[HughRunyan]

Environment

SWEET_python am-waste-flow, advanced_dst_city._with_other_landfills (added by #90).

Reproduction

A city-level ADST request with one listed landfill at share -0.01, plus other_landfills.

Expected vs actual

• Expected: the listed site takes nothing, and other landfills take 100% of what the city landfills.
• Actual: _validate_share_range tolerates shares down to −0.02, and the routing used them as given. The split became [-0.01, 1.01]. For 10,000 t of residual waste, that buried −100 t at the listed site and 10,100 t at other landfills.

Copilot flagged it on #95 (whose diff showed #90's code at the time). It was never fixed in #90.

Fix

Clip tolerated shares to [0, 1] before working out what other landfills take. A regression test fails on am-waste-flow and passes here.

Acceptance criteria

• A tolerated negative share sends no negative waste anywhere.
• pytest -m "not integration": 610 passed.

Definition of Done

Acceptance criteria met · tests pass · changelog updated · reviewed & merged.

🤖 Generated with Claude Code

[Copilot]

🟢 Approval recommended

The localized fix addresses the reported failure and includes regression coverage for both variants, with no unresolved findings.

No open issues 🎉

What changed in this PR

Fixes negative waste routing in advanced DST city requests with other_landfills.

Changes:

• Clips tolerated listed shares to [0, 1] before calculating remaining waste.
• Adds regression coverage for baseline and scenario routing.
• Documents the correction in the changelog.

File	Description
tests/​test_adst_city_other_landfills.py	Checks that tolerated negative shares produce no negative waste.
SWEET_python/​advanced_dst_city.py	Clips shares while preserving validation and normalization.
changelog/​2026-10.md	Documents the routing fix.

_{Review effort: Balanced}