Feat/benchmarks

[FBumann]

Internal performance benchmarks

Adds a benchmarks/ directory for tracking linopy's own build time, LP write speed, matrix generation, and peak memory across problem sizes.

Tools

• pytest-benchmark for timing (save/compare across branches)
• pytest-memray for peak memory (tracks C/numpy allocations)

Models

Model	Description	Sizes
basic	Dense NN model, 2N^2 vars/cons	10 — 1600
knapsack	N binary variables, 1 constraint	100 — 1M
expression_arithmetic	Broadcasting, scaling, summation across dims	10 — 1000
sparse_network	Ring network with mismatched bus/line coords	10 — 1000
pypsa_scigrid	Real power system (requires pypsa)	10 — 200 snapshots

Timing phases

Phase	File	What it measures
Build	test_build.py	Model construction (add_variables, add_constraints, add_objective)
LP write	test_lp_write.py	Writing the model to an LP file
Matrices	test_matrices.py	Generating sparse matrices (A, b, c, bounds) from the model

Memory benchmarks

memory.py runs each test in a separate process with pytest-memray for accurate per-test peak memory. Results are saved as JSON for cross-branch comparison.

python benchmarks/memory.py save master
python benchmarks/memory.py save my-feature
python benchmarks/memory.py compare master my-feature

Quick start

pip install -e ".[benchmark]"

# Timing
pytest benchmarks/ --quick
pytest benchmarks/test_build.py --benchmark-save=master

# Memory
python benchmarks/memory.py save master --quick

See benchmarks/README.md for full details.

Checklist

• Code changes are sufficiently documented
• A note for the release notes doc/release_notes.rst is included
• I consent to the release of this PR's code under the MIT license

[FBumann]

Maybe we should merge sth like this before merging #591