@namzu/sdk 2.0.0 → 3.0.0
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 +2265 -0
- package/README.md +49 -53
- package/dist/__tests__/default-path-conformance.test.d.ts +2 -0
- package/dist/__tests__/default-path-conformance.test.d.ts.map +1 -0
- package/dist/__tests__/default-path-conformance.test.js +262 -0
- package/dist/__tests__/default-path-conformance.test.js.map +1 -0
- package/dist/__tests__/shipped-defaults.test.d.ts +2 -0
- package/dist/__tests__/shipped-defaults.test.d.ts.map +1 -0
- package/dist/__tests__/shipped-defaults.test.js +97 -0
- package/dist/__tests__/shipped-defaults.test.js.map +1 -0
- package/dist/advisory/__tests__/budget-enforcement.test.d.ts +2 -0
- package/dist/advisory/__tests__/budget-enforcement.test.d.ts.map +1 -0
- package/dist/advisory/__tests__/budget-enforcement.test.js +159 -0
- package/dist/advisory/__tests__/budget-enforcement.test.js.map +1 -0
- package/dist/advisory/__tests__/structured-response.test.d.ts +2 -0
- package/dist/advisory/__tests__/structured-response.test.d.ts.map +1 -0
- package/dist/advisory/__tests__/structured-response.test.js +99 -0
- package/dist/advisory/__tests__/structured-response.test.js.map +1 -0
- package/dist/advisory/budget.d.ts +15 -0
- package/dist/advisory/budget.d.ts.map +1 -0
- package/dist/advisory/budget.js +21 -0
- package/dist/advisory/budget.js.map +1 -0
- package/dist/advisory/context.d.ts +2 -0
- package/dist/advisory/context.d.ts.map +1 -1
- package/dist/advisory/context.js +16 -0
- package/dist/advisory/context.js.map +1 -1
- package/dist/advisory/executor.d.ts +14 -11
- package/dist/advisory/executor.d.ts.map +1 -1
- package/dist/advisory/executor.js +36 -24
- package/dist/advisory/executor.js.map +1 -1
- package/dist/advisory/executor.test.d.ts +8 -6
- package/dist/advisory/executor.test.d.ts.map +1 -1
- package/dist/advisory/executor.test.js +12 -8
- package/dist/advisory/executor.test.js.map +1 -1
- package/dist/advisory/index.d.ts +2 -0
- package/dist/advisory/index.d.ts.map +1 -1
- package/dist/advisory/index.js +2 -0
- package/dist/advisory/index.js.map +1 -1
- package/dist/advisory/parse.d.ts +23 -0
- package/dist/advisory/parse.d.ts.map +1 -0
- package/dist/advisory/parse.js +56 -0
- package/dist/advisory/parse.js.map +1 -0
- package/dist/agents/AbstractAgent.d.ts +48 -0
- package/dist/agents/AbstractAgent.d.ts.map +1 -1
- package/dist/agents/AbstractAgent.js +77 -0
- package/dist/agents/AbstractAgent.js.map +1 -1
- package/dist/agents/PipelineAgent.d.ts +10 -0
- package/dist/agents/PipelineAgent.d.ts.map +1 -1
- package/dist/agents/PipelineAgent.js +12 -0
- package/dist/agents/PipelineAgent.js.map +1 -1
- package/dist/agents/ReactiveAgent.d.ts +10 -0
- package/dist/agents/ReactiveAgent.d.ts.map +1 -1
- package/dist/agents/ReactiveAgent.js +40 -0
- package/dist/agents/ReactiveAgent.js.map +1 -1
- package/dist/agents/RouterAgent.d.ts +10 -0
- package/dist/agents/RouterAgent.d.ts.map +1 -1
- package/dist/agents/RouterAgent.js +23 -2
- package/dist/agents/RouterAgent.js.map +1 -1
- package/dist/agents/SupervisorAgent.d.ts +11 -1
- package/dist/agents/SupervisorAgent.d.ts.map +1 -1
- package/dist/agents/SupervisorAgent.js +23 -1
- package/dist/agents/SupervisorAgent.js.map +1 -1
- package/dist/agents/__tests__/idempotency-key.test.d.ts +2 -0
- package/dist/agents/__tests__/idempotency-key.test.d.ts.map +1 -0
- package/dist/agents/__tests__/idempotency-key.test.js +149 -0
- package/dist/agents/__tests__/idempotency-key.test.js.map +1 -0
- package/dist/agents/__tests__/invocation-lock.test.d.ts +2 -0
- package/dist/agents/__tests__/invocation-lock.test.d.ts.map +1 -0
- package/dist/agents/__tests__/invocation-lock.test.js +126 -0
- package/dist/agents/__tests__/invocation-lock.test.js.map +1 -0
- package/dist/agents/__tests__/reachability.test.d.ts +2 -0
- package/dist/agents/__tests__/reachability.test.d.ts.map +1 -0
- package/dist/agents/__tests__/reachability.test.js +173 -0
- package/dist/agents/__tests__/reachability.test.js.map +1 -0
- package/dist/agents/__tests__/supervisor-ledger.test.js +1 -1
- package/dist/agents/__tests__/supervisor-ledger.test.js.map +1 -1
- package/dist/agents/index.d.ts +1 -1
- package/dist/agents/index.d.ts.map +1 -1
- package/dist/agents/lock.d.ts +0 -1
- package/dist/agents/lock.d.ts.map +1 -1
- package/dist/agents/lock.js.map +1 -1
- package/dist/bridge/a2a/late-events.test.d.ts +2 -0
- package/dist/bridge/a2a/late-events.test.d.ts.map +1 -0
- package/dist/bridge/a2a/late-events.test.js +70 -0
- package/dist/bridge/a2a/late-events.test.js.map +1 -0
- package/dist/bridge/a2a/mapper.d.ts.map +1 -1
- package/dist/bridge/a2a/mapper.js +55 -3
- package/dist/bridge/a2a/mapper.js.map +1 -1
- package/dist/bridge/a2a/mapper.test.js +1 -1
- package/dist/bridge/a2a/mapper.test.js.map +1 -1
- package/dist/bridge/a2a/message.d.ts.map +1 -1
- package/dist/bridge/a2a/message.js +5 -1
- package/dist/bridge/a2a/message.js.map +1 -1
- package/dist/bridge/sse/late-events.test.d.ts +2 -0
- package/dist/bridge/sse/late-events.test.d.ts.map +1 -0
- package/dist/bridge/sse/late-events.test.js +163 -0
- package/dist/bridge/sse/late-events.test.js.map +1 -0
- package/dist/bridge/sse/mapper.d.ts.map +1 -1
- package/dist/bridge/sse/mapper.js +100 -0
- package/dist/bridge/sse/mapper.js.map +1 -1
- package/dist/compaction/__tests__/cleared-result-recovery.test.d.ts +2 -0
- package/dist/compaction/__tests__/cleared-result-recovery.test.d.ts.map +1 -0
- package/dist/compaction/__tests__/cleared-result-recovery.test.js +101 -0
- package/dist/compaction/__tests__/cleared-result-recovery.test.js.map +1 -0
- package/dist/compaction/__tests__/context-window.test.d.ts +2 -0
- package/dist/compaction/__tests__/context-window.test.d.ts.map +1 -0
- package/dist/compaction/__tests__/context-window.test.js +62 -0
- package/dist/compaction/__tests__/context-window.test.js.map +1 -0
- package/dist/compaction/__tests__/dangling.test.js +16 -5
- package/dist/compaction/__tests__/dangling.test.js.map +1 -1
- package/dist/compaction/__tests__/lossiness.test.d.ts +2 -0
- package/dist/compaction/__tests__/lossiness.test.d.ts.map +1 -0
- package/dist/compaction/__tests__/lossiness.test.js +163 -0
- package/dist/compaction/__tests__/lossiness.test.js.map +1 -0
- package/dist/compaction/__tests__/state-across-resume.test.d.ts +2 -0
- package/dist/compaction/__tests__/state-across-resume.test.d.ts.map +1 -0
- package/dist/compaction/__tests__/state-across-resume.test.js +88 -0
- package/dist/compaction/__tests__/state-across-resume.test.js.map +1 -0
- package/dist/compaction/__tests__/tool-result-editing.test.d.ts +2 -0
- package/dist/compaction/__tests__/tool-result-editing.test.d.ts.map +1 -0
- package/dist/compaction/__tests__/tool-result-editing.test.js +184 -0
- package/dist/compaction/__tests__/tool-result-editing.test.js.map +1 -0
- package/dist/compaction/__tests__/trim-opens-on-user.test.d.ts +2 -0
- package/dist/compaction/__tests__/trim-opens-on-user.test.d.ts.map +1 -0
- package/dist/compaction/__tests__/trim-opens-on-user.test.js +109 -0
- package/dist/compaction/__tests__/trim-opens-on-user.test.js.map +1 -0
- package/dist/compaction/__tests__/verifier-model.test.d.ts +2 -0
- package/dist/compaction/__tests__/verifier-model.test.d.ts.map +1 -0
- package/dist/compaction/__tests__/verifier-model.test.js +58 -0
- package/dist/compaction/__tests__/verifier-model.test.js.map +1 -0
- package/dist/compaction/context-window.d.ts +44 -0
- package/dist/compaction/context-window.d.ts.map +1 -0
- package/dist/compaction/context-window.js +97 -0
- package/dist/compaction/context-window.js.map +1 -0
- package/dist/compaction/dangling.js +48 -1
- package/dist/compaction/dangling.js.map +1 -1
- package/dist/compaction/extractor.d.ts.map +1 -1
- package/dist/compaction/extractor.js +55 -1
- package/dist/compaction/extractor.js.map +1 -1
- package/dist/compaction/index.d.ts +5 -0
- package/dist/compaction/index.d.ts.map +1 -1
- package/dist/compaction/index.js +3 -0
- package/dist/compaction/index.js.map +1 -1
- package/dist/compaction/manager.d.ts +25 -0
- package/dist/compaction/manager.d.ts.map +1 -1
- package/dist/compaction/manager.js +42 -8
- package/dist/compaction/manager.js.map +1 -1
- package/dist/compaction/managers/structured.d.ts.map +1 -1
- package/dist/compaction/managers/structured.js +7 -3
- package/dist/compaction/managers/structured.js.map +1 -1
- package/dist/compaction/retention.d.ts +15 -0
- package/dist/compaction/retention.d.ts.map +1 -0
- package/dist/compaction/retention.js +89 -0
- package/dist/compaction/retention.js.map +1 -0
- package/dist/compaction/serializer.d.ts.map +1 -1
- package/dist/compaction/serializer.js +28 -11
- package/dist/compaction/serializer.js.map +1 -1
- package/dist/compaction/tool-result-editing.d.ts +54 -0
- package/dist/compaction/tool-result-editing.d.ts.map +1 -0
- package/dist/compaction/tool-result-editing.js +164 -0
- package/dist/compaction/tool-result-editing.js.map +1 -0
- package/dist/compaction/types.d.ts +10 -0
- package/dist/compaction/types.d.ts.map +1 -1
- package/dist/compaction/verifier.d.ts +20 -1
- package/dist/compaction/verifier.d.ts.map +1 -1
- package/dist/compaction/verifier.js +11 -2
- package/dist/compaction/verifier.js.map +1 -1
- package/dist/compaction/wire.d.ts +54 -0
- package/dist/compaction/wire.d.ts.map +1 -0
- package/dist/compaction/wire.js +45 -0
- package/dist/compaction/wire.js.map +1 -0
- package/dist/config/runtime.d.ts +161 -45
- package/dist/config/runtime.d.ts.map +1 -1
- package/dist/config/runtime.js +61 -19
- package/dist/config/runtime.js.map +1 -1
- package/dist/connector/builtins/http.d.ts.map +1 -1
- package/dist/connector/builtins/http.js +14 -1
- package/dist/connector/builtins/http.js.map +1 -1
- package/dist/connector/builtins/http.test.js +12 -0
- package/dist/connector/builtins/http.test.js.map +1 -1
- package/dist/connector/mcp/__tests__/discovery-policy.test.d.ts +2 -0
- package/dist/connector/mcp/__tests__/discovery-policy.test.d.ts.map +1 -0
- package/dist/connector/mcp/__tests__/discovery-policy.test.js +109 -0
- package/dist/connector/mcp/__tests__/discovery-policy.test.js.map +1 -0
- package/dist/connector/mcp/__tests__/list-pagination.test.d.ts +2 -0
- package/dist/connector/mcp/__tests__/list-pagination.test.d.ts.map +1 -0
- package/dist/connector/mcp/__tests__/list-pagination.test.js +81 -0
- package/dist/connector/mcp/__tests__/list-pagination.test.js.map +1 -0
- package/dist/connector/mcp/__tests__/policy.test.d.ts +2 -0
- package/dist/connector/mcp/__tests__/policy.test.d.ts.map +1 -0
- package/dist/connector/mcp/__tests__/policy.test.js +86 -0
- package/dist/connector/mcp/__tests__/policy.test.js.map +1 -0
- package/dist/connector/mcp/__tests__/protocol-negotiation.test.d.ts +2 -0
- package/dist/connector/mcp/__tests__/protocol-negotiation.test.d.ts.map +1 -0
- package/dist/connector/mcp/__tests__/protocol-negotiation.test.js +81 -0
- package/dist/connector/mcp/__tests__/protocol-negotiation.test.js.map +1 -0
- package/dist/connector/mcp/__tests__/schema-fidelity.test.d.ts +2 -0
- package/dist/connector/mcp/__tests__/schema-fidelity.test.d.ts.map +1 -0
- package/dist/connector/mcp/__tests__/schema-fidelity.test.js +206 -0
- package/dist/connector/mcp/__tests__/schema-fidelity.test.js.map +1 -0
- package/dist/connector/mcp/__tests__/structured-result.test.d.ts +2 -0
- package/dist/connector/mcp/__tests__/structured-result.test.d.ts.map +1 -0
- package/dist/connector/mcp/__tests__/structured-result.test.js +112 -0
- package/dist/connector/mcp/__tests__/structured-result.test.js.map +1 -0
- package/dist/connector/mcp/adapter.d.ts +14 -0
- package/dist/connector/mcp/adapter.d.ts.map +1 -1
- package/dist/connector/mcp/adapter.js +292 -18
- package/dist/connector/mcp/adapter.js.map +1 -1
- package/dist/connector/mcp/client.d.ts +38 -0
- package/dist/connector/mcp/client.d.ts.map +1 -1
- package/dist/connector/mcp/client.js +138 -12
- package/dist/connector/mcp/client.js.map +1 -1
- package/dist/connector/mcp/client.test.d.ts +2 -0
- package/dist/connector/mcp/client.test.d.ts.map +1 -0
- package/dist/connector/mcp/client.test.js +105 -0
- package/dist/connector/mcp/client.test.js.map +1 -0
- package/dist/connector/mcp/discovery.d.ts +32 -1
- package/dist/connector/mcp/discovery.d.ts.map +1 -1
- package/dist/connector/mcp/discovery.js +41 -4
- package/dist/connector/mcp/discovery.js.map +1 -1
- package/dist/connector/mcp/http-sse.d.ts.map +1 -1
- package/dist/connector/mcp/http-sse.js +15 -1
- package/dist/connector/mcp/http-sse.js.map +1 -1
- package/dist/connector/mcp/index.d.ts +3 -0
- package/dist/connector/mcp/index.d.ts.map +1 -1
- package/dist/connector/mcp/index.js +1 -0
- package/dist/connector/mcp/index.js.map +1 -1
- package/dist/connector/mcp/policy.d.ts +57 -0
- package/dist/connector/mcp/policy.d.ts.map +1 -0
- package/dist/connector/mcp/policy.js +84 -0
- package/dist/connector/mcp/policy.js.map +1 -0
- package/dist/connector/mcp/schema-refs.d.ts +26 -0
- package/dist/connector/mcp/schema-refs.d.ts.map +1 -0
- package/dist/connector/mcp/schema-refs.js +111 -0
- package/dist/connector/mcp/schema-refs.js.map +1 -0
- package/dist/constants/agent/index.d.ts.map +1 -1
- package/dist/constants/agent/index.js +3 -0
- package/dist/constants/agent/index.js.map +1 -1
- package/dist/constants/compaction/index.js +1 -1
- package/dist/constants/compaction/index.js.map +1 -1
- package/dist/constants/continuation.d.ts +4 -4
- package/dist/constants/continuation.js +4 -4
- package/dist/constants/mcp/index.d.ts +31 -0
- package/dist/constants/mcp/index.d.ts.map +1 -1
- package/dist/constants/mcp/index.js +31 -0
- package/dist/constants/mcp/index.js.map +1 -1
- package/dist/constants/plugin/index.d.ts +23 -2
- package/dist/constants/plugin/index.d.ts.map +1 -1
- package/dist/constants/plugin/index.js +23 -2
- package/dist/constants/plugin/index.js.map +1 -1
- package/dist/constants/sandbox/index.d.ts +2 -3
- package/dist/constants/sandbox/index.d.ts.map +1 -1
- package/dist/constants/sandbox/index.js +6 -5
- package/dist/constants/sandbox/index.js.map +1 -1
- package/dist/constants/telemetry/index.d.ts +7 -0
- package/dist/constants/telemetry/index.d.ts.map +1 -1
- package/dist/constants/telemetry/index.js +7 -0
- package/dist/constants/telemetry/index.js.map +1 -1
- package/dist/constants/tools/index.d.ts +16 -0
- package/dist/constants/tools/index.d.ts.map +1 -1
- package/dist/constants/tools/index.js +16 -0
- package/dist/constants/tools/index.js.map +1 -1
- package/dist/contracts/a2a.d.ts +94 -94
- package/dist/contracts/api.d.ts +40 -2
- package/dist/contracts/api.d.ts.map +1 -1
- package/dist/contracts/run-status.d.ts +17 -0
- package/dist/contracts/run-status.d.ts.map +1 -0
- package/dist/contracts/run-status.js +41 -0
- package/dist/contracts/run-status.js.map +1 -0
- package/dist/contracts/schemas.d.ts +6 -6
- package/dist/eval/__tests__/case-deadline.test.d.ts +2 -0
- package/dist/eval/__tests__/case-deadline.test.d.ts.map +1 -0
- package/dist/eval/__tests__/case-deadline.test.js +128 -0
- package/dist/eval/__tests__/case-deadline.test.js.map +1 -0
- package/dist/eval/__tests__/end-to-end.test.d.ts +2 -0
- package/dist/eval/__tests__/end-to-end.test.d.ts.map +1 -0
- package/dist/eval/__tests__/end-to-end.test.js +223 -0
- package/dist/eval/__tests__/end-to-end.test.js.map +1 -0
- package/dist/eval/__tests__/harness-honesty.test.d.ts +2 -0
- package/dist/eval/__tests__/harness-honesty.test.d.ts.map +1 -0
- package/dist/eval/__tests__/harness-honesty.test.js +85 -0
- package/dist/eval/__tests__/harness-honesty.test.js.map +1 -0
- package/dist/eval/__tests__/harness.test.d.ts +2 -0
- package/dist/eval/__tests__/harness.test.d.ts.map +1 -0
- package/dist/eval/__tests__/harness.test.js +207 -0
- package/dist/eval/__tests__/harness.test.js.map +1 -0
- package/dist/eval/__tests__/judge.test.d.ts +2 -0
- package/dist/eval/__tests__/judge.test.d.ts.map +1 -0
- package/dist/eval/__tests__/judge.test.js +255 -0
- package/dist/eval/__tests__/judge.test.js.map +1 -0
- package/dist/eval/__tests__/scorer-gates.test.d.ts +2 -0
- package/dist/eval/__tests__/scorer-gates.test.d.ts.map +1 -0
- package/dist/eval/__tests__/scorer-gates.test.js +136 -0
- package/dist/eval/__tests__/scorer-gates.test.js.map +1 -0
- package/dist/eval/experiment.d.ts +54 -0
- package/dist/eval/experiment.d.ts.map +1 -0
- package/dist/eval/experiment.js +263 -0
- package/dist/eval/experiment.js.map +1 -0
- package/dist/eval/from-run.d.ts +30 -0
- package/dist/eval/from-run.d.ts.map +1 -0
- package/dist/eval/from-run.js +46 -0
- package/dist/eval/from-run.js.map +1 -0
- package/dist/eval/index.d.ts +8 -0
- package/dist/eval/index.d.ts.map +1 -0
- package/dist/eval/index.js +5 -0
- package/dist/eval/index.js.map +1 -0
- package/dist/eval/judge.d.ts +63 -0
- package/dist/eval/judge.d.ts.map +1 -0
- package/dist/eval/judge.js +136 -0
- package/dist/eval/judge.js.map +1 -0
- package/dist/eval/scorers.d.ts +40 -0
- package/dist/eval/scorers.d.ts.map +1 -0
- package/dist/eval/scorers.js +151 -0
