@jinn-network/client 0.1.6 → 0.1.7-canary.0a586ca9
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +33 -0
- package/README.md +67 -1
- package/deployments/deployment-jinn-mvi-l1-sepolia-fast.json +23 -4
- package/deployments/deployment-jinn-mvi-l1-sepolia.json +23 -4
- package/deployments/deployment-jinn-mvi-l2-baseSepolia.json +5 -4
- package/dist/adapters/mech/adapter.d.ts +57 -2
- package/dist/adapters/mech/adapter.js +366 -63
- package/dist/adapters/mech/adapter.js.map +1 -1
- package/dist/adapters/mech/contracts.d.ts +17 -4
- package/dist/adapters/mech/contracts.js +19 -4
- package/dist/adapters/mech/contracts.js.map +1 -1
- package/dist/adapters/mech/safe-revert.d.ts +20 -0
- package/dist/adapters/mech/safe-revert.js +12 -4
- package/dist/adapters/mech/safe-revert.js.map +1 -1
- package/dist/adapters/mech/safe.d.ts +6 -2
- package/dist/adapters/mech/safe.js +32 -11
- package/dist/adapters/mech/safe.js.map +1 -1
- package/dist/adapters/mech/types.d.ts +6 -1
- package/dist/adapters/mech/types.js.map +1 -1
- package/dist/adapters/mech/verdict-code.d.ts +1 -0
- package/dist/adapters/mech/verdict-code.js +18 -0
- package/dist/adapters/mech/verdict-code.js.map +1 -1
- package/dist/agent/operator-claude.js +8 -0
- package/dist/agent/operator-claude.js.map +1 -1
- package/dist/api/activity-events-endpoint.d.ts +14 -0
- package/dist/api/activity-events-endpoint.js +59 -0
- package/dist/api/activity-events-endpoint.js.map +1 -0
- package/dist/api/admin-endpoint.d.ts +15 -3
- package/dist/api/admin-endpoint.js +24 -2
- package/dist/api/admin-endpoint.js.map +1 -1
- package/dist/api/bootstrap-endpoint.d.ts +1 -2
- package/dist/api/bootstrap-endpoint.js +85 -18
- package/dist/api/bootstrap-endpoint.js.map +1 -1
- package/dist/api/codex-doctor-endpoint.d.ts +90 -0
- package/dist/api/codex-doctor-endpoint.js +296 -0
- package/dist/api/codex-doctor-endpoint.js.map +1 -0
- package/dist/api/discovery-endpoint.d.ts +1 -0
- package/dist/api/discovery-endpoint.js +26 -0
- package/dist/api/discovery-endpoint.js.map +1 -1
- package/dist/api/fleet-build.d.ts +1 -0
- package/dist/api/fleet-build.js +2 -1
- package/dist/api/fleet-build.js.map +1 -1
- package/dist/api/gather-status.d.ts +37 -0
- package/dist/api/gather-status.js +572 -19
- package/dist/api/gather-status.js.map +1 -1
- package/dist/api/hermes-doctor-endpoint.d.ts +128 -3
- package/dist/api/hermes-doctor-endpoint.js +265 -22
- package/dist/api/hermes-doctor-endpoint.js.map +1 -1
- package/dist/api/launcher-status.d.ts +24 -17
- package/dist/api/launcher-status.js +13 -11
- package/dist/api/launcher-status.js.map +1 -1
- package/dist/api/launcher-tasks.d.ts +1 -1
- package/dist/api/launcher-tasks.js +12 -8
- package/dist/api/launcher-tasks.js.map +1 -1
- package/dist/api/operator-artifacts-endpoint.js +73 -6
- package/dist/api/operator-artifacts-endpoint.js.map +1 -1
- package/dist/api/portfolio-v0-build.d.ts +10 -0
- package/dist/api/portfolio-v0-build.js +24 -5
- package/dist/api/portfolio-v0-build.js.map +1 -1
- package/dist/api/prediction-v1-build.d.ts +10 -0
- package/dist/api/prediction-v1-build.js +7 -1
- package/dist/api/prediction-v1-build.js.map +1 -1
- package/dist/api/server.d.ts +31 -1
- package/dist/api/server.js +95 -2
- package/dist/api/server.js.map +1 -1
- package/dist/api/setup-endpoints.d.ts +7 -0
- package/dist/api/setup-endpoints.js +67 -135
- package/dist/api/setup-endpoints.js.map +1 -1
- package/dist/api/setup-retry-endpoint.d.ts +19 -0
- package/dist/api/setup-retry-endpoint.js +32 -0
- package/dist/api/setup-retry-endpoint.js.map +1 -0
- package/dist/api/solvernets-endpoints.d.ts +8 -0
- package/dist/api/solvernets-endpoints.js +100 -105
- package/dist/api/solvernets-endpoints.js.map +1 -1
- package/dist/api/status-build.d.ts +167 -2
- package/dist/api/status-build.js +118 -27
- package/dist/api/status-build.js.map +1 -1
- package/dist/api/status-harness-rollup.d.ts +35 -0
- package/dist/api/status-harness-rollup.js +45 -0
- package/dist/api/status-harness-rollup.js.map +1 -0
- package/dist/api/task-run-routing.d.ts +7 -0
- package/dist/api/task-run-routing.js +12 -0
- package/dist/api/task-run-routing.js.map +1 -0
- package/dist/api/task-runs-build.d.ts +21 -0
- package/dist/api/task-runs-build.js +14 -1
- package/dist/api/task-runs-build.js.map +1 -1
- package/dist/build-info.json +4 -4
- package/dist/build-meta.json +1 -1
- package/dist/captures/live-publisher.js +24 -4
- package/dist/captures/live-publisher.js.map +1 -1
- package/dist/captures/publish.d.ts +1 -1
- package/dist/chain-read-errors.d.ts +22 -0
- package/dist/chain-read-errors.js +41 -1
- package/dist/chain-read-errors.js.map +1 -1
- package/dist/cli/commands/auth.js +1 -1
- package/dist/cli/commands/auth.js.map +1 -1
- package/dist/cli/commands/codedigest-revert-check.d.ts +33 -0
- package/dist/cli/commands/codedigest-revert-check.js +249 -0
- package/dist/cli/commands/codedigest-revert-check.js.map +1 -0
- package/dist/cli/commands/create.js +3 -2
- package/dist/cli/commands/create.js.map +1 -1
- package/dist/cli/commands/doctor.d.ts +2 -0
- package/dist/cli/commands/doctor.js +2 -0
- package/dist/cli/commands/doctor.js.map +1 -1
- package/dist/cli/commands/rewards.js +11 -7
- package/dist/cli/commands/rewards.js.map +1 -1
- package/dist/cli/commands/solver-nets.d.ts +1 -0
- package/dist/cli/commands/solver-nets.js +179 -16
- package/dist/cli/commands/solver-nets.js.map +1 -1
- package/dist/cli/commands/solver-plugins-block.d.ts +33 -0
- package/dist/cli/commands/solver-plugins-block.js +118 -0
- package/dist/cli/commands/solver-plugins-block.js.map +1 -0
- package/dist/cli/commands/solver-plugins-feedback.d.ts +72 -0
- package/dist/cli/commands/solver-plugins-feedback.js +262 -0
- package/dist/cli/commands/solver-plugins-feedback.js.map +1 -0
- package/dist/cli/commands/solver-plugins-read.d.ts +54 -0
- package/dist/cli/commands/solver-plugins-read.js +259 -0
- package/dist/cli/commands/solver-plugins-read.js.map +1 -0
- package/dist/cli/commands/solver-plugins.d.ts +35 -0
