perf: speed up LP file writing (2.5-3.9x on large models, no regressions on small)#564

Merged

FabianHofmann merged 19 commits intoPyPSA:masterfrom

FBumann:perf/lp-write-speed-combined

Feb 6, 2026

Commits on Jan 31, 2026

perf: use Polars streaming engine for LP file writing
FBumann
committed
fix: log warning with traceback when Polars streaming fallback triggers
FBumann
committed
perf: speed up LP constraint writing by replacing concat+sort with join
FBumann
committed
fix: missing space in lp file
FBumann
committed
perf: skip group_terms when unnecessary and avoid xarray broadcast for short DataFrame

FBumann
and
claude
committed
perf: skip group_terms in LinearExpression.to_polars when no duplicate vars

FBumann
and
claude
committed
perf: reduce per-constraint overhead in Constraint.to_polars()

FBumann
and
claude
committed
fix: handle empty constraint slices in sign_flat check

FBumann
and
claude
committed
docs: add LP write speed improvement to release notes

FBumann
and
claude
committed
bench: add LP write benchmark script with plotting
FBumann
committed
bench: larger model
FBumann
committed
perf: Add maybe_group_terms_polars() helper in common.py that checks for duplicate (labels, vars) pairs before calling group_terms_polars. Use it in both Constraint.to_polars() and LinearExpression…
FBumann
committed
Add variance to plot
FBumann
committed
test: add coverage for streaming fallback and maybe_group_terms_polars
FBumann
committed

Commits on Feb 2, 2026

fix: mypy
FBumann
committed
fix: mypy
FBumann
committed

Commits on Feb 6, 2026