3.1.4 - 11/22/24

• Feature: Enable initializing a population of all newborns
• Feature: Enable setting data sources directly on DiseaseModel components
• Refactor: Update DiseaseModel to leverage the on_initialize_simulants method of Machine

3.1.3 - 12/03/24

• Feature: Create relative risk pipeline

3.1.2 - 11/13/24

• Modernize type hinting
• Refactor: Update DiseaseModel to rely on Machine's time-step and time-step-cleanup methods

3.1.1 - 11/12/24

• Feature: Rename DiseaseModel initial_state to residual_state
• Refactor: Use callables rather than strings to define lookup table builders

3.1.0 - 11/07/24

• Drop support for python 3.9

3.0.11 - 11/07/24

• Add ScaledPopulation component to population package

3.0.10 - 10/29/24

• Import subpackage classes to top-level namespace

3.0.9 - 09/12/24

• Fix broken import in scale_up module

3.0.8 - 09/12/24

• Fix broken tests in RiskAttributableDisease due to changes in how to set population_view

3.0.7 - 09/10/24

• Properly create the population_view in RateTransition

3.0.6 - 09/04/24

• Fix bug that was occurring when RiskEffect's rr_source was a float or DataFrame
• Better handle exposure column creation in Risk component
• Rename LBWSRisk 'exposure_column_name()' staticmethod to not collide with Risk attr

3.0.5 - 08/29/24

• Add missing results module-level docstrings
• Strengthen the SimpleCause docstring
• Rename the SimpleCause class method

3.0.4 - 08/28/24

• Strengthen results docstrings and clean up others throughout
• Add mention of NonLogLinearRiskEffect class to risk exposure tutorial

3.0.3 - 08/20/24

• Use external script to find matching dependency branches
• Fix bug when excluding all but one causes of disability from results
• Update FertilityAgeSpecificRates for new data format of age specific fertility rates

3.0.2 - 08/19/24

• Parameterize the script to find matching dependency branches

3.0.1 - 08/19/24

• Fix bug in build script when the upstream vivarium branch should be main

3.0.0 - 08/13/24

Breaking changes:

• Update components to use new Vivarium lookup table configuration
• Update Observers to use new Vivarium results interface classes and methods
• Change the "all_causes" disability weight pipeline from "disability_weight" to "all_causes.disability_weight"

Major changes:

• Refactor risk components to standardize/simplify risk exposure distributions
• Enable the setting of CSMR using a causes configuration parser
• Implement a PublicHealthObserver to help enforce the existence of "measure", "entity_type", "sub_entity", and "entity" columns in the results
• Refactor Risk component to create exposure column in state table
• Add support for non-loglinear relative risks

Other changes:

• Use external layered-config-tree package instead of Vivarium ConfigTree class
• Install matching upstream branches in github builds
• Stop writing seed and draw columns to the results
• Automatically run Jenkins builds on push or pull request
• Various other minor updates

2.3.3 - 06/17/24

• Hotfix pin numpy below 2.0

2.3.2 - 3/13/24

• Update to Mortality Observer to make subclassing easier.

2.3.1 - 3/11/24

• Update Mortality Observer to include tracked in population filter
• Fix bug in get_initialization_parameters to only remove existing keys if necessary

2.3.0 - 3/7/24

• Update population configuration keys to be more descriptive

2.2.3 - 3/6/24

• Update Mortality Observer to allow running with cause specific or total deaths and ylls.

