calibrated-agent-evaluation 0.1.0__tar.gz
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.
- calibrated_agent_evaluation-0.1.0/.gitignore +61 -0
- calibrated_agent_evaluation-0.1.0/CHANGELOG.md +35 -0
- calibrated_agent_evaluation-0.1.0/LICENSE +175 -0
- calibrated_agent_evaluation-0.1.0/NOTICE +2 -0
- calibrated_agent_evaluation-0.1.0/PKG-INFO +237 -0
- calibrated_agent_evaluation-0.1.0/README.md +163 -0
- calibrated_agent_evaluation-0.1.0/THIRD-PARTY-LICENSES +446 -0
- calibrated_agent_evaluation-0.1.0/pyproject.toml +402 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/__init__.py +16 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/__main__.py +11 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_conversation.py +52 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_errors.py +47 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_filelock.py +55 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_finite.py +121 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_install.py +174 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_model.py +84 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_provenance.py +239 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/__init__.py +132 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/bias-probes-1/compassion_fade.json +272 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/bias-probes-1/length.json +272 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/bias-probes-1/self_preference.json +380 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/bias-probes-1/sycophancy.json +284 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/bias-probes-2/compassion_fade.json +668 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/bias-probes-2/length.json +668 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/bias-probes-2/self_preference.json +938 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/bias-probes-2/sycophancy.json +698 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/bias-thresholds.json +11 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/gold.jsonl.gz +0 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/judge-v1.json +8 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/judge-v2.json +8 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/judge.cassette.json.gz +0 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/manifest.json +43 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/certify/traces.jsonl.gz +0 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/labels/gold.json +132 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/normalized/CLM-4471.json +162 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/normalized/CLM-4473.json +150 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/normalized/CLM-4475.json +151 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/normalized/CLM-4477.json +158 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/normalized/ORD-1001.json +71 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_quickstart/data/normalized/ORD-1002.json +71 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/_recordstore.py +314 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/__init__.py +379 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/adot.py +330 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/arms.py +807 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/audit_report.py +347 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/audit_types.py +536 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/bias.py +320 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/calibrate.py +162 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/code_interpreter.py +260 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/cost.py +162 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/datasets.py +506 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/datasets_codec.py +219 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/detectors/__init__.py +11 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/detectors/_payload.py +134 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/detectors/detect.py +457 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/detectors/models.py +767 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/detectors/namer.py +376 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/detectors/session.py +347 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/detectors/types.py +216 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/detectors/vocabulary.py +373 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/doctor.py +513 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/endpoint_label.py +31 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/errors.py +568 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/evaluate.py +707 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/evaluate_payload.py +475 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/evaluator_create.py +403 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/guards.py +127 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/ingest.py +151 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/judge.py +583 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/metrics.py +254 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/online.py +807 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/online_record.py +331 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/online_version.py +15 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/optimization.py +525 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/policy_spans.py +466 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/pricing.py +349 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/readback.py +440 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/readers.py +795 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/replay.py +695 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/run_manifest.py +190 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/run_spans.py +106 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/runtime.py +1055 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/runtime_boundary.py +798 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/runtime_cassette.py +382 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/runtime_errors.py +49 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/runtime_executor.py +985 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/runtime_probe.py +306 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/runtime_response.py +307 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/scores.py +147 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/session.py +250 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/settle.py +96 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/simulated_user.py +139 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/sources.py +324 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/agentcore/span_reader.py +399 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/attribution/__init__.py +90 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/attribution/agreement.py +107 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/attribution/codebook.py +94 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/attribution/definition.py +112 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/attribution/errors.py +34 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/attribution/labels.py +48 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/attribution/rates.py +86 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/attribution/relocation.py +139 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/attribution/types.py +207 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/attribution/walkback.py +190 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/backends/__init__.py +79 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/backends/_bounded.py +62 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/backends/_export_tap.py +86 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/backends/base.py +220 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/backends/dashboard.py +334 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/backends/factory.py +85 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/backends/langfuse.py +1061 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/backends/langfuse_foreign.py +210 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/backends/local.py +322 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/__init__.py +80 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/battery.py +284 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/errors.py +131 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/gate.py +103 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/loader.py +120 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/mitigate.py +23 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/probe.py +170 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/record.py +123 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/store.py +62 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/thresholds.py +52 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/bias/version.py +16 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/calibrate/__init__.py +162 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/calibrate/engine.py +570 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/calibrate/errors.py +179 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/calibrate/ledger.py +200 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/calibrate/pairing.py +137 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/calibrate/plan.py +411 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/calibrate/record.py +282 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/calibrate/store.py +67 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/__init__.py +31 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/agentcore_route.py +435 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/agentspec.py +43 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/args.py +1714 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/atoms.py +75 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/backends.py +168 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/config.py +146 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/consistency.py +461 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/conversation_route.py +107 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/dispatch.py +167 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/drift.py +206 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/errors.py +111 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/goldagreement.py +200 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/handlers.py +119 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/judgefile.py +233 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/judgekit.py +260 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/judging.py +213 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/pairing.py +324 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/parser.py +290 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/plain.py +558 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/present.py +104 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/proposals.py +58 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/records.py +655 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/scope.py +70 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/__init__.py +9 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/agreement.py +196 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/audit.py +860 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/bias.py +200 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/calibrate.py +466 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/calibrate_agentcore.py +311 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/calibrate_check.py +269 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/capture.py +52 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/cluster.py +248 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/codebook.py +110 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/dataset.py +353 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/dataset_run.py +481 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/define.py +79 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/doctor.py +259 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/dumpload.py +186 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/generate.py +679 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/improve.py +1049 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/ingest.py +434 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/judge.py +336 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/label.py +225 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/label_pairwise.py +111 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/online.py +359 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/online_read.py +739 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/plan.py +86 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/prefer.py +232 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/probe.py +656 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/probe_policy.py +305 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/quickstart.py +226 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/ratify.py +131 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/results.py +80 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/run.py +851 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/stages/score.py +395 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cli/taxonomy.py +120 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cluster/__init__.py +49 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cluster/codebook.py +188 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cluster/coverage.py +64 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cluster/draft.py +73 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cluster/embed.py +40 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cluster/errors.py +16 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cluster/group.py +38 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cluster/ratify_map.py +302 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cluster/saturation.py +243 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/cluster/types.py +87 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/dumpload/__init__.py +76 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/dumpload/io.py +576 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/dumpload/registry.py +107 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/dumpload/results.py +609 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/__init__.py +107 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/admission.py +130 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/checklist.py +54 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/context.py +28 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/decision.py +67 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/display.py +118 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/errors.py +106 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/gate.py +80 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/judgedef.py +94 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/lint.py +214 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/store.py +101 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/enforce/version.py +13 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/__init__.py +137 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/base.py +37 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/bedrock_embedder.py +123 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/bedrock_renderer.py +64 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/benchmark_tasks.py +50 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/cassette.py +459 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/cassette_provider.py +157 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/errors.py +43 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/inspect_executor.py +221 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/judge_model.py +459 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/probe_bridge.py +207 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/sandbox.py +274 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/strands_capture.py +323 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/strands_cassette_model.py +132 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/strands_executor.py +281 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/execute/task_builders.py +142 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/flagship/__init__.py +110 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/flagship/errors.py +57 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/flagship/oracle.py +112 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/flagship/orchestrator.py +362 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/flagship/reference.py +525 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/flagship/report.py +199 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/flagship/runner.py +183 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/flagship/stages.py +292 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/flagship/store.py +62 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/improve/__init__.py +132 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/improve/confirm.py +95 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/improve/crew.py +344 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/improve/errors.py +75 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/improve/guard.py +244 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/improve/paired.py +230 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/improve/panel.py +196 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/improve/prereg.py +263 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/improve/report.py +311 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/improve/store.py +161 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/ingest/__init__.py +60 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/ingest/core.py +1537 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/ingest/detect.py +199 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/ingest/errors.py +38 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/ingest/sample.py +147 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/__init__.py +87 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/cache.py +140 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/errors.py +35 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/executable.py +92 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/levels.py +60 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/probe_runner.py +152 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/process.py +90 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/run_record.py +124 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/runtime.py +300 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/targets.py +163 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/types.py +56 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/verdict.py +83 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/verdict_schema.py +65 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/judges/version.py +14 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/__init__.py +185 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/_depth.py +34 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/agent_boundaries.py +65 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/content.py +177 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/content_placements.py +176 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/conventions/__init__.py +41 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/conventions/detect.py +121 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/conventions/dispatch.py +31 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/conventions/messages_adapter.py +269 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/conventions/openinference_adapter.py +222 