- package/dist/eval/scorers.js.map +1 -0
- package/dist/eval/types.d.ts +150 -0
- package/dist/eval/types.d.ts.map +1 -0
- package/dist/eval/types.js +2 -0
- package/dist/eval/types.js.map +1 -0
- package/dist/execution/local.d.ts.map +1 -1
- package/dist/execution/local.js +6 -1
- package/dist/execution/local.js.map +1 -1
- package/dist/execution/local.test.d.ts +2 -0
- package/dist/execution/local.test.d.ts.map +1 -0
- package/dist/execution/local.test.js +37 -0
- package/dist/execution/local.test.js.map +1 -0
- package/dist/gateway/__tests__/list-survives-eviction.test.d.ts +2 -0
- package/dist/gateway/__tests__/list-survives-eviction.test.d.ts.map +1 -0
- package/dist/gateway/__tests__/list-survives-eviction.test.js +102 -0
- package/dist/gateway/__tests__/list-survives-eviction.test.js.map +1 -0
- package/dist/gateway/local.d.ts +42 -2
- package/dist/gateway/local.d.ts.map +1 -1
- package/dist/gateway/local.js +110 -3
- package/dist/gateway/local.js.map +1 -1
- package/dist/manager/agent/__tests__/dispose-cancels.test.d.ts +2 -0
- package/dist/manager/agent/__tests__/dispose-cancels.test.d.ts.map +1 -0
- package/dist/manager/agent/__tests__/dispose-cancels.test.js +88 -0
- package/dist/manager/agent/__tests__/dispose-cancels.test.js.map +1 -0
- package/dist/manager/agent/__tests__/lifecycle.test.js +230 -0
- package/dist/manager/agent/__tests__/lifecycle.test.js.map +1 -1
- package/dist/manager/agent/lifecycle.d.ts +35 -0
- package/dist/manager/agent/lifecycle.d.ts.map +1 -1
- package/dist/manager/agent/lifecycle.js +117 -7
- package/dist/manager/agent/lifecycle.js.map +1 -1
- package/dist/manager/run/persistence.d.ts +90 -1
- package/dist/manager/run/persistence.d.ts.map +1 -1
- package/dist/manager/run/persistence.js +120 -0
- package/dist/manager/run/persistence.js.map +1 -1
- package/dist/plugin/__tests__/enable-contributions.test.js +3 -3
- package/dist/plugin/__tests__/enable-contributions.test.js.map +1 -1
- package/dist/plugin/__tests__/hook-ordering.test.d.ts +2 -0
- package/dist/plugin/__tests__/hook-ordering.test.d.ts.map +1 -0
- package/dist/plugin/__tests__/hook-ordering.test.js +227 -0
- package/dist/plugin/__tests__/hook-ordering.test.js.map +1 -0
- package/dist/plugin/__tests__/install-time-refusal.test.d.ts +2 -0
- package/dist/plugin/__tests__/install-time-refusal.test.d.ts.map +1 -0
- package/dist/plugin/__tests__/install-time-refusal.test.js +49 -0
- package/dist/plugin/__tests__/install-time-refusal.test.js.map +1 -0
- package/dist/plugin/__tests__/lifecycle.test.js +95 -68
- package/dist/plugin/__tests__/lifecycle.test.js.map +1 -1
- package/dist/plugin/lifecycle.d.ts +15 -1
- package/dist/plugin/lifecycle.d.ts.map +1 -1
- package/dist/plugin/lifecycle.js +67 -20
- package/dist/plugin/lifecycle.js.map +1 -1
- package/dist/plugin/loader.d.ts +15 -0
- package/dist/plugin/loader.d.ts.map +1 -1
- package/dist/plugin/loader.js +33 -1
- package/dist/plugin/loader.js.map +1 -1
- package/dist/probe/registry.d.ts.map +1 -1
- package/dist/probe/registry.js +19 -1
- package/dist/probe/registry.js.map +1 -1
- package/dist/probe/registry.test.d.ts +9 -3
- package/dist/probe/registry.test.d.ts.map +1 -1
- package/dist/probe/registry.test.js +44 -6
- package/dist/probe/registry.test.js.map +1 -1
- package/dist/provider/__tests__/mock-script.test.d.ts +2 -0
- package/dist/provider/__tests__/mock-script.test.d.ts.map +1 -0
- package/dist/provider/__tests__/mock-script.test.js +151 -0
- package/dist/provider/__tests__/mock-script.test.js.map +1 -0
- package/dist/provider/__tests__/retry-visibility.test.d.ts +2 -0
- package/dist/provider/__tests__/retry-visibility.test.d.ts.map +1 -0
- package/dist/provider/__tests__/retry-visibility.test.js +137 -0
- package/dist/provider/__tests__/retry-visibility.test.js.map +1 -0
- package/dist/provider/__tests__/retry.test.d.ts +2 -0
- package/dist/provider/__tests__/retry.test.d.ts.map +1 -0
- package/dist/provider/__tests__/retry.test.js +259 -0
- package/dist/provider/__tests__/retry.test.js.map +1 -0
- package/dist/provider/capabilities.d.ts +1 -0
- package/dist/provider/capabilities.d.ts.map +1 -1
- package/dist/provider/capabilities.js +2 -0
- package/dist/provider/capabilities.js.map +1 -1
- package/dist/provider/collect.d.ts +1 -1
- package/dist/provider/collect.js +1 -1
- package/dist/provider/collect.test.d.ts +1 -1
- package/dist/provider/collect.test.js +1 -1
- package/dist/provider/index.d.ts +3 -0
- package/dist/provider/index.d.ts.map +1 -1
- package/dist/provider/index.js +2 -0
- package/dist/provider/index.js.map +1 -1
- package/dist/provider/mock-register.d.ts.map +1 -1
- package/dist/provider/mock-register.js +5 -2
- package/dist/provider/mock-register.js.map +1 -1
- package/dist/provider/mock.d.ts +35 -5
- package/dist/provider/mock.d.ts.map +1 -1
- package/dist/provider/mock.js +104 -32
- package/dist/provider/mock.js.map +1 -1
- package/dist/provider/retry.d.ts +42 -0
- package/dist/provider/retry.d.ts.map +1 -0
- package/dist/provider/retry.js +161 -0
- package/dist/provider/retry.js.map +1 -0
- package/dist/public-runtime.d.ts +26 -7
- package/dist/public-runtime.d.ts.map +1 -1
- package/dist/public-runtime.js +50 -7
- package/dist/public-runtime.js.map +1 -1
- package/dist/public-types.d.ts +5 -1
- package/dist/public-types.d.ts.map +1 -1
- package/dist/rag/__tests__/retrieval-math.test.d.ts +2 -0
- package/dist/rag/__tests__/retrieval-math.test.d.ts.map +1 -0
- package/dist/rag/__tests__/retrieval-math.test.js +46 -0
- package/dist/rag/__tests__/retrieval-math.test.js.map +1 -0
- package/dist/rag/chunking.d.ts +12 -0
- package/dist/rag/chunking.d.ts.map +1 -1
- package/dist/rag/chunking.js +32 -1
- package/dist/rag/chunking.js.map +1 -1
- package/dist/rag/embedding.d.ts +4 -4
- package/dist/rag/embedding.d.ts.map +1 -1
- package/dist/rag/embedding.js +23 -3
- package/dist/rag/embedding.js.map +1 -1
- package/dist/rag/embedding.test.d.ts +1 -1
- package/dist/rag/embedding.test.js +59 -10
- package/dist/rag/embedding.test.js.map +1 -1
- package/dist/rag/index.d.ts +1 -1
- package/dist/rag/index.d.ts.map +1 -1
- package/dist/rag/index.js +1 -1
- package/dist/rag/index.js.map +1 -1
- package/dist/rag/retriever.d.ts +0 -1
- package/dist/rag/retriever.d.ts.map +1 -1
- package/dist/rag/retriever.js +75 -22
- package/dist/rag/retriever.js.map +1 -1
- package/dist/rag/vector-store.test.d.ts +1 -1
- package/dist/rag/vector-store.test.js +2 -2
- package/dist/rag/vector-store.test.js.map +1 -1
- package/dist/registry/ManagedRegistry.test.d.ts +1 -1
- package/dist/registry/ManagedRegistry.test.js +1 -1
- package/dist/registry/Registry.test.d.ts +1 -1
- package/dist/registry/Registry.test.js +1 -1
- package/dist/registry/connector/definitions.test.d.ts +1 -1
- package/dist/registry/connector/definitions.test.js +1 -1
- package/dist/registry/tool/__tests__/schema.test.d.ts +2 -0
- package/dist/registry/tool/__tests__/schema.test.d.ts.map +1 -0
- package/dist/registry/tool/__tests__/schema.test.js +176 -0
- package/dist/registry/tool/__tests__/schema.test.js.map +1 -0
- package/dist/registry/tool/__tests__/tool-name-validation.test.d.ts +2 -0
- package/dist/registry/tool/__tests__/tool-name-validation.test.d.ts.map +1 -0
- package/dist/registry/tool/__tests__/tool-name-validation.test.js +70 -0
- package/dist/registry/tool/__tests__/tool-name-validation.test.js.map +1 -0
- package/dist/registry/tool/execute.d.ts +37 -0
- package/dist/registry/tool/execute.d.ts.map +1 -1
- package/dist/registry/tool/execute.js +91 -21
- package/dist/registry/tool/execute.js.map +1 -1
- package/dist/registry/tool/schema.d.ts +26 -0
- package/dist/registry/tool/schema.d.ts.map +1 -0
- package/dist/registry/tool/schema.js +77 -0
- package/dist/registry/tool/schema.js.map +1 -0
- package/dist/registry/toolset/catalog.d.ts.map +1 -1
- package/dist/registry/toolset/catalog.js +0 -1
- package/dist/registry/toolset/catalog.js.map +1 -1
- package/dist/registry/toolset/catalog.test.js +0 -1
- package/dist/registry/toolset/catalog.test.js.map +1 -1
- package/dist/run/reporter.d.ts.map +1 -1
- package/dist/run/reporter.js +76 -0
- package/dist/run/reporter.js.map +1 -1
- package/dist/runtime/bidi/__tests__/session.test.d.ts +2 -0
- package/dist/runtime/bidi/__tests__/session.test.d.ts.map +1 -0
- package/dist/runtime/bidi/__tests__/session.test.js +201 -0
- package/dist/runtime/bidi/__tests__/session.test.js.map +1 -0
- package/dist/runtime/bidi/index.d.ts +5 -0
- package/dist/runtime/bidi/index.d.ts.map +1 -0
- package/dist/runtime/bidi/index.js +3 -0
- package/dist/runtime/bidi/index.js.map +1 -0
- package/dist/runtime/bidi/mock.d.ts +34 -0
- package/dist/runtime/bidi/mock.d.ts.map +1 -0
- package/dist/runtime/bidi/mock.js +60 -0
- package/dist/runtime/bidi/mock.js.map +1 -0
- package/dist/runtime/bidi/session.d.ts +44 -0
- package/dist/runtime/bidi/session.d.ts.map +1 -0
- package/dist/runtime/bidi/session.js +154 -0
- package/dist/runtime/bidi/session.js.map +1 -0
- package/dist/runtime/decision/parser.d.ts.map +1 -1
- package/dist/runtime/decision/parser.js +8 -1
- package/dist/runtime/decision/parser.js.map +1 -1
- package/dist/runtime/decision/parser.test.d.ts +2 -0
- package/dist/runtime/decision/parser.test.d.ts.map +1 -0
- package/dist/runtime/decision/parser.test.js +37 -0
- package/dist/runtime/decision/parser.test.js.map +1 -0
- package/dist/runtime/query/__tests__/answer-review.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/answer-review.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/answer-review.test.js +141 -0
- package/dist/runtime/query/__tests__/answer-review.test.js.map +1 -0
- package/dist/runtime/query/__tests__/capability-negotiation.test.js +24 -111
- package/dist/runtime/query/__tests__/capability-negotiation.test.js.map +1 -1
- package/dist/runtime/query/__tests__/checkpoint-store.test.js +12 -44
- package/dist/runtime/query/__tests__/checkpoint-store.test.js.map +1 -1
- package/dist/runtime/query/__tests__/checkpoint.test.js +2 -10
- package/dist/runtime/query/__tests__/checkpoint.test.js.map +1 -1
- package/dist/runtime/query/__tests__/citations.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/citations.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/citations.test.js +85 -0
- package/dist/runtime/query/__tests__/citations.test.js.map +1 -0
- package/dist/runtime/query/__tests__/content-budget.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/content-budget.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/content-budget.test.js +69 -0
- package/dist/runtime/query/__tests__/content-budget.test.js.map +1 -0
- package/dist/runtime/query/__tests__/context.test.js +3 -2
- package/dist/runtime/query/__tests__/context.test.js.map +1 -1
- package/dist/runtime/query/__tests__/deferred-tools.test.js +21 -130
- package/dist/runtime/query/__tests__/deferred-tools.test.js.map +1 -1
- package/dist/runtime/query/__tests__/delegation-resume.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/delegation-resume.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/delegation-resume.test.js +111 -0
- package/dist/runtime/query/__tests__/delegation-resume.test.js.map +1 -0
- package/dist/runtime/query/__tests__/document-attachments.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/document-attachments.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/document-attachments.test.js +92 -0
- package/dist/runtime/query/__tests__/document-attachments.test.js.map +1 -0
- package/dist/runtime/query/__tests__/durable-park-and-trace.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/durable-park-and-trace.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/durable-park-and-trace.test.js +194 -0
- package/dist/runtime/query/__tests__/durable-park-and-trace.test.js.map +1 -0
- package/dist/runtime/query/__tests__/durable-question-park.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/durable-question-park.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/durable-question-park.test.js +253 -0
- package/dist/runtime/query/__tests__/durable-question-park.test.js.map +1 -0
- package/dist/runtime/query/__tests__/durable-run-state.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/durable-run-state.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/durable-run-state.test.js +158 -0
- package/dist/runtime/query/__tests__/durable-run-state.test.js.map +1 -0
- package/dist/runtime/query/__tests__/executor-plugin-hooks.test.js +24 -20
- package/dist/runtime/query/__tests__/executor-plugin-hooks.test.js.map +1 -1
- package/dist/runtime/query/__tests__/finish-reason-vs-tool-calls.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/finish-reason-vs-tool-calls.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/finish-reason-vs-tool-calls.test.js +187 -0
- package/dist/runtime/query/__tests__/finish-reason-vs-tool-calls.test.js.map +1 -0
- package/dist/runtime/query/__tests__/grant-cannot-overrule-the-gate.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/grant-cannot-overrule-the-gate.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/grant-cannot-overrule-the-gate.test.js +95 -0
- package/dist/runtime/query/__tests__/grant-cannot-overrule-the-gate.test.js.map +1 -0
- package/dist/runtime/query/__tests__/guardrails-e2e.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/guardrails-e2e.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/guardrails-e2e.test.js +132 -0
- package/dist/runtime/query/__tests__/guardrails-e2e.test.js.map +1 -0
- package/dist/runtime/query/__tests__/guardrails.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/guardrails.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/guardrails.test.js +163 -0
- package/dist/runtime/query/__tests__/guardrails.test.js.map +1 -0
- package/dist/runtime/query/__tests__/hook-model-visibility.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/hook-model-visibility.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/hook-model-visibility.test.js +119 -0
- package/dist/runtime/query/__tests__/hook-model-visibility.test.js.map +1 -0
- package/dist/runtime/query/__tests__/long-document-flow.test.js +9 -2
- package/dist/runtime/query/__tests__/long-document-flow.test.js.map +1 -1
- package/dist/runtime/query/__tests__/memory-promotion.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/memory-promotion.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/memory-promotion.test.js +119 -0
- package/dist/runtime/query/__tests__/memory-promotion.test.js.map +1 -0
- package/dist/runtime/query/__tests__/orphan-task-turn-end.test.js +5 -27
- package/dist/runtime/query/__tests__/orphan-task-turn-end.test.js.map +1 -1
- package/dist/runtime/query/__tests__/overflow-relief.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/overflow-relief.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/overflow-relief.test.js +119 -0
- package/dist/runtime/query/__tests__/overflow-relief.test.js.map +1 -0
- package/dist/runtime/query/__tests__/partial-batch-resume.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/partial-batch-resume.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/partial-batch-resume.test.js +205 -0
- package/dist/runtime/query/__tests__/partial-batch-resume.test.js.map +1 -0
- package/dist/runtime/query/__tests__/plugin-hooks.test.js +0 -1
- package/dist/runtime/query/__tests__/plugin-hooks.test.js.map +1 -1
- package/dist/runtime/query/__tests__/prepare-step-chain.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/prepare-step-chain.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/prepare-step-chain.test.js +104 -0
- package/dist/runtime/query/__tests__/prepare-step-chain.test.js.map +1 -0
- package/dist/runtime/query/__tests__/prepare-step.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/prepare-step.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/prepare-step.test.js +194 -0
- package/dist/runtime/query/__tests__/prepare-step.test.js.map +1 -0
- package/dist/runtime/query/__tests__/probe-veto-is-error.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/probe-veto-is-error.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/probe-veto-is-error.test.js +111 -0
- package/dist/runtime/query/__tests__/probe-veto-is-error.test.js.map +1 -0
- package/dist/runtime/query/__tests__/prompt.test.js +1 -1
- package/dist/runtime/query/__tests__/prompt.test.js.map +1 -1
- package/dist/runtime/query/__tests__/provider-fault-pauses.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/provider-fault-pauses.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/provider-fault-pauses.test.js +86 -0
- package/dist/runtime/query/__tests__/provider-fault-pauses.test.js.map +1 -0
- package/dist/runtime/query/__tests__/reasoning-stream.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/reasoning-stream.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/reasoning-stream.test.js +146 -0
- package/dist/runtime/query/__tests__/reasoning-stream.test.js.map +1 -0