2.2.2 - 2/28/24

• Fix bug in rescale_binned_proportions to update midpoitn for new age bins

2.2.1 - 2/26/24

• Update LinearScaleUp configuration defaults

2.2.0 - 02/14/24

• Implement CausesConfigurationParser to parse causes configuration into DiseaseModel components
• Bugfix assign sex-location-age demographic proportions by year when only one year in dataset

2.1.4 - 01/10/24

• Exclude undesirable arguments from the return of BaseDiseaseState name and __repr__ methods

2.1.3 - 01/09/24

• Update PyPI to 2FA with trusted publisher

2.1.2 - 12/21/23

• Fix tests failing due to Vivarium 2.3.0 release

2.1.1 - 10/13/23

• Perform actions in DiseaseState setup using class methods rather than hardcoding to allow for cleaner subclassing

2.1.0 - 10/05/23

• Remove explicit support for Python 3.8
• Minor bugfix to ensure default remission rate calls the right artifact key

2.0.1 - 09/27/23

• Address a CopyWithSettingWarning in results stratifier

2.0.0 - 09/22/23

• Refactor all components to inherit from vivarium.Component
• Refactor components inheriting from another vivarium class to be compatible with vivarium 2.0.0

1.0.5 - 09/19/23

• Update unit test for dtypes

1.0.4 - 09/15/23

• Address Pandas 2.1 FutureWarnings

1.0.3 - 08/10/23

• Pass BaseDiseaseState constructor kwargs to its super-class's constructor

1.0.2 - 08/10/23

• Minor bugfix to ensure dead simulants do not get observed transitions

1.0.1 - 08/07/23

• Minor bugfix to improve handling of excess mortality rate data

1.0.0 - 08/02/23

• Performance and architectural improvements to results manager, including observers
• Updates versioning to use setuptools_scm
• Other bugfixes

0.11.0 - 06/01/23

• Support Python 3.8-3.11
• Update vivarium pin
• Handle FutureWarning
• Refactor to create a 'get_transition_names' function

0.10.24 - 05/11/23

• Standardize builder, cause argument order in state get data functions
• Mends a bug where configured key_columns for randomness were not used in register_simulants

0.10.23 - 05/03/23

• Throw error when artifact doesn't contain relative risk data for desired target
• Rename for_initialization argument to match updated argument name in vivarium

0.10.22 - 12/27/22

• Update CI and setup for building python 3.7-3.10

0.10.21 - 11/16/22

• Fix bug in timing of disease transition observations
• Add logging when adding risks with a relative risk less than 1 from artifact

0.10.20 - 07/25/22

• Update Vivarium pin

0.10.19 - 06/29/22

• Create new LBWSG components
• Fix a bug when stratifying newly born simulants
• Fix pandas deprecation warnings
• Fix a bug when stratifying an empty population
• Allow configuration of sex subsetting of the population
• Add support for empty populations
• Fix a bug in counting deaths and ylls
• Refactor DiseaseState to be compatible with latest release of vivarium
• Add CODEOWNERS

0.10.18 - 04/22/22

• Improve ScaleUp component configuration
• Enable Mortality component to handle affected unmodeled causes
• Refactor RiskEffect calculation for clarity and extensibility
• Implement ResultsStratifier to stratify outputs
• Refactor all observers to be compatible with the ResultsStratifier component

0.10.17 - 02/15/22

• Autoformat code with black and isort.
• Add black and isort checks to CI.

0.10.16 - 02/13/22

• Update CI

0.10.15 - 01/25/22

• Implement LinearScaleUp component
• Refactor Risk, RiskEffect, and Mortality components for inheritance
• Added pull request template
• Fix bug in excess mortality pipeline name
• Fix bug in risk propensity pipeline

0.10.14 - 10/29/21

• Update license to BSD 3-clause
• Add .zenodo.json metadata replacing AUTHORS.rst

0.10.13 - 08/31/21

• implement categorical risk observer
• fix pandas warning in application of risk effect
• fix column name bug
• improve performance of reshaping
• require 2.0.6 and later of risk_distributions

0.10.12 - 08/10/21

• Fix bugs in DiseaseState
• Improve functionality of ensemble distributions
• Improve CI

0.10.11 - 05/18/21

• Fix bug in computing ages from an age distribution

0.10.10 - 05/10/21

• Improve standard DiseaseObserver
• Add 'transition rate' to the RateTransition object
• Add state and transition names to DiseaseModel and RiskAttributableDisease
• Get location from artifact rather than config file
• Fix bug that resulted in non-unique initializations of populations

