- updated nzshm-common dependency
- dependencies for compatibility with openquake-engine v3.19
- 0.5% in 50 years PoE for disaggregations
- fix publication workflow
- update pandas dependency to ~2.0.3
- faster queries for THS_OpenquakeMeta table
- faster queries for THS_OpenquakeRealization table
- support for python 3.8
- faster queries for THS_HazardAggregation table
- query optimisation to gridded_hazard_query
- query optimisation to disagg_querys
- mypy 1.5.0
- pynamodb 5.5.0
- update mkdocs toolchain
- GHA scripts install with
--extra openquake
- ths_testing script for evaluation of performance changes
- python 3.11
- use poetry 1.4.2 for release build workflow
- update nzshm-common dependency 0.6.0
- mock cache when testing hazard aggregation query
- remove version control for ToshiOpenquakeMeta
- update openquake dependency 3.16
- update nzshm-common dependency 0.5.0
- fix SA(0.7) value
- script ths_cache to prepopulate and test caching
- local caching feature
- more spectral periods in constraint enum
- new constraints to existing THS models
- fix enum validations and apply to model fields
- remove v2 type options from batch save
- refactor model package
- refactor model.attributes package
- more test coverage
- two new models for DisaggAggregations
- validation via Enum for aggregation values
- new enumerations and constraints for probabilities, IMTS and VS30s
- fix for queries with mixed length vs30 index keys
- migrate more print statements to logging.debug
- new query get_one_gridded_hazard
- -m option to script to export meta tables only
- migrated print statements to logging.debug
- removed monkey patch for BASE183 - it iss in oqengine 3.15 now
- more test cover
- using nzshm-common==0.3.2 from pypi.
- specify poetry==1.2.0b3 in all the GHA yml files.
- THS_HazardAggregation table support for csv serialisation.
- refactoring/renaming openquake import modules.
- one openquake test no longer works as expected. It's off-piste so skipping it for now.
- data_functions migrated to THP
- branch_combinator migrated to THP
- V3 THS table models with improved indexing and and performance (esp. THS_HazardAggregation table)
- using latest CodedLocation API to manage gridded lcoations and resampling.
- realisation aggregration computations. These have moving to toshi-hazard-post
- multi_batch module for parallelised batch saves
- DESIGN.md capture notes on the experiments, test and mods to the package
- new switch on V2 queries to force normalised_location_id
- new '-f' switch on store_hazard script to force normalised_location_id
- lat, lon Float fields to support numeric range filtering in queries
- created timestamp field on stas, rlzs v2
- added pynamodb_attributes for FloatAttribute, TimestampAttribute types
- V2 store queries will automatically use nomralised location if custom sites aren't available.
- refactored model modules.
- new V2 models for stats and rlzs.
- new get_hazard script for manual testing.
- extra test coverage with optional openquake install as DEV dependency.
- meta dataframes are cut back to dstore defaults to minimise size.
- meta.aggs attribute
- meta.inv_tme attribute
- store hazard can create tables.
- store hazard adds extra meta.
- store hazard truncates values for rlz and agg fields.
- make stats & rlz queries tolerant to ID-only form (fails with REAL dynamodb & not in mocks).
- updated usage.
- store_hazard script for openquake systems.
- tightened up model attributes names.
- query api improvements
- added meta table
- new query methods for meta and rlzs
- moved vs30 from curves to meta
- updated docs
- fixed mkdoc rendering of python & markdown.
- fix poetry lockfile
- First release on PyPI.
- query and model modules providing basic support for openquake hazard stats curves only.