- package/dist/cli/commands/solver-plugins.js +399 -2
- package/dist/cli/commands/solver-plugins.js.map +1 -1
- package/dist/cli/commands/status.js +1 -1
- package/dist/cli/commands/status.js.map +1 -1
- package/dist/cli/commands/tasks.js +101 -11
- package/dist/cli/commands/tasks.js.map +1 -1
- package/dist/cli/commands/update.d.ts +10 -0
- package/dist/cli/commands/update.js +36 -0
- package/dist/cli/commands/update.js.map +1 -1
- package/dist/cli/index.js +2 -0
- package/dist/cli/index.js.map +1 -1
- package/dist/cli/introspection-context.js +5 -0
- package/dist/cli/introspection-context.js.map +1 -1
- package/dist/cli/task-native-readiness.d.ts +10 -1
- package/dist/cli/task-native-readiness.js +30 -6
- package/dist/cli/task-native-readiness.js.map +1 -1
- package/dist/config.d.ts +287 -235
- package/dist/config.js +318 -114
- package/dist/config.js.map +1 -1
- package/dist/daemon/checkpoint-loop.d.ts +48 -0
- package/dist/daemon/checkpoint-loop.js +76 -0
- package/dist/daemon/checkpoint-loop.js.map +1 -0
- package/dist/daemon/creator.d.ts +1 -1
- package/dist/daemon/creator.js +20 -3
- package/dist/daemon/creator.js.map +1 -1
- package/dist/daemon/daemon.d.ts +22 -0
- package/dist/daemon/daemon.js +174 -31
- package/dist/daemon/daemon.js.map +1 -1
- package/dist/daemon/eviction-loop.d.ts +40 -0
- package/dist/daemon/eviction-loop.js +67 -0
- package/dist/daemon/eviction-loop.js.map +1 -0
- package/dist/daemon/gate-logger.d.ts +9 -0
- package/dist/daemon/gate-logger.js +2 -0
- package/dist/daemon/gate-logger.js.map +1 -0
- package/dist/daemon/jinn-claim-loop-wiring.d.ts +33 -0
- package/dist/daemon/jinn-claim-loop-wiring.js +40 -0
- package/dist/daemon/jinn-claim-loop-wiring.js.map +1 -0
- package/dist/daemon/jinn-claim-loop.d.ts +24 -17
- package/dist/daemon/jinn-claim-loop.js +77 -23
- package/dist/daemon/jinn-claim-loop.js.map +1 -1
- package/dist/daemon/readiness-gate.d.ts +1 -4
- package/dist/daemon/readiness-gate.js.map +1 -1
- package/dist/daemon/skip-log-dedup.d.ts +69 -0
- package/dist/daemon/skip-log-dedup.js +106 -0
- package/dist/daemon/skip-log-dedup.js.map +1 -0
- package/dist/daemon/spend-cap-gate.d.ts +40 -0
- package/dist/daemon/spend-cap-gate.js +46 -0
- package/dist/daemon/spend-cap-gate.js.map +1 -0
- package/dist/dashboard/assets/index-BNs_ewI6.js +345 -0
- package/dist/dashboard/assets/index-C4huIsUW.css +32 -0
- package/dist/dashboard/index.html +2 -2
- package/dist/discovery/factory.d.ts +17 -5
- package/dist/discovery/factory.js +46 -18
- package/dist/discovery/factory.js.map +1 -1
- package/dist/discovery/http.js +469 -3
- package/dist/discovery/http.js.map +1 -1
- package/dist/discovery/onchain.d.ts +5 -0
- package/dist/discovery/onchain.js +448 -18
- package/dist/discovery/onchain.js.map +1 -1
- package/dist/discovery/types.d.ts +174 -1
- package/dist/discovery/types.js +8 -10
- package/dist/discovery/types.js.map +1 -1
- package/dist/discovery/with-fallback.d.ts +7 -0
- package/dist/discovery/with-fallback.js +37 -0
- package/dist/discovery/with-fallback.js.map +1 -1
- package/dist/earning/bootstrap.d.ts +100 -4
- package/dist/earning/bootstrap.js +239 -76
- package/dist/earning/bootstrap.js.map +1 -1
- package/dist/earning/contracts.d.ts +14 -0
- package/dist/earning/contracts.js +17 -5
- package/dist/earning/contracts.js.map +1 -1
- package/dist/earning/funding-plan.js +27 -18
- package/dist/earning/funding-plan.js.map +1 -1
- package/dist/earning/jinn-rewards.d.ts +46 -0
- package/dist/earning/jinn-rewards.js +32 -0
- package/dist/earning/jinn-rewards.js.map +1 -1
- package/dist/earning/safe-adapter.d.ts +2 -0
- package/dist/earning/safe-adapter.js +37 -11
- package/dist/earning/safe-adapter.js.map +1 -1
- package/dist/earning/store.d.ts +8 -0
- package/dist/earning/store.js.map +1 -1
- package/dist/earning/testnet-setup-migration.d.ts +12 -0
- package/dist/earning/testnet-setup-migration.js +27 -1
- package/dist/earning/testnet-setup-migration.js.map +1 -1
- package/dist/earning/types.d.ts +21 -6
- package/dist/earning/viem-clients.d.ts +11 -4
- package/dist/earning/viem-clients.js +14 -5
- package/dist/earning/viem-clients.js.map +1 -1
- package/dist/erc8004/identity.d.ts +19 -3
- package/dist/erc8004/identity.js +21 -6
- package/dist/erc8004/identity.js.map +1 -1
- package/dist/erc8004/index.d.ts +1 -1
- package/dist/erc8004/index.js.map +1 -1
- package/dist/erc8004/reputation.d.ts +8 -0
- package/dist/erc8004/reputation.js +22 -3
- package/dist/erc8004/reputation.js.map +1 -1
- package/dist/events/types.d.ts +2 -2
- package/dist/harnesses/cost-estimates.d.ts +124 -0
- package/dist/harnesses/cost-estimates.js +265 -0
- package/dist/harnesses/cost-estimates.js.map +1 -0
- package/dist/harnesses/engine/engine.d.ts +78 -0
- package/dist/harnesses/engine/engine.js +153 -11
- package/dist/harnesses/engine/engine.js.map +1 -1
- package/dist/harnesses/engine/persistence.d.ts +51 -1
- package/dist/harnesses/engine/persistence.js +118 -5
- package/dist/harnesses/engine/persistence.js.map +1 -1
- package/dist/harnesses/engine/work-dir-reaper.d.ts +65 -0
- package/dist/harnesses/engine/work-dir-reaper.js +100 -0
- package/dist/harnesses/engine/work-dir-reaper.js.map +1 -0
- package/dist/harnesses/impls/hermes-agent/adapter.js +40 -0
- package/dist/harnesses/impls/hermes-agent/adapter.js.map +1 -1
- package/dist/harnesses/impls/hermes-agent/bootstrap.d.ts +20 -0
- package/dist/harnesses/impls/hermes-agent/bootstrap.js +44 -8
- package/dist/harnesses/impls/hermes-agent/bootstrap.js.map +1 -1
- package/dist/harnesses/impls/hermes-agent/config-builder.d.ts +1 -1
- package/dist/harnesses/impls/hermes-agent/config-builder.js +4 -2
- package/dist/harnesses/impls/hermes-agent/config-builder.js.map +1 -1
- package/dist/harnesses/impls/hermes-agent/harness.d.ts +59 -1
- package/dist/harnesses/impls/hermes-agent/harness.js +104 -0
- package/dist/harnesses/impls/hermes-agent/harness.js.map +1 -1
- package/dist/harnesses/impls/hermes-agent/prompt.d.ts +6 -6
- package/dist/harnesses/impls/hermes-agent/prompt.js +6 -6
