v0.9.13

• Add a step to report in RoutineBase.run
• Add more tests to cover DG and ES related routines
• Improve formulation for DG and ESD involved routines
• Improve module Report and method RoutineBase.export_csv
• Support TimedEvent in ANDES case conversion
• Add Var vBus in DCOPF for placeholder

v0.9.12

• Refactor OModel.initialized as a property method

• Add a demo to show using Constraint.e for debugging

• Fix ams.opt.omodel.Param.evaluate when its value is a number

• Improve ams.opt.omodel.ExpressionCalc for better performance

• Refactor module ams.opt

• Add class ams.opt.Expression

• Switch from PYPOWER to ANDES in routine PFlow

• Switch from PYPOWER to regular formulation in routine DCPF

• Refactor routines DCPF and DCOPF

• In RDocumenter, set Srouce to be owner if there is no src

• Specify multiprocess<=0.70.16 in requirements as 0.70.17 does not support Linux

• Reset setup.py to ensure compatibility

v0.9.11

• Add pyproject.toml for PEP 517 and PEP 518 compliance
• Add model Jumper
• Fix deprecation warning related to pandas.fillna and newshape in NumPy
• Minor refactor on solvers information in the module shared
• Change default values of minimum ON/OFF duration time of generators to be 1 and 0.5 hours
• Add parameter uf for enforced generator on/off status
• In servicee LoadScale, consider load online status
• Consider line online status in routine ED
• Add methods evaluate and finalize in the class OModel to handle optimization elements generation and assembling
• Refactor OModel.init() and Routine.init()
• Add ANDES paper as a citation file for now
• Add more routine tests for generator trip, line trip, and load trip
• Add a README to overview built-in cases
• Rename methods v2 as e for classes Constraint and Objective
• Add benchmark functions
• Improve using of eval() in module omodel
• Refactor module interop.andes as module interface for simplicity

v0.9.10

Hotfix of import issue in v0.9.9.

• In module MatProcessor, add two parameters permc_spec and use_umfpack in function build_ptdf
• Follow RTD's deprecation of Sphinx context injection at build time
• In MATPOWER conversion, set devices name as None
• Skip macOS tests in azure-pipelines due to failure in fixing its configuration
• Improve the logic of setting Optz value
• Support NumPy v2.0.0

v0.9.8

• Assign MParam.owner when declaring
• In MatProcessor, improve build_ptdf and build_lodf to allow partial building and incremental building
• Add in 'cases/matpower/Benchmark.json' for benchmark with MATPOWER
• Improve known_good test
• Minor fix in main.py selftest part
• Set dependency NumPy version to be <2.0.0 to avoid CVXPY compatibility issues