- package/dist/runtime/query/__tests__/recoverable-settle.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/recoverable-settle.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/recoverable-settle.test.js +119 -0
- package/dist/runtime/query/__tests__/recoverable-settle.test.js.map +1 -0
- package/dist/runtime/query/__tests__/resume-budget.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/resume-budget.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/resume-budget.test.js +130 -0
- package/dist/runtime/query/__tests__/resume-budget.test.js.map +1 -0
- package/dist/runtime/query/__tests__/resume-pending.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/resume-pending.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/resume-pending.test.js +226 -0
- package/dist/runtime/query/__tests__/resume-pending.test.js.map +1 -0
- package/dist/runtime/query/__tests__/rich-content-budget-wiring.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/rich-content-budget-wiring.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/rich-content-budget-wiring.test.js +89 -0
- package/dist/runtime/query/__tests__/rich-content-budget-wiring.test.js.map +1 -0
- package/dist/runtime/query/__tests__/stop-condition.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/stop-condition.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/stop-condition.test.js +205 -0
- package/dist/runtime/query/__tests__/stop-condition.test.js.map +1 -0
- package/dist/runtime/query/__tests__/structured-output.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/structured-output.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/structured-output.test.js +219 -0
- package/dist/runtime/query/__tests__/structured-output.test.js.map +1 -0
- package/dist/runtime/query/__tests__/structured-run-failure.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/structured-run-failure.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/structured-run-failure.test.js +117 -0
- package/dist/runtime/query/__tests__/structured-run-failure.test.js.map +1 -0
- package/dist/runtime/query/__tests__/terminal-tool.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/terminal-tool.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/terminal-tool.test.js +119 -0
- package/dist/runtime/query/__tests__/terminal-tool.test.js.map +1 -0
- package/dist/runtime/query/__tests__/tool-grants.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/tool-grants.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/tool-grants.test.js +85 -0
- package/dist/runtime/query/__tests__/tool-grants.test.js.map +1 -0
- package/dist/runtime/query/__tests__/tool-output-budget.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/tool-output-budget.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/tool-output-budget.test.js +0 -0
- package/dist/runtime/query/__tests__/tool-output-budget.test.js.map +1 -0
- package/dist/runtime/query/__tests__/tool-pause.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/tool-pause.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/tool-pause.test.js +165 -0
- package/dist/runtime/query/__tests__/tool-pause.test.js.map +1 -0
- package/dist/runtime/query/__tests__/tool-progress.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/tool-progress.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/tool-progress.test.js +122 -0
- package/dist/runtime/query/__tests__/tool-progress.test.js.map +1 -0
- package/dist/runtime/query/__tests__/tool-repair-retry.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/tool-repair-retry.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/tool-repair-retry.test.js +196 -0
- package/dist/runtime/query/__tests__/tool-repair-retry.test.js.map +1 -0
- package/dist/runtime/query/__tests__/tool-result-wire.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/tool-result-wire.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/tool-result-wire.test.js +106 -0
- package/dist/runtime/query/__tests__/tool-result-wire.test.js.map +1 -0
- package/dist/runtime/query/__tests__/tool-timeout.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/tool-timeout.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/tool-timeout.test.js +184 -0
- package/dist/runtime/query/__tests__/tool-timeout.test.js.map +1 -0
- package/dist/runtime/query/__tests__/trace-parenting.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/trace-parenting.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/trace-parenting.test.js +111 -0
- package/dist/runtime/query/__tests__/trace-parenting.test.js.map +1 -0
- package/dist/runtime/query/__tests__/wiring-seams.test.d.ts +2 -0
- package/dist/runtime/query/__tests__/wiring-seams.test.d.ts.map +1 -0
- package/dist/runtime/query/__tests__/wiring-seams.test.js +237 -0
- package/dist/runtime/query/__tests__/wiring-seams.test.js.map +1 -0
- package/dist/runtime/query/checkpoint.d.ts +124 -6
- package/dist/runtime/query/checkpoint.d.ts.map +1 -1
- package/dist/runtime/query/checkpoint.js +207 -6
- package/dist/runtime/query/checkpoint.js.map +1 -1
- package/dist/runtime/query/events.js +2 -2
- package/dist/runtime/query/events.js.map +1 -1
- package/dist/runtime/query/executor.d.ts +224 -7
- package/dist/runtime/query/executor.d.ts.map +1 -1
- package/dist/runtime/query/executor.js +662 -35
- package/dist/runtime/query/executor.js.map +1 -1
- package/dist/runtime/query/guard.d.ts +18 -0
- package/dist/runtime/query/guard.d.ts.map +1 -1
- package/dist/runtime/query/guard.js +16 -1
- package/dist/runtime/query/guard.js.map +1 -1
- package/dist/runtime/query/guardrail-presets.d.ts +36 -0
- package/dist/runtime/query/guardrail-presets.d.ts.map +1 -0
- package/dist/runtime/query/guardrail-presets.js +96 -0
- package/dist/runtime/query/guardrail-presets.js.map +1 -0
- package/dist/runtime/query/guardrails.d.ts +35 -0
- package/dist/runtime/query/guardrails.d.ts.map +1 -0
- package/dist/runtime/query/guardrails.js +102 -0
- package/dist/runtime/query/guardrails.js.map +1 -0
- package/dist/runtime/query/index.d.ts +213 -2
- package/dist/runtime/query/index.d.ts.map +1 -1
- package/dist/runtime/query/index.js +458 -10
- package/dist/runtime/query/index.js.map +1 -1
- package/dist/runtime/query/iteration/__tests__/cancel-settles-turn.test.d.ts +2 -0
- package/dist/runtime/query/iteration/__tests__/cancel-settles-turn.test.d.ts.map +1 -0
- package/dist/runtime/query/iteration/__tests__/cancel-settles-turn.test.js +131 -0
- package/dist/runtime/query/iteration/__tests__/cancel-settles-turn.test.js.map +1 -0
- package/dist/runtime/query/iteration/index.d.ts +76 -1
- package/dist/runtime/query/iteration/index.d.ts.map +1 -1
- package/dist/runtime/query/iteration/index.js +577 -34
- package/dist/runtime/query/iteration/index.js.map +1 -1
- package/dist/runtime/query/iteration/phases/advisory.d.ts.map +1 -1
- package/dist/runtime/query/iteration/phases/advisory.js +14 -2
- package/dist/runtime/query/iteration/phases/advisory.js.map +1 -1
- package/dist/runtime/query/iteration/phases/advisory.test.d.ts +1 -1
- package/dist/runtime/query/iteration/phases/advisory.test.js +4 -1
- package/dist/runtime/query/iteration/phases/advisory.test.js.map +1 -1
- package/dist/runtime/query/iteration/phases/checkpoint.js +2 -2
- package/dist/runtime/query/iteration/phases/checkpoint.js.map +1 -1
- package/dist/runtime/query/iteration/phases/compaction-safe-cut.test.js +1 -1
- package/dist/runtime/query/iteration/phases/compaction-safe-cut.test.js.map +1 -1
- package/dist/runtime/query/iteration/phases/compaction.d.ts +22 -1
- package/dist/runtime/query/iteration/phases/compaction.d.ts.map +1 -1
- package/dist/runtime/query/iteration/phases/compaction.js +297 -64
- package/dist/runtime/query/iteration/phases/compaction.js.map +1 -1
- package/dist/runtime/query/iteration/phases/context-measurement.test.d.ts +21 -0
- package/dist/runtime/query/iteration/phases/context-measurement.test.d.ts.map +1 -0
- package/dist/runtime/query/iteration/phases/context-measurement.test.js +161 -0
- package/dist/runtime/query/iteration/phases/context-measurement.test.js.map +1 -0
- package/dist/runtime/query/iteration/phases/context.d.ts +62 -2
- package/dist/runtime/query/iteration/phases/context.d.ts.map +1 -1
- package/dist/runtime/query/iteration/phases/context.js +87 -0
- package/dist/runtime/query/iteration/phases/context.js.map +1 -1
- package/dist/runtime/query/iteration/phases/hitl-answer-question.test.js +3 -1
- package/dist/runtime/query/iteration/phases/hitl-answer-question.test.js.map +1 -1
- package/dist/runtime/query/iteration/phases/overflow-escalation.test.d.ts +15 -0
- package/dist/runtime/query/iteration/phases/overflow-escalation.test.d.ts.map +1 -0
- package/dist/runtime/query/iteration/phases/overflow-escalation.test.js +130 -0
- package/dist/runtime/query/iteration/phases/overflow-escalation.test.js.map +1 -0
- package/dist/runtime/query/iteration/phases/plan.d.ts.map +1 -1
- package/dist/runtime/query/iteration/phases/plan.js +8 -2
- package/dist/runtime/query/iteration/phases/plan.js.map +1 -1
- package/dist/runtime/query/iteration/phases/retained-messages.test.d.ts +2 -0
- package/dist/runtime/query/iteration/phases/retained-messages.test.d.ts.map +1 -0
- package/dist/runtime/query/iteration/phases/retained-messages.test.js +180 -0
- package/dist/runtime/query/iteration/phases/retained-messages.test.js.map +1 -0
- package/dist/runtime/query/iteration/phases/tool-review-denials.test.d.ts +2 -0
- package/dist/runtime/query/iteration/phases/tool-review-denials.test.d.ts.map +1 -0
- package/dist/runtime/query/iteration/phases/tool-review-denials.test.js +224 -0
- package/dist/runtime/query/iteration/phases/tool-review-denials.test.js.map +1 -0
- package/dist/runtime/query/iteration/phases/tool-review.d.ts +29 -1
- package/dist/runtime/query/iteration/phases/tool-review.d.ts.map +1 -1
- package/dist/runtime/query/iteration/phases/tool-review.js +114 -44
- package/dist/runtime/query/iteration/phases/tool-review.js.map +1 -1
- package/dist/runtime/query/iteration/phases/working-memory-compaction.test.js +20 -9
- package/dist/runtime/query/iteration/phases/working-memory-compaction.test.js.map +1 -1
- package/dist/runtime/query/iteration/stream-turn.d.ts +2 -26
- package/dist/runtime/query/iteration/stream-turn.d.ts.map +1 -1
- package/dist/runtime/query/iteration/stream-turn.js +254 -16
- package/dist/runtime/query/iteration/stream-turn.js.map +1 -1
- package/dist/runtime/query/plugin-hooks.d.ts.map +1 -1
- package/dist/runtime/query/plugin-hooks.js +6 -2
- package/dist/runtime/query/plugin-hooks.js.map +1 -1
- package/dist/runtime/query/prompt.d.ts.map +1 -1
- package/dist/runtime/query/prompt.js +25 -1
- package/dist/runtime/query/prompt.js.map +1 -1
- package/dist/runtime/query/question-park.d.ts +85 -0
- package/dist/runtime/query/question-park.d.ts.map +1 -0
- package/dist/runtime/query/question-park.js +71 -0
- package/dist/runtime/query/question-park.js.map +1 -0
- package/dist/runtime/query/result.d.ts +9 -0
- package/dist/runtime/query/result.d.ts.map +1 -1
- package/dist/runtime/query/result.js +49 -0
- package/dist/runtime/query/result.js.map +1 -1
- package/dist/runtime/query/resume-pending.d.ts +106 -0
- package/dist/runtime/query/resume-pending.d.ts.map +1 -0
- package/dist/runtime/query/resume-pending.js +312 -0
- package/dist/runtime/query/resume-pending.js.map +1 -0
- package/dist/runtime/query/run-state.d.ts +52 -0
- package/dist/runtime/query/run-state.d.ts.map +1 -0
- package/dist/runtime/query/run-state.js +87 -0
- package/dist/runtime/query/run-state.js.map +1 -0
- package/dist/runtime/query/tool-grants.d.ts +44 -0
- package/dist/runtime/query/tool-grants.d.ts.map +1 -0
- package/dist/runtime/query/tool-grants.js +63 -0
- package/dist/runtime/query/tool-grants.js.map +1 -0
- package/dist/runtime/query/tool-output-budget.d.ts +83 -0
- package/dist/runtime/query/tool-output-budget.d.ts.map +1 -0
- package/dist/runtime/query/tool-output-budget.js +147 -0
- package/dist/runtime/query/tool-output-budget.js.map +1 -0
- package/dist/runtime/query/tool-pause.d.ts +41 -0
- package/dist/runtime/query/tool-pause.d.ts.map +1 -0
- package/dist/runtime/query/tool-pause.js +94 -0
- package/dist/runtime/query/tool-pause.js.map +1 -0
- package/dist/runtime/query/tooling.d.ts +10 -1
- package/dist/runtime/query/tooling.d.ts.map +1 -1
- package/dist/runtime/query/tooling.js +13 -0
- package/dist/runtime/query/tooling.js.map +1 -1
- package/dist/sandbox/__tests__/isolation.test.d.ts +2 -0
- package/dist/sandbox/__tests__/isolation.test.d.ts.map +1 -0
- package/dist/sandbox/__tests__/isolation.test.js +139 -0
- package/dist/sandbox/__tests__/isolation.test.js.map +1 -0
- package/dist/sandbox/__tests__/output-truncation.test.d.ts +2 -0
- package/dist/sandbox/__tests__/output-truncation.test.d.ts.map +1 -0
- package/dist/sandbox/__tests__/output-truncation.test.js +88 -0
- package/dist/sandbox/__tests__/output-truncation.test.js.map +1 -0
- package/dist/sandbox/__tests__/resource-limits.test.d.ts +2 -0
- package/dist/sandbox/__tests__/resource-limits.test.d.ts.map +1 -0
- package/dist/sandbox/__tests__/resource-limits.test.js +86 -0
- package/dist/sandbox/__tests__/resource-limits.test.js.map +1 -0
- package/dist/sandbox/factory.js +1 -1
- package/dist/sandbox/factory.js.map +1 -1
- package/dist/sandbox/index.d.ts +2 -0
- package/dist/sandbox/index.d.ts.map +1 -1
- package/dist/sandbox/index.js +1 -0
- package/dist/sandbox/index.js.map +1 -1
- package/dist/sandbox/isolation.d.ts +15 -0
- package/dist/sandbox/isolation.d.ts.map +1 -0
- package/dist/sandbox/isolation.js +53 -0
- package/dist/sandbox/isolation.js.map +1 -0
- package/dist/sandbox/provider/local.d.ts +51 -2
- package/dist/sandbox/provider/local.d.ts.map +1 -1
- package/dist/sandbox/provider/local.js +139 -69
- package/dist/sandbox/provider/local.js.map +1 -1
- package/dist/session/__tests__/integration/spawn-rollback.test.d.ts +1 -1
- package/dist/session/__tests__/integration/spawn-rollback.test.js +1 -1
- package/dist/session/handoff/broadcast.js +3 -3
- package/dist/session/handoff/broadcast.js.map +1 -1
- package/dist/session/migration/__tests__/filesystem.test.js +5 -1
- package/dist/session/migration/__tests__/filesystem.test.js.map +1 -1
- package/dist/session/migration/filesystem.d.ts.map +1 -1
- package/dist/session/migration/filesystem.js +7 -13
- package/dist/session/migration/filesystem.js.map +1 -1
- package/dist/session/migration/marker.d.ts.map +1 -1
- package/dist/session/migration/marker.js +5 -10
- package/dist/session/migration/marker.js.map +1 -1
- package/dist/session/retention/disk-backend.d.ts.map +1 -1
- package/dist/session/retention/disk-backend.js +5 -12
- package/dist/session/retention/disk-backend.js.map +1 -1
- package/dist/session/workspace/__tests__/git-worktree.test.js +6 -2
- package/dist/session/workspace/__tests__/git-worktree.test.js.map +1 -1
- package/dist/session/workspace/__tests__/shared-run.test.js +15 -0
- package/dist/session/workspace/__tests__/shared-run.test.js.map +1 -1
- package/dist/session/workspace/shared-run.js +8 -1
- package/dist/session/workspace/shared-run.js.map +1 -1
- package/dist/skills/__tests__/frontmatter.test.d.ts +2 -0
- package/dist/skills/__tests__/frontmatter.test.d.ts.map +1 -0
- package/dist/skills/__tests__/frontmatter.test.js +106 -0
- package/dist/skills/__tests__/frontmatter.test.js.map +1 -0
- package/dist/skills/loader.d.ts.map +1 -1
- package/dist/skills/loader.js +42 -3
- package/dist/skills/loader.js.map +1 -1
- package/dist/store/__tests__/checkpoint-integrity.test.d.ts +2 -0
- package/dist/store/__tests__/checkpoint-integrity.test.d.ts.map +1 -0
- package/dist/store/__tests__/checkpoint-integrity.test.js +112 -0
- package/dist/store/__tests__/checkpoint-integrity.test.js.map +1 -0
- package/dist/store/__tests__/checkpoint-versioning.test.d.ts +2 -0
- package/dist/store/__tests__/checkpoint-versioning.test.d.ts.map +1 -0
- package/dist/store/__tests__/checkpoint-versioning.test.js +151 -0
- package/dist/store/__tests__/checkpoint-versioning.test.js.map +1 -0
- package/dist/store/__tests__/schema-adoption.test.d.ts +2 -0
- package/dist/store/__tests__/schema-adoption.test.d.ts.map +1 -0
- package/dist/store/__tests__/schema-adoption.test.js +90 -0
- package/dist/store/__tests__/schema-adoption.test.js.map +1 -0
- package/dist/store/__tests__/schema.test.d.ts +2 -0
- package/dist/store/__tests__/schema.test.d.ts.map +1 -0
- package/dist/store/__tests__/schema.test.js +93 -0
- package/dist/store/__tests__/schema.test.js.map +1 -0
- package/dist/store/memory/disk.d.ts.map +1 -1
- package/dist/store/memory/disk.js +16 -16
- package/dist/store/memory/disk.js.map +1 -1
- package/dist/store/run/disk.d.ts +23 -0
- package/dist/store/run/disk.d.ts.map +1 -1
- package/dist/store/run/disk.js +148 -23
- package/dist/store/run/disk.js.map +1 -1
- package/dist/store/schema.d.ts +73 -0
- package/dist/store/schema.d.ts.map +1 -0
- package/dist/store/schema.js +103 -0
- package/dist/store/schema.js.map +1 -0
- package/dist/store/session/disk.d.ts.map +1 -1
- package/dist/store/session/disk.js +25 -13
- package/dist/store/session/disk.js.map +1 -1
