Convert DHS dataset flat file data dictionaries to UTF-8. This addresses parsing parsing issue that arose following R 4.3.0 (#146).
In doing so, it addresses issue of accents in variable and value labels not being parsed correctly.
sf::read_sf
(previously rgdal::readOGR)download_boundaries patch (#138) for pause when
downloading survey boundaries. Set default sleep to 5 seconds to avoid
timeout.available_datasets patch (#131) for new version of
qdapRegex on CRANmodel_datasets with
rdhs::model_datasets to avoid errors if rdhs
namespace is not loaded.available_datasets patch (#115)Remove class "dhs_dataset" from downloaded micro
data sets. This class is not anywhere and it creates an error for
dplyr_v1.0.
Cached datasets will need to be re-downloaded after updating to clear the dhs_dataset clas.
Replace readLines() with
brio::read_lines() to make parsers robust to Windows
encoding issues (similar to
https://stackoverflow.com/questions/18789330/r-on-windows-character-encoding-hell).
Use "sf" as default download method for
download_boundaries(..., method = "sf"). Add arguments
quiet_download and quiet_parse = TRUE to
download_boundaries(). quiet_download (default
FALSE) controls download.file() messages.
quiet_parse (default TRUE) controls messages
from sf::st_read() when
method = "sf".
download_boundaries for downloading spatial
boundaries using (#71)dhs_gps_data_format for DHS GPS Information
(#74)unzip_special correctly detects 4Gb files (#43)delabel_df for converting labelled data frames to
characters (#54)extraction(add_geo=TRUE) correction for Kenya 2014
surveys (#67)New as_factor.labelled for backward compatibility
with haven <2.0.0 labelled
classes.
model_datasets now internal and exported dataset
(#60).
New vignettes: country_codes
Documentation typos corrected (#55)
New set_rdhs_config for providing login credentials.
This deprecates set_dhs_credentials.
New get_rdhs_config shows the credentials currently
used by rdhs
Permissionn from user now required for file saving.
API requests can now ignore any cached responses
(force = TRUE) argument (#23):
dat <- dhs_countries(force = TRUE)get_datasets(clear_cache = TRUE) will clear the
cached available datasets, enabling newly requested datasets to be
downloaded (@kaisero,
#29).
geojson objects can now be requested from the API so that you can return geojson objects for mapping purposes (#28) e.g. :
d <- dhs_data(countryIds = "SN", surveyYearStart = 2014,
breakdown = "subnational", returnGeometry = TRUE,
f = "geojson")
# convert to spatial object
sp <- geojsonio::as.json(d) %>% geojsonio::geojson_spdhs_data(), dhs_countries() and other
API functions (dhs_x()).authenticate_dhs() now works with short project
names.New vignettes: anemia
New read_dhs_flat() for reading flat
datasets.