- package/dist/harnesses/impls/index.d.ts +7 -0
- package/dist/harnesses/impls/index.js +16 -1
- package/dist/harnesses/impls/index.js.map +1 -1
- package/dist/harnesses/impls/learner/adapters/codex-code.d.ts +9 -0
- package/dist/harnesses/impls/learner/adapters/codex-code.js +30 -8
- package/dist/harnesses/impls/learner/adapters/codex-code.js.map +1 -1
- package/dist/harnesses/impls/learner/harness.d.ts +47 -4
- package/dist/harnesses/impls/learner/harness.js +105 -2
- package/dist/harnesses/impls/learner/harness.js.map +1 -1
- package/dist/harnesses/impls/learner/harvest.d.ts +1 -1
- package/dist/harnesses/impls/learner/harvest.js +23 -5
- package/dist/harnesses/impls/learner/harvest.js.map +1 -1
- package/dist/harnesses/impls/learner/plugin-path.d.ts +0 -13
- package/dist/harnesses/impls/learner/plugin-path.js +35 -15
- package/dist/harnesses/impls/learner/plugin-path.js.map +1 -1
- package/dist/harnesses/impls/learner/restoration-patch.d.ts +2 -2
- package/dist/harnesses/impls/learner/restoration-patch.js +25 -6
- package/dist/harnesses/impls/learner/restoration-patch.js.map +1 -1
- package/dist/harnesses/impls/learner/types.d.ts +11 -0
- package/dist/harnesses/impls/stub.d.ts +58 -0
- package/dist/harnesses/impls/stub.js +89 -0
- package/dist/harnesses/impls/stub.js.map +1 -0
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/eval-runner.d.ts +69 -50
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/eval-runner.js +199 -94
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/eval-runner.js.map +1 -1
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/harness.d.ts +12 -1
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/harness.js +121 -7
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/harness.js.map +1 -1
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/hf-fetcher.d.ts +88 -4
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/hf-fetcher.js +143 -22
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/hf-fetcher.js.map +1 -1
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/index.d.ts +6 -0
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/index.js +1 -1
- package/dist/harnesses/impls/swe-rebench-v2-evaluator/index.js.map +1 -1
- package/dist/harnesses/readiness-registry.d.ts +7 -0
- package/dist/harnesses/readiness-registry.js +18 -1
- package/dist/harnesses/readiness-registry.js.map +1 -1
- package/dist/learner/revert-decision.d.ts +59 -0
- package/dist/learner/revert-decision.js +53 -0
- package/dist/learner/revert-decision.js.map +1 -0
- package/dist/learner/revert-stats.d.ts +24 -0
- package/dist/learner/revert-stats.js +44 -0
- package/dist/learner/revert-stats.js.map +1 -0
- package/dist/main.js +470 -142
- package/dist/main.js.map +1 -1
- package/dist/mcp/get-codedigest-reward.d.ts +13 -0
- package/dist/mcp/get-codedigest-reward.js +23 -0
- package/dist/mcp/get-codedigest-reward.js.map +1 -0
- package/dist/mcp/server.js +23 -0
- package/dist/mcp/server.js.map +1 -1
- package/dist/observability/emit-event.d.ts +3 -2
- package/dist/observability/emit-event.js +22 -1
- package/dist/observability/emit-event.js.map +1 -1
- package/dist/operator-errors.d.ts +7 -0
- package/dist/operator-errors.js +13 -1
- package/dist/operator-errors.js.map +1 -1
- package/dist/plugins/learner/.claude-plugin/plugin.json +9 -0
- package/dist/plugins/learner/.codex-plugin/plugin.json +39 -0
- package/dist/plugins/learner/AGENTS.md +40 -0
- package/dist/plugins/learner/CLAUDE.md +33 -0
- package/dist/plugins/learner/README.md +59 -0
- package/dist/plugins/learner/hooks/hooks.json +16 -0
- package/dist/plugins/learner/hooks/session-start +38 -0
- package/dist/plugins/learner/skills/learn/SKILL.md +412 -0
- package/dist/plugins/learner/skills/learn/analyst-prompt.md +68 -0
- package/dist/plugins/learner/skills/learn/consolidator-prompt.md +111 -0
- package/dist/plugins/learner/skills/learn/explorer-prompt.md +53 -0
- package/dist/plugins/learner/skills/learn/planner-prompt.md +87 -0
- package/dist/plugins/learner/skills/learn/promoter-prompt.md +184 -0
- package/dist/plugins/learner/skills/learn/step-worker-prompt.md +47 -0
- package/dist/plugins/learner/skills/learn/strategist-prompt.md +85 -0
- package/dist/preflight/pidfile-liveness.d.ts +44 -0
- package/dist/preflight/pidfile-liveness.js +103 -0
- package/dist/preflight/pidfile-liveness.js.map +1 -0
- package/dist/preflight/rpc-network.d.ts +40 -0
- package/dist/preflight/rpc-network.js +67 -1
- package/dist/preflight/rpc-network.js.map +1 -1
- package/dist/restart-daemon.d.ts +90 -0
- package/dist/restart-daemon.js +95 -0
- package/dist/restart-daemon.js.map +1 -0
- package/dist/rpc/transport.d.ts +109 -0
- package/dist/rpc/transport.js +220 -0
- package/dist/rpc/transport.js.map +1 -0
- package/dist/scripts/donation-consumption-acceptance.js +7 -28
- package/dist/scripts/donation-consumption-acceptance.js.map +1 -1
- package/dist/scripts/swe-rebench-v2-pytest-missing.json +16 -0
- package/dist/setup/halt-mode.d.ts +14 -0
- package/dist/setup/halt-mode.js +17 -0
- package/dist/setup/halt-mode.js.map +1 -0
- package/dist/solver-nets/prediction-operator-ux.d.ts +1 -2
- package/dist/solver-nets/prediction-operator-ux.js +90 -47
- package/dist/solver-nets/prediction-operator-ux.js.map +1 -1
- package/dist/solver-nets/registry.d.ts +20 -1
- package/dist/solver-nets/registry.js +38 -25
- package/dist/solver-nets/registry.js.map +1 -1
- package/dist/solver-types/_swe-rebench-v2-pool-cache.d.ts +58 -0
- package/dist/solver-types/_swe-rebench-v2-pool-cache.js +87 -0
- package/dist/solver-types/_swe-rebench-v2-pool-cache.js.map +1 -0
- package/dist/solver-types/_swe-rebench-v2-pool.d.ts +9 -2
- package/dist/solver-types/_swe-rebench-v2-pool.js +15 -20
- package/dist/solver-types/_swe-rebench-v2-pool.js.map +1 -1
- package/dist/solver-types/_swe-rebench-v2-state.d.ts +15 -0
- package/dist/solver-types/_swe-rebench-v2-state.js +19 -0
- package/dist/solver-types/_swe-rebench-v2-state.js.map +1 -1
- package/dist/solver-types/_swe-rebench-v2-substrate.d.ts +1 -0