- package/dist/store/task/__tests__/lookup-across-runs.test.d.ts +2 -0
- package/dist/store/task/__tests__/lookup-across-runs.test.d.ts.map +1 -0
- package/dist/store/task/__tests__/lookup-across-runs.test.js +60 -0
- package/dist/store/task/__tests__/lookup-across-runs.test.js.map +1 -0
- package/dist/store/task/disk.d.ts +18 -0
- package/dist/store/task/disk.d.ts.map +1 -1
- package/dist/store/task/disk.js +61 -16
- package/dist/store/task/disk.js.map +1 -1
- package/dist/store/thread/disk.d.ts.map +1 -1
- package/dist/store/thread/disk.js +20 -11
- package/dist/store/thread/disk.js.map +1 -1
- package/dist/streaming/coalesce.d.ts +1 -1
- package/dist/streaming/coalesce.js +1 -1
- package/dist/telemetry/__tests__/latency-instruments.test.d.ts +2 -0
- package/dist/telemetry/__tests__/latency-instruments.test.d.ts.map +1 -0
- package/dist/telemetry/__tests__/latency-instruments.test.js +82 -0
- package/dist/telemetry/__tests__/latency-instruments.test.js.map +1 -0
- package/dist/telemetry/__tests__/metrics.test.d.ts +2 -0
- package/dist/telemetry/__tests__/metrics.test.d.ts.map +1 -0
- package/dist/telemetry/__tests__/metrics.test.js +112 -0
- package/dist/telemetry/__tests__/metrics.test.js.map +1 -0
- package/dist/telemetry/__tests__/span-hierarchy.test.d.ts +2 -0
- package/dist/telemetry/__tests__/span-hierarchy.test.d.ts.map +1 -0
- package/dist/telemetry/__tests__/span-hierarchy.test.js +84 -0
- package/dist/telemetry/__tests__/span-hierarchy.test.js.map +1 -0
- package/dist/telemetry/attributes.d.ts +44 -0
- package/dist/telemetry/attributes.d.ts.map +1 -1
- package/dist/telemetry/attributes.js +80 -0
- package/dist/telemetry/attributes.js.map +1 -1
- package/dist/telemetry/metrics.d.ts +41 -0
- package/dist/telemetry/metrics.d.ts.map +1 -0
- package/dist/telemetry/metrics.js +160 -0
- package/dist/telemetry/metrics.js.map +1 -0
- package/dist/test-setup.d.ts +3 -3
- package/dist/test-setup.js +3 -3
- package/dist/test-support/paths.d.ts +14 -0
- package/dist/test-support/paths.d.ts.map +1 -0
- package/dist/test-support/paths.js +16 -0
- package/dist/test-support/paths.js.map +1 -0
- package/dist/tools/__tests__/path-containment.test.d.ts +2 -0
- package/dist/tools/__tests__/path-containment.test.d.ts.map +1 -0
- package/dist/tools/__tests__/path-containment.test.js +133 -0
- package/dist/tools/__tests__/path-containment.test.js.map +1 -0
- package/dist/tools/__tests__/posix-path.test.d.ts +2 -0
- package/dist/tools/__tests__/posix-path.test.d.ts.map +1 -0
- package/dist/tools/__tests__/posix-path.test.js +67 -0
- package/dist/tools/__tests__/posix-path.test.js.map +1 -0
- package/dist/tools/__tests__/sandboxed-search.test.d.ts +2 -0
- package/dist/tools/__tests__/sandboxed-search.test.d.ts.map +1 -0
- package/dist/tools/__tests__/sandboxed-search.test.js +114 -0
- package/dist/tools/__tests__/sandboxed-search.test.js.map +1 -0
- package/dist/tools/builtins/__tests__/computer-use.test.js +12 -2
- package/dist/tools/builtins/__tests__/computer-use.test.js.map +1 -1
- package/dist/tools/builtins/__tests__/edit.test.js +10 -216
- package/dist/tools/builtins/__tests__/edit.test.js.map +1 -1
- package/dist/tools/builtins/__tests__/payload-budget.test.js +16 -30
- package/dist/tools/builtins/__tests__/payload-budget.test.js.map +1 -1
- package/dist/tools/builtins/__tests__/read-file.test.js +23 -1
- package/dist/tools/builtins/__tests__/read-file.test.js.map +1 -1
- package/dist/tools/builtins/__tests__/write-file.test.js +13 -37
- package/dist/tools/builtins/__tests__/write-file.test.js.map +1 -1
- package/dist/tools/builtins/bash.d.ts.map +1 -1
- package/dist/tools/builtins/bash.js +33 -4
- package/dist/tools/builtins/bash.js.map +1 -1
- package/dist/tools/builtins/computer-use.d.ts +22 -22
- package/dist/tools/builtins/computer-use.d.ts.map +1 -1
- package/dist/tools/builtins/computer-use.js +7 -1
- package/dist/tools/builtins/computer-use.js.map +1 -1
- package/dist/tools/builtins/edit.d.ts +5 -2
- package/dist/tools/builtins/edit.d.ts.map +1 -1
- package/dist/tools/builtins/edit.js +125 -105
- package/dist/tools/builtins/edit.js.map +1 -1
- package/dist/tools/builtins/glob.d.ts +1 -1
- package/dist/tools/builtins/glob.d.ts.map +1 -1
- package/dist/tools/builtins/glob.js +63 -6
- package/dist/tools/builtins/glob.js.map +1 -1
- package/dist/tools/builtins/grep.d.ts.map +1 -1
- package/dist/tools/builtins/grep.js +61 -8
- package/dist/tools/builtins/grep.js.map +1 -1
- package/dist/tools/builtins/index.js +6 -6
- package/dist/tools/builtins/index.js.map +1 -1
- package/dist/tools/builtins/ls.d.ts.map +1 -1
- package/dist/tools/builtins/ls.js +4 -2
- package/dist/tools/builtins/ls.js.map +1 -1
- package/dist/tools/builtins/read-file.d.ts +1 -1
- package/dist/tools/builtins/read-file.d.ts.map +1 -1
- package/dist/tools/builtins/read-file.js +38 -4
- package/dist/tools/builtins/read-file.js.map +1 -1
- package/dist/tools/builtins/write-file.d.ts +2 -1
- package/dist/tools/builtins/write-file.d.ts.map +1 -1
- package/dist/tools/builtins/write-file.js +40 -66
- package/dist/tools/builtins/write-file.js.map +1 -1
- package/dist/tools/coordinator/__tests__/agent.test.d.ts +1 -1
- package/dist/tools/coordinator/__tests__/agent.test.js +2 -2
- package/dist/tools/coordinator/__tests__/agent.test.js.map +1 -1
- package/dist/tools/coordinator/agent.d.ts +20 -9
- package/dist/tools/coordinator/agent.d.ts.map +1 -1
- package/dist/tools/coordinator/agent.js +2 -1
- package/dist/tools/coordinator/agent.js.map +1 -1
- package/dist/tools/coordinator/index.d.ts +20 -0
- package/dist/tools/coordinator/index.d.ts.map +1 -1
- package/dist/tools/coordinator/index.js +80 -32
- package/dist/tools/coordinator/index.js.map +1 -1
- package/dist/tools/defineTool.d.ts +16 -0
- package/dist/tools/defineTool.d.ts.map +1 -1
- package/dist/tools/defineTool.js +4 -0
- package/dist/tools/defineTool.js.map +1 -1
- package/dist/tools/glob-match.d.ts +18 -0
- package/dist/tools/glob-match.d.ts.map +1 -0
- package/dist/tools/glob-match.js +50 -0
- package/dist/tools/glob-match.js.map +1 -0
- package/dist/tools/paths.d.ts +20 -0
- package/dist/tools/paths.d.ts.map +1 -0
- package/dist/tools/paths.js +41 -0
- package/dist/tools/paths.js.map +1 -0
- package/dist/tools/posix-path.d.ts +24 -0
- package/dist/tools/posix-path.d.ts.map +1 -0
- package/dist/tools/posix-path.js +67 -0
- package/dist/tools/posix-path.js.map +1 -0
- package/dist/types/advisory/config.d.ts +21 -3
- package/dist/types/advisory/config.d.ts.map +1 -1
- package/dist/types/advisory/result.d.ts +12 -10
- package/dist/types/advisory/result.d.ts.map +1 -1
- package/dist/types/agent/base.d.ts +20 -2
- package/dist/types/agent/base.d.ts.map +1 -1
- package/dist/types/agent/factory.d.ts +15 -9
- package/dist/types/agent/factory.d.ts.map +1 -1
- package/dist/types/agent/gateway.d.ts +18 -0
- package/dist/types/agent/gateway.d.ts.map +1 -1
- package/dist/types/agent/manager.d.ts +19 -0
- package/dist/types/agent/manager.d.ts.map +1 -1
- package/dist/types/agent/reactive.d.ts +48 -0
- package/dist/types/agent/reactive.d.ts.map +1 -1
- package/dist/types/agent/router.d.ts +11 -0
- package/dist/types/agent/router.d.ts.map +1 -1
- package/dist/types/agent/supervisor.d.ts +2 -2
- package/dist/types/agent/task.d.ts +24 -0
- package/dist/types/agent/task.d.ts.map +1 -1
- package/dist/types/bidi/index.d.ts +140 -0
- package/dist/types/bidi/index.d.ts.map +1 -0
- package/dist/types/bidi/index.js +2 -0
- package/dist/types/bidi/index.js.map +1 -0
- package/dist/types/common/__tests__/numeric-correctness.test.d.ts +2 -0
- package/dist/types/common/__tests__/numeric-correctness.test.d.ts.map +1 -0
- package/dist/types/common/__tests__/numeric-correctness.test.js +108 -0
- package/dist/types/common/__tests__/numeric-correctness.test.js.map +1 -0
- package/dist/types/common/index.d.ts +1 -1
- package/dist/types/common/index.d.ts.map +1 -1
- package/dist/types/common/index.js +15 -4
- package/dist/types/common/index.js.map +1 -1
- package/dist/types/connector/mcp.d.ts +29 -0
- package/dist/types/connector/mcp.d.ts.map +1 -1
- package/dist/types/errors/__tests__/catalog.test.d.ts +2 -0
- package/dist/types/errors/__tests__/catalog.test.d.ts.map +1 -0
- package/dist/types/errors/__tests__/catalog.test.js +106 -0
- package/dist/types/errors/__tests__/catalog.test.js.map +1 -0
- package/dist/types/errors/__tests__/errors.test.d.ts +2 -0
- package/dist/types/errors/__tests__/errors.test.d.ts.map +1 -0
- package/dist/types/errors/__tests__/errors.test.js +104 -0
- package/dist/types/errors/__tests__/errors.test.js.map +1 -0
- package/dist/types/errors/catalog.d.ts +83 -0
- package/dist/types/errors/catalog.d.ts.map +1 -0
- package/dist/types/errors/catalog.js +148 -0
- package/dist/types/errors/catalog.js.map +1 -0
- package/dist/types/errors/index.d.ts +73 -0
- package/dist/types/errors/index.d.ts.map +1 -0
- package/dist/types/errors/index.js +70 -0
- package/dist/types/errors/index.js.map +1 -0
- package/dist/types/guardrail/index.d.ts +57 -0
- package/dist/types/guardrail/index.d.ts.map +1 -0
- package/dist/types/guardrail/index.js +2 -0
- package/dist/types/guardrail/index.js.map +1 -0
- package/dist/types/hitl/index.d.ts +97 -14
- package/dist/types/hitl/index.d.ts.map +1 -1
- package/dist/types/hitl/index.js.map +1 -1
- package/dist/types/ids/index.d.ts +2 -2
- package/dist/types/message/content.d.ts +22 -0
- package/dist/types/message/content.d.ts.map +1 -0
- package/dist/types/message/content.js +49 -0
- package/dist/types/message/content.js.map +1 -0
- package/dist/types/message/index.d.ts +193 -8
- package/dist/types/message/index.d.ts.map +1 -1
- package/dist/types/message/index.js +19 -4
- package/dist/types/message/index.js.map +1 -1
- package/dist/types/plugin/index.d.ts +78 -3
- package/dist/types/plugin/index.d.ts.map +1 -1
- package/dist/types/plugin/index.js +0 -1
- package/dist/types/plugin/index.js.map +1 -1
- package/dist/types/provider/__tests__/cause-chain.test.d.ts +2 -0
- package/dist/types/provider/__tests__/cause-chain.test.d.ts.map +1 -0
- package/dist/types/provider/__tests__/cause-chain.test.js +108 -0
- package/dist/types/provider/__tests__/cause-chain.test.js.map +1 -0
- package/dist/types/provider/__tests__/declared-retryable.test.d.ts +2 -0
- package/dist/types/provider/__tests__/declared-retryable.test.d.ts.map +1 -0
- package/dist/types/provider/__tests__/declared-retryable.test.js +70 -0
- package/dist/types/provider/__tests__/declared-retryable.test.js.map +1 -0
- package/dist/types/provider/__tests__/overflow-classification.test.d.ts +2 -0
- package/dist/types/provider/__tests__/overflow-classification.test.d.ts.map +1 -0
- package/dist/types/provider/__tests__/overflow-classification.test.js +80 -0
- package/dist/types/provider/__tests__/overflow-classification.test.js.map +1 -0
- package/dist/types/provider/chat.d.ts +27 -1
- package/dist/types/provider/chat.d.ts.map +1 -1
- package/dist/types/provider/config.d.ts +83 -2
- package/dist/types/provider/config.d.ts.map +1 -1
- package/dist/types/provider/errors.d.ts +102 -0
- package/dist/types/provider/errors.d.ts.map +1 -0
- package/dist/types/provider/errors.js +393 -0
- package/dist/types/provider/errors.js.map +1 -0
- package/dist/types/provider/index.d.ts +3 -2
- package/dist/types/provider/index.d.ts.map +1 -1
- package/dist/types/provider/stream.d.ts +59 -3
- package/dist/types/provider/stream.d.ts.map +1 -1
- package/dist/types/rag/embedding.d.ts +11 -2
- package/dist/types/rag/embedding.d.ts.map +1 -1
- package/dist/types/run/__tests__/step.test.d.ts +2 -0
- package/dist/types/run/__tests__/step.test.d.ts.map +1 -0
- package/dist/types/run/__tests__/step.test.js +107 -0
- package/dist/types/run/__tests__/step.test.js.map +1 -0
- package/dist/types/run/answer-review.d.ts +46 -0
- package/dist/types/run/answer-review.d.ts.map +1 -0
- package/dist/types/run/answer-review.js +2 -0
- package/dist/types/run/answer-review.js.map +1 -0
- package/dist/types/run/config.d.ts +33 -0
- package/dist/types/run/config.d.ts.map +1 -1
- package/dist/types/run/derive-status.d.ts +29 -0
- package/dist/types/run/derive-status.d.ts.map +1 -0
- package/dist/types/run/derive-status.js +49 -0
- package/dist/types/run/derive-status.js.map +1 -0
- package/dist/types/run/entity.d.ts +16 -0
- package/dist/types/run/entity.d.ts.map +1 -1
- package/dist/types/run/events.d.ts +228 -4
- package/dist/types/run/events.d.ts.map +1 -1
- package/dist/types/run/events.js +7 -0
- package/dist/types/run/events.js.map +1 -1
- package/dist/types/run/index.d.ts +5 -0
- package/dist/types/run/index.d.ts.map +1 -1
- package/dist/types/run/index.js +3 -0
- package/dist/types/run/index.js.map +1 -1
- package/dist/types/run/memory-promotion.d.ts +90 -0
- package/dist/types/run/memory-promotion.d.ts.map +1 -0
- package/dist/types/run/memory-promotion.js +38 -0
- package/dist/types/run/memory-promotion.js.map +1 -0
- package/dist/types/run/prepare-step.d.ts +100 -0
- package/dist/types/run/prepare-step.d.ts.map +1 -0
- package/dist/types/run/prepare-step.js +2 -0
- package/dist/types/run/prepare-step.js.map +1 -0
- package/dist/types/run/state.d.ts +85 -0
- package/dist/types/run/state.d.ts.map +1 -0
- package/dist/types/run/state.js +33 -0
- package/dist/types/run/state.js.map +1 -0
- package/dist/types/run/step.d.ts +77 -0
- package/dist/types/run/step.d.ts.map +1 -0
- package/dist/types/run/step.js +26 -0
- package/dist/types/run/step.js.map +1 -0
- package/dist/types/run/stop-reason.d.ts +31 -3
- package/dist/types/run/stop-reason.d.ts.map +1 -1
- package/dist/types/sandbox/index.d.ts +82 -10
- package/dist/types/sandbox/index.d.ts.map +1 -1
- package/dist/types/sandbox/index.js +13 -1
- package/dist/types/sandbox/index.js.map +1 -1
- package/dist/types/structured-output/index.d.ts +6 -5
- package/dist/types/structured-output/index.d.ts.map +1 -1
- package/dist/types/thread/entity.d.ts +2 -3
- package/dist/types/thread/entity.d.ts.map +1 -1
- package/dist/types/tool/index.d.ts +193 -1
- package/dist/types/tool/index.d.ts.map +1 -1
- package/dist/types/tool/index.js +1 -1
- package/dist/types/tool/index.js.map +1 -1
- package/dist/types/tool/repair.d.ts +49 -0
- package/dist/types/tool/repair.d.ts.map +1 -0
- package/dist/types/tool/repair.js +2 -0
- package/dist/types/tool/repair.js.map +1 -0
- package/dist/types/toolset/index.d.ts +2 -5
- package/dist/types/toolset/index.d.ts.map +1 -1
- package/dist/utils/__tests__/atomic-write.test.d.ts +2 -0
- package/dist/utils/__tests__/atomic-write.test.d.ts.map +1 -0
- package/dist/utils/__tests__/atomic-write.test.js +91 -0
- package/dist/utils/__tests__/atomic-write.test.js.map +1 -0
- package/dist/utils/__tests__/shell-compress.test.d.ts +2 -0
- package/dist/utils/__tests__/shell-compress.test.d.ts.map +1 -0
- package/dist/utils/__tests__/shell-compress.test.js +41 -0
- package/dist/utils/__tests__/shell-compress.test.js.map +1 -0
- package/dist/utils/atomic-write.d.ts +10 -0
- package/dist/utils/atomic-write.d.ts.map +1 -0
- package/dist/utils/atomic-write.js +83 -0
- package/dist/utils/atomic-write.js.map +1 -0
- package/dist/utils/shell-compress.d.ts.map +1 -1
- package/dist/utils/shell-compress.js +12 -1
- package/dist/utils/shell-compress.js.map +1 -1
- package/dist/verification/presets.d.ts +6 -3
- package/dist/verification/presets.d.ts.map +1 -1
- package/dist/verification/presets.js +6 -3
- package/dist/verification/presets.js.map +1 -1
- package/package.json +3 -3
- package/src/__tests__/default-path-conformance.test.ts +306 -0
- package/src/__tests__/shipped-defaults.test.ts +111 -0
- package/src/advisory/__tests__/budget-enforcement.test.ts +208 -0
- package/src/advisory/__tests__/structured-response.test.ts +128 -0
- package/src/advisory/budget.ts +27 -0
- package/src/advisory/context.ts +19 -0
- package/src/advisory/executor.test.ts +12 -8
- package/src/advisory/executor.ts +37 -25
- package/src/advisory/index.ts +2 -0
- package/src/advisory/parse.ts +63 -0
- package/src/agents/AbstractAgent.ts +82 -0
- package/src/agents/PipelineAgent.ts +19 -0
- package/src/agents/ReactiveAgent.ts +47 -0
- package/src/agents/RouterAgent.ts +32 -2
- package/src/agents/SupervisorAgent.ts +31 -1
- package/src/agents/__tests__/idempotency-key.test.ts +177 -0
- package/src/agents/__tests__/invocation-lock.test.ts +154 -0
- package/src/agents/__tests__/reachability.test.ts +224 -0
- package/src/agents/__tests__/supervisor-ledger.test.ts +1 -1
- package/src/agents/index.ts +1 -1
- package/src/agents/lock.ts +0 -2
- package/src/bridge/a2a/late-events.test.ts +84 -0
- package/src/bridge/a2a/mapper.test.ts +1 -1
- package/src/bridge/a2a/mapper.ts +70 -5
- package/src/bridge/a2a/message.ts +5 -1
- package/src/bridge/sse/late-events.test.ts +182 -0
- package/src/bridge/sse/mapper.ts +109 -0
- package/src/compaction/__tests__/cleared-result-recovery.test.ts +118 -0
- package/src/compaction/__tests__/context-window.test.ts +77 -0
- package/src/compaction/__tests__/dangling.test.ts +16 -6
- package/src/compaction/__tests__/lossiness.test.ts +187 -0
