Changes in version 0.9.0.9000 Breaking changes - Removed run_document_extractor() app (previously deprecated in 0.9.0). UI overhaul: migration to bslib - Migrated UI framework from shinydashboard to bslib. - Sidebar reorganized into accordion panels for cleaner grouping. - Drag-to-resize column layout; column split also configurable via new corpus_map_column_width argument in explore(). - Added find-in-text bar for searching within the document viewer, with unicode support and sticky positioning while the corpus map scrolls. - Added prev/next navigation arrows for browsing days and documents. - Added checkbox to include active search terms in corpus filtering. - Redesigned document visualisation with color legend for search/highlight terms. - Duplicate search patterns now allowed. - Calendar days without documents are now ignored when clicked. Other - Removed dependency on shinyWidgets. - Introduced code formatting with air. Changes in version 0.9.0 (2024-09-02) 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) Changes in version 0.8.6 (2022-06-20) - Fix bug introduced when re2 was reintroduced (4eeb2f3), which caused app to crash when searching in custom columns/metadata Changes in version 0.8.5 (2022-04-08) - re2 regex engine again included, now using the 're2' package. This significantly speeds up many search operations. Thank you, @girishji! See https://github.com/girishji/re2 - Small UI fixes Changes in version 0.8.4 (2021-03-18) - Fixed bug where time selection slider widget in calendar view did not work (#25) Changes in version 0.8.3 (2020-11-13) - 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) Changes in version 0.8.2 (2020-03-07) re2r regex engine disabled - Because 're2r' will be removed/archived by the CRAN maintainers. See https://github.com/qinwf/re2r/issues/22. Other - the apps now allow any mix of whitespace in input terms Changes in version 0.8.1.1 (2020-02-07) 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. Changes in version 0.8.0 (2019-12-02) Breaking API changes - run_corpus_explorer() is deprecated and replaced with explore(). - explore() works in the precise same way as run_corpus_explorer(), but is faster to type and makes a nice pair with the new explore0() function (see below). - run_corpus_explorer() still works (as a thin wrapper to explore()), but yields a warning. - Removed normalise argument from prepare_data() - Added within_group_identifier and tile_length_range arguments in prepare_data() (leading to new order of arguments) 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. Changes in version 0.7.0 (2019-11-15) 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. Changes in version 0.6.3 (2019-10-06) 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 - Documentation moved to https://kgjerde.github.io/corporaexplorer