- package/dist/solver-types/_swe-rebench-v2-substrate.js +10 -0
- package/dist/solver-types/_swe-rebench-v2-substrate.js.map +1 -1
- package/dist/solver-types/_swe-rebench-v2-validated-pool.d.ts +145 -2
- package/dist/solver-types/_swe-rebench-v2-validated-pool.js +482 -44
- package/dist/solver-types/_swe-rebench-v2-validated-pool.js.map +1 -1
- package/dist/solver-types/swe-rebench-v2-auto.d.ts +38 -14
- package/dist/solver-types/swe-rebench-v2-auto.js +87 -28
- package/dist/solver-types/swe-rebench-v2-auto.js.map +1 -1
- package/dist/solver-types/swe-rebench-v2.d.ts +19 -2
- package/dist/solver-types/swe-rebench-v2.js +351 -96
- package/dist/solver-types/swe-rebench-v2.js.map +1 -1
- package/dist/solvernets/daemon-init.d.ts +10 -2
- package/dist/solvernets/daemon-init.js +22 -2
- package/dist/solvernets/daemon-init.js.map +1 -1
- package/dist/solvernets/launched-record-dispatcher.d.ts +4 -0
- package/dist/solvernets/launched-record-dispatcher.js +41 -7
- package/dist/solvernets/launched-record-dispatcher.js.map +1 -1
- package/dist/solvernets/registry-client-erc8004.js +11 -0
- package/dist/solvernets/registry-client-erc8004.js.map +1 -1
- package/dist/solvernets/store.d.ts +7 -2
- package/dist/solvernets/store.js +1 -0
- package/dist/solvernets/store.js.map +1 -1
- package/dist/spend/cost-surface-status.d.ts +10 -0
- package/dist/spend/cost-surface-status.js +22 -0
- package/dist/spend/cost-surface-status.js.map +1 -0
- package/dist/spend/credential.d.ts +8 -0
- package/dist/spend/credential.js +30 -0
- package/dist/spend/credential.js.map +1 -0
- package/dist/spend/daemon-config.d.ts +13 -0
- package/dist/spend/daemon-config.js +24 -0
- package/dist/spend/daemon-config.js.map +1 -0
- package/dist/spend/pricing.d.ts +16 -0
- package/dist/spend/pricing.js +26 -0
- package/dist/spend/pricing.js.map +1 -0
- package/dist/spend/record.d.ts +13 -0
- package/dist/spend/record.js +30 -0
- package/dist/spend/record.js.map +1 -0
- package/dist/spend/usage.d.ts +27 -0
- package/dist/spend/usage.js +113 -0
- package/dist/spend/usage.js.map +1 -0
- package/dist/store/store.d.ts +61 -0
- package/dist/store/store.js +302 -7
- package/dist/store/store.js.map +1 -1
- package/dist/tasks/sources.d.ts +18 -1
- package/dist/tasks/sources.js +33 -5
- package/dist/tasks/sources.js.map +1 -1
- package/dist/trajectory/transcript-parsers/codex-session.d.ts +12 -6
- package/dist/trajectory/transcript-parsers/codex-session.js +114 -13
- package/dist/trajectory/transcript-parsers/codex-session.js.map +1 -1
- package/dist/trajectory/transcript-parsers/types.d.ts +8 -8
- package/dist/trajectory/transcript-session-dirs.d.ts +18 -0
- package/dist/trajectory/transcript-session-dirs.js +85 -0
- package/dist/trajectory/transcript-session-dirs.js.map +1 -0
- package/dist/trajectory/transcript-watcher.d.ts +20 -1
- package/dist/trajectory/transcript-watcher.js +98 -32
- package/dist/trajectory/transcript-watcher.js.map +1 -1
- package/dist/tx-retry.d.ts +166 -19
- package/dist/tx-retry.js +310 -32
- package/dist/tx-retry.js.map +1 -1
- package/dist/types/payloads/portfolio-v0.d.ts +3 -3
- package/dist/types/payloads/prediction-apy-v0.d.ts +8 -8
- package/dist/types/payloads/prediction-v0.d.ts +17 -17
- package/dist/types/task-document.d.ts +392 -0
- package/dist/types/task-document.js +10 -0
- package/dist/types/task-document.js.map +1 -1
- package/dist/types/task.d.ts +28 -0
- package/dist/util/extract-tx-hash.d.ts +14 -0
- package/dist/util/extract-tx-hash.js +19 -0
- package/dist/util/extract-tx-hash.js.map +1 -0
- package/dist/vendor/@jinn-network/sdk/dist/contracts.js +1 -1
- package/dist/vendor/@jinn-network/sdk/dist/solvernets/manifest-schema.d.ts +3 -0
- package/dist/vendor/@jinn-network/sdk/dist/solvernets/manifest-schema.js +1 -0
- package/package.json +37 -13
- package/plugins/learner/skills/learn/consolidator-prompt.md +18 -1
- package/plugins/learner/skills/learn/promoter-prompt.md +72 -1
- package/plugins/swe-rebench-v2-diffmin/README.md +10 -9
- package/plugins/swe-rebench-v2-diffmin/jinn.plugin.json +1 -1
- package/plugins/swe-rebench-v2-diffmin/skills/diffmin/SKILL.md +15 -10
- package/plugins/swe-rebench-v2-diffmin/skills/test-map/SKILL.md +10 -12
- package/plugins/swe-rebench-v2-runtime/.claude-plugin/plugin.json +1 -1
- package/plugins/swe-rebench-v2-runtime/.codex-plugin/plugin.json +3 -3
- package/plugins/swe-rebench-v2-runtime/README.md +6 -6
- package/plugins/swe-rebench-v2-runtime/jinn.plugin.json +2 -3
- package/plugins/swe-rebench-v2-runtime/skills/task/SKILL.md +69 -0
- package/dist/dashboard/assets/index-DOlzFN8a.css +0 -32
- package/dist/dashboard/assets/index-NkZ7CTAT.js +0 -140
- package/plugins/swe-rebench-v2-runtime/skills/orient/SKILL.md +0 -29
- package/plugins/swe-rebench-v2-runtime/skills/plan/SKILL.md +0 -53
|
@@ -42,6 +42,13 @@ export declare class HarnessReadinessRegistry {
|
|
|
42
42
|
start(): void;
|
|
43
43
|
stop(): void;
|
|
44
44
|
getSnapshot(): HarnessReadinessSnapshot;
|
|
45
|
+
/**
|
|
46
|
+
* Frozen-at-construction `joinedHarnessesByCid` map. Exposed so callers
|
|
47
|
+
* that already have a registry reference (e.g. the `/v1/status` handler
|
|
48
|
+
* building a HarnessRollup) can pair it with `getSnapshot()` without
|
|
49
|
+
* reaching past the registry's encapsulation boundary.
|
|
50
|
+
*/
|
|
51
|
+
getJoinedHarnessesByCid(): Record<string, JoinedHarnessSpec>;
|
|
45
52
|
isReadyForClaim(manifestCid: string): ReadyStatus;
|
|
46
53
|
refreshNow(): Promise<void>;
|
|
47
54
|
private _doRefresh;
|
|
@@ -40,6 +40,15 @@ export class HarnessReadinessRegistry {
|
|
|
40
40
|
getSnapshot() {
|
|
41
41
|
return this.snapshot;
|
|
42
42
|
}
|
|
43
|
+
/**
|
|
44
|
+
* Frozen-at-construction `joinedHarnessesByCid` map. Exposed so callers
|
|
45
|
+
* that already have a registry reference (e.g. the `/v1/status` handler
|
|
46
|
+
* building a HarnessRollup) can pair it with `getSnapshot()` without
|
|
47
|
+
* reaching past the registry's encapsulation boundary.
