get_party_colors() provides a reusable plotting palette for Austrian parties and parliamentary groups, including common aliases and historical ÖVP colors.
Breaking changes
All exported get_*() functions now return a zero-row tibble with their documented columns instead of NULL (or invisible(NULL)) when the API finds no results, accompanied by an informative message. Code that checked is.null(result) should check nrow(result) == 0 instead.
All user-facing errors, warnings, and messages are now signalled via the cli package; message wording may differ slightly.
Bug fixes
get_committees(): zero-row results now retain the exact default or details_type = "members" schema, including typed dates and the members list-column where requested. Member details now return one row per committee, combine the PDF and HTML membership links, ignore illustrated member lists, and select National Council, Federal Council, Hauptausschuss, and special committee layouts defensively. Unsupported layouts warn and yield an empty member tibble instead of a fabricated failure row. Exact citations accept both "1/SA-BU" and canonical "SA-BU/1" order. Echo output now reports the number of rows actually returned, including zero after citation filtering.
get_events(): leaving the legislative period and dates unset now returns events across all available dates, including a completely unfiltered call. The echoed Parliament website URL derives its date and availability filters from the returned rows so it reproduces the API results instead of applying the website’s current-events defaults. Austrian civil-date boundaries are serialized consistently across operating systems. Non-empty results now return a tibble, matching the documented class and the empty-result path.
get_items(): combining person with institution now resolves the person across all institutional categories and applies the institution filter only to the returned items. institution accepts both "NR"/"BR" and the full German names "Nationalrat"/"Bundesrat". When legis_period = NULL, the echoed Parliament website URL now explicitly selects all available periods so it reproduces the package results instead of defaulting to the current period. The historical period codes "PN" and "KN" can now be selected explicitly. An unmatched person returns a typed zero-row result instead of sending an unfiltered item request.
get_mandates(): all empty-result paths now return the complete documented schema, including party and substitute columns. Non-empty results use the same column order and types when optional upstream fields are absent.
get_mps(): empty date-filtered searches now retain the date column as a Date, matching non-empty results.
get_mps(): removed ~160 lines of unreachable dead code after an early return().
get_mps_current(): no longer errors when a search returns no members (previously NULL was piped into mutate()). Empty National Council and Federal Council searches now retain their respective result schemas.
get_mps_details(detail_type = "activities"): the institution filter silently matched nothing, and the institution column leaked the raw German names "Nationalrat"/"Bundesrat" instead of the documented "NR"/"BR". The upstream API changed its gremium vocabulary from short codes to full names. Both directions now speak the new vocabulary, so the filter works again and the column values are as documented. Code that worked around this by matching "Nationalrat" should match "NR".
get_mps_details(detail_type = "committees"): no longer errors for MPs with multiple name variants.
get_names(): no longer returns a bare NA when the person-detail fetch fails, which could break its own vectorized path. Empty results and people without previous-name records now retain the complete documented schema.
get_participation(): the graceful empty-result path was unreachable due to an internal assertion that errored first; empty results now return a typed zero-row tibble.
get_persons(): empty searches with mandates = TRUE, and matching people without mandates, now retain the complete typed mandates_* schema.
get_plenary_meetings(): when legis_period = NULL, the echoed Parliament website URL now explicitly selects every available period instead of defaulting to the current period.
get_transcripts(): the row-limit error message incorrectly said the limit was 10,000 (the actual limit is 100,000); a misplaced sprintf() argument in the PDF-export error message was fixed. All-period searches now echo a website URL that explicitly selects every available period instead of defaulting to the current period.
Enhancements
aux_check_pad_intern_exists() now uses httr2::req_retry() for its preliminary person-page check, allowing at most three total attempts for retryable responses before an identifier is treated as unavailable.
echo = TRUE no longer prints the raw JSON request body; it prints the URL to the corresponding search results on the Parliament website and the number of results. The URL carries the same filter information in a directly usable form, with reserved characters in query values safely encoded.
All Parliament HTTP requests now use the ParlAT package user agent and allow at most three total attempts for retryable responses (httr2::req_retry(max_tries = 3)).
