#305,
  #693, @charliejhadley).

• Message are now properly capitalized (#565, @jsta)

• Fields withheld_scope, withheld_copyright, withheld_in_countries are
  now correctly parsed (#647, @alexpghayes).

• Functions like search_tweets(), lookup_statuses() and others return the
  appropriate date time format for the right columns (created_at mostly)
  (#653, #657, #660, @alexpghayes, @Rickpat).

• Premium/sandbox environments are supported in search_fullarchive() and
  search_30day() (#578, #713).

• The vignette must be pre-computed before submission (#609, @maelle).

rtweet-v0.7.0

Change version number

rtweet-v0.6.9

udpate docs

rtweet: Collecting Twitter data

Development version of {rtweet}

rtweet-v0.6.7

rtweet CRAN release 2018-06-20

rtweet: Collecting Twitter Data (github v0.6.13)

Dev version 0.6.13 of rtweet.