|
|
48
|
+
*/
|
|
49
|
+
getJoinedHarnessesByCid() {
|
|
50
|
+
return this.opts.joinedHarnessesByCid;
|
|
51
|
+
}
|
|
43
52
|
isReadyForClaim(manifestCid) {
|
|
44
53
|
const joined = this.opts.joinedHarnessesByCid[manifestCid];
|
|
45
54
|
if (!joined) {
|
|
@@ -70,8 +79,16 @@ export class HarnessReadinessRegistry {
|
|
|
70
79
|
}
|
|
71
80
|
}
|
|
72
81
|
async _doRefresh() {
|
|
73
|
-
// Group joined entries by harnessName so we only call isReady() once per
|
|
82
|
+
// Group joined entries by harnessName so we only call isReady() once per
|
|
83
|
+
// harness. Seed the map with EVERY registered harness — not just joined
|
|
84
|
+
// ones — so the snapshot covers harnesses the operator could pick but
|
|
85
|
+
// has not joined yet. The SPA join form (#332) consults this snapshot to
|
|
86
|
+
// disable not-ready harness options before any join exists, so an
|
|
87
|
+
// unjoined harness must still appear (with an empty `manifestCids`).
|
|
74
88
|
const harnessToCids = new Map();
|
|
89
|
+
for (const name of Object.keys(this.opts.harnessesByName)) {
|
|
90
|
+
harnessToCids.set(name, []);
|
|
91
|
+
}
|
|
75
92
|
for (const [cid, joined] of Object.entries(this.opts.joinedHarnessesByCid)) {
|
|
76
93
|
const list = harnessToCids.get(joined.harnessName) ?? [];
|
|
77
94
|
list.push(cid);
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"readiness-registry.js","sourceRoot":"","sources":["../../src/harnesses/readiness-registry.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AA+BH,MAAM,wBAAwB,GAAG,KAAK,CAAC;AACvC,MAAM,2BAA2B,GAAG,KAAK,CAAC;AAE1C,MAAM,OAAO,wBAAwB;IAClB,IAAI,CAA4C;IACzD,QAAQ,GAA6B;QAC3C,eAAe,EAAE,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE;QAC1C,SAAS,EAAE,EAAE;KACd,CAAC;IACM,KAAK,GAA0C,IAAI,CAAC;IACpD,eAAe,GAAG,KAAK,CAAC;IAEhC,YAAY,IAAqC;QAC/C,IAAI,CAAC,IAAI,GAAG;YACV,cAAc,EAAE,IAAI,CAAC,cAAc,IAAI,wBAAwB;YAC/D,gBAAgB,EAAE,IAAI,CAAC,gBAAgB,IAAI,2BAA2B;YACtE,eAAe,EAAE,IAAI,CAAC,eAAe;YACrC,oBAAoB,EAAE,IAAI,CAAC,oBAAoB;SAChD,CAAC;IACJ,CAAC;IAED,KAAK;QACH,IAAI,IAAI,CAAC,KAAK;YAAE,OAAO;QACvB,IAAI,CAAC,KAAK,GAAG,WAAW,CAAC,GAAG,EAAE,GAAG,KAAK,IAAI,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC;IACxF,CAAC;IAED,IAAI;QACF,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YACf,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAC1B,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC;QACpB,CAAC;IACH,CAAC;IAED,WAAW;QACT,OAAO,IAAI,CAAC,QAAQ,CAAC;IACvB,CAAC;IAED,eAAe,CAAC,WAAmB;QACjC,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,oBAAoB,CAAC,WAAW,CAAC,CAAC;QAC3D,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,OAAO;gBACL,KAAK,EAAE,KAAK;gBACZ,MAAM,EAAE,eAAe,WAAW,0BAA0B;aAC7D,CAAC;QACJ,CAAC;QACD,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,KAAK,MAAM,CAAC,WAAW,CAAC,CAAC;QACxF,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,sCAAsC,EAAE,CAAC;QAC1E,CAAC;QACD,OAAO;YACL,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,MAAM,EAAE,KAAK,CAAC,MAAM;YACpB,QAAQ,EAAE,KAAK,CAAC,QAAQ;SACzB,CAAC;IACJ,CAAC;IAED,KAAK,CAAC,UAAU;QACd,IAAI,IAAI,CAAC,eAAe;YAAE,OAAO;QACjC,IAAI,CAAC,eAAe,GAAG,IAAI,CAAC;QAC5B,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;QAC1B,CAAC;gBAAS,CAAC;YACT,IAAI,CAAC,eAAe,GAAG,KAAK,CAAC;QAC/B,CAAC;IACH,CAAC;IAEO,KAAK,CAAC,UAAU;QACtB,
|
|
1
|
+
{"version":3,"file":"readiness-registry.js","sourceRoot":"","sources":["../../src/harnesses/readiness-registry.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AA+BH,MAAM,wBAAwB,GAAG,KAAK,CAAC;AACvC,MAAM,2BAA2B,GAAG,KAAK,CAAC;AAE1C,MAAM,OAAO,wBAAwB;IAClB,IAAI,CAA4C;IACzD,QAAQ,GAA6B;QAC3C,eAAe,EAAE,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE;QAC1C,SAAS,EAAE,EAAE;KACd,CAAC;IACM,KAAK,GAA0C,IAAI,CAAC;IACpD,eAAe,GAAG,KAAK,CAAC;IAEhC,YAAY,IAAqC;QAC/C,IAAI,CAAC,IAAI,GAAG;YACV,cAAc,EAAE,IAAI,CAAC,cAAc,IAAI,wBAAwB;YAC/D,gBAAgB,EAAE,IAAI,CAAC,gBAAgB,IAAI,2BAA2B;YACtE,eAAe,EAAE,IAAI,CAAC,eAAe;YACrC,oBAAoB,EAAE,IAAI,CAAC,oBAAoB;SAChD,CAAC;IACJ,CAAC;IAED,KAAK;QACH,IAAI,IAAI,CAAC,KAAK;YAAE,OAAO;QACvB,IAAI,CAAC,KAAK,GAAG,WAAW,CAAC,GAAG,EAAE,GAAG,KAAK,IAAI,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC;IACxF,CAAC;IAED,IAAI;QACF,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;YACf,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAC1B,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC;QACpB,CAAC;IACH,CAAC;IAED,WAAW;QACT,OAAO,IAAI,CAAC,QAAQ,CAAC;IACvB,CAAC;IAED;;;;;OAKG;IACH,uBAAuB;QACrB,OAAO,IAAI,CAAC,IAAI,CAAC,oBAAoB,CAAC;IACxC,CAAC;IAED,eAAe,CAAC,WAAmB;QACjC,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,oBAAoB,CAAC,WAAW,CAAC,CAAC;QAC3D,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,OAAO;gBACL,KAAK,EAAE,KAAK;gBACZ,MAAM,EAAE,eAAe,WAAW,0BAA0B;aAC7D,CAAC;QACJ,CAAC;QACD,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,KAAK,MAAM,CAAC,WAAW,CAAC,CAAC;QACxF,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,sCAAsC,EAAE,CAAC;QAC1E,CAAC;QACD,OAAO;YACL,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,MAAM,EAAE,KAAK,CAAC,MAAM;YACpB,QAAQ,EAAE,KAAK,CAAC,QAAQ;SACzB,CAAC;IACJ,CAAC;IAED,KAAK,CAAC,UAAU;QACd,IAAI,IAAI,CAAC,eAAe;YAAE,OAAO;QACjC,IAAI,CAAC,eAAe,GAAG,IAAI,CAAC;QAC5B,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,UAAU,EAAE,CAAC;QAC1B,CAAC;gBAAS,CAAC;YACT,IAAI,CAAC,eAAe,GAAG,KAAK,CAAC;QAC/B,CAAC;IACH,CAAC;IAEO,KAAK,CAAC,UAAU;QACtB,yEAAyE;QACzE,wEAAwE;QACxE,sEAAsE;QACtE,yEAAyE;QACzE,kEAAkE;QAClE,qEAAqE;QACrE,MAAM,aAAa,GAAG,IAAI,GAAG,EAAoB,CAAC;QAClD,KAAK,MAAM,IAAI,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,CAAC,EAAE,CAAC;YAC1D,aAAa,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;QAC9B,CAAC;QACD,KAAK,MAAM,CAAC,GAAG,EAAE,MAAM,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,oBAAoB,CAAC,EAAE,CAAC;YAC3E,MAAM,IAAI,GAAG,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,WAAW,CAAC,IAAI,EAAE,CAAC;YACzD,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACf,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,WAAW,EAAE,IAAI,CAAC,CAAC;QAC9C,CAAC;QAED,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,GAAG,CAC/B,KAAK,CAAC,IAAI,CAAC,aAAa,CAAC,OAAO,EAAE,CAAC,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,EAAE;YAC7D,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,CAAC;YAChD,IAAI,CAAC,OAAO,EAAE,CAAC;gBACb,OAAO;oBACL,WAAW,EAAE,IAAI;oBACjB,YAAY,EAAE,IAAI;oBAClB,KAAK,EAAE,KAAK;oBACZ,MAAM,EAAE,WAAW,IAAI,sCAAsC;oBAC7D,QAAQ,EAAE;wBACR,WAAW,EAAE,sDAAsD;qBACpE;iBACF,CAAC;YACJ,CAAC;YACD,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,CAAC;gBACrB,iEAAiE;gBACjE,OAAO;oBACL,WAAW,EAAE,IAAI;oBACjB,YAAY,EAAE,IAAI;oBAClB,KAAK,EAAE,IAAI;iBACZ,CAAC;YACJ,CAAC;YACD,IAAI,CAAC;gBACH,IAAI,aAAwD,CAAC;gBAC7D,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC;oBAChC,OAAO,CAAC,OAAO,CAAC,EAAE,UAAU,EAAE,GAAG,EAAE,CAAC;oBACpC,IAAI,OAAO,CAAc,CAAC,CAAC,EAAE,MAAM,EAAE,EAAE;wBACrC,aAAa,GAAG,UAAU,CACxB,GAAG,EAAE,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,mBAAmB,CAAC,CAAC,EAC5C,IAAI,CAAC,IAAI,CAAC,gBAAgB,CAC3B,CAAC;oBACJ,CAAC,CAAC;iBACH,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE;oBACd,IAAI,aAAa,KAAK,SAAS;wBAAE,YAAY,CAAC,aAAa,CAAC,CAAC;gBAC/D,CAAC,CAAC,CAAC;gBACH,OAAO;oBACL,WAAW,EAAE,IAAI;oBACjB,YAAY,EAAE,IAAI;oBAClB,KAAK,EAAE,MAAM,CAAC,KAAK;oBACnB,MAAM,EAAE,MAAM,CAAC,MAAM;oBACrB,QAAQ,EAAE,MAAM,CAAC,QAAQ;iBAC1B,CAAC;YACJ,CAAC;YAAC,OAAO,GAAG,EAAE,CAAC;gBACb,MAAM,GAAG,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;gBAC7D,OAAO;oBACL,WAAW,EAAE,IAAI;oBACjB,YAAY,EAAE,IAAI;oBAClB,KAAK,EAAE,KAAK;oBACZ,MAAM,EAAE,kBAAkB,GAAG,EAAE;iBAChC,CAAC;YACJ,CAAC;QACH,CAAC,CAAC,CACH,CAAC;QAEF,IAAI,CAAC,QAAQ,GAAG;YACd,eAAe,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACzC,SAAS,EAAE,OAAO;SACnB,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -0,0 +1,59 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Revert-decision logic for the learner's Memory-consolidation phase (#764).
|
|
3
|
+
*
|
|
4
|
+
* Given per-codeDigest pass-rate aggregates for a candidate Improve commit
|
|
5
|
+
* (`withCommit`) and its parent (`atParent`), decide whether the commit
|
|
6
|
+
* significantly regressed the frozen-eval pass rate and should be reverted.
|
|
7
|
+
*
|
|
8
|
+
* Thresholds are explicit and documented here (AC4) — not magic constants:
|
|
9
|
+
* - minSamplesPerArm: minimum indexed attempts required in EACH arm before a
|
|
10
|
+
* statistical test is meaningful. Below it, plateau is expected Level-1
|
|
11
|
+
* behaviour, so we abstain (reason 'insufficient_samples').
|
|
12
|
+
* - alpha: significance threshold (revert only when p < alpha, two-sided).
|
|
13
|
+
* - recentAttemptsWindow: how many most-recent attempts per codeDigest the
|
|
14
|
+
* aggregate is computed over. Overridable via implStateDir/policy.json
|
|
15
|
+
* (`policy.revert.recentAttemptsWindow`), mirroring policy.maxNotesBytes.
|
|
16
|
+
*/
|
|
17
|
+
export interface RevertPolicy {
|
|
18
|
+
/** Minimum indexed attempts required per arm. Default 30. */
|
|
19
|
+
minSamplesPerArm: number;
|
|
20
|
+
/** Two-sided significance threshold. Default 0.05 (95% confidence). */
|
|
21
|
+
alpha: number;
|
|
22
|
+
/** Recent-attempts window per codeDigest. Default 200. */
|
|
23
|
+
recentAttemptsWindow: number;
|
|
24
|
+
}
|
|
25
|
+
export declare const DEFAULT_REVERT_POLICY: RevertPolicy;
|
|
26
|
+
export interface CodeDigestAggregate {
|
|
27
|
+
codeDigest: string;
|
|
28
|
+
/** Total indexed attempts for this codeDigest (within the window). */
|
|
29
|
+
attempts: number;
|
|
30
|
+
/** Pass count (verdictEnvelopeMeta.actualPassed === true). */
|
|
31
|
+
passes: number;
|
|
32
|
+
/** passes / attempts; 0 when attempts === 0. */
|
|
33
|
+
passRate: number;
|
|
34
|
+
}
|
|
35
|
+
export type RevertReason = 'significant_regression' | 'insufficient_samples' | 'not_significant' | 'no_regression';
|
|
36
|
+
export interface RevertDecisionInput {
|
|
37
|
+
withCommit: CodeDigestAggregate;
|
|
38
|
+
atParent: CodeDigestAggregate;
|
|
39
|
+
}
|
|
40
|
+
export interface RevertDecision {
|
|
41
|
+
withCommit: {
|
|
42
|
+
codeDigest: string;
|
|
43
|
+
n: number;
|
|
44
|
+
passRate: number;
|
|
45
|
+
};
|
|
46
|
+
atParent: {
|
|
47
|
+
codeDigest: string;
|
|
48
|
+
n: number;
|
|
49
|
+
passRate: number;
|
|
50
|
+
};
|
|
51
|
+
delta: number;
|
|
52
|
+
pValue: number;
|
|
53
|
+
significant: boolean;
|
|
54
|
+
recommendRevert: boolean;
|
|
55
|
+
reason: RevertReason;
|
|
56
|
+
}
|
|
57
|
+
export declare function decideRevert(input: RevertDecisionInput, policy?: RevertPolicy): RevertDecision;
|
|
58
|
+
/** Merge a partial policy (e.g. from implStateDir/policy.json) over the defaults. */
|
|
59
|
+
export declare function resolveRevertPolicy(override?: Partial<RevertPolicy>): RevertPolicy;
|
|
@@ -0,0 +1,53 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Revert-decision logic for the learner's Memory-consolidation phase (#764).
