Closed GregorDeCillia closed 3 years ago
The commits so far should cover all necessary changes to R/
except some functions for the resources article.
Missing articles
Additional TODOs
pkgdown/
od_downloads()
and od_cache_summary()
print()
method for the return value of od_json()
Reminder to self: The current articles assume that you have a persistent cache set up in ~/.cache/STATcubeR/open_data
. There might be room for improvement.
Github is currently experiencing some issues so I'll take the time to summarize the progress so far
This branch is now "feature complete". There are two changes that will make sure devtools::check()
runs without errors
sc_meta()
print(x, ...)
to print.od_json
After that, I need to check if the old articles about the STATcube REST API are still working properly. They did some days ago so only minor changes are not to be expected.
The current version passes devtools::check()
with no errors, warnings or notes. Furthermore, the pkgdown site can be rendered without any issues.
@bernhard-da if you want to do updates on the docs, build them locally and transfer them to the webserver with our internal useSTAT
package.
pkgdown::build_site()
useSTAT:::deploy_pkgdown_dev()
I'd recommend to set up caching for development by modifying .Renviron
. Otherwise, you'll download the same cubes many times when building the articles and man pages.
usethis::edit_r_environ()
STATCUBE_KEY = YOUR_API_KEY_GOES_HERE
STATCUBE_CACHE = TRUE
OD_CACHE_DIR = "~/.cache/STATcubeR/open_data/"
STATCUBE_CACHE_DIR = "~/.cache/STATcubeR/api/"
On that note - some possible improvements
od_cache_dir
for consistencyAn updated version of the docs is available at https://statistikat.github.io/STATcubeR/dev/
@bernhard-da could you please provide some feedback about the current state of the pkgdown documentation? I don't think it needs to be perfect but having a second opinion would be very useful.
Some remarks about the links
dev/
in the url. This will be resolved when this PR is merged and the pkgdown site is deployed to https://statistikat.github.io/STATcubeR/I suggest you start with article index and the reference index and decide if we should update any names or descriptions. The articles linked in the navbar are the most important ones.
I added you as an author and linked to your gh-account. Just let me know if this is okay https://github.com/statistikat/STATcubeR/blob/8c564510aac366794b67e8ddffe4eda60e8444e2/pkgdown/_pkgdown.yml#L13-L14
@GregorDeCillia very nice work, all articles are very well written and very clear; just some very minor things
https://statistikat.github.io/STATcubeR/dev/articles/sc_table.html
you could remove the part with the raw
response as you will probably remove it in the future I guess@bernhard-da thank you. The link is now fixed in the VCS. Yes: the sc_table class
is still a little bit weird: see the full class documentation. It is probably a good idea not to mention that in the introduction docs.
I am not sure if $raw
will be fully removed because there are certain parts that are "lost" when we do so, but it is certainly something for advanced users.
Maybe this PR should also be used to refactor the dependencies
magrittr
is not really necessary but makes for prettier codehttr
is needed for API calls. jsonlite
is necessary to parse the httr
responses.rmarkdown
is pretty much irrelevant since the vignettes
are only available onlinedata.tree
provides a print method for schema responses but nothing else.rstudioapi
allows sc_table()$json$edit
but nothing elsexml2
is only used in od_list
to do some webscrapingpillar
is just used in .onLoad()
to register pretty printing in case the pkg is availableI'm still wondering why devtools::check()
doesn't complain about {R6}
not being in the dependencies.
This PR will provide some documentation updates regarding the new Open Government Data functionalities
vignettes/
vignettes/
man/