wpgdata 0.3.0

Breaking changes

• peg_get(), peg_query(), and peg_all() have been removed and
  consolidated into a single peg_data() function. Update existing code:
  • peg_get("id") → peg_data("id")
  • peg_query("id", filter = x > 1, select = "col") → peg_data("id", filter = x > 1, select = "col")
  • peg_all("id") → peg_data("id")

New functions

• peg_data() — replaces peg_get(), peg_query(), and peg_all() with a
  unified interface for fetching dataset rows. Supports server-side filtering,
  column selection, row ordering, and offset pagination via top and skip.
  All pages are fetched in parallel using httr2::req_perform_parallel(),
  delivering significant throughput improvements over the previous sequential
  approach.

peg_catalogue()

• Catalogue pages and per-dataset metadata are now fetched in parallel using
  httr2::req_perform_parallel(), with the count request and first page fired
  simultaneously to save a full round trip.
• Metadata is now sourced from the /api/views/{id}.json endpoint per dataset,
  replacing the Socrata Discovery API as the sole data source.
• Added new columns: row_count, col_count, view_count, group,
  department, update_frequency, quality_rank, license, license_link,
  tags, license_id, rows_updated_at, view_last_modified,
  publication_date, index_updated_at.
• row_count is now sourced from cached column statistics (cachedContents),
  making it fast without requiring a separate OData count request per dataset.
• Removed the default limit of 200 — all available datasets are fetched by
  default.
• Results are sorted by rows_updated_at descending.
• HTTP errors and network failures for individual datasets are caught and
  skipped with a warning rather than aborting the entire fetch.

peg_info()

• Added input validation for dataset_id (rejects NULL, NA, empty string,
  non-character, and length > 1).
• Added httr2::req_timeout() and httr2::req_retry() with exponential
  backoff for transient 429, 500, and 503 responses.
• Fixed double-read of the response body; the empty-body check is now handled
  by a single tryCatch around resp_body_json().

peg_metadata()

• Added input validation for dataset_id matching peg_info().
• Added httr2::req_timeout() and httr2::req_retry() with exponential
  backoff for transient 429, 500, and 503 responses.
• Fixed double-read of the response body.
• Added %||% NA_character_ fallbacks for name, field_name, and type —
  previously these would silently insert NULL for malformed column entries.
• Replaced soft-deprecated purrr::map_dfr() with purrr::map() +
  purrr::list_rbind().

wpgdata 0.2.0

New functions

• peg_catalogue() — list all datasets available on the City of Winnipeg
  Open Data Portal with name, ID, category, and last updated date.

Internal changes

• Added get_catalogue_count(), fetch_catalogue_page(), and
  parse_catalogue() internal helpers to utils.R.
• Imported rlang::.data pronoun to suppress R CMD check notes in
  dplyr::mutate() calls.

wpgdata 0.1.0

Initial release

• peg_get() — fetch the first page of a dataset
• peg_query() — filter, select, sort and paginate datasets
• peg_all() — download all rows with automatic pagination
• peg_metadata() — retrieve column names and types
• peg_info() — retrieve dataset-level information
• Supports both R expressions and raw OData strings in peg_query()
• Automatic pagination detection and progress bar in peg_all()