tests · sds · Tangled

bayes for days

sds tests

44 files

did:web:projective-bichuwa.test.pds.did.bot · fix(harness): refuse a host-only bot path in one and play · 4d ago

feat: match harness with confidence intervals

test(training): a fixture where one machine of two did all the work

refactor(scoring)!: delete the intent scorer and stance

test_afford_probe.py

feat(candidates): ask MegaMek when a mover may spend more than it has

test_baseline.py

fix(training): one explore rate drives both phases, and displacement is not a fault

test_battle_armor.py

test(battle-armor): every design in the library through the observation

test_beliefs.py

feat(observability): render the threat map, rounds posterior and target damage

feat(candidates): the sweep generates every reachable stand, and the wire carries what it needs

test_c3_pricing.py

feat(c3): the generator prices the network it writes

test_compressed_logs.py

feat(harness)!: compress a match's decision logs when its container exits

test_conformance.py

fix(protocol): potential counts a Streak SRM's loaded round

feat(harness): pin each seat to a physical core of its own

perf(harness): shard the corpus save scan and default reads to a pool

test_correlate.py

feat(scenarios): correlate by the matchup, from the bot's side

test_counterfactual.py

feat(harness): split a counterfactual by how much the column varied

test_crossing.py

feat(scenarios)!: drop River Delta 1 from the map pool

test_decision_record.py

fix(harness): restart a live bot after a decision timeout

test_explain.py

feat(candidates): the sweep generates every reachable stand, and the wire carries what it needs

test_figures.py

docs(tactics): what the matrix said, and a figure for every tactic

test_firingaudit.py

feat(observability): the decision log, and the tools that read it

feat: match harness with confidence intervals

test_integration.py

test(hierarchy): prove the force tree end to end

test_labeldoc.py

docs(training): LABELS.md, every label over one worked match

test_labelrank.py

fix(training)!: read the engagement gap from one side

feat(observability): render the threat map, rounds posterior and target damage

test_los_corpus.py

test(los): 2383 MegaMek cases the Rust port must match

test_loscache.py

test(observability): a reading that went backwards is a cleared cache

feat(scenarios)!: drop the four boards that carry a bridge

fix(scenarios): refuse a run that plays an asymmetric suite one way

feat(melee): count the physicals a run resolved, from MegaMek's log

test_move_probe.py

feat(candidates): walk every step of a board, and stand where the mover stands

test_movestates.py

feat(hierarchy): incoming damage, optionality, scored deployment, and the firing superset

fix(harness): refuse a host-only bot path in one and play

test_runreport.py

feat(observability): one report per run, in one place

test_runstats.py

fix(harness): the by-round table averaged totals, not turns

test_scenario.py

feat(hierarchy): declare a force tree in generated scenarios

feat(observability): what each feature actually reported

fix(candidates)!: ask MegaMek whether a prone Mek may stand, and keep no rule

feat(scenario): battle value tiers, rolled crews, and the tools to correlate a result

test_step_cost.py

feat(candidates): ask MegaMek what a step costs each mover

test_tactic_coverage.py

test(tactics): a lance that reaches bombard, and no uncovered opening left

test_tactics.py

refactor(scoring)!: delete the intent scorer and stance

test(beliefs): pin the host's per-round BV records

feat(observability): render the threat map, rounds posterior and target damage