|
|
3
|
+
*
|
|
4
|
+
* Given per-codeDigest pass-rate aggregates for a candidate Improve commit
|
|
5
|
+
* (`withCommit`) and its parent (`atParent`), decide whether the commit
|
|
6
|
+
* significantly regressed the frozen-eval pass rate and should be reverted.
|
|
7
|
+
*
|
|
8
|
+
* Thresholds are explicit and documented here (AC4) — not magic constants:
|
|
9
|
+
* - minSamplesPerArm: minimum indexed attempts required in EACH arm before a
|
|
10
|
+
* statistical test is meaningful. Below it, plateau is expected Level-1
|
|
11
|
+
* behaviour, so we abstain (reason 'insufficient_samples').
|
|
12
|
+
* - alpha: significance threshold (revert only when p < alpha, two-sided).
|
|
13
|
+
* - recentAttemptsWindow: how many most-recent attempts per codeDigest the
|
|
14
|
+
* aggregate is computed over. Overridable via implStateDir/policy.json
|
|
15
|
+
* (`policy.revert.recentAttemptsWindow`), mirroring policy.maxNotesBytes.
|
|
16
|
+
*/
|
|
17
|
+
import { twoProportionZTest } from './revert-stats.js';
|
|
18
|
+
export const DEFAULT_REVERT_POLICY = {
|
|
19
|
+
minSamplesPerArm: 30,
|
|
20
|
+
alpha: 0.05,
|
|
21
|
+
recentAttemptsWindow: 200,
|
|
22
|
+
};
|
|
23
|
+
export function decideRevert(input, policy = DEFAULT_REVERT_POLICY) {
|
|
24
|
+
const { withCommit, atParent } = input;
|
|
25
|
+
const base = {
|
|
26
|
+
withCommit: { codeDigest: withCommit.codeDigest, n: withCommit.attempts, passRate: withCommit.passRate },
|
|
27
|
+
atParent: { codeDigest: atParent.codeDigest, n: atParent.attempts, passRate: atParent.passRate },
|
|
28
|
+
};
|
|
29
|
+
// Sample floor first — a zero-attempt codeDigest is "insufficient_samples",
|
|
30
|
+
// NOT pass-rate zero (a fresh promotion that has not run yet is not a regression).
|
|
31
|
+
if (withCommit.attempts < policy.minSamplesPerArm || atParent.attempts < policy.minSamplesPerArm) {
|
|
32
|
+
return { ...base, delta: withCommit.passRate - atParent.passRate, pValue: 1, significant: false, recommendRevert: false, reason: 'insufficient_samples' };
|
|
33
|
+
}
|
|
34
|
+
const stats = twoProportionZTest({
|
|
35
|
+
passesA: withCommit.passes,
|
|
36
|
+
totalA: withCommit.attempts,
|
|
37
|
+
passesB: atParent.passes,
|
|
38
|
+
totalB: atParent.attempts,
|
|
39
|
+
});
|
|
40
|
+
const significant = stats.pValue < policy.alpha;
|
|
41
|
+
if (stats.delta >= 0) {
|
|
42
|
+
return { ...base, delta: stats.delta, pValue: stats.pValue, significant, recommendRevert: false, reason: 'no_regression' };
|
|
43
|
+
}
|
|
44
|
+
if (!significant) {
|
|
45
|
+
return { ...base, delta: stats.delta, pValue: stats.pValue, significant, recommendRevert: false, reason: 'not_significant' };
|
|
46
|
+
}
|
|
47
|
+
return { ...base, delta: stats.delta, pValue: stats.pValue, significant: true, recommendRevert: true, reason: 'significant_regression' };
|
|
48
|
+
}
|
|
49
|
+
/** Merge a partial policy (e.g. from implStateDir/policy.json) over the defaults. */
|
|
50
|
+
export function resolveRevertPolicy(override) {
|
|
51
|
+
return { ...DEFAULT_REVERT_POLICY, ...(override ?? {}) };
|
|
52
|
+
}
|
|
53
|
+
//# sourceMappingURL=revert-decision.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"revert-decision.js","sourceRoot":"","sources":["../../src/learner/revert-decision.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAEH,OAAO,EAAE,kBAAkB,EAAE,MAAM,mBAAmB,CAAC;AAWvD,MAAM,CAAC,MAAM,qBAAqB,GAAiB;IACjD,gBAAgB,EAAE,EAAE;IACpB,KAAK,EAAE,IAAI;IACX,oBAAoB,EAAE,GAAG;CAC1B,CAAC;AAiCF,MAAM,UAAU,YAAY,CAC1B,KAA0B,EAC1B,SAAuB,qBAAqB;IAE5C,MAAM,EAAE,UAAU,EAAE,QAAQ,EAAE,GAAG,KAAK,CAAC;IACvC,MAAM,IAAI,GAAG;QACX,UAAU,EAAE,EAAE,UAAU,EAAE,UAAU,CAAC,UAAU,EAAE,CAAC,EAAE,UAAU,CAAC,QAAQ,EAAE,QAAQ,EAAE,UAAU,CAAC,QAAQ,EAAE;QACxG,QAAQ,EAAE,EAAE,UAAU,EAAE,QAAQ,CAAC,UAAU,EAAE,CAAC,EAAE,QAAQ,CAAC,QAAQ,EAAE,QAAQ,EAAE,QAAQ,CAAC,QAAQ,EAAE;KACjG,CAAC;IAEF,4EAA4E;IAC5E,mFAAmF;IACnF,IAAI,UAAU,CAAC,QAAQ,GAAG,MAAM,CAAC,gBAAgB,IAAI,QAAQ,CAAC,QAAQ,GAAG,MAAM,CAAC,gBAAgB,EAAE,CAAC;QACjG,OAAO,EAAE,GAAG,IAAI,EAAE,KAAK,EAAE,UAAU,CAAC,QAAQ,GAAG,QAAQ,CAAC,QAAQ,EAAE,MAAM,EAAE,CAAC,EAAE,WAAW,EAAE,KAAK,EAAE,eAAe,EAAE,KAAK,EAAE,MAAM,EAAE,sBAAsB,EAAE,CAAC;IAC5J,CAAC;IAED,MAAM,KAAK,GAAG,kBAAkB,CAAC;QAC/B,OAAO,EAAE,UAAU,CAAC,MAAM;QAC1B,MAAM,EAAE,UAAU,CAAC,QAAQ;QAC3B,OAAO,EAAE,QAAQ,CAAC,MAAM;QACxB,MAAM,EAAE,QAAQ,CAAC,QAAQ;KAC1B,CAAC,CAAC;IACH,MAAM,WAAW,GAAG,KAAK,CAAC,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC;IAEhD,IAAI,KAAK,CAAC,KAAK,IAAI,CAAC,EAAE,CAAC;QACrB,OAAO,EAAE,GAAG,IAAI,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,MAAM,EAAE,KAAK,CAAC,MAAM,EAAE,WAAW,EAAE,eAAe,EAAE,KAAK,EAAE,MAAM,EAAE,eAAe,EAAE,CAAC;IAC7H,CAAC;IACD,IAAI,CAAC,WAAW,EAAE,CAAC;QACjB,OAAO,EAAE,GAAG,IAAI,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,MAAM,EAAE,KAAK,CAAC,MAAM,EAAE,WAAW,EAAE,eAAe,EAAE,KAAK,EAAE,MAAM,EAAE,iBAAiB,EAAE,CAAC;IAC/H,CAAC;IACD,OAAO,EAAE,GAAG,IAAI,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,MAAM,EAAE,KAAK,CAAC,MAAM,EAAE,WAAW,EAAE,IAAI,EAAE,eAAe,EAAE,IAAI,EAAE,MAAM,EAAE,wBAAwB,EAAE,CAAC;AAC3I,CAAC;AAED,qFAAqF;AACrF,MAAM,UAAU,mBAAmB,CAAC,QAAgC;IAClE,OAAO,EAAE,GAAG,qBAAqB,EAAE,GAAG,CAAC,QAAQ,IAAI,EAAE,CAAC,EAAE,CAAC;AAC3D,CAAC"}
|
|
@@ -0,0 +1,24 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Pure statistics for the per-codeDigest revert decision (issue #764).
