Changelog • semfindr

semfindr 0.2.1

CRAN release: 2026-07-27

Updated index_plot(). The argument column can now be omitted if the object has only one column (only one user statistic). (0.2.0.1, 0.2.0.3)
Updated roxygen2 to 8.0.0 and also updated the documentation. (0.2.0.3)
Fixed an issue related to the coming lavaan 0.7-3. (0.2.0.4)
Fixed a bug when case_id is set but to_rerun not set. (0.2.0.5)
Rebuilt some vignettes. (0.2.0.6)

CRAN release: 2026-03-06

Used precomputed vignettes to reduce the package size and also the speed to build vignettes. (0.1.9.1)
est_change_raw() will now keep the original order of parameters. (0.1.9.2)
Updated terms to match those used in the manuscript. (0.1.9.3)

CRAN release: 2025-03-03

Updated user_change_raw() to allow for user functions that return an unnamed vector. Also added a pkgdown website article for user_change_raw(). (0.1.8.5)
For lavaan_rerun(), added the argument ncores to make it easier to set the number of CPU cores to use in parallel processing. (0.1.8.7)
For lavaan_rerun(), added the option to show a progress bar (generated by pbapply) by setting progress to TRUE. Enabled by default. (0.1.8.8)

Added on.exit() such that workers will be stopped if lavaan_rerun() terminates abnormally. (0.1.8.2)
Updated some default values of approx_check() to reflect new features added to other functions. (0.1.8.4)
Fixed a few Rd issues on CRAN. (0.l.8.6)
Print a more conservative estimate of the expected time. (0.1.8.7)
Remove an invalid link in the reference. (0.1.8.9)

Fixed a bug with index_plot() when the extracted column has no names. (0.1.8.1)

CRAN release: 2024-04-08

Updated the two tests for nonconvergence. Will check against results generated by directly calling lavaan functions. (0.1.6.1)

CRAN release: 2023-11-12

Added index_plot() for generating an index plot for an arbitrary statistic. (0.1.5.5)
Added user_change_raw() for computing case influence on user statistics. (0.1.5.8)

Fixed typos in the article on multiple-group models. (0.1.5.1)
Fixed pars_id(). Users can now use the internal lavaan group labels (e.g., ".g2", ".g3") to denote parameters in a group. This method is not recommended (noted in the help page) but is included as an option. (0.1.5.2)
Fixed an error at CRAN check due to suggested packages not installed. (0.1.5.3)
Use lavaan::lavCor() to estimate means and correlations for Mahalanobis distance when missing data is present. The package norm2 is no longer needed nor suggested. (0.1.5.4)
Made plot functions to work with NAs. (0.1.5.6)
Updated README.md. (0.1.5.7)
Finalized 0.1.6. (0.1.6)

CRAN release: 2023-06-19

Added a print method for the output of est_change(), est_change_raw(), est_change_approx(), and est_change_raw_approx(). (0.1.4.3, 0.1.4.16, 0.1.4.18, 0.1.4.19, 0.1.4.20)
Added a print method for the output of fit_measures_change() and fit_measures_change_approx(). (0.1.4.4, 0.1.4.16)
Added a print method for the output of mahalanobis_rerun() and mahalanobis_predictors() (0.1.4.5-0.1.4.6, 0.1.4.16, 0.1.4.18)
Added a print method for the output of influence_stat(). (0.1.4.7)
Provided initial support for multiple-group models and models with linearity constraints. (0.1.4.8 - 0.1.4.14)
Added support for labels and ":=" to pars_id(). (0.1.4.13)

Fixed the pkgdown site. (0.1.4.1)
approx_check() will check if the model has at least one equality constraint. (0.1.4.2)
Added a “Limitations” section to the vignette of the approximate approach. (0.1.4.2)
Fixed the documentation of pars_id(). (0.1.4.12)
Added an article for multiple-group models. (0.1.4.15)
Restructured articles and references in the pgkdown websites. (0.1.4.15)
In print.lavaan_rerun(), keep the format of lavaan warnings. (0.1.4.17)
Remark that fit_measures_change_approx() only supports selected fit measures. (0.1.4.21)
Add cutoff_change and largest_change to est_change_gcd_plot(). (0.1.4.22)
Updated a few tests on parameters which are fixed in the model but their standardized versions are free. (0.1.4.23)
Updated vignettes and articles with links to the package website. (0.1.4.24)
Updated all vignettes for the new print methods. (0.1.4.25)

CRAN release: 2023-02-17

Added DOI of Pek and MacCallum in the DESCRIPTION. (0.1.1.1)
Added approx_check() to check whether the input object is supported by the approximate method. (0.1.1.2)
Added est_change_plot() and est_change_gcd_plot(), diagnostic plots for casewise influence on parameter estimates. (0.1.1.3)
Diagnostic plot functions revised to allow users to fully control elements drawn. (0.1.1.4)
Fixed typos and grammatical mistakes in help pages and vignettes. (0.1.1.5)
Fixed an invalid URI in a vignette (casewise_scores). (0.1.3)
Added the documentation for the return value of pars_id_to_lorg(). (0.1.4)

First public release.
Added skip_all_checks to lavaan_rerun(), allowing users to experiment lavaan_rerun() and other functions on models not officially supported.
Revised est_change() and est_change_raw() to support the use of operators (e.g., ~, =~) to select parameters.
Added badges and R CMD Check Action.
Updated est_change(), est_change_raw() and est_change_approx() to support models with labelled parameters. (0.1.0.9005)
Added pars_id() and pars_id_to_lorg() for converting parameter specification to identification numbers ( positions in the vector of coefficients or row numbers in the parameter tables). (0.1.0.9006)
Updated est_change_* functions to use pars_id() and pars_id_to_lorg(). (0.1.0.9007)
Modified lavaan_rerun() to use lavaan::lavaan() instead of update() as the default way to rerun. (0.1.0.9008).
Updated some of the tests. (0.1.0.9009)
Added more examples. (0.1.0.9010)
Updated documentation (e.g., README and DESCRIPTION). (0.1.0.9011)
Updated influence_stat() and the plot functions to support the approximate approach. (0.1.0.9012)
Updated documentation.

Added more vignettes.
lavaan_rerun() can accept an output with inadmissible estimates. Disabled by default. Can be enabled by setting allow_inadmissible to TRUE.

Added a print method for the lavaan_rerun() class.
Added mahalanobis_predictors() to compute the Mahalanobis distance using only the observed predictors.
Both mahalanobis_predictors() and mahalanobis_rerun() support datasets with missing data.
lavaan_rerun() can specify cases to exclude and rerun by specifying the case IDs or selecting cases based on Mahalanobis distance on all observed variables or on residuals of observed variables in a path model.

Used lavaan::update() in lavaan_rerun. This is more reliable than recreating the call.
Added implied_scores(). It supports only single-group path analysis models for now.