Changes in version 0.4.0 (2025-09-13)                  

New features

  - Users can now supply a list of vcov arguments to compare multiple SE
    adjustments on the fly. (#65)
    
    est = est = feols(flipper_len ~ bill_len | species, penguins)
    ggcoefplot(est, vcov = list("iid", "hc1", ~species))

Bug fixes

  - The keep and drop arguments now work correctly with a list of
    models. Thanks to @femdias for the report. (#60)
  - If a (named) list of models if provided to either ggcoefplot or
    ggiplot then the original order is preserved for grouping and facet
    behaviour (#63)

Internals

  - Bump ggplot2 dependency to v4.0.0 and update test snapshots. (#55,
    #59)
  - Bump fixest dependency to v0.13.0 and update tests. (#58, #59)

                 Changes in version 0.3.0 (2025-05-14)                  

New features

  - The new aggr_es(..., period = "diff") convenience keyword argument
    allows users to estimate the difference between the (mean) post- and
    pre-treatment periods. Thanks to @FBrunamonti for the suggestion.
    (#52).

Internals

  - Bump svglite dependency version and update test snapshots. (#51)

                 Changes in version 0.2.0 (2025-01-22)                  

Bug fixes

  - Fix missing "hypothesis" attribute for aggr_es objects. (#43)
  - Fix dodge misalignment between points and lines with multi fixest
    objects. (#44)

Internals

  - Replace ggh4x dependency with legendry. (#41 @teunbrand)

Misc

  - Minor website and documentation improvements.

                 Changes in version 0.1.0 (2023-12-14)                  

First CRAN release!

New features

  - The aggr_es function now supports numeric sequences for aggregating
    a specific subset of periods, in addition to the existing keyword
    strings like "pre" or "post". This functionality also passes through
    to the higher order plotting functions that call aggr_es under the
    hood. For example, ggiplot(est, aggr_eff = 6:8). (#33)
  - Users can now adjust standard errors for model objects on-the-fly at
    plot time, by passing an appropriate argument, e.g. ggcoefplot(est,
    vcov = "hc1"). These on-the-fly adjustments are done via
    summary.fixest, and so the effect is just the same as passing an
    adjusted object directly, e.g. ggcoefplot(summary(est, vcov =
    "hc1")). However, it may prove more convenient for simultaneously
    adjusting a list of multiple models, e.g. ggcoefplot(list(est1,
    est2, est3), vcov = "hc1"). (#35)

                        Changes in version 0.0.3                        

Breaking change

  - The package name has been changed to ggfixest (#29).

                        Changes in version 0.0.2                        

New features

  - Support for ggcoefplot, a ggplot equivalent of coefplot (#28).
  - Support pt.size argument for controlling the size of point markers
    (#27). Thanks @jcvdav.
  - Support keep and drop arguments for subsetting coefficients (#22).

Bug fixes and breaking changes

  - Fix naming mismatch in multiple estimation with different time
    periods (#10). Thanks @brockmwilson.
  - Slight tweak to default theme, which now uses dotted grid lines to
    more closely match iplot() (#e5cf0b0).
  - Correctly parse formula-transformed dependent variable names, e.g.
    log(y (#20).
  - The confidence intervals for some figures may be slightly wider due
    to upstream changes in fixest (#25; see also
    https://github.com/lrberge/fixest/pull/408).

Internals

  - Add a (visual) test suite (#12 with several increments thereafter).
    Thanks to @vincentarelbundock for ~~tinyviztest~~ tinysnapshot!
  - Switch to marginaleffects::hypotheses() internally for aggr_es() to
    match the upstream changes in marginaleffects.
  - Simplify multi_fixest object parsing (#19).
  - Minor documentation improvements.

                        Changes in version 0.0.1                        

  - Tweaks to plot output, including integer breaks on x-axis (where
    appropriate) and allow additional user-level control (e.g. CI alpha
    or width levels)
  - Support multiple confidence levels (#2, #5)
  - Support multiple LHS variables (#1)
  - Added a NEWS.md file to track changes to the package.