- package/src/compaction/__tests__/state-across-resume.test.ts +99 -0
- package/src/compaction/__tests__/tool-result-editing.test.ts +213 -0
- package/src/compaction/__tests__/trim-opens-on-user.test.ts +118 -0
- package/src/compaction/__tests__/verifier-model.test.ts +83 -0
- package/src/compaction/context-window.ts +107 -0
- package/src/compaction/dangling.ts +48 -1
- package/src/compaction/extractor.ts +56 -1
- package/src/compaction/index.ts +16 -0
- package/src/compaction/manager.ts +53 -8
- package/src/compaction/managers/structured.ts +7 -6
- package/src/compaction/retention.ts +89 -0
- package/src/compaction/serializer.ts +42 -11
- package/src/compaction/tool-result-editing.ts +229 -0
- package/src/compaction/types.ts +10 -0
- package/src/compaction/verifier.ts +24 -1
- package/src/compaction/wire.ts +91 -0
- package/src/config/runtime.ts +63 -19
- package/src/connector/builtins/http.test.ts +13 -0
- package/src/connector/builtins/http.ts +15 -1
- package/src/connector/mcp/__tests__/discovery-policy.test.ts +133 -0
- package/src/connector/mcp/__tests__/list-pagination.test.ts +98 -0
- package/src/connector/mcp/__tests__/policy.test.ts +100 -0
- package/src/connector/mcp/__tests__/protocol-negotiation.test.ts +105 -0
- package/src/connector/mcp/__tests__/schema-fidelity.test.ts +250 -0
- package/src/connector/mcp/__tests__/structured-result.test.ts +143 -0
- package/src/connector/mcp/adapter.ts +306 -19
- package/src/connector/mcp/client.test.ts +159 -0
- package/src/connector/mcp/client.ts +160 -13
- package/src/connector/mcp/discovery.ts +74 -5
- package/src/connector/mcp/http-sse.ts +16 -1
- package/src/connector/mcp/index.ts +4 -0
- package/src/connector/mcp/policy.ts +134 -0
- package/src/connector/mcp/schema-refs.ts +126 -0
- package/src/constants/agent/index.ts +3 -0
- package/src/constants/compaction/index.ts +1 -1
- package/src/constants/continuation.ts +4 -4
- package/src/constants/mcp/index.ts +34 -0
- package/src/constants/plugin/index.ts +24 -2
- package/src/constants/sandbox/index.ts +6 -5
- package/src/constants/telemetry/index.ts +8 -0
- package/src/constants/tools/index.ts +20 -0
- package/src/contracts/api.ts +40 -0
- package/src/contracts/run-status.ts +43 -0
- package/src/eval/__tests__/case-deadline.test.ts +147 -0
- package/src/eval/__tests__/end-to-end.test.ts +259 -0
- package/src/eval/__tests__/harness-honesty.test.ts +99 -0
- package/src/eval/__tests__/harness.test.ts +249 -0
- package/src/eval/__tests__/judge.test.ts +306 -0
- package/src/eval/__tests__/scorer-gates.test.ts +156 -0
- package/src/eval/experiment.ts +335 -0
- package/src/eval/from-run.ts +48 -0
- package/src/eval/index.ts +21 -0
- package/src/eval/judge.ts +240 -0
- package/src/eval/scorers.ts +169 -0
- package/src/eval/types.ts +157 -0
- package/src/execution/local.test.ts +46 -0
- package/src/execution/local.ts +6 -1
- package/src/gateway/__tests__/list-survives-eviction.test.ts +124 -0
- package/src/gateway/local.ts +117 -3
- package/src/manager/agent/__tests__/dispose-cancels.test.ts +104 -0
- package/src/manager/agent/__tests__/lifecycle.test.ts +297 -0
- package/src/manager/agent/lifecycle.ts +127 -7
- package/src/manager/run/persistence.ts +132 -1
- package/src/plugin/__tests__/enable-contributions.test.ts +3 -3
- package/src/plugin/__tests__/hook-ordering.test.ts +252 -0
- package/src/plugin/__tests__/install-time-refusal.test.ts +68 -0
- package/src/plugin/__tests__/lifecycle.test.ts +95 -73
- package/src/plugin/lifecycle.ts +76 -22
- package/src/plugin/loader.ts +37 -1
- package/src/probe/registry.test.ts +49 -6
- package/src/probe/registry.ts +21 -1
- package/src/provider/__tests__/mock-script.test.ts +179 -0
- package/src/provider/__tests__/registry.test.ts +1 -1
- package/src/provider/__tests__/retry-visibility.test.ts +160 -0
- package/src/provider/__tests__/retry.test.ts +296 -0
- package/src/provider/capabilities.ts +4 -0
- package/src/provider/collect.test.ts +1 -1
- package/src/provider/collect.ts +1 -1
- package/src/provider/index.ts +8 -0
- package/src/provider/mock-register.ts +5 -2
- package/src/provider/mock.ts +117 -34
- package/src/provider/retry.ts +199 -0
- package/src/public-runtime.ts +111 -7
- package/src/public-types.ts +5 -1
- package/src/rag/__tests__/retrieval-math.test.ts +54 -0
- package/src/rag/chunking.ts +32 -1
- package/src/rag/embedding.test.ts +65 -10
- package/src/rag/embedding.ts +26 -5
- package/src/rag/index.ts +1 -1
- package/src/rag/retriever.ts +92 -21
- package/src/rag/vector-store.test.ts +2 -2
- package/src/registry/ManagedRegistry.test.ts +1 -1
- package/src/registry/Registry.test.ts +1 -1
- package/src/registry/connector/definitions.test.ts +1 -1
- package/src/registry/tool/__tests__/schema.test.ts +208 -0
- package/src/registry/tool/__tests__/tool-name-validation.test.ts +88 -0
- package/src/registry/tool/execute.ts +104 -22
- package/src/registry/tool/schema.ts +80 -0
- package/src/registry/toolset/catalog.test.ts +0 -1
- package/src/registry/toolset/catalog.ts +0 -1
- package/src/run/reporter.ts +84 -0
- package/src/runtime/bidi/__tests__/session.test.ts +232 -0
- package/src/runtime/bidi/index.ts +4 -0
- package/src/runtime/bidi/mock.ts +90 -0
- package/src/runtime/bidi/session.ts +205 -0
- package/src/runtime/decision/parser.test.ts +46 -0
- package/src/runtime/decision/parser.ts +8 -1
- package/src/runtime/query/__tests__/answer-review.test.ts +172 -0
- package/src/runtime/query/__tests__/capability-negotiation.test.ts +25 -129
- package/src/runtime/query/__tests__/checkpoint-store.test.ts +13 -50
- package/src/runtime/query/__tests__/checkpoint.test.ts +2 -10
- package/src/runtime/query/__tests__/citations.test.ts +111 -0
- package/src/runtime/query/__tests__/content-budget.test.ts +83 -0
- package/src/runtime/query/__tests__/context.test.ts +3 -2
- package/src/runtime/query/__tests__/deferred-tools.test.ts +23 -148
- package/src/runtime/query/__tests__/delegation-resume.test.ts +159 -0
- package/src/runtime/query/__tests__/document-attachments.test.ts +125 -0
- package/src/runtime/query/__tests__/durable-park-and-trace.test.ts +245 -0
- package/src/runtime/query/__tests__/durable-question-park.test.ts +305 -0
- package/src/runtime/query/__tests__/durable-run-state.test.ts +194 -0
- package/src/runtime/query/__tests__/executor-plugin-hooks.test.ts +24 -20
- package/src/runtime/query/__tests__/finish-reason-vs-tool-calls.test.ts +206 -0
- package/src/runtime/query/__tests__/grant-cannot-overrule-the-gate.test.ts +115 -0
- package/src/runtime/query/__tests__/guardrails-e2e.test.ts +164 -0
- package/src/runtime/query/__tests__/guardrails.test.ts +236 -0
- package/src/runtime/query/__tests__/hook-model-visibility.test.ts +143 -0
- package/src/runtime/query/__tests__/long-document-flow.test.ts +9 -2
- package/src/runtime/query/__tests__/memory-promotion.test.ts +148 -0
- package/src/runtime/query/__tests__/orphan-task-turn-end.test.ts +5 -31
- package/src/runtime/query/__tests__/overflow-relief.test.ts +139 -0
- package/src/runtime/query/__tests__/partial-batch-resume.test.ts +264 -0
- package/src/runtime/query/__tests__/plugin-hooks.test.ts +0 -1
- package/src/runtime/query/__tests__/prepare-step-chain.test.ts +127 -0
- package/src/runtime/query/__tests__/prepare-step.test.ts +233 -0
- package/src/runtime/query/__tests__/probe-veto-is-error.test.ts +145 -0
- package/src/runtime/query/__tests__/prompt.test.ts +1 -1
- package/src/runtime/query/__tests__/provider-fault-pauses.test.ts +104 -0
- package/src/runtime/query/__tests__/reasoning-stream.test.ts +177 -0
- package/src/runtime/query/__tests__/recoverable-settle.test.ts +145 -0
- package/src/runtime/query/__tests__/resume-budget.test.ts +158 -0
- package/src/runtime/query/__tests__/resume-pending.test.ts +273 -0
- package/src/runtime/query/__tests__/rich-content-budget-wiring.test.ts +109 -0
- package/src/runtime/query/__tests__/stop-condition.test.ts +265 -0
- package/src/runtime/query/__tests__/structured-output.test.ts +268 -0
- package/src/runtime/query/__tests__/structured-run-failure.test.ts +144 -0
- package/src/runtime/query/__tests__/terminal-tool.test.ts +148 -0
- package/src/runtime/query/__tests__/tool-grants.test.ts +102 -0
- package/src/runtime/query/__tests__/tool-output-budget.test.ts +0 -0
- package/src/runtime/query/__tests__/tool-pause.test.ts +209 -0
- package/src/runtime/query/__tests__/tool-progress.test.ts +154 -0
- package/src/runtime/query/__tests__/tool-repair-retry.test.ts +255 -0
- package/src/runtime/query/__tests__/tool-result-wire.test.ts +135 -0
- package/src/runtime/query/__tests__/tool-timeout.test.ts +230 -0
- package/src/runtime/query/__tests__/trace-parenting.test.ts +128 -0
- package/src/runtime/query/__tests__/wiring-seams.test.ts +288 -0
- package/src/runtime/query/checkpoint.ts +242 -10
- package/src/runtime/query/events.ts +2 -2
- package/src/runtime/query/executor.ts +847 -39
- package/src/runtime/query/guard.ts +26 -1
- package/src/runtime/query/guardrail-presets.ts +124 -0
- package/src/runtime/query/guardrails.ts +144 -0
- package/src/runtime/query/index.ts +759 -12
- package/src/runtime/query/iteration/__tests__/cancel-settles-turn.test.ts +177 -0
- package/src/runtime/query/iteration/index.ts +656 -40
- package/src/runtime/query/iteration/phases/advisory.test.ts +4 -1
- package/src/runtime/query/iteration/phases/advisory.ts +14 -2
- package/src/runtime/query/iteration/phases/checkpoint.ts +2 -2
- package/src/runtime/query/iteration/phases/compaction-safe-cut.test.ts +3 -1
- package/src/runtime/query/iteration/phases/compaction.ts +329 -64
- package/src/runtime/query/iteration/phases/context-measurement.test.ts +196 -0
- package/src/runtime/query/iteration/phases/context.ts +155 -2
- package/src/runtime/query/iteration/phases/hitl-answer-question.test.ts +3 -1
- package/src/runtime/query/iteration/phases/overflow-escalation.test.ts +167 -0
- package/src/runtime/query/iteration/phases/plan.ts +10 -2
- package/src/runtime/query/iteration/phases/retained-messages.test.ts +226 -0
- package/src/runtime/query/iteration/phases/tool-review-denials.test.ts +270 -0
- package/src/runtime/query/iteration/phases/tool-review.ts +140 -51
- package/src/runtime/query/iteration/phases/working-memory-compaction.test.ts +21 -9
- package/src/runtime/query/iteration/stream-turn.ts +285 -15
- package/src/runtime/query/plugin-hooks.ts +6 -2
- package/src/runtime/query/prompt.ts +26 -1
- package/src/runtime/query/question-park.ts +120 -0
- package/src/runtime/query/result.ts +64 -0
- package/src/runtime/query/resume-pending.ts +394 -0
- package/src/runtime/query/run-state.ts +124 -0
- package/src/runtime/query/tool-grants.ts +83 -0
- package/src/runtime/query/tool-output-budget.ts +183 -0
- package/src/runtime/query/tool-pause.ts +119 -0
- package/src/runtime/query/tooling.ts +23 -1
- package/src/sandbox/__tests__/isolation.test.ts +164 -0
- package/src/sandbox/__tests__/output-truncation.test.ts +106 -0
- package/src/sandbox/__tests__/resource-limits.test.ts +106 -0
- package/src/sandbox/factory.ts +1 -1
- package/src/sandbox/index.ts +2 -0
- package/src/sandbox/isolation.ts +72 -0
- package/src/sandbox/provider/local.ts +173 -74
- package/src/session/__tests__/integration/spawn-rollback.test.ts +1 -1
- package/src/session/handoff/broadcast.ts +3 -3
- package/src/session/migration/__tests__/filesystem.test.ts +26 -19
- package/src/session/migration/filesystem.ts +7 -12
- package/src/session/migration/marker.ts +5 -9
- package/src/session/retention/disk-backend.ts +5 -11
- package/src/session/workspace/__tests__/git-worktree.test.ts +6 -2
- package/src/session/workspace/__tests__/shared-run.test.ts +18 -0
- package/src/session/workspace/shared-run.ts +7 -1
- package/src/skills/__tests__/frontmatter.test.ts +135 -0
- package/src/skills/loader.ts +47 -3
- package/src/store/__tests__/checkpoint-integrity.test.ts +134 -0
- package/src/store/__tests__/checkpoint-versioning.test.ts +194 -0
- package/src/store/__tests__/schema-adoption.test.ts +118 -0
- package/src/store/__tests__/schema.test.ts +119 -0
- package/src/store/memory/disk.ts +17 -16
- package/src/store/run/disk.ts +169 -22
- package/src/store/schema.ts +135 -0
- package/src/store/session/disk.ts +30 -21
- package/src/store/task/__tests__/lookup-across-runs.test.ts +77 -0
- package/src/store/task/disk.ts +59 -15
- package/src/store/thread/disk.ts +21 -10
- package/src/streaming/coalesce.ts +1 -1
- package/src/telemetry/__tests__/latency-instruments.test.ts +120 -0
- package/src/telemetry/__tests__/metrics.test.ts +160 -0
- package/src/telemetry/__tests__/span-hierarchy.test.ts +94 -0
- package/src/telemetry/attributes.ts +109 -0
- package/src/telemetry/metrics.ts +190 -0
- package/src/test-setup.ts +3 -3
- package/src/test-support/paths.ts +17 -0
- package/src/tools/__tests__/path-containment.test.ts +167 -0
- package/src/tools/__tests__/posix-path.test.ts +83 -0
- package/src/tools/__tests__/sandboxed-search.test.ts +140 -0
- package/src/tools/builtins/__tests__/computer-use.test.ts +15 -2
- package/src/tools/builtins/__tests__/edit.test.ts +10 -270
- package/src/tools/builtins/__tests__/payload-budget.test.ts +20 -40
- package/src/tools/builtins/__tests__/read-file.test.ts +29 -1
- package/src/tools/builtins/__tests__/write-file.test.ts +13 -52
- package/src/tools/builtins/bash.ts +34 -4
- package/src/tools/builtins/computer-use.ts +7 -1
- package/src/tools/builtins/edit.ts +162 -116
- package/src/tools/builtins/glob.ts +71 -6
- package/src/tools/builtins/grep.ts +82 -8
- package/src/tools/builtins/index.ts +6 -6
- package/src/tools/builtins/ls.ts +4 -2
- package/src/tools/builtins/read-file.ts +40 -4
- package/src/tools/builtins/write-file.ts +44 -68
- package/src/tools/coordinator/__tests__/agent.test.ts +2 -2
- package/src/tools/coordinator/agent.ts +23 -10
- package/src/tools/coordinator/index.ts +111 -31
- package/src/tools/defineTool.ts +20 -0
- package/src/tools/glob-match.ts +50 -0
- package/src/tools/paths.ts +44 -0
- package/src/tools/posix-path.ts +73 -0
- package/src/types/advisory/config.ts +21 -3
- package/src/types/advisory/result.ts +12 -7
- package/src/types/agent/base.ts +22 -2
- package/src/types/agent/factory.ts +15 -9
- package/src/types/agent/gateway.ts +20 -0
- package/src/types/agent/manager.ts +19 -0
- package/src/types/agent/reactive.ts +50 -0
- package/src/types/agent/router.ts +11 -0
- package/src/types/agent/supervisor.ts +2 -2
- package/src/types/agent/task.ts +26 -0
- package/src/types/bidi/index.ts +119 -0
- package/src/types/common/__tests__/numeric-correctness.test.ts +123 -0
- package/src/types/common/index.ts +16 -4
- package/src/types/computer-use/index.ts +1 -1
- package/src/types/connector/mcp.ts +29 -0
- package/src/types/errors/__tests__/catalog.test.ts +130 -0
- package/src/types/errors/__tests__/errors.test.ts +118 -0
- package/src/types/errors/catalog.ts +211 -0
- package/src/types/errors/index.ts +121 -0
- package/src/types/guardrail/index.ts +62 -0
- package/src/types/hitl/index.ts +102 -20
- package/src/types/ids/index.ts +2 -2
- package/src/types/message/content.ts +50 -0
- package/src/types/message/index.ts +210 -9
- package/src/types/plugin/index.ts +82 -2
- package/src/types/provider/__tests__/cause-chain.test.ts +132 -0
- package/src/types/provider/__tests__/declared-retryable.test.ts +85 -0
- package/src/types/provider/__tests__/overflow-classification.test.ts +113 -0
- package/src/types/provider/chat.ts +29 -1
- package/src/types/provider/config.ts +83 -2
- package/src/types/provider/errors.ts +472 -0
- package/src/types/provider/index.ts +5 -0
- package/src/types/provider/stream.ts +63 -3
- package/src/types/rag/embedding.ts +11 -2
- package/src/types/run/__tests__/step.test.ts +123 -0
- package/src/types/run/answer-review.ts +51 -0
- package/src/types/run/config.ts +35 -0
- package/src/types/run/derive-status.ts +62 -0
- package/src/types/run/entity.ts +18 -0
- package/src/types/run/events.ts +230 -3
- package/src/types/run/index.ts +5 -0
- package/src/types/run/memory-promotion.ts +112 -0
- package/src/types/run/prepare-step.ts +109 -0
- package/src/types/run/state.ts +111 -0
- package/src/types/run/step.ts +95 -0
- package/src/types/run/stop-reason.ts +29 -1
- package/src/types/sandbox/index.ts +93 -12
- package/src/types/structured-output/index.ts +6 -5
- package/src/types/thread/entity.ts +2 -3
- package/src/types/tool/index.ts +204 -1
- package/src/types/tool/repair.ts +54 -0
- package/src/types/toolset/index.ts +2 -6
- package/src/utils/__tests__/atomic-write.test.ts +108 -0
- package/src/utils/__tests__/shell-compress.test.ts +49 -0
- package/src/utils/atomic-write.ts +85 -0
- package/src/utils/shell-compress.ts +12 -1
- package/src/verification/presets.ts +6 -3
|
@@ -0,0 +1,136 @@
|
|
|
1
|
+
import { describe, expect, it } from 'vitest';
|
|
2
|
+
import { formatReport, runExperiment } from '../experiment.js';
|
|
3
|
+
/**
|
|
4
|
+
* A case's verdict was one unweighted mean over every scorer against one
|
|
5
|
+
* suite-wide threshold, and the two halves of that fought each other.
