Changes in version 0.3.3 (2025-08-25)                  

  - Fixes an issue with testing plot_missing() with the future ggplot2
    release (#28).

                 Changes in version 0.3.2 (2023-09-05)                  

Several small fixes.

Bug fixes

  - Manually add a "states-packge" alias for the package doc since
    roxygen2 does not do this by default anymore.
  - Switch from the deprecated ggplot2::aes_string() to ggplot2::aes() +
    the .data pronoun in plot_missing().
  - Fix a dplyr join warning in state_panel() due to the new
    "relationship" argument.
  - Update old URLs in README and other documentation.

                 Changes in version 0.3.1 (2021-04-14)                  

  - Strip readr "spec" attribute from the cowstates and gwstates data.
    They are both now plain data frames with no attributes. (#22)
  - Expand list of short country names in prettyc(), e.g. instead of
    "Belarus (Byelorussia)", just "Belarus".
  - Move to testthat 3rd edition (3e).

Bug fixes

  - all.equal() in R 4.1.x (R-devel at this point) will check
    environments for equivalence as well. This breaks one test since 2e
    testthat relies on all.equal() for expect_equal(). Moving to
    testthat 3e fixes this bug. (#26)

                 Changes in version 0.3.0 (2020-12-11)                  

  - Although the original COW state list does not distinguish
    micro-states, like the G&W list does, the cowstates data now has a
    microstates coding derived from the G&W coding (#24). This makes it
    easier, for example, to filter out micro-states from a state panel
    dataset.
  - In the gwstates and cowstates data, the "iso3c" columns have been
    renamed to "gwn" and "cown", as they are not in fact ISO 3-character
    country abbreviations (#19).
  - plot_missing()'s "space" argument has been renamed "ccode".
  - id_date_sequence() now works with integer years, there is no need to
    convert them to Date class anymore (#20).

Improved support for state_panel shortcuts

state_panel() has further improved support for input shortcuts (#3):

  - Accepted formats for the "start" and "end" arguments now
    include 2006, "2006", "2000-06", and "2006-06-01". See parse_date()
    for details.
  - The implied time period is now also inferred from shortcut input
    when "by" is left at the new default of NULL.
  - The "partial" option has gained new "first" and "last" values that
    also support shortcut input. Previously a shortcut input like 2006
    could only be used with the "any" partial option, and to get
    country-years for states that were independent on the first day of a
    year one had to use state_panel("YYYY-01-01", partial = "exact",
    ...). This can now instead be done with state_panel(2006, partial =
    "first", ...).

New functions

  - Added country_names() to translate country codes to country names
    and prettyc() to shorten some of the longer country names, like
    "Macedonia, the former Yugoslav Republic of".
  - Updated vignette on G&W and COW differences to take into account
    when trying to translate between codes.
  - Added compare() helper to compare two statelist data frames prior to
    merging.
  - Added parse_date() to handle more flexible "start" and "end" date
    input (#12).

Bug fixes

  - fixed a misspelled country name, "Rumania" to "Romania"

                 Changes in version 0.2.2 (2019-01-11)                  

  - Added option to use simpler input when generating a yearly panel,
    i.e. state_panel(2001, 2005) instead of state_panel("2001-01-01",
    "2005-01-01", by = "year", partial = "any").
  - Changed sfind() so that it ignores capitalization and treats
    character numbers as numbers. These pairs of queries now produce the
    same results, unlike before:
      - "Gabon" and "gabon",
      - 260 and "260"
  - Changes in plot_missing() and missing_info():
      - make "data" the first argument and "x" the second argument so
        that it works nicer in pipes.
      - add more intelligent defaults so that not all arguments have to
        be explicitly specified.
      - rename the "time_unit" argument to "period".
      - added "partial" argument, it works the same as in state_panel().
  - Fixed plot_missing() and missing_info() error when "data" is a
    tibble ("tbl_df").

                 Changes in version 0.2.1 (2018-05-04)                  

  - plot_missing improvements: examples with Polity data, added a
    skip_labels option that will plot every n-th label for countries
    instead of all labels to avoid overplotting on the y-axis. The
    latter is inspired by https://github.com/xuyiqing/panelView.
  - Added Polity combined score as example data set.
  - Fixed encoding issue for gwstates$country_names like Cote d'Ivoire
    and Wüttemberg.

                 Changes in version 0.2.0 (2018-02-23)                  

  - Initial version on CRAN.
  - Refactor state_panel and add a partial option to allow including
    partial state-periods (as opposed to the default which is based on
    the exact start date, for all period resolutions in "by").
  - Rework plot_missing: fix a bug that would leave out independent
    states with no cases in the input data; add missing_info function
    that generates the data underlying the plots.
  - Add pkgdown website for states package documentation.

                        Changes in version 0.1.0                        

  - Initial version on Github.
  - Added a NEWS.md file to track changes to the package.