process_data() no longer prints a note about FCA’s 2024
data when it hits an error. The 2024 files in the AWS S3 bucket are now
fixed, so the note is out of date. Errors from
process_data() are now raised as normal errors instead of
being printed (#47).download_data() now fails gracefully, per CRAN policy
on packages that use Internet resources. If the data can’t be downloaded
(e.g., there is no internet connection, the AWS S3 bucket is
unavailable, or there is no data for the requested year and
month) or unzipped, download_data() prints an
informative message and returns FALSE (invisibly) instead
of throwing an error. It returns TRUE (invisibly) on
success (#47).download_data() documentation now explains that the
2024 RCR7 data files in the AWS S3 bucket differ from FCA’s
files. Ketchbrook added the rows that FCA’s files are missing (#23,
#46).Sometime around 2025-12-08, the Farm Credit Administration (FCA)
added a layer to it’s website to “Verify you are a human” before hitting
the page where the data can be downloaded via hyperlinks. This caused
the download_data() function (which relies on
utils::download.file()) to fail, including in CRAN
checks/builds.
To remediate this issue, Ketchbrook Analytics has replicated the
files on the FCA website (from 2000-present) in a public AWS S3 bucket.
download_data() now retrieves files from the public AWS S3
bucket instead of the FCA website directly.
Addresses CRAN comments regarding:
DESCRIPTIONDESCRIPTIONcompare_metadata()compare_metadata()
outputAddresses CRAN comments regarding:
DESCRIPTION< > characters
in download_data() roxygen commentsget() in get_codes_dict(),
removing need to call library(fcall) before calling
get_codes_dict()LICENSEprocess_data()download_data() has a new (logical) quiet
argument that controls console messaging during download processappend_data()add_user_read_access()R CMD check GitHub Action for automated package
checksImports