Stale hardcoded browser cookies/session IDs, Chrome/Edge client hints, Fetch Metadata headers, and browser user-agent strings were removed from requests.
get_pad_intern() now removes academic titles from name searches before looking up matching Parliament person identifiers. Previously academic titles could cause searches to fail.
Documentation
Reference pages gained Examples sections, and the get_party_colors() examples now render the actual color swatches.
Fixed dropdown styling in the pkgdown navbar on the package website.
Internal changes
New shared internal helpers (R/utils-shared.R) replace duplicated rename-map and echo/URL-reconstruction blocks across the package.
Debug print() calls, commented browser() calls, no-op req_verbose() blocks, and two large dead functions were removed.
New unit tests for get_names(), get_mps_current(), the detail-page JSON helpers, and the pure auxiliary converters.
Live tests skip count-sensitive assertions while the Parliament search index is degraded (skip_if_api_index_degraded()), and row-count assertions tolerate ±10% drift in live mode (expect_row_count()), so upstream outages no longer look like package regressions. Adds tools/check_api_index.sh for a manual health check and a daily workflow that monitors index health.
Non-standard-evaluation pronouns declared via globalVariables(), resolving the related code-analysis R CMD check NOTEs.
ParlAT 0.0.6
Bug fixes
Fixed errors in detail-page parsing caused by upstream changes on the Austrian Parliament website.
ParlAT 0.0.5
get_item_details() now accepts votes = TRUE by default and returns vote information from National Council item pages in a votes list-column. Set votes = FALSE to omit vote extraction.
get_item_details() now returns the raw item and document type codes as item_type and type_doc.
get_item_details() now returns the human-readable document type as type_doc_long instead of type.
get_item_details() now returns the legislative period as legis_period instead of gp_code.
ParlAT 0.0.4
Bug fixes
get_items(): Restored fallback mappings for legacy and renamed Parliament API export columns (his_url, personen, fraktionen, themen, sw, gp_code, nrbr), ensuring live queries return stable item_url, topic, keyword, person, parliamentary group, and institution columns (#27).
get_plenary_meeting_details(): Fixed handling of sparse plenary meetings that are missing speakers, decisions, or timeline data.
Live API testing: Updated the scheduled workflow against master, and relaxed the duplicate-row assertion for get_items() to reflect non-deterministic upstream duplicates.
Documentation
get_mps(): Documented inclusive date boundary edge case.
ParlAT 0.0.3
Breaking changes
get_plenary_sessions() has been renamed to get_plenary_meetings() to align with the Austrian Parliament API terminology.
get_item_details(): Retrieve detailed information for parliamentary items, including speech/debate data extraction and stage-level metadata.
get_plenary_meeting_details(): Retrieve detailed data for individual plenary meetings, with support for both URL-based and parameter-based input.
Enhancements
get_items(): Significant performance improvements and better handling of compact vs. sparse API export rows. Added CLI progress messages for long-running requests.
get_mps(): New date column in output when date parameter is provided. Improved documentation for gender and mandate_detail fields.
get_mandates(): Enforced mutual exclusivity of name and pad_intern parameters with clearer documentation.
Internal changes
Removed tidyselect from dependencies.
Removed R/sysdata.rda.
R CMD check compliance: replaced T/F with TRUE/FALSE, converted \dontrun{} to \donttest{}, fixed non-ASCII string literals.
Expanded test coverage for new and existing functions.
Added inst/WORDLIST for spell checking and CONTRIBUTING.md.
ParlAT 0.0.2
Minor improvements to input checks and error handling in several data retrieval functions.
Progress messages during longer requests are now shown with cli, for more consistent feedback.
Tests were significantly expanded and cleaned up to improve package reliability.
ParlAT 0.0.1
Initial release of the ParlAT package providing wrappers around the Austrian Parliament Open Data API, including helpers to retrieve committees, events, mandates, members of parliament, plenary meetings, transcripts, and related datasets.
Added convenience utilities to harmonize names, expand legislative period identifiers, and query participation or item-specific datasets with consistent parameter handling.
Published introductory documentation and reference listings to guide users through available endpoints and usage patterns.
