Compatibility with stringr v.1.6.0 (#218, @hadley).
Transfer of Dates to SQL Server now works again
(#220).
DiseasystoreBase$key_join_features() no longer accept
character stratifications – must use rlang::quos()
(#203).The $observable_regex field is added, which shows
the regex that demarcates observables from stratifications
(#204).
Improved clean up of temporary tables (#207). Tables created
internally in diseasystore are now labelled and removed
when no longer needed.
In DiseasystoreBase$key_join_features(),
stratifications no longer need to do computation on other features
(#203).
...) are now passed to the
?FeatureHandler$compute() and
?FeatureHandler$get() functions in
?FeatureHandler (#162). Furthermore, a reference to the
diseasystore is now passed to
?FeatureHandler$compute() as ds to give the
?FeatureHandler access to other features via
?DiseasystoreBase$get_feature().New age helpers add_years() and
age_on_date() has been added to help compute features for
individual data (#125).
Added the fields
?DiseasystoreBase$available_observables and
?DiseasystoreBase$available_features for easier overview
(#139).
The data availability period for each diseasystore
is now exposed via the ?DiseasystoreBase$min_start_date and
?DiseasystoreBase$max_end_date fields (#138).
diseasystores can now have variable backend support
(#175). In case a backend is insufficient to support the computations,
the backend can be blocked for the diseasystore.
test_diseasystore() now have a skip_backend
argument to skip tests for the disallowed backends.
Added the ?DiseasystoreSimulist which implements a
synthetic, individual-level, data set generated with
simulist.
A bug was fixed where data was duplicated when features were not divided into separate tables (#192).
Two bugs were fixed in
?DiseasystoreBase$determine_new_ranges() where existing
tables were not detected:
{SCDB} > v0.4) (#158).slice_ts not being correctly matched with
existing data on some backends (#172).Long stratification expression are now properly parsed in
?DiseasystoreBase$key_join_features() (#161).
When creating your own diseasystore it is now easier
to inherit from the base module (#189).
When creating your own diseasystore it is now easier
to inherit from the base module (#189).
Warnings about existing stratifications now only appear if the requested stratification contains a named expression matching an existing stratification. Simply requesting the stratification from the feature store will no longer produce a warning (#192).
diseasyoption() can now be called without the
option argument to return all
diseasy/diseasystore options (#159).
In addition, a new namespace argument can restrict the
option look-up to a specific package
(e.g. diseasystore).
test_diseasystore() now also checks the
?FeatureHandler return data directly:
valid_from and valid_until has
class Date (#154).valid_from and valid_until
columns are chronologically ordered (#176).An example has been added for building a
diseasystore with individual level data (#162). See
vignette("extending-diseasystore-example").
Added benchmarking vignette vignette("benchmarks")
(#144).
%.% operator is made more flexible to function as a
drop-in replacement for $ (#145).{SCDB} v0.3 is removed.?DiseasystoreEcdcRespiratoryViruses: A feature store
that uses the ECDC Respiratory viruses weekly repository
(#124).
With the release of {SCDB} v0.4, we now support more
database backends:
SQLite with attached schemas (#121).
Microsoft SQL Server (#128).
PostgreSQL (#128).
DuckDB (#127).
test_diseasystore() is added to provide a
standardised method for testing new diseasystores (#123).
diseasyoption() now allows a default option to be set
with the .default argument (#122).Improved the stability of internal functions.
Adapted to release of {SCDB} v0.3.
Improved test stability when internet is unavailable.
Reduced the data footprint during tests.
?DiseasystoreBase: A base class for the
diseasystores.
?DiseasystoreGoogleCovid19: A feature store that
uses the Google COVID-19 Open Data repository.
?FeatureHandler: A simple class to handle individual
feature computations.
Aggregators: A set of aggregators for the
?FeatureHandler to use.
drop_diseasystore(): A function to remove data from
a feature store.
Added a informative operator to access data %.%.
NULL if element does not
exist.age_labels(): A function to generate human-readable
and sortable age groupings.
The functions are fully documented.
Vignettes for the use of the package is included.
vignette("quick-start")vignette("extending-diseasystore")Vignette for the Google COVID-19 data is included.
vignette("diseasystore-google-covid-19")