Breaking changes

New arguments (and new order of arguments) in prepare_data():
* text_column: Any character column can be specified as as the column containing the main textual data (not only ‘Text’ as before). Thanks to @discoleo for suggestion (#31).

Changes in reserved column names (see ?prepare_data for details). In particular, Year is now allowed.

Other changes

  • Non-date-based corpora can now contain column called ‘Date’.
  • Introduced nicer display of searches with (exactly) two search terms (feebc552, feebc552e4, 325aeb3b).

Future removal

Added warning that the ‘run_document_extractor()’ app will be removed in a future version of corporaexplorer.

Bug fix:

Fixed filtering by dates (8c1a175b)

Other

  • Changed internals to avoid ggplot2 warning
  • Changing testing library to shinytest2
  • Change in testing due to changes in data.table (thanks for heads-up, @MichaelChirico)
  • Fix bug introduced when re2 was reintroduced (4eeb2f3), which caused app to crash when searching in custom columns/metadata
  • Fixed bug where time selection slider widget in calendar view did not work (#25)
  • Fixed bugs and warning messages caused by changes in other packages (tibble and dplyr).
  • Fixed special case of warnings occurring when visualising documents with zero hits (2270046)
  • Changed the way colours are represented in document visualisation (63b3857)

re2r regex engine disabled

Other

  • the apps now allow any mix of whitespace in input terms

New arguments

  • Added columns_for_ui_checkboxes argument to prepare_data() in order to include sidebar checkboxes for convenient filtering by categorical variable/column (not necessarily factor type, but most helpful when limited number of values in the variable) in the explorer app (explore()).

Other

  • Removed dependencies on packages forcats, shinycssloaders, and zoo.
  • Fixed reactive issue leading to plots rendering twice when size was different from previous plot (now one has to click button to update plot size).
  • Other minor bug fixes.
  • Internal changes related to extension packages.

Breaking API changes

New function

  • Added explore0(), a convenience function to directly explore a data frame or character vector without first creating a ‘corporaexplorerobject’ with prepare_data().

Demo apps

  • Included demo apps: Jane Austen’s novels and State of the Union addresses, with data from the janeaustenr and sotu packages, respectively. See run_janeausten_app() and run_sotu_app()

Other

  • Minimum corpus plot height is again 100.

Breaking API changes

  • Added search_options argument to run_corpus_explorer(), and moved the arguments use_matrix, regex_engine, optional_info, allow_unreasonable_patterns there.
  • Added plot_options argument (with several possibilities to customise plot colours etc.) to run_corpus_explorer(), and moved the max_docs_in_wall_view setting there.

Other

  • Greatly simplified deployment to Shiny Server and shinyapps.io etc. (#19). See article on the package website.
  • Improved app experience on small screens (avoiding cluttering of ui elements).
  • Updated LICENSE file with license for jQuery.scrollTo.
  • Plot minimum height set to 50.
  • Minor app CSS twitches.

API changes

  • Add ui_options argument to run_corpus_explorer().
  • Add search_input argument to run_corpus_explorer().

Bug fixes

  • Print method for corporaexplorerobject now works properly
  • Open but empty “Filter corpus” field no longer disables time filtering

Other