|
|
3
|
+
*
|
|
4
|
+
* Two-proportion z-test on pass/total for "arm A" (codeDigest WITH a candidate
|
|
5
|
+
* Improve commit) vs "arm B" (codeDigest AT the commit's parent). No I/O; unit-
|
|
6
|
+
* tested directly with hand-computed z-values. `delta = pA - pB` (negative means
|
|
7
|
+
* the commit made the pass rate worse).
|
|
8
|
+
*/
|
|
9
|
+
export interface TwoProportionInput {
|
|
10
|
+
passesA: number;
|
|
11
|
+
totalA: number;
|
|
12
|
+
passesB: number;
|
|
13
|
+
totalB: number;
|
|
14
|
+
}
|
|
15
|
+
export interface TwoProportionResult {
|
|
16
|
+
/** pA - pB (negative => arm A is worse). */
|
|
17
|
+
delta: number;
|
|
18
|
+
/** Test statistic; sign matches `delta`. 0 when either arm has no samples. */
|
|
19
|
+
z: number;
|
|
20
|
+
/** Two-sided p-value in [0, 1]. 1 when there is no signal. */
|
|
21
|
+
pValue: number;
|
|
22
|
+
}
|
|
23
|
+
/** Two-proportion z-test. Returns no-signal (z=0, p=1) if either total is 0. */
|
|
24
|
+
export declare function twoProportionZTest(input: TwoProportionInput): TwoProportionResult;
|
|
@@ -0,0 +1,44 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Pure statistics for the per-codeDigest revert decision (issue #764).
|
|
3
|
+
*
|
|
4
|
+
* Two-proportion z-test on pass/total for "arm A" (codeDigest WITH a candidate
|
|
5
|
+
* Improve commit) vs "arm B" (codeDigest AT the commit's parent). No I/O; unit-
|
|
6
|
+
* tested directly with hand-computed z-values. `delta = pA - pB` (negative means
|
|
7
|
+
* the commit made the pass rate worse).
|
|
8
|
+
*/
|
|
9
|
+
/** Two-proportion z-test. Returns no-signal (z=0, p=1) if either total is 0. */
|
|
10
|
+
export function twoProportionZTest(input) {
|
|
11
|
+
const { passesA, totalA, passesB, totalB } = input;
|
|
12
|
+
if (totalA <= 0 || totalB <= 0) {
|
|
13
|
+
const pA = totalA > 0 ? passesA / totalA : 0;
|
|
14
|
+
const pB = totalB > 0 ? passesB / totalB : 0;
|
|
15
|
+
return { delta: pA - pB, z: 0, pValue: 1 };
|
|
16
|
+
}
|
|
17
|
+
const pA = passesA / totalA;
|
|
18
|
+
const pB = passesB / totalB;
|
|
19
|
+
const delta = pA - pB;
|
|
20
|
+
const pooled = (passesA + passesB) / (totalA + totalB);
|
|
21
|
+
const se = Math.sqrt(pooled * (1 - pooled) * (1 / totalA + 1 / totalB));
|
|
22
|
+
if (se === 0) {
|
|
23
|
+
// Both arms 0% or both 100% — no measurable difference.
|
|
24
|
+
return { delta, z: 0, pValue: 1 };
|
|
25
|
+
}
|
|
26
|
+
const z = delta / se;
|
|
27
|
+
const pValue = 2 * (1 - standardNormalCdf(Math.abs(z)));
|
|
28
|
+
return { delta, z, pValue };
|
|
29
|
+
}
|
|
30
|
+
/** Standard normal CDF via the Abramowitz & Stegun 7.1.26 erf approximation. */
|
|
31
|
+
function standardNormalCdf(x) {
|
|
32
|
+
return 0.5 * (1 + erf(x / Math.SQRT2));
|
|
33
|
+
}
|
|
34
|
+
function erf(x) {
|
|
35
|
+
// Abramowitz & Stegun 7.1.26, max abs error ~1.5e-7.
|
|
36
|
+
const t = 1 / (1 + 0.3275911 * Math.abs(x));
|
|
37
|
+
const y = 1 -
|
|
38
|
+
((((1.061405429 * t - 1.453152027) * t + 1.421413741) * t - 0.284496736) * t +
|
|
39
|
+
0.254829592) *
|
|
40
|
+
t *
|
|
41
|
+
Math.exp(-x * x);
|
|
42
|
+
return x >= 0 ? y : -y;
|
|
43
|
+
}
|
|
44
|
+
//# sourceMappingURL=revert-stats.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"revert-stats.js","sourceRoot":"","sources":["../../src/learner/revert-stats.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAkBH,gFAAgF;AAChF,MAAM,UAAU,kBAAkB,CAAC,KAAyB;IAC1D,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,GAAG,KAAK,CAAC;IACnD,IAAI,MAAM,IAAI,CAAC,IAAI,MAAM,IAAI,CAAC,EAAE,CAAC;QAC/B,MAAM,EAAE,GAAG,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;QAC7C,MAAM,EAAE,GAAG,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;QAC7C,OAAO,EAAE,KAAK,EAAE,EAAE,GAAG,EAAE,EAAE,CAAC,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;IAC7C,CAAC;IACD,MAAM,EAAE,GAAG,OAAO,GAAG,MAAM,CAAC;IAC5B,MAAM,EAAE,GAAG,OAAO,GAAG,MAAM,CAAC;IAC5B,MAAM,KAAK,GAAG,EAAE,GAAG,EAAE,CAAC;IACtB,MAAM,MAAM,GAAG,CAAC,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,MAAM,GAAG,MAAM,CAAC,CAAC;IACvD,MAAM,EAAE,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,GAAG,MAAM,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC;IACxE,IAAI,EAAE,KAAK,CAAC,EAAE,CAAC;QACb,wDAAwD;QACxD,OAAO,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;IACpC,CAAC;IACD,MAAM,CAAC,GAAG,KAAK,GAAG,EAAE,CAAC;IACrB,MAAM,MAAM,GAAG,CAAC,GAAG,CAAC,CAAC,GAAG,iBAAiB,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACxD,OAAO,EAAE,KAAK,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC;AAC9B,CAAC;AAED,gFAAgF;AAChF,SAAS,iBAAiB,CAAC,CAAS;IAClC,OAAO,GAAG,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;AACzC,CAAC;AAED,SAAS,GAAG,CAAC,CAAS;IACpB,qDAAqD;IACrD,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,GAAG,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IAC5C,MAAM,CAAC,GACL,CAAC;QACD,CAAC,CAAC,CAAC,CAAC,WAAW,GAAG,CAAC,GAAG,WAAW,CAAC,GAAG,CAAC,GAAG,WAAW,CAAC,GAAG,CAAC,GAAG,WAAW,CAAC,GAAG,CAAC;YAC1E,WAAW,CAAC;YACZ,CAAC;YACD,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IACrB,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AACzB,CAAC"}
|