Development plan

[tsmbland]

Putting together a rough plan to help organise the work that needs to be done. This will also help when it comes to planning releases, which has been a bit of a mess up until now.

This is very much a live document/work in progress.

Major changes

These are changes that would warrant a bump in the minor version number (e.g. 1.2.1 -> 1.3.0). Each of these milestones should be developed in their own dev branch, and merged into main once ready to create a release.

Simplify/improve the interface

• Deleting parameters that don't need to be there, simplifying the example models etc.
• Remove optional parameters from example models #499
• Output warnings to a file #498
• ScalingSize field in Technodata.csv isn't/shouldn't be required #494
• Fuel and EndUse fields in Technodata.csv aren't/shouldn't be required #493
• Get rid of the foresight parameter #476
• overwrite parameter doesn't behave as expected #430
• BaseYearImport.csv and BaseYearExport.csv are not used by any of the example models #370

Rewrite the input layer

• Create minimal example database with SQLAlchemy and/or DuckDB #331

Rewrite the scipy solver (/major changes)

• Production calculated by the solver is never used #423
• Memory error for transport model using scipy solver #389

Add support for early decommissioning

• Allow early decommissioning #502

Changes to the calculation of costs and prices

• ScalingSize field in Technodata.csv isn't/shouldn't be required #494
• Commodity prices in MUSE #466
• Change logic of how LCOE is used to calculate commodity prices #288

Change the way foresight is dealt with

• Current year, investment year and forecast year #477
• Foresight feature in MUSE requires revision #196
• Error in description of foresight in section 5.6 [BUG]  #195

Dropping outputs

• Removing MCAcapacity_supply.csv etc.
• Audit of model outputs #413

Drop support for Python 3.9

• Incompatible with Numpy 2.1.0 #459

Minor changes

These are changes that would warrant a bump in the patch version number (e.g. 1.2.1 -> 1.2.2). These can be merged directly into main as individual PRs, and a new release created if/when significant enough changes are made

Documentation improvements

• Confusion about the CommodityType parameter #492
• [QUESTION] How is decommission implemented? #487
• dispatch_production parameter is not explained in the documentation #474
• Formatting issues with wide tables in the documentation #473
• Add a tutorial with retrofit agents #393
• interpolation_mode parameter needs clarification (9.2) #265
• Explain links between investment and dispatch #218
• Documentation section title "9.2. Simulation settings" is not clear #208
• More advanced examples are needed  #207
• Documentation "9.3.7. Agents" has a number of poorly explained parameters #206
• The PJ/y unit is confusing people and needs explanation #202
• ScalingSize parameter in techno data needs clarification #201
• "Unit" description in "9.3.2. Commodity Description" has important information buried in it which should be separate #200
• 9.3.2. Commodity Description in the documentation needs to be clarified #199
• "Efficiency" parameter in techno data is misleading #197
• Add a section in the documentation on configuring and interpreting basic MUSE outputs #192
• Trade-related steps are missing in the flow diagrams section of the documentation #188

Bug fixes

• [BUG] Unexpected behaviour in multi-region models #475
• Model fails when only one timeslice [BUG]  #446
• Production dispatch method "match" not working #431
• Model fails when existing capacity for all technologies in a sector is set to zero [BUG]  #427
• Emission costs output doesn't work #421
• Fuel costs output gives zero for every value #419
• Odd results when existing capacity data doesn't match the time framework [BUG] #371

Infrastructure

• Explore the use of hypothesis in unit tests #465
• Remove develop branch #441
• Add documentation for making new release #436
• Remove use of bump2version #435
• Better unit tests for model outputs #422
• test_minimum_service_factor marked as xfail #400
• [BUG] Version number for MUSE in CITATION.cff is out of date #399
• Add script for regenerating data in tests/example_outputs #361

Add support for Python 3.13

• Add support for Python 3.13 #504

Refactoring changes

• Some of the other work would benefit from major refactoring changes to the code (e.g. creating a solvers module to clearly separate code for the scipy and adhoc solvers)
• A bit more thought required