1.6.0 (2024-05-16) ================== * added translation parameter to mp_corpus function to allow for retrieving (english) translations * added mp_parties function to access the party datasets * added mp_corpus_df convenience function to retrieve the corpus as tibble/data.frame * added mp_corpus_df_bilingual convenience function to retrieve the corpus with both the original language text and the (english) translated text as tibble/data.frame * added as_tibble parameter to mp_corpus function to retrieve corpus as tibble/data.frame * added as_tibble function for ManifestoDocument and ManifestoCorpus objects * added tibble_metadata parameter to mp_corpus function to select document level metadata variables * count_codes now excludes "H" codes by default to avoid unintended effects on category percentages and subsequent problems in scaling functions * added detailed missing documents explanation for mp_corpus* requests * added na.rm.voteshares and na.rm.positions parameters to median_voter function * adaptations for the integration of the south america dataset into the normal dataset * added missing function export of recode_cee_codes.ManifestoDocument * updated vignette with the new features and a clarification for enhanced reproducability * added only_stable parameter to mp_corpusversions to hide corpus development versions by default * Bugfix and clarification in aggregate_pers function regarding handling of NAs/overwriting * Bugfix in mp_scale.ManifestoCorpus to handle handbook v5 documents correctly * Bugfix in baeck_policy_dimensions (thus meyer_miller_2013_policy_dimensions and mp_nicheness) by adding 110 category to foreign dimension * Bugfix in as.data.frame to avoid all metadata columns being cast to character class * Bugfix in mp_metadata regarding the ordering of the results in a specific scenario * Bugfix of a small cache handling inconstency * Bugfix for the rare scenario requesting the same item multiple times in a single mp_corpus request * removed some unnecessary warnings when filtering mp_corpus by variables with NA values in the dataset * removed some unnecessary type casting warnings * changed the default manifesto project api root url 1.5.0 (2020-11-29) ================== * added 2019 version of franzmann-kaiser issue structures (while allowing for also using older ones) * Bugfixes in mp_availability to avoid errors when using more modern dplyr versions * Bugfix in scale_ratio and subsequently split into two distinct new functions to avoid confusion * Bugfix in mp_bootstrap and added feature to auto-rescale to the sum of the columns to be permuted * refactoring of code to prevent deprecations when using more modern versions of package dependencies * some smaller things (added cee_categories function; updated v5_categories, mp_describe_code, vignette) 1.4.0 (2020-04-20) ================== * added mp_dedication function in commemoration of Jirka Lewandowski 1.3.0 (2018-05-22) ================== * added codebook interface: three functions mp_codebook, mp_view_codebook, and mp_describe_code to investigate the codebook and the descriptions of individual categories * Important bugfix in mp_clarity: incorrect results when using election-level saliency weighting * rename of aggregate_cee_codes to recode_cee_codes to be more consistent in function naming * new function aggregate_pers_cee which is a wrapper of aggregate_pers using cee_aggregation_relations to aggregate cee categories to the main categories in the main dataset * v5_categories function is now also exported * Bugfix in mp_maindataset: id_perm variable should not be converted to numeric (and thus become NA) 1.2.4 (2017-05-08) ================== * Bugfix in mp_maindataset: variable "corpusversion" (exists from dataset version 2016b onwards) should not be set NA * Bugfix in mp_interpolate when country-timeseries starts with NAs * changes in the of dplyr functions to be compatible with dplyr 0.5 and dplyr 0.6 * Dependency on readxl 1.0.0 for better column type guessing in downloading xls dataset file 1.2.3 (2016-07-22) ================== * changes in the use of dplyr function to be compatible with dplyr 0.5 and tibble 1.1 1.2.2 (2016-06-28) ================== * small but important bugfix to give full access to new dataset version 2016a: new variables partyabbrev and datasetversion should not be converted to numeric automatically 1.2.1 (2016-06-20) ================== * Important bugfix in nicheness_meyer_miller (also mp_nicheness): Nicheness values according to Meyer & Miller were wrongly assigned to datapoints! * Bugfix in issue_attention_diversity: Error list of length ... not meaningful * franzmann_kaiser scaling now works also for future versions of haven (> 0.2.0) * Fixed that mp_coreversions returned corpus versions instead under certain circumstances * mp_cite now not only prints a message but returns a data frame with citations * mp_maindataset now takes version also in format "2012a" and not only "MPDS2012a" 1.2 (2016-03-21) ================ * South America Dataset and Corpus are now available in manifestoR. The respective documents are availabe since corpus version 2016-2 and require manifestoR 1.2 to be accessible. * mp_availability now returns a data.frame with attributes and not a list. This might break code backwards, you cannot acces mp_availability()$availability any longer. What was there before is now directly the return value of mp_availability. * You can download the Manifesto Project Main Dataset now also in binary formats (dta, xlsx, sav) and read it from there in order to obtain labelled columns as in STATA. * codes() now returns a vector and not a data.frame() * main dataset uses strings not factors * programmatic clarity measure implemented by Sven Regel * Franzmann & Kaiser 2009 scaling procedure included * With this a new function to split the Belgium party system is introduced * Subsetting metadata now works also by year * Citation handling for MZES documents implemented 1.1-1 (2015-11-12) ================== * Correct and extended startup messages * Nicheness measures by Meyer & Miller 2013 and Bischof 2015 implemented * Corpus documents can now be accessed via their manifesto_id (= party_date) and not only by number * implementation change of tm Source and reader for ManifestoCorpus * counting codes reproduces dataset style for handbook version 5: subcategories + aggregated percentages * issue attention diversity measure after Greene implemented in issue_attention_diversity * More user-friendly error messages in case of incorrect requests (metadata, API key) * fixes among others https://github.com/ManifestoProject/manifestoR/issues/1 1.0-5 (2015-09-14) ================== * Dependency increased to dplyr 0.4.3, since 0.4.2 contains a bug affecting manifestoR code * Codings are now treated as character vectors to work with handbook version 5 codings * Citation information is now requesting the API * Several small bugfixes