0.10.9 - 01/25/21

• Improve performance of polytomous risk ppf calculations

0.10.8 - 1/5/21

• Fix deploy script

0.10.7 - 1/5/21

• Github actions replaces Travis for CI
• Unpin pandas and numpy

0.10.6 - 11/5/20

• Fix bug when risk effects are defined by a distribution

0.10.5 - 10/2/20

• Remove code from shigella vaccine
• Remove sample history observer
• Update randomness implementation to be consistent with latest version of vivarium
• Make prevalence sampling configurable
• Refactor to avoid warnings
• Clarify cut age bin math
• Pin to pandas 0.24.x
• Fix Travis validation issues

0.10.4 - 01/14/20

• Fix regression bug in RiskAttributableDisease
• Introduce low birth weight and short gestation risk and risk effect

0.10.3 - 12/13/19

• Fix regression bug in SIR_fixed_duration.

0.10.2 - 11/29/19

• Fix disease observer bug that prevented it from loading its configuration.

0.10.1 - 11/27/19

• Update MSLT components to new vivarium APIs.

0.10.0 - 11/18/19

• Update vivarium event system usage to no longer require explicit use of events.
• Move Artifact to vivarium.
• Clean up utility functions location and usage.
• Consistent preference of pathlib over os.path
• Small API updates for configuration.
• Restructure components to allow all subcomponents to be created during initialization.
• Remove healthcare access component.
• Restructure mortality calculation in a style more consistent with risk-disease pairs.
• Update to new API for simulation creation.
• Remove usages of 'omit_missing_columns' in favor of population subviews.
• Be consistent about rate naming conventions.
• Rename Disability component to DisabilityObserver.
• Rename 'age_group_start' and 'age_group_end' to 'age_start' and 'age_end' in data and lookup table usage.
• Have components specify all necessary dependencies for the resources (pipelines, state table columns, and randomness streams) that they manage.
• Update risk effect to make it easier to extend.
• Allow lookup table specification without naming bin columns in data.
• Update joint_value_postprocessor to union_postprocessor
• Clean up some of the MSLT calculations
• Dichotomous distribution bugfix

0.9.19 - 09/30/19

• Add python and vivarium to the intersphinx mapping.
• Bring in docs for non-standard risks.
• Bugfix in parameterized risk component.
• Update MSLT code to appropriate names/data artifact usage.

0.9.18 - 07/29/19

• Pin pandas version to be compatible with tables.
• Fix in RiskAttributableDisease disability calculation.

0.9.17 - 07/17/19

• Add names to mslt components.
• Clip non-ensemble distribution percentiles.

0.9.16 - 07/16/19

• Update observers to not report ages younger than those modeled.

0.9.15 - 07/03/19

• Fix docstring formatting.

0.9.14 - 07/03/19

• Update api documentation format.
• Bring in MSLT components.

0.9.13 - 06/18/19

• Move VivariumError to the correct place.
• Add names to all public health components.
• Add several missing ``__repr__``s.
• Modify the artifact to accept data that is wide on draws.
• Update components to new component manager api.
• Bugfix in SimulationDistribution

0.9.12 - 04/23/19

• Update docstring for categorical risk observer.
• Fix pipeline names in risk attributable disease.

0.9.11 - 04/22/19

• Add documentation for the data artifact.
• Bugfix in parameterized risk for covariates.
• Make disease observers work with paf of one risks.
• Make mortality and disability observers work with risk attributable diseases.
• Add simulation info to simulant creator.

0.9.10 - 03/29/19

• Bugfix in disease observer.

0.9.9 - 03/28/19

• Bugfix in data free risk components when using a covariate for coverage.
• Bugfix for simulations that start in a future year with extrapolate.

0.9.8 - 03/19/19

• Bugfix in mortality observer.

0.9.7 - 03/17/19

• Bugfixes in disease and treatment observers.
• Remove unnecessary output metrics.