|
|
6
|
+
*
|
|
7
|
+
* At the default threshold of 1 the harness never reports a false pass —
|
|
8
|
+
* but a trajectory F1 and a graded judge can essentially never reach 1, so
|
|
9
|
+
* every real suite lowers it. And every step down buys tolerance for the
|
|
10
|
+
* fuzzy scorers by buying exactly the same tolerance for the deterministic
|
|
11
|
+
* ones: at 0.75, a trajectory score of 0 alongside three perfect scores
|
|
12
|
+
* averages to 0.75 and reports PASSED. The regression the harness exists
|
|
13
|
+
* to catch comes back green.
|
|
14
|
+
*
|
|
15
|
+
* The workarounds were lossy too — folding every hard check into one
|
|
16
|
+
* custom scorer destroys per-dimension reporting, and two experiments at
|
|
17
|
+
* two thresholds re-runs the dataset and splits one case's verdict across
|
|
18
|
+
* two reports.
|
|
19
|
+
*/
|
|
20
|
+
const run = () => ({
|
|
21
|
+
output: 'answer',
|
|
22
|
+
steps: [],
|
|
23
|
+
toolCalls: [],
|
|
24
|
+
totalTokens: 0,
|
|
25
|
+
totalCostUsd: 0,
|
|
26
|
+
durationMs: 0,
|
|
27
|
+
});
|
|
28
|
+
const scorer = (name, score, extra = {}) => ({
|
|
29
|
+
name,
|
|
30
|
+
score: () => ({ score, reason: `${name} scored ${score}` }),
|
|
31
|
+
...extra,
|
|
32
|
+
});
|
|
33
|
+
const experiment = (scorers, passThreshold = 0.75) => runExperiment({
|
|
34
|
+
name: 'gates',
|
|
35
|
+
cases: [{ name: 'one', input: null }],
|
|
36
|
+
scorers,
|
|
37
|
+
passThreshold,
|
|
38
|
+
run: async () => run(),
|
|
39
|
+
});
|
|
40
|
+
describe('a gate scorer', () => {
|
|
41
|
+
it('fails the case even when the mean clears the threshold', async () => {
|
|
42
|
+
const report = await experiment([
|
|
43
|
+
scorer('trajectory', 0, { severity: 'gate' }),
|
|
44
|
+
scorer('completion', 1),
|
|
45
|
+
scorer('contains', 1),
|
|
46
|
+
scorer('judge', 1),
|
|
47
|
+
]);
|
|
48
|
+
// Mean is exactly 0.75 against a 0.75 threshold — reported passed
|
|
49
|
+
// before, with the one deterministic check at zero.
|
|
50
|
+
expect(report.cases[0]?.mean).toBeCloseTo(0.75);
|
|
51
|
+
expect(report.cases[0]?.status).toBe('failed');
|
|
52
|
+
});
|
|
53
|
+
it('names itself, so the reader does not have to guess', async () => {
|
|
54
|
+
const report = await experiment([
|
|
55
|
+
scorer('trajectory', 0, { severity: 'gate' }),
|
|
56
|
+
scorer('judge', 1),
|
|
57
|
+
]);
|
|
58
|
+
expect(report.cases[0]?.failedGates).toEqual(['trajectory']);
|
|
59
|
+
});
|
|
60
|
+
it('honours its own threshold when it sets one', async () => {
|
|
61
|
+
// "Good enough" is not one number across dimensions: a trajectory
|
|
62
|
+
// match at 0.8 may be fine while a completion check at 0.8 is
|
|
63
|
+
// meaningless.
|
|
64
|
+
const lenient = await experiment([
|
|
65
|
+
scorer('trajectory', 0.8, { severity: 'gate', threshold: 0.7 }),
|
|
66
|
+
]);
|
|
67
|
+
expect(lenient.cases[0]?.status).toBe('passed');
|
|
68
|
+
const strict = await experiment([
|
|
69
|
+
scorer('trajectory', 0.8, { severity: 'gate', threshold: 0.9 }),
|
|
70
|
+
]);
|
|
71
|
+
expect(strict.cases[0]?.status).toBe('failed');
|
|
72
|
+
});
|
|
73
|
+
it('falls back to the suite threshold when it sets none', async () => {
|
|
74
|
+
const report = await experiment([scorer('trajectory', 0.5, { severity: 'gate' })], 0.75);
|
|
75
|
+
expect(report.cases[0]?.failedGates).toEqual(['trajectory']);
|
|
76
|
+
});
|
|
77
|
+
it('does not fail the case when it could not judge at all', async () => {
|
|
78
|
+
// Unavailable is not zero. A gate that threw did not judge the run
|
|
79
|
+
// badly, it failed to judge it — the inconclusive path, not a
|
|
80
|
+
// failure, and the two demand opposite responses.
|
|
81
|
+
const report = await runExperiment({
|
|
82
|
+
name: 'gates',
|
|
83
|
+
cases: [{ name: 'one', input: null }],
|
|
84
|
+
scorers: [
|
|
85
|
+
{
|
|
86
|
+
name: 'judge-gate',
|
|
87
|
+
severity: 'gate',
|
|
88
|
+
score: () => {
|
|
89
|
+
throw new Error('provider unreachable');
|
|
90
|
+
},
|
|
91
|
+
},
|
|
92
|
+
],
|
|
93
|
+
run: async () => run(),
|
|
94
|
+
});
|
|
95
|
+
expect(report.cases[0]?.status).toBe('inconclusive');
|
|
96
|
+
expect(report.cases[0]?.failedGates).toBeUndefined();
|
|
97
|
+
});
|
|
98
|
+
it('lets a passing gate through', async () => {
|
|
99
|
+
const report = await experiment([scorer('trajectory', 1, { severity: 'gate' })]);
|
|
100
|
+
expect(report.cases[0]?.status).toBe('passed');
|
|
101
|
+
});
|
|
102
|
+
});
|
|
103
|
+
describe('a soft scorer', () => {
|
|
104
|
+
it('only moves the mean', async () => {
|
|
105
|
+
// The fuzzy dimensions, where a number below 1 is normal and a
|
|
106
|
+
// threshold is a judgement call.
|
|
107
|
+
const report = await experiment([scorer('judge', 0.5), scorer('other', 1)]);
|
|
108
|
+
expect(report.cases[0]?.status).toBe('passed');
|
|
109
|
+
expect(report.cases[0]?.failedGates).toBeUndefined();
|
|
110
|
+
});
|
|
111
|
+
it('is the default, so a suite that sets nothing is unchanged', async () => {
|
|
112
|
+
const report = await experiment([
|
|
113
|
+
scorer('a', 0),
|
|
114
|
+
scorer('b', 1),
|
|
115
|
+
scorer('c', 1),
|
|
116
|
+
scorer('d', 1),
|
|
117
|
+
]);
|
|
118
|
+
expect(report.cases[0]?.status).toBe('passed');
|
|
119
|
+
});
|
|
120
|
+
it('can still fail the case through the mean', async () => {
|
|
121
|
+
const report = await experiment([scorer('a', 0), scorer('b', 0.5)]);
|
|
122
|
+
expect(report.cases[0]?.status).toBe('failed');
|
|
123
|
+
expect(report.cases[0]?.failedGates).toBeUndefined();
|
|
124
|
+
});
|
|
125
|
+
});
|
|
126
|
+
describe('the report', () => {
|
|
127
|
+
it('says which gate missed, above the individual scores', async () => {
|
|
128
|
+
const report = await experiment([
|
|
129
|
+
scorer('trajectory', 0, { severity: 'gate' }),
|
|
130
|
+
scorer('judge', 1),
|
|
131
|
+
]);
|
|
132
|
+
const text = formatReport(report);
|
|
133
|
+
expect(text).toContain('gate missed: trajectory');
|
|
134
|
+
});
|
|
135
|
+
});
|
|
136
|
+
//# sourceMappingURL=scorer-gates.test.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"scorer-gates.test.js","sourceRoot":"","sources":["../../../src/eval/__tests__/scorer-gates.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,QAAQ,CAAA;AAE7C,OAAO,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAA;AAG9D;;;;;;;;;;;;;;;;GAgBG;AAEH,MAAM,GAAG,GAAG,GAAY,EAAE,CAAC,CAAC;IAC3B,MAAM,EAAE,QAAQ;IAChB,KAAK,EAAE,EAAE;IACT,SAAS,EAAE,EAAE;IACb,WAAW,EAAE,CAAC;IACd,YAAY,EAAE,CAAC;IACf,UAAU,EAAE,CAAC;CACb,CAAC,CAAA;AAEF,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,KAAa,EAAE,QAAyB,EAAE,EAAU,EAAE,CAAC,CAAC;IACrF,IAAI;IACJ,KAAK,EAAE,GAAU,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,IAAI,WAAW,KAAK,EAAE,EAAE,CAAC;IAClE,GAAG,KAAK;CACR,CAAC,CAAA;AAEF,MAAM,UAAU,GAAG,CAAC,OAAiB,EAAE,aAAa,GAAG,IAAI,EAAE,EAAE,CAC9D,aAAa,CAAC;IACb,IAAI,EAAE,OAAO;IACb,KAAK,EAAE,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC;IACrC,OAAO;IACP,aAAa;IACb,GAAG,EAAE,KAAK,IAAI,EAAE,CAAC,GAAG,EAAE;CACtB,CAAC,CAAA;AAEH,QAAQ,CAAC,eAAe,EAAE,GAAG,EAAE;IAC9B,EAAE,CAAC,wDAAwD,EAAE,KAAK,IAAI,EAAE;QACvE,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC;YAC/B,MAAM,CAAC,YAAY,EAAE,CAAC,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC;YAC7C,MAAM,CAAC,YAAY,EAAE,CAAC,CAAC;YACvB,MAAM,CAAC,UAAU,EAAE,CAAC,CAAC;YACrB,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC;SAClB,CAAC,CAAA;QAEF,kEAAkE;QAClE,oDAAoD;QACpD,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,WAAW,CAAC,IAAI,CAAC,CAAA;QAC/C,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IAC/C,CAAC,CAAC,CAAA;IAEF,EAAE,CAAC,oDAAoD,EAAE,KAAK,IAAI,EAAE;QACnE,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC;YAC/B,MAAM,CAAC,YAAY,EAAE,CAAC,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC;YAC7C,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC;SAClB,CAAC,CAAA;QACF,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,WAAW,CAAC,CAAC,OAAO,CAAC,CAAC,YAAY,CAAC,CAAC,CAAA;IAC7D,CAAC,CAAC,CAAA;IAEF,EAAE,CAAC,4CAA4C,EAAE,KAAK,IAAI,EAAE;QAC3D,kEAAkE;QAClE,8DAA8D;QAC9D,eAAe;QACf,MAAM,OAAO,GAAG,MAAM,UAAU,CAAC;YAChC,MAAM,CAAC,YAAY,EAAE,GAAG,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,SAAS,EAAE,GAAG,EAAE,CAAC;SAC/D,CAAC,CAAA;QACF,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;QAE/C,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC;YAC/B,MAAM,CAAC,YAAY,EAAE,GAAG,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,SAAS,EAAE,GAAG,EAAE,CAAC;SAC/D,CAAC,CAAA;QACF,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IAC/C,CAAC,CAAC,CAAA;IAEF,EAAE,CAAC,qDAAqD,EAAE,KAAK,IAAI,EAAE;QACpE,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC,CAAC,MAAM,CAAC,YAAY,EAAE,GAAG,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QACxF,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,WAAW,CAAC,CAAC,OAAO,CAAC,CAAC,YAAY,CAAC,CAAC,CAAA;IAC7D,CAAC,CAAC,CAAA;IAEF,EAAE,CAAC,uDAAuD,EAAE,KAAK,IAAI,EAAE;QACtE,mEAAmE;QACnE,8DAA8D;QAC9D,kDAAkD;QAClD,MAAM,MAAM,GAAG,MAAM,aAAa,CAAC;YAClC,IAAI,EAAE,OAAO;YACb,KAAK,EAAE,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC;YACrC,OAAO,EAAE;gBACR;oBACC,IAAI,EAAE,YAAY;oBAClB,QAAQ,EAAE,MAAM;oBAChB,KAAK,EAAE,GAAG,EAAE;wBACX,MAAM,IAAI,KAAK,CAAC,sBAAsB,CAAC,CAAA;oBACxC,CAAC;iBACD;aACD;YACD,GAAG,EAAE,KAAK,IAAI,EAAE,CAAC,GAAG,EAAE;SACtB,CAAC,CAAA;QAEF,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,cAAc,CAAC,CAAA;QACpD,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,WAAW,CAAC,CAAC,aAAa,EAAE,CAAA;IACrD,CAAC,CAAC,CAAA;IAEF,EAAE,CAAC,6BAA6B,EAAE,KAAK,IAAI,EAAE;QAC5C,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC,CAAC,MAAM,CAAC,YAAY,EAAE,CAAC,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,CAAA;QAChF,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IAC/C,CAAC,CAAC,CAAA;AACH,CAAC,CAAC,CAAA;AAEF,QAAQ,CAAC,eAAe,EAAE,GAAG,EAAE;IAC9B,EAAE,CAAC,qBAAqB,EAAE,KAAK,IAAI,EAAE;QACpC,+DAA+D;QAC/D,iCAAiC;QACjC,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC,CAAC,MAAM,CAAC,OAAO,EAAE,GAAG,CAAC,EAAE,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,CAAC,CAAA;QAC3E,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;QAC9C,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,WAAW,CAAC,CAAC,aAAa,EAAE,CAAA;IACrD,CAAC,CAAC,CAAA;IAEF,EAAE,CAAC,2DAA2D,EAAE,KAAK,IAAI,EAAE;QAC1E,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC;YAC/B,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC;YACd,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC;YACd,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC;YACd,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC;SACd,CAAC,CAAA;QACF,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IAC/C,CAAC,CAAC,CAAA;IAEF,EAAE,CAAC,0CAA0C,EAAE,KAAK,IAAI,EAAE;QACzD,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC,CAAC,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC,EAAE,MAAM,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC,CAAA;QACnE,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;QAC9C,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,WAAW,CAAC,CAAC,aAAa,EAAE,CAAA;IACrD,CAAC,CAAC,CAAA;AACH,CAAC,CAAC,CAAA;AAEF,QAAQ,CAAC,YAAY,EAAE,GAAG,EAAE;IAC3B,EAAE,CAAC,qDAAqD,EAAE,KAAK,IAAI,EAAE;QACpE,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC;YAC/B,MAAM,CAAC,YAAY,EAAE,CAAC,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC;YAC7C,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC;SAClB,CAAC,CAAA;QAEF,MAAM,IAAI,GAAG,YAAY,CAAC,MAAM,CAAC,CAAA;QACjC,MAAM,CAAC,IAAI,CAAC,CAAC,SAAS,CAAC,yBAAyB,CAAC,CAAA;IAClD,CAAC,CAAC,CAAA;AACH,CAAC,CAAC,CAAA"}
|
|
@@ -0,0 +1,54 @@
|
|
|
1
|
+
import type { CaseResult, EvalCase, EvalRun, ExperimentReport, Scorer } from './types.js';
|
|
2
|
+
export interface ExperimentConfig<TInput = unknown> {
|
|
3
|
+
name: string;
|
|
4
|
+
cases: readonly EvalCase<TInput>[];
|
|
5
|
+
/** Applied to every case unless the case overrides them. */
|
|
6
|
+
scorers: readonly Scorer[];
|
|
7
|
+
/**
|
|
8
|
+
* Execute one case. Returning an `EvalRun` rather than driving `query()`
|
|
9
|
+
* here keeps the harness independent of how you construct a run —
|
|
10
|
+
* scripted mock, real provider, or a whole agent behind a facade.
|
|
11
|
+
*/
|
|
12
|
+
run: (input: TInput, evalCase: EvalCase<TInput>, signal: AbortSignal) => Promise<EvalRun>;
|
|
13
|
+
/**
|
|
14
|
+
* Deadline for a single case. Unset means no deadline.
|
|
15
|
+
*
|
|
16
|
+
* `executeCase` used to be a bare await, so a `run` closure that never
|
|
17
|
+
* settled blocked its worker and `runExperiment` never returned — no
|
|
18
|
+
* report, no partial results, nothing to read. The documented path
|
|
19
|
+
* inherits deadlines from the runtime it drives, so the residual is a
|
|
20
|
+
* closure that does not go through `query()` and a mid-iteration
|
|
21
|
+
* provider stall the between-iterations guard cannot see. Both are
|
|
22
|
+
* reachable, and neither is the suite's fault to absorb silently.
|
|
23
|
+
*
|
|
24
|
+
* A timed-out case is REPORTED and the suite continues, exactly like a
|
|
25
|
+
* case that threw: forty cases should not be lost to one that hung.
|
|
26
|
+
*/
|
|
27
|
+
timeoutMs?: number;
|
|
28
|
+
/** Mean score a case must reach to count as passed. Default 1. */
|
|
29
|
+
passThreshold?: number;
|
|
30
|
+
/** Cases to run at once. Default 1 — deterministic ordering by default. */
|
|
31
|
+
concurrency?: number;
|
|
32
|
+
onCaseFinish?: (result: CaseResult) => void;
|
|
33
|
+
}
|
|
34
|
+
/**
|
|
35
|
+
* Run a dataset and score it.
|
|
36
|
+
*
|
|
37
|
+
* There was no evaluation harness of any kind: no dataset, no scorer, no
|
|
38
|
+
* judge, no trajectory assertion. So every behavior change in this SDK —
|
|
39
|
+
* a tool description, the `search_tools` top-k, the compaction threshold —
|
|
40
|
+
* shipped with no regression signal behind it, and the only way to notice
|
|
41
|
+
* a degradation was for a user to hit it.
|
|
42
|
+
*
|
|
43
|
+
* The scorer contract requires a `reason` on every score, so a report says
|
|
44
|
+
* what got worse rather than only that something did.
|
|
45
|
+
*/
|
|
46
|
+
export declare function runExperiment<TInput>(config: ExperimentConfig<TInput>): Promise<ExperimentReport>;
|
|
47
|
+
/**
|
|
48
|
+
* Render a report as text.
|
|
49
|
+
*
|
|
50
|
+
* Failures print their scorer reasons, because a CI log that says
|
|
51
|
+
* "0.62" is a log that sends someone back to reproduce it by hand.
