Changelog • tmtyro

tmtyro 0.4.1

get_gutenberg_corpus() should do less, and now it does. Other functionality is available via gutenbergr.

New function contextualize() shows terms in a window of context
New function add_index() adds a column showing word indices within each document
load_texts() adds support to keep original capitalization and punctuation alongside the tokenized word column with the keep_original argument. This process does not work in all instances, so the option defaults to FALSE.
add_dictionary() includes an option to keep original terms. This is useful for n-gram dictionaries, where a match might otherwise span multiple rows.
add_ngrams() supports negative ranges, for building context windows
add_partitions() supports overlapping partitions
standardize_titles() capitalizes words after terminal punctuation

add_dictionary() now supports n-gram dictionaries, matching across multiple words
make_dictionary() has a slightly changed syntax, with clearer argument names definitions and name
Along with its related visualize() methods, plot_doc_word_bars() improves support for color_y = TRUE and reorder_y = TRUE
When naming colors, change_color() now allows setting colors for unnamed values
standardize_titles() capitalizes Roman numerals
load_texts() adds support for custom tokenization using the dots parameter from tidytext::unnest_tokens()

New function add_partitions() adds a partition column, useful for getting same-sized samples
identify_by() now works with multiple columns, and it keeps existing metadata columns. This is especially useful with the new add_partitions() column, using something like my_corpus() |> add_partitions() |> identify_by(title, partition) before continuing to work with partitioned documents. To return framing to unpartitioned data, used identify_by(title) or whatever other column is most relevant.
New visualization and tabulization methods for expand_documents()
Functions now imported: count() and drop_na()
When the ggraph package is loaded, plot_bigrams() now uses a color scale on edges, rather than spot color on nodes, with full support for change_color()
Improved documentation with website articles for customizing colors and showing code comparisons

New tabulize() generic function for preparing tables with supported methods
Standardizing argument names between visualize() and tabulize()
New package documentation for getting started
New collapse_rows() function for clean tables using gt::gt()
New feature in standardize_titles() to keep initial articles
New options in plot_doc_word_bars() to keep the order of Y-axis values consistent and to color by Y-axis value instead of by facet
Rename add_lexical_diversity() to add_vocabulary()
Add option for renaming existing doc_id column when using identify_by()

get_gutenberg_corpus() now retrieves HTML versions of texts from Project Gutenberg and parses header tags for section markers
New function parse_html() for reading headers in an HTML file
New function move_header_to_text() for converting header to text
New function identify_by() to simplify using something other than doc_id
Improved internal linking within documentation

Better working visualize() function as generic with supported methods
Improved change_colors() with added support for the Okabe-Ito colorset and the option of starting with something other than the first color of a palette. With these changes, color options have been removed from other visualization functions to consolidate them within change_colors().
When a data set includes only one unique doc_id, visualizations are no longer divided into facets.
In an effort to reduce the number of dependencies, many packages have been removed from “Imports” (geomtextpath, ggrepel, glue, NLP, openNLP, plotly, RColorBrewer, stopwords, textstem, wordcloud). Where appropriate, these have been shifted to “Suggests” or dropped entirely.