- update dependencies to latest versions (see diff of requirements.txt)
- DataSpec and conversions support nullable and non-nullable data types in numpy and pandas
- Previously undefined behaviour of nullable data types must now be defined in conversions
- fixed make_epw_file and adding test_make_epw_file
- fixed DataClient issue with data type conversion after filling nulls
- fixed DataClient.py removal of columns that get truncated to all NA
- simplify
.env
setup and usage - adding archetypal .idf building model geometries
- remove pytest logging of dependency libraries and only show logs from failed tests
- adding
DataClient.generate_dummy_data()
to simplify demo and testing data - improve
demo_LocalSource.ipynb
- DataSpecs can now be partially filled to allow for multiple data sources for same specification
- ControllerModel now has
options
attribute that defines deadband size - no longer include outdoor temperature in null check columns to allow for automatic filling of missing weather data
- fixed
make_data_directories
usage when no local_cache is given - fixed
get_local_cache_file
usage when no local_cache is given - fixed
DataSpec.py
null_check_columns and units - conditional skipping of
DataClient
tests that use external data sources if those sources are not configured