This is a minor release including bug repairs and elimination of a defunct package dependency.
YRmisc dependency from namespace and repaired compute_lagcorr function that relied upon on it, replacing the original with a base R alternativeread_dyads and read_1file functions to repair column header name TextRaw so that it matched text_raw that is used in subsequent cleaning stepsThis is a major release including several new features and small changes.
generate_shams to shuffle the order of interlocutor turns within each conversation. This creates a group of control conversations.remove_backchannel in prep_dyads allows users to remove turns composed entirely of stopwords, preventing them from being filled in later stages.summarize_dyads now computes dAUC for a lagged time series in addition to the unchanged. This is designed to account for both ways that interlocutors can be paired in each exchange.This is a minor release involving one bug repair
This is a minor release involving mostly optimization of functions.
This is a major release. We have included many new features and fixed numerous bugs.
prep_dyads() that combines two previous preparation steps clean_dyads() and align_dyads()read_1file() that formats a conversation transcript already in your R environmentomit_stops to prep_dyads specifying optional stopword removalwhich_stoplist to `prep_dyads`` specifying a four stopword list options (SMART_stops, MIT_stops, Temple_stops25, CA_orig_stops).ConversationAlignread_dyads() will throw error and warning message if any transcript has more than or less than two participants for any individual conversation (Event_ID)corpus_analytics() function that produces a variety of descriptive summary statistics (e.g., n-conversations, type token ratio, average turn length) in a table format.summarize_dyads() restructured in significant ways including the addition of three new arguments: custom_lags default lags for correlation are set at -2.0,2 users are free to specifiy additional lags , default is NULL, sumdat_only produces a summary dataframe with values averaged to two rows per conversation (one for each participant, corr_type specifies correlation to apply to lagged datasummarize_dyads() has been omitted in favor of proportionally rescaling dAUC to a standardized/fixed number of turns (100)lookup_db, replaced with lookup_Jul25, includes many more dimensions and options for scaled vs. raw versions of variablesCA_OriginalStops. However, the default stopword list is new - Temple_Stops25align_dyads()