|
|
52
|
+
*/
|
|
53
|
+
export declare function formatReport(report: ExperimentReport): string;
|
|
54
|
+
//# sourceMappingURL=experiment.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"experiment.d.ts","sourceRoot":"","sources":["../../src/eval/experiment.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACX,UAAU,EAEV,QAAQ,EACR,OAAO,EACP,gBAAgB,EAEhB,MAAM,EACN,MAAM,YAAY,CAAA;AAEnB,MAAM,WAAW,gBAAgB,CAAC,MAAM,GAAG,OAAO;IACjD,IAAI,EAAE,MAAM,CAAA;IACZ,KAAK,EAAE,SAAS,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAA;IAClC,4DAA4D;IAC5D,OAAO,EAAE,SAAS,MAAM,EAAE,CAAA;IAC1B;;;;OAIG;IACH,GAAG,EAAE,CAAC,KAAK,EAAE,MAAM,EAAE,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,EAAE,MAAM,EAAE,WAAW,KAAK,OAAO,CAAC,OAAO,CAAC,CAAA;IACzF;;;;;;;;;;;;;OAaG;IACH,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,kEAAkE;IAClE,aAAa,CAAC,EAAE,MAAM,CAAA;IACtB,2EAA2E;IAC3E,WAAW,CAAC,EAAE,MAAM,CAAA;IACpB,YAAY,CAAC,EAAE,CAAC,MAAM,EAAE,UAAU,KAAK,IAAI,CAAA;CAC3C;AAED;;;;;;;;;;;GAWG;AACH,wBAAsB,aAAa,CAAC,MAAM,EACzC,MAAM,EAAE,gBAAgB,CAAC,MAAM,CAAC,GAC9B,OAAO,CAAC,gBAAgB,CAAC,CAsG3B;AAmHD;;;;;GAKG;AACH,wBAAgB,YAAY,CAAC,MAAM,EAAE,gBAAgB,GAAG,MAAM,CAsD7D"}
|
|
@@ -0,0 +1,263 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Run a dataset and score it.
|
|
3
|
+
*
|
|
4
|
+
* There was no evaluation harness of any kind: no dataset, no scorer, no
|
|
5
|
+
* judge, no trajectory assertion. So every behavior change in this SDK —
|
|
6
|
+
* a tool description, the `search_tools` top-k, the compaction threshold —
|
|
7
|
+
* shipped with no regression signal behind it, and the only way to notice
|
|
8
|
+
* a degradation was for a user to hit it.
|
|
9
|
+
*
|
|
10
|
+
* The scorer contract requires a `reason` on every score, so a report says
|
|
11
|
+
* what got worse rather than only that something did.
|
|
12
|
+
*/
|
|
13
|
+
export async function runExperiment(config) {
|
|
14
|
+
const startedAt = Date.now();
|
|
15
|
+
const threshold = config.passThreshold ?? 1;
|
|
16
|
+
const concurrency = Math.max(1, config.concurrency ?? 1);
|
|
17
|
+
const results = new Array(config.cases.length);
|
|
18
|
+
let cursor = 0;
|
|
19
|
+
const worker = async () => {
|
|
20
|
+
for (;;) {
|
|
21
|
+
const index = cursor++;
|
|
22
|
+
const evalCase = config.cases[index];
|
|
23
|
+
if (!evalCase)
|
|
24
|
+
return;
|
|
25
|
+
const run = await executeCase(config, evalCase);
|
|
26
|
+
const scorers = evalCase.scorers ?? config.scorers;
|
|
27
|
+
const scores = {};
|
|
28
|
+
// Scores are keyed by name, so two scorers sharing one collapse:
|
|
29
|
+
// the mean's denominator becomes the count of distinct NAMES and
|
|
30
|
+
// the surviving score is whichever ran last. Two
|
|
31
|
+
// `containsScorer(...)` instances are both called 'contains', so
|
|
32
|
+
// this is easy to hit by accident and silently halves the
|
|
33
|
+
// evidence. Ambiguous results are worse than a loud failure.
|
|
34
|
+
const seen = new Set();
|
|
35
|
+
for (const scorer of scorers) {
|
|
36
|
+
if (seen.has(scorer.name)) {
|
|
37
|
+
throw new Error(`Duplicate scorer name "${scorer.name}" for case "${evalCase.name}". Scores are keyed by name, so the second would overwrite the first and the case mean would be computed over the wrong denominator. Give each scorer a distinct name.`);
|
|
38
|
+
}
|
|
39
|
+
seen.add(scorer.name);
|
|
40
|
+
scores[scorer.name] = await safeScore(scorer, run, evalCase);
|
|
41
|
+
}
|
|
42
|
+
// Only scores that were actually produced count — an unavailable
|
|
43
|
+
// scorer leaves the denominator alone rather than dragging the
|
|
44
|
+
// mean toward zero with a measurement that never happened.
|
|
45
|
+
const values = Object.values(scores)
|
|
46
|
+
.filter((s) => s.unavailable !== true)
|
|
47
|
+
.map((s) => s.score);
|
|
48
|
+
const mean = values.length === 0 ? 0 : values.reduce((a, b) => a + b, 0) / values.length;
|
|
49
|
+
// A gate miss fails the case whatever the mean says. Averaging a
|
|
50
|
+
// hard check together with a fuzzy one lets three good scores
|
|
51
|
+
// carry a zero: trajectory 0 + completion 1 + contains 1 +
|
|
52
|
+
// judge 1 averages to 0.75 and reports passed at a threshold of
|
|
53
|
+
// 0.75 — the exact regression the harness exists to catch,
|
|
54
|
+
// reported green. An UNAVAILABLE gate does not fail the case;
|
|
55
|
+
// it did not judge the run at all, which is the inconclusive
|
|
56
|
+
// path, not a failure.
|
|
57
|
+
const failedGates = scorers
|
|
58
|
+
.filter((s) => s.severity === 'gate')
|
|
59
|
+
.filter((s) => {
|
|
60
|
+
const score = scores[s.name];
|
|
61
|
+
if (!score || score.unavailable === true)
|
|
62
|
+
return false;
|
|
63
|
+
return score.score < (s.threshold ?? threshold);
|
|
64
|
+
})
|
|
65
|
+
.map((s) => s.name);
|
|
66
|
+
const status = Object.keys(scores).length > 0 && values.length === 0
|
|
67
|
+
? 'inconclusive'
|
|
68
|
+
: failedGates.length > 0
|
|
69
|
+
? 'failed'
|
|
70
|
+
: mean >= threshold
|
|
71
|
+
? 'passed'
|
|
72
|
+
: 'failed';
|
|
73
|
+
const result = {
|
|
74
|
+
case: evalCase.name,
|
|
75
|
+
run,
|
|
76
|
+
scores,
|
|
77
|
+
mean,
|
|
78
|
+
status,
|
|
79
|
+
passed: status === 'passed',
|
|
80
|
+
// Named, not just counted: "failed" with a mean of 0.75 sends
|
|
81
|
+
// somebody to read four scores and guess which one mattered.
|
|
82
|
+
...(failedGates.length > 0 ? { failedGates } : {}),
|
|
83
|
+
};
|
|
84
|
+
results[index] = result;
|
|
85
|
+
config.onCaseFinish?.(result);
|
|
86
|
+
}
|
|
87
|
+
};
|
|
88
|
+
await Promise.all(Array.from({ length: Math.min(concurrency, config.cases.length) }, worker));
|
|
89
|
+
const settled = results.filter(Boolean);
|
|
90
|
+
// An inconclusive case has no mean to contribute; averaging its zero in
|
|
91
|
+
// would report a suite-wide regression caused by the harness.
|
|
92
|
+
const scored = settled.filter((r) => r.status !== 'inconclusive');
|
|
93
|
+
const mean = scored.length === 0 ? 0 : scored.reduce((sum, r) => sum + r.mean, 0) / scored.length;
|
|
94
|
+
return {
|
|
95
|
+
name: config.name,
|
|
96
|
+
cases: settled,
|
|
97
|
+
mean,
|
|
98
|
+
passed: settled.filter((r) => r.status === 'passed').length,
|
|
99
|
+
failed: settled.filter((r) => r.status === 'failed').length,
|
|
100
|
+
inconclusive: settled.filter((r) => r.status === 'inconclusive').length,
|
|
101
|
+
byScorer: meanByScorer(settled),
|
|
102
|
+
durationMs: Date.now() - startedAt,
|
|
103
|
+
};
|
|
104
|
+
}
|
|
105
|
+
/**
|
|
106
|
+
* A case that throws is a RESULT, not a crash. An eval suite whose first
|
|
107
|
+
* broken case aborts the run tells you nothing about the other forty.
|
|
108
|
+
*/
|
|
109
|
+
async function executeCase(config, evalCase) {
|
|
110
|
+
const startedAt = Date.now();
|
|
111
|
+
const controller = new AbortController();
|
|
112
|
+
const timeoutMs = config.timeoutMs;
|
|
113
|
+
// The signal is handed to `run` so a closure that drives `query()` can
|
|
114
|
+
// pass it through and actually stop working. A closure that ignores it
|
|
115
|
+
// is merely detached rather than stopped — the same bargain every other
|
|
116
|
+
// deadline in the SDK makes — but the SUITE is unblocked either way,
|
|
117
|
+
// which is the part that was missing.
|
|
118
|
+
const timer = timeoutMs !== undefined && timeoutMs > 0
|
|
119
|
+
? setTimeout(() => controller.abort(new Error(`case timed out after ${timeoutMs}ms`)), timeoutMs)
|
|
120
|
+
: undefined;
|
|
121
|
+
const deadline = timer === undefined
|
|
122
|
+
? undefined
|
|
123
|
+
: new Promise((_resolve, reject) => {
|
|
124
|
+
controller.signal.addEventListener('abort', () => reject(controller.signal.reason), {
|
|
125
|
+
once: true,
|
|
126
|
+
});
|
|
127
|
+
});
|
|
128
|
+
try {
|
|
129
|
+
const work = config.run(evalCase.input, evalCase, controller.signal);
|
|
130
|
+
// A rejection of the loser must never surface as an unhandled
|
|
131
|
+
// rejection after the race has already been decided.
|
|
132
|
+
if (deadline)
|
|
133
|
+
work.catch(() => { });
|
|
134
|
+
return await (deadline ? Promise.race([work, deadline]) : work);
|
|
135
|
+
}
|
|
136
|
+
catch (err) {
|
|
137
|
+
return {
|
|
138
|
+
output: null,
|
|
139
|
+
steps: [],
|
|
140
|
+
toolCalls: [],
|
|
141
|
+
totalTokens: 0,
|
|
142
|
+
totalCostUsd: 0,
|
|
143
|
+
// Real elapsed time, not zero: a case that burned its whole
|
|
144
|
+
// deadline is the most interesting number in the report, and
|
|
145
|
+
// zero would hide it.
|
|
146
|
+
durationMs: Date.now() - startedAt,
|
|
147
|
+
error: err instanceof Error ? err.message : String(err),
|
|
148
|
+
};
|
|
149
|
+
}
|
|
150
|
+
finally {
|
|
151
|
+
// Only the timer needs clearing. Aborting here would fire a spurious
|
|
152
|
+
// signal at a `run` that has already settled — the deadline branch
|
|
153
|
+
// has aborted already, and every other path is done.
|
|
154
|
+
clearTimeout(timer);
|
|
155
|
+
}
|
|
156
|
+
}
|
|
157
|
+
/** A throwing scorer scores zero with the throw as its reason. */
|
|
158
|
+
async function safeScore(scorer, run, evalCase) {
|
|
159
|
+
// A run that THREW scores zero, whatever the scorer would have said.
|
|
160
|
+
// `executeCase` catches the failure and returns an empty run, and an
|
|
161
|
+
// empty run walks straight into every scorer's happy path:
|
|
162
|
+
// `stepBudgetScorer` sees 0 steps against its allowance and returns 1,
|
|
163
|
+
// `trajectoryScorer` sees no tools expected and none called and returns
|
|
164
|
+
// 1. So a suite whose runs were all dying reported green. The failure is
|
|
165
|
+
// recorded on `run.error` and nothing consulted it.
|
|
166
|
+
if (run.error !== undefined) {
|
|
167
|
+
return { score: 0, reason: `run failed: ${run.error}`, details: { error: run.error } };
|
|
168
|
+
}
|
|
169
|
+
try {
|
|
170
|
+
return await scorer.score(run, evalCase);
|
|
171
|
+
}
|
|
172
|
+
catch (err) {
|
|
173
|
+
// UNAVAILABLE, not zero. A scorer that threw did not judge the run
|
|
174
|
+
// badly — it failed to judge it at all, and the two call for
|
|
175
|
+
// opposite responses. Scoring the throw zero was survivable while
|
|
176
|
+
// every scorer was a pure function; a scorer that reaches a provider
|
|
177
|
+
// makes it actively misleading, because one rate limit would look
|
|
178
|
+
// exactly like a behavioural regression.
|
|
179
|
+
return {
|
|
180
|
+
score: 0,
|
|
181
|
+
unavailable: true,
|
|
182
|
+
reason: `scorer "${scorer.name}" could not judge this run: ${err instanceof Error ? err.message : String(err)}`,
|
|
183
|
+
};
|
|
184
|
+
}
|
|
185
|
+
}
|
|
186
|
+
function meanByScorer(results) {
|
|
187
|
+
const sums = new Map();
|
|
188
|
+
for (const result of results) {
|
|
189
|
+
for (const [name, score] of Object.entries(result.scores)) {
|
|
190
|
+
if (score.unavailable === true)
|
|
191
|
+
continue;
|
|
192
|
+
const entry = sums.get(name) ?? { total: 0, count: 0 };
|
|
193
|
+
entry.total += score.score;
|
|
194
|
+
entry.count++;
|
|
195
|
+
sums.set(name, entry);
|
|
196
|
+
}
|
|
197
|
+
}
|
|
198
|
+
const out = {};
|
|
199
|
+
// A scorer that was never available anywhere is omitted rather than
|
|
200
|
+
// reported as 0 — a dimension with no measurements is not a dimension
|
|
201
|
+
// that scored badly.
|
|
202
|
+
for (const [name, { total, count }] of sums) {
|
|
203
|
+
if (count > 0)
|
|
204
|
+
out[name] = total / count;
|
|
205
|
+
}
|
|
206
|
+
return out;
|
|
207
|
+
}
|
|
208
|
+
/**
|
|
209
|
+
* Render a report as text.
|
|
210
|
+
*
|
|
211
|
+
* Failures print their scorer reasons, because a CI log that says
|
|
212
|
+
* "0.62" is a log that sends someone back to reproduce it by hand.
|
|
213
|
+
*/
|
|
214
|
+
export function formatReport(report) {
|
|
215
|
+
const lines = [
|
|
216
|
+
`${report.name}: ${report.passed}/${report.cases.length} passed (mean ${report.mean.toFixed(2)}) in ${report.durationMs}ms`,
|
|
217
|
+
'',
|
|
218
|
+
];
|
|
219
|
+
// Surfaced above the failures, because an inconclusive case means the
|
|
220
|
+
// harness is broken and every number below it is measured over less
|
|
221
|
+
// evidence than it looks like.
|
|
222
|
+
if (report.inconclusive > 0) {
|
|
223
|
+
lines.push(` ${report.inconclusive} case${report.inconclusive === 1 ? '' : 's'} could not be judged — the numbers below cover the rest`, '');
|
|
224
|
+
}
|
|
225
|
+
for (const [name, mean] of Object.entries(report.byScorer)) {
|
|
226
|
+
lines.push(` ${name}: ${mean.toFixed(2)}`);
|
|
227
|
+
}
|
|
228
|
+
const failures = report.cases.filter((c) => c.status === 'failed');
|
|
229
|
+
if (failures.length > 0) {
|
|
230
|
+
lines.push('', 'Failures:');
|
|
231
|
+
for (const failure of failures) {
|
|
232
|
+
lines.push(` ✗ ${failure.case} (${failure.mean.toFixed(2)})`);
|
|
233
|
+
if (failure.failedGates && failure.failedGates.length > 0) {
|
|
234
|
+
// First line under the case, because it says WHY this failed.
|
|
235
|
+
// A mean of 0.75 next to four scores leaves the reader
|
|
236
|
+
// guessing which one mattered.
