Changelog • poldis

poldis 0.2.0

2026-09-17

Added a US presidential speeches vignette presenting a case study with plots
Improved documentation and examples across exported functions
Improved README to be shorter and clearer
Fixed tests to skip gracefully when spaCy (Python) is unavailable, rather than erroring
Removed ORCID field from DESCRIPTION to avoid CRAN submission issues

Renamed extract_title() to extract_first_sentence() to be more descriptive
Removed extract_date()
Removed sim_urgency() for now, as it does not work well with the updated get_urgency()
Improved annotate_text() to parse large text vectors in chunks, avoiding memory issues and R crashes
Improved annotate_text() to initialise the spaCy backend only once, increasing efficiency
Improved get_urgency() to score from word stems instead of lemmas, increasing accuracy and speed
Improved get_urgency() scoring so that the urgency score is the sum of the other dimensions divided by the “must” score
Added a “min” option to the summarise argument of get_urgency()
Fixed get_urgency() to return 0s, rather than NAs, when scores are genuinely zero
Fixed get_urgency() where spacing was being replaced by dots
Improved gather_topics() to match on word stems instead of lemmas, increasing accuracy and speed
Fixed a bug in priority identification in select_priorities()
Improved read_pdf() detection of non-readable PDFs

Renamed and replaced US_News_Conferences_1960_1980 with US_inaugural_addresses_1993_2025
Improved internal data to a single dataset with scores from the second urgency survey
Removed the urgency codebook (please refer to the paper for details)
Moved raw replication data to a separate data branch to reduce package size

CRAN release: 2024-09-04

2024-09-03

Updated documentation for get_urgency() to include more details about urgency scores
Updated internal data and data raw files to increase transparency

Updated get_urgency() and gather_topics() to open codebooks for urgency or topic when no arguments are declared
Updated urgency scores in get_urgency() according to survey results.
Added sim_urgency() function to simulate urgent priorities
Updated select_priorities() to be more strict and accurate about selection of priorities

CRAN release: 2024-07-21

2024-07-17

Removed redundant terms from description for CRAN submission
Removed commented out code lines and examples in documentation for CRAN submission

2024-07-15

Updated text tools
- Renamed text tools functions to start with “extract_” (extract_names(), extract_title(), extract_context(), extract_date(), extract_locations(), extract_match())
- Closed #11 by adding extract_similarities() function to fuzzy match texts
- Added annotate_text() function to parse words or sentences using NLP
- Added read_pdf() function to load readable and non-readable text files from PDFs
Closed #14 by adding select_promises() function to extract future promises in text
Closed #15 by adding gather_topics() and gather_related_terms() for assigning topics to texts
Added get_urgency() function for coding urgency from text
Added summary and plotting methods for “urgency” and “topics” classes

2022-09-25

Renamed all functions to consistently start with extract_*
Made the package smaller and more concise by removing extract_dates() and additional commented codes
Closed #6 by updating extract_context() function to make it more flexible for users
Updated extract_speaker() to only return a list of speakers for texts
Closed #4 and #5 by removing str_translate() function for translating strings

2022-02-07

2021-11-04

2021-03-02