0.9.6 - 03/13/19

• Generic observers for mortality, disability, person time, and treatment counts.
• Bugfix for large propensities when using risk distributions.
• Bugfix for rr distribution parameter name.

0.9.5 - 03/01/19

• Bugfix in validating rebinning risks for continuous risks.

0.9.4 - 03/01/19

• Added neonatal models and support for birth prevalence in DiseaseModel.
• Added a risk attributable disease model.
• Added support for rebinning polytomous risks into dichotomous risks.

0.9.3 - 02/26/19

• Bugfix in checking relative risk source type from configuration.

0.9.2 - 02/22/19

• Pin numpy and tables dependencies.
• Remove forecast flags
• Update crude birth rate fertility component
• Allow parameterization of RiskEffect components with normal and lognormal distributions.
• New observers for disease and treatment.

0.9.1 - 02/14/19

• Update dependencies

0.9.0 - 02/12/19

• Dataset manager logging.
• Added an SIR with duration model.
• Built observer for death counts and person years by age and year.
• Updated population and crude birth rate models for GBD 2017.
• Built an observer to point sample categorical risk exposure.
• Updated risk distribution and effect to work with the updated risk_distributions package.
• Updated healthcare access component.
• Added component for therapeutic inertia.
• Exposed individual cause disability weights as pipelines.
• Various bugfixes and api updates.

0.8.13 - 01/04/19

• Added support for multi-location data artifacts.
• Added CI branch synchronization

0.8.12 - 12/27/18

• Bugfix in categorical paf calculation

0.8.11 - 12/20/18

• Bugfix for mock_artifact testing data to include newly added columns.
• Bugfix to handle single-value sequela disability weight data.

0.8.10 - 12/20/18

• Added a replace function to the artifact class.
• Fixed a bug in age-specific fertility rate component.
• Added data free risk and risk effect components
• Removed the autogeneration of risk effects.
• Updated the risk and risk effect API.
• Added a configuration flag and component updates for limited forecasting data usage.
• Put in cause-level disability weights.
• Updated the population API.
• Added in standard epi disease models.
• Added support for morbidity only diseases.
• Expanded risk effects to target excess mortality.
• A host of model fixes and updates for the MSLT subpackage.

0.8.9 - 11/15/18

• Update documentation dependencies.

0.8.8 - 11/15/18

• Fix bug in population age generation.
• Assign initial event time for prevalent cases of disease with a dwell time.
• Set up artifact filter terms.
• Remove mean year and age columns.

0.8.7 - 11/07/18

• Switch to calculating pafs on the fly for non-continuous risks.
• Adding components for mslt.
• Pulled out distributions into separate package.

0.8.6 - 11/05/18

• Extend interactive api to package up data artifact manager in standard sims.
• Exposed disease prevalence propensity as a pipeline
• Added logic to rebin polytomous risks to dichotomous risks.
• Cleaned up confusing naming in metrics pipelines.
• Allow open cohorts to extrapolate birth rate data into the future.

0.8.5 - 10/23/18

• Update mass treatment campaign configuration for easier distributed runs.
• Fix leaking global state in mock artifact.
• Correctly implement order 0 interpolation.

0.8.4 - 10/09/18

• Fix bug that caused dead people to still experience disease transitions.
• Switch risk components to use pipelines for exposure/propensity
• Cleaned up return types from distribution.ppf
• Added indirect effects

0.8.3 - 09/27/18

• Remove caching from artifact writes as it causes bugs.

0.8.2 - 09/05/18

• Fix bug where the artifact manager assumed the data to be dataframe
• Fix bug where the hdf applied filters even where it is not valid.

0.8.1 - 08/22/18

• Fix various deployment things
• Add badges
• Remove unused metrics components
• Use __about__ in docs
• Extracted Artifact as an abstraction over hdf files.
• Cleaned up Artifact manager plugin
• Updated mock artifact

0.8.0 - 07/24/18

• Initial Release