Releases: ONSBigData/precon
Releases · ONSBigData/precon
Release v0.7.0-alpha.5
API changes
- Changed
base_price_ffill_shiftto justffill_shift - Added quality adjustment to
index_calculator - Modified
index_calculatorso that it now acceptsbase_pricesas a parameter, and fills and shifts them when it does.
Bugs
- Fixed in
get_base_priceswhere.eq()method was being used instead ofisin().
Release v0.7.0-alpha.4
- Fixed bug in get_base_prices
- Added function base_price_fill_shift
Release v0.7.0-alpha.3
- Changed
get_base_pricesso it now fills within year, and also shifts the values as default behaviour. - The
aggregatefunction now sets weights where indices are 'inf' to zero. - The
aggregatefunction ensures periods where all indices are invalid are set to NaN rather than zero to avoid
aZeroDivisionErrorand return NaN instead.
Release v0.7.0-alpha.2
- Fixed bug in subset_shared_axis: now using method
.equals()instead of non-existent method.eq().
Release v0.7.0-alpha.1
- Added a subset step to
aggregateandimpute_base_pricesto reduce DataFrame arg sizes when functions are applied using groupby.
Release v0.7.0-alpha.0
Alpha testing for...
Added new aggregation functionality:
- Added the
aggregate_levelfunction to aggregate by a grouping. - Added the
aggregate_up_hierarchyfunction to aggregate up hierarchy given by MultiIndex levels.
Release v0.6.2
Bug fix: fixed an issue with the round_and_adjust function.
Release v0.6.1
Bug fixes
- Fix affecting impute_base_prices function
- Fixed the API definition within init.py
Other
- Updated install instructions in README
Release v0.6.0
New features
- Added functionality for
impute_base_pricesfunction accepting theto_imputeargument aggregationfunction now works with mean or geometric mean depending onmethodargument- The function
calculate_indexintroduced offering various different index methods - The
index_calculatorpipeline function offers an end-to-end pipeline for calculating indices with optional base price imputation