statnipokladna: Use Data from the Czech Public Finance Database

[x] sp_get_dataset should accept multiple time periods, as sp_get_table does now should return a vector of paths to ZIP files (not unzip outputs)

export sp_get_dataset_url

This should be the only breaking change for the user - sp_get_dataset no longer returns

sp_get_table [do we need sp_get_table_file to return path to the right local unzipped file?] should call sp_get_dataset on its inputs (multiple time periods if exist) should then iterate over the output of sp_get_dataset using sp_load_table(), which takes an n >= 1 vector of paths to ZIP files as an input
New: sp_get_table_path: new function that takes ZIP path and returns CSV path
New: sp_load_table: new function that takes file path and table ID as input (or just dest_dir? That would be sufficient but would not improve targets-style workflow)

new sp_get_codelist_url fn broken out, takes codelist ID as input and returns URL to XML as output new sp_get_codelist_file fn broken out of sp_get_codelist, takes codelist ID as input and returns XML file path as output new sp_load_codelist fn broken out; takes codelist XML file path as input and outputs tibble

sp_get_codelist becomes a wrapper around these two

No changes visible to users of sp_[get|add]_codelist

for tables and you can either call the two/three steps sequentially or use a higher-level wrapper

note in readme and in getting started vignette
cross refs between functions in See also (if section can be edited) - or better label the lower-level fns with "Fine-grained control workflow"

set time periods, don't rely on defaults
think about whether you need to keep intermediate files:
if so,
use sp_get_dataset as input to your make/targets workflow
set dest_dir in function call or sp.dest_dir in your project-specific options
if not,
still build custom functions that save output of sp_get_table|codelist into file and use the file path in your make/targets workflow - that gives you visibility of the data files on which your analysis depends
make sure you do not rerun sp_get_table unnecessarily - that takes time and bandwidth and more importantly can break your results if the source data change
think about how you use redownload