Changes in version 1.0.2 (2026-05-15) - correcting test file to resolve CRAN checks error - CRAN release Changes in version 1.0.1 (2026-05-12) - removing duplicate test files - CRAN release Changes in version 1.0.0 - major re-factoring, stable release Changes in version 0.10.30 - Cleaned helpfile to hide internal utility functions from end users. - Removed obsolete function structureData() and related documentation. - Marked classificationEndpoint() and correspondenceList() as deprecated, in favour of classificationList() and correspondenceTableList(). - Harmonised user-facing extraction functions: all now expose a showQuery argument with consistent ordering of parameters. - Clarified in the documentation which functions accept both data.frames and CSV file paths as input (e.g. classificationQC()). - Added basic unit tests to ensure internal utilities are not exported and that the interface of extraction functions is stable. - Reworked classificationList() to: - use a JSON configuration file for endpoint URLs with sensible fallbacks, - provide a stable Prefix (lower-case, no spaces) compatible with retrieveClassificationTable(), - add a Languages column summarising all languages used in concept labels, - support offline mode via the global option useLocalDataForVignettes. - Made classificationEndpoint() a thin deprecated wrapper around classificationList(), with dedicated tests checking argument names, returned values and deprecation warnings. - Updated aggregateCorrespondenceTable() and analyseCorrespondenceTable() to: - accept both CSV file paths and data.frames for all table inputs (AB, A, B), - standardise input validation through testInputTable(), - improve error messages for duplicate codes, missing values and hierarchy inconsistencies, - keep CSVout as the last argument and validate it via testCsvParameter(). - Simplified retrieveClassificationTable(): - removed the obsolete localData argument (behaviour now governed solely by options(useLocalDataForVignettes = TRUE/FALSE)), - ensured that both offline (embedded CSV) and online (SPARQL) modes respect the showQuery contract (data.frame vs list with SPARQL.query), - documented the argument order (endpoint, prefix, conceptScheme, language, level, CSVout, showQuery). - Aligned the interface of retrieveCorrespondenceTable() so that endpoint precedes prefix and updated the documentation and tests accordingly. - Updated lengthsFile() to: - clarify that correction is a Boolean argument, - handle special classifications (CN, PRODICOM, CPA, etc.) in a more explicit and documented way, - issue clearer warnings when code lengths are inconsistent across levels. - Reordered arguments of newCorrespondenceTable() so that: - the logical/semantic arguments come first (Tables, Reference, MismatchTolerance, Redundancy_trim), - CSVout is now the final, optional output argument, - examples and tests now call the function with explicit argument names to avoid ambiguity. - Updated updateCorrespondenceTable() so that: - all classification-related inputs (A, B, Astar, AB, AAstar) accept either CSV file paths or data.frames, - CSVout is the last argument, - input validation and CSV export are routed through the same shared helpers as the other correspondence-table functions. Changes in version 0.10.29 New features - Added correspondenceList() to list available correspondence tables in CELLAR and FAO, with a structured output and improved robustness against malformed SPARQL responses. - Extended classificationList() with: - endpoint ("CELLAR", "FAO", "ALL") - showQuery to return SPARQL queries. Now supports offline mode via options(useLocalDataForVignettes = TRUE) using bundled CSV files for reproducible vignette builds and offline work. Improvements and API changes - correctionClassification(): - Added new prefix argument to enable prefix-specific corrections. - Replaced SPARQL example with static demo data (inst/extdata/Nace2.csv) so that examples work offline and during CRAN checks. - Documentation enhanced with clearer input/output description. - lengthsFile(): - Improved detection of inconsistent code lengths across levels. - More explicit warnings when code hierarchies do not behave as expected. - retrieveClassificationTable(): - Cleaned all non-ASCII characters in comments and messages. - Strengthened error handling when SPARQL queries fail or return unexpected number of columns. - Added explicit imports for utils::capture.output and utils::str. Quality control module - classificationQC() functionality has been temporarily disabled for this release (tests skipped), but its supporting imports (stringr, tools) remain in place for the upcoming version. Testing improvements - Added new unit tests for: - dataStructure() : validation of returned level/size structure. - prefixList() : consistency of prefix extraction for each endpoint. - retrieveCorrespondenceTable() : correct parsing, duplicate checks, error handling. - retrieveClassificationTable() : correct structure, column naming, error detection. - Updated existing tests for: - analyseCorrespondenceTable() : better handling of missing files, mismatches, duplicates, and filtering options. - aggregateCorrespondenceTable() : improved robustness when validating AB tables. Changes in version 0.10.22 Changes and Improvements - Rewrote the logic for classificationList() and correspondenceTableList() to handle the "ALL" endpoint through recursive calls, ensuring consistent and robust behavior. - Internal function prefixList() is now no longer exported, as it is not meant for end-users. It remains accessible via correspondenceTables:::prefixList() if needed. - Cleaned up the "RetrieveClassificationsAndCorrespondenceTables" vignette: - Restored uncommented example code (e.g., correspondenceTableList("ALL")) to improve readability and reproducibility. - Removed internal calls to prefixList() from examples to avoid confusion. - Fixed vignette rendering errors by defining missing intermediate objects (e.g., result_ls). - Improved @examples sections and corrected Rd documentation accordingly. Package Infrastructure - Reduced export clutter by making utility-only functions internal. - Ensured full R CMD check compliance (0 errors, 0 warnings, 1 note on line width). Minor - Cleaned up documentation and fixed formatting for clarity. - Updated vignettes titles to ensure consistency between YAML and \VignetteIndexEntry. Changes in version 0.8.1 - adding the functions prefixList(), classEndpoint(), correspondenceList(), retrieveCorrespondenceTable(), retrieveClassificationTable(), structureData() Changes in version 0.8.0 - adding test cases - corrections for the newCorrespondenceTable() and updateCorrespondenceTable() functions Changes in version 0.7.5 - updating the vignette showing the initializing code part Changes in version 0.7.4 (2022-09-27) - aligning to the second round of comments from CRAN - replacing readr::write_excel_csv() with data.table::fwrite() - removing absolute paths from vignettes and examples - the example of newCorrespondenceTable() using only the 100 hundred rows of the classifications - more details in the error messages - CRAN release Changes in version 0.7.3 - aligning to the comments from CRAN - using txtProgressBar() and setTxtProgressBar() functions - replacing cat() with message() function - examples updated to be able to test it automatically Changes in version 0.7.2 - adjustment for CRAN approval Changes in version 0.7.1 - initial public release on GitHub