|
|
237
|
+
lines.push(` gate missed: ${failure.failedGates.join(', ')}`);
|
|
238
|
+
}
|
|
239
|
+
for (const [name, score] of Object.entries(failure.scores)) {
|
|
240
|
+
if (score.unavailable === true) {
|
|
241
|
+
lines.push(` ${name}: not judged — ${score.reason}`);
|
|
242
|
+
}
|
|
243
|
+
else if (score.score < 1) {
|
|
244
|
+
lines.push(` ${name}: ${score.reason}`);
|
|
245
|
+
}
|
|
246
|
+
}
|
|
247
|
+
if (failure.run.error)
|
|
248
|
+
lines.push(` error: ${failure.run.error}`);
|
|
249
|
+
}
|
|
250
|
+
}
|
|
251
|
+
const unjudged = report.cases.filter((c) => c.status === 'inconclusive');
|
|
252
|
+
if (unjudged.length > 0) {
|
|
253
|
+
lines.push('', 'Not judged:');
|
|
254
|
+
for (const item of unjudged) {
|
|
255
|
+
lines.push(` ? ${item.case}`);
|
|
256
|
+
for (const [name, score] of Object.entries(item.scores)) {
|
|
257
|
+
lines.push(` ${name}: ${score.reason}`);
|
|
258
|
+
}
|
|
259
|
+
}
|
|
260
|
+
}
|
|
261
|
+
return lines.join('\n');
|
|
262
|
+
}
|
|
263
|
+
//# sourceMappingURL=experiment.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"experiment.js","sourceRoot":"","sources":["../../src/eval/experiment.ts"],"names":[],"mappings":"AA2CA;;;;;;;;;;;GAWG;AACH,MAAM,CAAC,KAAK,UAAU,aAAa,CAClC,MAAgC;IAEhC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAA;IAC5B,MAAM,SAAS,GAAG,MAAM,CAAC,aAAa,IAAI,CAAC,CAAA;IAC3C,MAAM,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,MAAM,CAAC,WAAW,IAAI,CAAC,CAAC,CAAA;IAExD,MAAM,OAAO,GAAiB,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAA;IAC5D,IAAI,MAAM,GAAG,CAAC,CAAA;IAEd,MAAM,MAAM,GAAG,KAAK,IAAmB,EAAE;QACxC,SAAS,CAAC;YACT,MAAM,KAAK,GAAG,MAAM,EAAE,CAAA;YACtB,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;YACpC,IAAI,CAAC,QAAQ;gBAAE,OAAM;YAErB,MAAM,GAAG,GAAG,MAAM,WAAW,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAA;YAC/C,MAAM,OAAO,GAAG,QAAQ,CAAC,OAAO,IAAI,MAAM,CAAC,OAAO,CAAA;YAClD,MAAM,MAAM,GAA0B,EAAE,CAAA;YAExC,iEAAiE;YACjE,iEAAiE;YACjE,iDAAiD;YACjD,iEAAiE;YACjE,0DAA0D;YAC1D,6DAA6D;YAC7D,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;YAC9B,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;gBAC9B,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;oBAC3B,MAAM,IAAI,KAAK,CACd,0BAA0B,MAAM,CAAC,IAAI,eAAe,QAAQ,CAAC,IAAI,wKAAwK,CACzO,CAAA;gBACF,CAAC;gBACD,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;gBACrB,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,MAAM,SAAS,CAAC,MAAM,EAAE,GAAG,EAAE,QAAQ,CAAC,CAAA;YAC7D,CAAC;YAED,iEAAiE;YACjE,+DAA+D;YAC/D,2DAA2D;YAC3D,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC;iBAClC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,KAAK,IAAI,CAAC;iBACrC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAA;YACrB,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;YAExF,iEAAiE;YACjE,8DAA8D;YAC9D,2DAA2D;YAC3D,gEAAgE;YAChE,2DAA2D;YAC3D,8DAA8D;YAC9D,6DAA6D;YAC7D,uBAAuB;YACvB,MAAM,WAAW,GAAG,OAAO;iBACzB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,MAAM,CAAC;iBACpC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;gBACb,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;gBAC5B,IAAI,CAAC,KAAK,IAAI,KAAK,CAAC,WAAW,KAAK,IAAI;oBAAE,OAAO,KAAK,CAAA;gBACtD,OAAO,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,SAAS,IAAI,SAAS,CAAC,CAAA;YAChD,CAAC,CAAC;iBACD,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;YAEpB,MAAM,MAAM,GACX,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;gBACpD,CAAC,CAAC,cAAc;gBAChB,CAAC,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC;oBACvB,CAAC,CAAC,QAAQ;oBACV,CAAC,CAAC,IAAI,IAAI,SAAS;wBAClB,CAAC,CAAC,QAAQ;wBACV,CAAC,CAAC,QAAQ,CAAA;YACd,MAAM,MAAM,GAAe;gBAC1B,IAAI,EAAE,QAAQ,CAAC,IAAI;gBACnB,GAAG;gBACH,MAAM;gBACN,IAAI;gBACJ,MAAM;gBACN,MAAM,EAAE,MAAM,KAAK,QAAQ;gBAC3B,8DAA8D;gBAC9D,6DAA6D;gBAC7D,GAAG,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;aAClD,CAAA;YACD,OAAO,CAAC,KAAK,CAAC,GAAG,MAAM,CAAA;YACvB,MAAM,CAAC,YAAY,EAAE,CAAC,MAAM,CAAC,CAAA;QAC9B,CAAC;IACF,CAAC,CAAA;IAED,MAAM,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,WAAW,EAAE,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC,CAAA;IAE7F,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;IACvC,wEAAwE;IACxE,8DAA8D;IAC9D,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAA;IACjE,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;IAEjG,OAAO;QACN,IAAI,EAAE,MAAM,CAAC,IAAI;QACjB,KAAK,EAAE,OAAO;QACd,IAAI;QACJ,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC,MAAM;QAC3D,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC,MAAM;QAC3D,YAAY,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAC,MAAM;QACvE,QAAQ,EAAE,YAAY,CAAC,OAAO,CAAC;QAC/B,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS;KAClC,CAAA;AACF,CAAC;AAED;;;GAGG;AACH,KAAK,UAAU,WAAW,CACzB,MAAgC,EAChC,QAA0B;IAE1B,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAA;IAC5B,MAAM,UAAU,GAAG,IAAI,eAAe,EAAE,CAAA;IACxC,MAAM,SAAS,GAAG,MAAM,CAAC,SAAS,CAAA;IAClC,uEAAuE;IACvE,uEAAuE;IACvE,wEAAwE;IACxE,qEAAqE;IACrE,sCAAsC;IACtC,MAAM,KAAK,GACV,SAAS,KAAK,SAAS,IAAI,SAAS,GAAG,CAAC;QACvC,CAAC,CAAC,UAAU,CACV,GAAG,EAAE,CAAC,UAAU,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,wBAAwB,SAAS,IAAI,CAAC,CAAC,EACxE,SAAS,CACT;QACF,CAAC,CAAC,SAAS,CAAA;IAEb,MAAM,QAAQ,GACb,KAAK,KAAK,SAAS;QAClB,CAAC,CAAC,SAAS;QACX,CAAC,CAAC,IAAI,OAAO,CAAQ,CAAC,QAAQ,EAAE,MAAM,EAAE,EAAE;YACxC,UAAU,CAAC,MAAM,CAAC,gBAAgB,CAAC,OAAO,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,MAAM,CAAC,EAAE;gBACnF,IAAI,EAAE,IAAI;aACV,CAAC,CAAA;QACH,CAAC,CAAC,CAAA;IAEL,IAAI,CAAC;QACJ,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,EAAE,QAAQ,EAAE,UAAU,CAAC,MAAM,CAAC,CAAA;QACpE,8DAA8D;QAC9D,qDAAqD;QACrD,IAAI,QAAQ;YAAE,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,GAAE,CAAC,CAAC,CAAA;QAClC,OAAO,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;IAChE,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACd,OAAO;YACN,MAAM,EAAE,IAAI;YACZ,KAAK,EAAE,EAAE;YACT,SAAS,EAAE,EAAE;YACb,WAAW,EAAE,CAAC;YACd,YAAY,EAAE,CAAC;YACf,4DAA4D;YAC5D,6DAA6D;YAC7D,sBAAsB;YACtB,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS;YAClC,KAAK,EAAE,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC;SACvD,CAAA;IACF,CAAC;YAAS,CAAC;QACV,qEAAqE;QACrE,mEAAmE;QACnE,qDAAqD;QACrD,YAAY,CAAC,KAAK,CAAC,CAAA;IACpB,CAAC;AACF,CAAC;AAED,kEAAkE;AAClE,KAAK,UAAU,SAAS,CAAC,MAAc,EAAE,GAAY,EAAE,QAAkB;IACxE,qEAAqE;IACrE,qEAAqE;IACrE,2DAA2D;IAC3D,uEAAuE;IACvE,wEAAwE;IACxE,yEAAyE;IACzE,oDAAoD;IACpD,IAAI,GAAG,CAAC,KAAK,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,EAAE,KAAK,EAAE,CAAC,EAAE,MAAM,EAAE,eAAe,GAAG,CAAC,KAAK,EAAE,EAAE,OAAO,EAAE,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,EAAE,CAAA;IACvF,CAAC;IAED,IAAI,CAAC;QACJ,OAAO,MAAM,MAAM,CAAC,KAAK,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAA;IACzC,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACd,mEAAmE;QACnE,6DAA6D;QAC7D,kEAAkE;QAClE,qEAAqE;QACrE,kEAAkE;QAClE,yCAAyC;QACzC,OAAO;YACN,KAAK,EAAE,CAAC;YACR,WAAW,EAAE,IAAI;YACjB,MAAM,EAAE,WAAW,MAAM,CAAC,IAAI,+BAC7B,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAChD,EAAE;SACF,CAAA;IACF,CAAC;AACF,CAAC;AAED,SAAS,YAAY,CAAC,OAA8B;IACnD,MAAM,IAAI,GAAG,IAAI,GAAG,EAA4C,CAAA;IAChE,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC9B,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC;YAC3D,IAAI,KAAK,CAAC,WAAW,KAAK,IAAI;gBAAE,SAAQ;YACxC,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,CAAA;YACtD,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,KAAK,CAAA;YAC1B,KAAK,CAAC,KAAK,EAAE,CAAA;YACb,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAA;QACtB,CAAC;IACF,CAAC;IACD,MAAM,GAAG,GAA2B,EAAE,CAAA;IACtC,oEAAoE;IACpE,sEAAsE;IACtE,qBAAqB;IACrB,KAAK,MAAM,CAAC,IAAI,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,IAAI,IAAI,EAAE,CAAC;QAC7C,IAAI,KAAK,GAAG,CAAC;YAAE,GAAG,CAAC,IAAI,CAAC,GAAG,KAAK,GAAG,KAAK,CAAA;IACzC,CAAC;IACD,OAAO,GAAG,CAAA;AACX,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,YAAY,CAAC,MAAwB;IACpD,MAAM,KAAK,GAAa;QACvB,GAAG,MAAM,CAAC,IAAI,KAAK,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM,iBAAiB,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,MAAM,CAAC,UAAU,IAAI;QAC3H,EAAE;KACF,CAAA;IAED,sEAAsE;IACtE,oEAAoE;IACpE,+BAA+B;IAC/B,IAAI,MAAM,CAAC,YAAY,GAAG,CAAC,EAAE,CAAC;QAC7B,KAAK,CAAC,IAAI,CACT,KAAK,MAAM,CAAC,YAAY,QAAQ,MAAM,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,yDAAyD,EAC7H,EAAE,CACF,CAAA;IACF,CAAC;IAED,KAAK,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC5D,KAAK,CAAC,IAAI,CAAC,KAAK,IAAI,KAAK,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAA;IAC5C,CAAC;IAED,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAA;IAClE,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CAAC,EAAE,EAAE,WAAW,CAAC,CAAA;QAC3B,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;YAChC,KAAK,CAAC,IAAI,CAAC,OAAO,OAAO,CAAC,IAAI,KAAK,OAAO,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;YAC9D,IAAI,OAAO,CAAC,WAAW,IAAI,OAAO,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC3D,8DAA8D;gBAC9D,uDAAuD;gBACvD,+BAA+B;gBAC/B,KAAK,CAAC,IAAI,CAAC,sBAAsB,OAAO,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;YACnE,CAAC;YACD,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC5D,IAAI,KAAK,CAAC,WAAW,KAAK,IAAI,EAAE,CAAC;oBAChC,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,kBAAkB,KAAK,CAAC,MAAM,EAAE,CAAC,CAAA;gBAC1D,CAAC;qBAAM,IAAI,KAAK,CAAC,KAAK,GAAG,CAAC,EAAE,CAAC;oBAC5B,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,KAAK,KAAK,CAAC,MAAM,EAAE,CAAC,CAAA;gBAC7C,CAAC;YACF,CAAC;YACD,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK;gBAAE,KAAK,CAAC,IAAI,CAAC,gBAAgB,OAAO,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,CAAA;QACvE,CAAC;IACF,CAAC;IAED,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,KAAK,cAAc,CAAC,CAAA;IACxE,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CAAC,EAAE,EAAE,aAAa,CAAC,CAAA;QAC7B,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;YAC7B,KAAK,CAAC,IAAI,CAAC,OAAO,IAAI,CAAC,IAAI,EAAE,CAAC,CAAA;YAC9B,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;gBACzD,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,KAAK,KAAK,CAAC,MAAM,EAAE,CAAC,CAAA;YAC7C,CAAC;QACF,CAAC;IACF,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACxB,CAAC"}
|
|
@@ -0,0 +1,30 @@
|
|
|
1
|
+
import type { Run } from '../types/run/entity.js';
|
|
2
|
+
import type { EvalRun } from './types.js';
|
|
3
|
+
/**
|
|
4
|
+
* Project a finished `Run` into the shape scorers consume.
|
|
5
|
+
*
|
|
6
|
+
* This is the whole bridge between the harness and the runtime, and it is
|
|
7
|
+
* three lines of mapping because `Run.steps` now exists. Before that, a
|
|
8
|
+
* trajectory scorer would have had to correlate raw `RunEvent`s by
|
|
9
|
+
* iteration number and diff cumulative counters — which is why the step
|
|
10
|
+
* record had to land before the harness could.
|
|
11
|
+
*/
|
|
12
|
+
export declare function evalRunFromRun(run: Run, opts?: {
|
|
13
|
+
durationMs?: number;
|
|
14
|
+
}): EvalRun;
|
|
15
|
+
/**
|
|
16
|
+
* Drain a `query()` generator to its returned `Run`, then project it.
|
|
17
|
+
*
|
|
18
|
+
* The common case, so it does not need writing per suite:
|
|
19
|
+
*
|
|
20
|
+
* ```ts
|
|
21
|
+
* runExperiment({
|
|
22
|
+
* name: 'file-editing',
|
|
23
|
+
* cases,
|
|
24
|
+
* scorers: [trajectoryScorer(), completionScorer()],
|
|
25
|
+
* run: (input) => evalRunFromQuery(query({ provider, tools, messages: input, … })),
|
|
26
|
+
* })
|
|
27
|
+
* ```
|
|
28
|
+
*/
|
|
29
|
+
export declare function evalRunFromQuery(generator: AsyncGenerator<unknown, Run>): Promise<EvalRun>;
|
|
30
|
+
//# sourceMappingURL=from-run.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"from-run.d.ts","sourceRoot":"","sources":["../../src/eval/from-run.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,GAAG,EAAE,MAAM,wBAAwB,CAAA;AACjD,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,YAAY,CAAA;AAEzC;;;;;;;;GAQG;AACH,wBAAgB,cAAc,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,GAAE;IAAE,UAAU,CAAC,EAAE,MAAM,CAAA;CAAO,GAAG,OAAO,CAcpF;AAED;;;;;;;;;;;;;GAaG;AACH,wBAAsB,gBAAgB,CAAC,SAAS,EAAE,cAAc,CAAC,OAAO,EAAE,GAAG,CAAC,GAAG,OAAO,CAAC,OAAO,CAAC,CAKhG"}
|
|
@@ -0,0 +1,46 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Project a finished `Run` into the shape scorers consume.
|
|
3
|
+
*
|
|
4
|
+
* This is the whole bridge between the harness and the runtime, and it is
|
|
5
|
+
* three lines of mapping because `Run.steps` now exists. Before that, a
|
|
6
|
+
* trajectory scorer would have had to correlate raw `RunEvent`s by
|
|
7
|
+
* iteration number and diff cumulative counters — which is why the step
|
|
8
|
+
* record had to land before the harness could.
|
|
9
|
+
*/
|
|
10
|
+
export function evalRunFromRun(run, opts = {}) {
|
|
11
|
+
const steps = run.steps ?? [];
|
|
12
|
+
return {
|
|
13
|
+
output: run.result ?? null,
|
|
14
|
+
...(run.structuredOutput !== undefined ? { structuredOutput: run.structuredOutput } : {}),
|
|
15
|
+
steps,
|
|
16
|
+
// Flattened in call order across steps — the trajectory.
|
|
17
|
+
toolCalls: steps.flatMap((s) => s.toolCalls.map((c) => c.function.name)),
|
|
18
|
+
...(run.stopReason ? { stopReason: run.stopReason } : {}),
|
|
19
|
+
totalTokens: run.tokenUsage.totalTokens,
|
|
20
|
+
totalCostUsd: run.costInfo.totalCost,
|
|
21
|
+
durationMs: opts.durationMs ?? (run.endedAt ?? Date.now()) - run.startedAt,
|
|
22
|
+
...(run.lastError ? { error: run.lastError } : {}),
|
|
23
|
+
};
|
|
24
|
+
}
|
|
25
|
+
/**
|
|
26
|
+
* Drain a `query()` generator to its returned `Run`, then project it.
|
|
27
|
+
*
|
|
28
|
+
* The common case, so it does not need writing per suite:
|
|
29
|
+
*
|
|
30
|
+
* ```ts
|
|
31
|
+
* runExperiment({
|
|
32
|
+
* name: 'file-editing',
|
|
33
|
+
* cases,
|
|
34
|
+
* scorers: [trajectoryScorer(), completionScorer()],
|
|
35
|
+
* run: (input) => evalRunFromQuery(query({ provider, tools, messages: input, … })),
|
|
36
|
+
* })
|
|
37
|
+
* ```
|
|
38
|
+
*/
|
|
39
|
+
export async function evalRunFromQuery(generator) {
|
|
40
|
+
const startedAt = Date.now();
|
|
41
|
+
let next = await generator.next();
|
|
42
|
+
while (!next.done)
|
|
43
|
+
next = await generator.next();
|
|
44
|
+
return evalRunFromRun(next.value, { durationMs: Date.now() - startedAt });
|
|
45
|
+
}
|
|
46
|
+
//# sourceMappingURL=from-run.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"from-run.js","sourceRoot":"","sources":["../../src/eval/from-run.ts"],"names":[],"mappings":"AAGA;;;;;;;;GAQG;AACH,MAAM,UAAU,cAAc,CAAC,GAAQ,EAAE,OAAgC,EAAE;IAC1E,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,IAAI,EAAE,CAAA;IAC7B,OAAO;QACN,MAAM,EAAE,GAAG,CAAC,MAAM,IAAI,IAAI;QAC1B,GAAG,CAAC,GAAG,CAAC,gBAAgB,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,gBAAgB,EAAE,GAAG,CAAC,gBAAgB,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACzF,KAAK;QACL,yDAAyD;QACzD,SAAS,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QACxE,GAAG,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,GAAG,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACzD,WAAW,EAAE,GAAG,CAAC,UAAU,CAAC,WAAW;QACvC,YAAY,EAAE,GAAG,CAAC,QAAQ,CAAC,SAAS;QACpC,UAAU,EAAE,IAAI,CAAC,UAAU,IAAI,CAAC,GAAG,CAAC,OAAO,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC,GAAG,GAAG,CAAC,SAAS;QAC1E,GAAG,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAClD,CAAA;AACF,CAAC;AAED;;;;;;;;;;;;;GAaG;AACH,MAAM,CAAC,KAAK,UAAU,gBAAgB,CAAC,SAAuC;IAC7E,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAA;IAC5B,IAAI,IAAI,GAAG,MAAM,SAAS,CAAC,IAAI,EAAE,CAAA;IACjC,OAAO,CAAC,IAAI,CAAC,IAAI;QAAE,IAAI,GAAG,MAAM,SAAS,CAAC,IAAI,EAAE,CAAA;IAChD,OAAO,cAAc,CAAC,IAAI,CAAC,KAAK,EAAE,EAAE,UAAU,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,CAAC,CAAA;AAC1E,CAAC"}
|
|
@@ -0,0 +1,8 @@
|
|
|
1
|
+
export { evalRunFromQuery, evalRunFromRun } from './from-run.js';
|
|
2
|
+
export { formatReport, runExperiment } from './experiment.js';
|
|
3
|
+
export type { ExperimentConfig } from './experiment.js';
|
|
4
|
+
export { judgeScorer } from './judge.js';
|
|
5
|
+
export type { JudgeScorerConfig } from './judge.js';
|
|
6
|
+
export { completionScorer, containsScorer, customScorer, stepBudgetScorer, trajectoryScorer, } from './scorers.js';
|
|
7
|
+
export type { CaseResult, CaseStatus, EvalCase, EvalRun, ExperimentReport, Score, Scorer, } from './types.js';
|
|
8
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/eval/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,cAAc,EAAE,MAAM,eAAe,CAAA;AAChE,OAAO,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAA;AAC7D,YAAY,EAAE,gBAAgB,EAAE,MAAM,iBAAiB,CAAA;AACvD,OAAO,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;AACxC,YAAY,EAAE,iBAAiB,EAAE,MAAM,YAAY,CAAA;AACnD,OAAO,EACN,gBAAgB,EAChB,cAAc,EACd,YAAY,EACZ,gBAAgB,EAChB,gBAAgB,GAChB,MAAM,cAAc,CAAA;AACrB,YAAY,EACX,UAAU,EACV,UAAU,EACV,QAAQ,EACR,OAAO,EACP,gBAAgB,EAChB,KAAK,EACL,MAAM,GACN,MAAM,YAAY,CAAA"}
|
|
@@ -0,0 +1,5 @@
|
|
|
1
|
+
export { evalRunFromQuery, evalRunFromRun } from './from-run.js';
|
|
2
|
+
export { formatReport, runExperiment } from './experiment.js';
|
|
3
|
+
export { judgeScorer } from './judge.js';
|
|
4
|
+
export { completionScorer, containsScorer, customScorer, stepBudgetScorer, trajectoryScorer, } from './scorers.js';
|
|
5
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/eval/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,cAAc,EAAE,MAAM,eAAe,CAAA;AAChE,OAAO,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAA;AAE7D,OAAO,EAAE,WAAW,EAAE,MAAM,YAAY,CAAA;AAExC,OAAO,EACN,gBAAgB,EAChB,cAAc,EACd,YAAY,EACZ,gBAAgB,EAChB,gBAAgB,GAChB,MAAM,cAAc,CAAA"}
|