Changes in version 0.1.0                        

  - Initial release.
  - New package hex logo: the major road network of the São Paulo
    Metropolitan Region in the INFOSIGA-SP dark blue (data-raw/logo.R).
  - read_infosiga() downloads (with caching) and imports the three
    INFOSIGA-SP datasets (sinistros, pessoas, veiculos) as tidy tibbles,
    handling the source encoding, decimal marks and date formats.
  - read_infosiga() returns a processed dataset by default (clean =
    TRUE): text columns are whitespace-trimmed (the source pads some
    fields to a fixed width), "NAO DISPONIVEL" becomes NA, ordinal
    columns (dia_da_semana, turno, gravidade_lesao, age bands) become
    ordered factors, the ano_mes_* year-month columns are parsed to
    first-of-month Dates, the binary tp_sinistro_* crash-type flags
    become logical, tempo_sinistro_obito becomes integer, the spurious
    ".0" on numero_logradouro is stripped, and coordinates outside the
    São Paulo state bounding box are dropped. Use clean = FALSE for the
    raw data as published, or clean_infosiga() to process a raw import
    afterwards.
  - infosiga_download() pre-fetches the source archive into a local
    cache. It tries the official DETRAN-SP endpoint first and falls back
    to a GitHub-release mirror if it is unavailable; additional mirrors
    can be supplied via the infosigasp.zip_url option (a character
    vector tried in order).
  - read_infosiga() and infosiga_download() warn when a cached archive
    is reused that is older than the infosigasp.stale_days option (30
    days by default; set to Inf to disable), since DETRAN-SP refreshes
    the data monthly under the same file name.
  - infosiga_cache_dir(), infosiga_cache_list() and
    infosiga_cache_clear() manage the on-disk cache.
  - infosiga_datasets() lists the available datasets and their keys.
  - infosiga_dictionary() downloads the official data dictionary.