statnipokladna 0.7.0

BREAKING CHANGES

sp_get_dataset() no longer unzips the downloaded archive. It therefore returns the path to the downloaded zip file, not to the unzipped files. This is needed to support a more modular workflow (see below and vignette("workflow", package = "statnipokladna")).
The year param in sp_get_table() and sp_get_dataset() now has no default. This is more sensible when there is no easy way to determine the latest available table/dataset and in any case better for reproducibility. The new sp_get_dataset_url() also has this updated behaviour.
the columns signifying the time period of the result of sp_get_table() have been renamed for more clarity and consistency:
- period_vykaz to vykaz_date
- per_m to vykaz_month
- per_yr to vykaz_year
sp_add_codelist() no longer creates messy column names in the form of [codelist name]_nazev_nazev

the core functions have been rewritten into a more modular architecture and their constituent modules exported. This allows more fine-grained control over workflows using lower-level functions to accommodate caching and reproducibility e.g. via {targets} or {drake}.
- sp_get_[dataset|table|codelist] are now effectively wrappers around several lower-level functions
- those previously using these core functions should see no change except for one breaking change in sp_get_dataset() (see above).
- the lower-level functions enable step-by-step workflows with transparency of intermediate steps (URLs, downloaded ZIP archives, pointers to specific CSV files, etc.) See vignette("workflow", package = "statnipokladna")

better error messages around mismatches between table ID and file in archive
more informative errors when online files are not available
improve date parsing in sp_get_codelist() to handle inconsistent date formats in some codelists
faster loading of tables from previously downloaded and unzipped datasets
empty strings in codelists are now returned as NA
more columns are now properly named in the output so they can be linked to codelists, incl. columns related to programme spend and rifngfenced ('purpose'-marked) spend
rozprog (programme spend ID) as well as nastroj and nastrojanal codelists are now recognised

when setting dest_dir in sp_get_*() functions, files are now put into the right directory even if dest_dir does not contain a trailing slash
sp_get_dataset_doc() now creates dest_dir if it does not exist, as advertised in the documentation
upgrade dplyr dependency to aviod mysterious "unused argument" error in codelist functions
balance sheets of city districts can now be loaded for all years where they are available
examples are now safe to run on CRAN