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/conventions/openllmetry_adapter.py +260 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/conventions/pairing.py +118 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/conventions/registry.py +80 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/datasets/__init__.py +154 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/datasets/hf.py +78 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/datasets/jsonl.py +44 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/datasets/parquet.py +58 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/errors.py +28 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/guidance.py +70 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/inspect_adapter.py +259 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/otel_adapter.py +626 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/otlp_json.py +363 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/ownership.py +122 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/sources.py +212 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/strands_adapter.py +440 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/normalize/tree.py +284 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/__init__.py +156 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/adapter.py +328 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/authorize.py +54 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/battery.py +180 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/benchmarks.py +214 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/capability.py +36 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/connection.py +139 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/errors.py +82 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/fingerprint.py +20 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/framing.py +29 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/harness.py +263 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/policy.py +561 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/policy_battery.py +296 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/rate.py +72 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/record.py +85 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/redact.py +32 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/robustness.py +174 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/store.py +61 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/version.py +8 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/probe/ws_adapter.py +207 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/py.typed +0 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/__init__.py +150 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/addresses.py +55 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/agreement.py +255 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/app.py +161 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/context.py +254 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/errors.py +45 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/gate.py +36 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/guards.py +152 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/labeling.py +144 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/overlap.py +241 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/pairwise.py +339 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/prefill.py +46 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/proposals.py +43 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/record.py +332 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/relabel.py +36 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/renderers/__init__.py +49 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/renderers/base.py +97 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/renderers/default.py +107 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/renderers/trace_view.py +358 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/routes.py +465 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/session.py +269 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/session_log.py +66 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/static/hotkeys.js +385 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/static/pairwise.js +28 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/static/styles.css +130 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/static/vendor/README.md +13 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/static/vendor/htmx.min.js +1 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/review/store.py +99 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/__init__.py +143 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/aggregate.py +75 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/drift_report.py +333 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/errors.py +54 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/features.py +88 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/flags.py +40 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/gate.py +210 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/judge_check.py +161 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/multi.py +41 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/report.py +457 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/staleness.py +81 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/store.py +61 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/run/trend.py +161 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/sampling/__init__.py +49 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/sampling/discovery.py +161 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/sampling/errors.py +26 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/sampling/production.py +192 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/sampling/stats.py +44 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/sampling/types.py +48 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/sampling/validate.py +53 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scenarios/__init__.py +56 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scenarios/build.py +224 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scenarios/errors.py +14 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scenarios/labels.py +49 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scenarios/slicefile.py +101 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scenarios/types.py +212 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/schema/__init__.py +70 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/schema/errors.py +58 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/schema/serde.py +182 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/schema/types.py +174 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/schema/validate.py +213 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scorers/__init__.py +86 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scorers/base.py +47 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scorers/batch.py +51 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scorers/errors.py +25 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scorers/execution.py +475 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scorers/heuristic.py +312 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scorers/lane.py +28 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/scorers/passk.py +129 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/__init__.py +55 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/agreement.py +311 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/drift.py +353 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/kappa.py +106 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/mcnemar.py +178 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/multiplicity.py +60 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/proportions.py +137 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/rogan_gladen.py +342 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/studies/__init__.py +19 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/studies/baseline_arm.py +143 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/studies/coverage.py +325 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/studies/planner_floor.py +160 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/stats/studies/rate_sample_minimum.py +147 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/__init__.py +70 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/atomfile.py +86 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/bradley_terry.py +192 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/classify.py +68 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/elo.py +96 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/errors.py +20 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/gate.py +29 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/pairfile.py +91 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/preference.py +49 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/store.py +142 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/subjective/types.py +99 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/synth/__init__.py +50 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/synth/borrow/__init__.py +13 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/synth/borrow/simulator.py +247 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/synth/conversation.py +460 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/synth/conversation_cassette.py +248 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/synth/errors.py +83 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/synth/generate.py +352 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/synth/guards.py +115 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/synth/risk.py +75 -0
- calibrated_agent_evaluation-0.1.0/src/calibrated_agent_evaluation/synth/types.py +78 -0
|
@@ -0,0 +1,61 @@
|
|
|
1
|
+
# Byte-compiled / optimized
|
|
2
|
+
__pycache__/
|
|
3
|
+
*.py[cod]
|
|
4
|
+
*$py.class
|
|
5
|
+
|
|
6
|
+
# Distribution / packaging
|
|
7
|
+
build/
|
|
8
|
+
dist/
|
|
9
|
+
*.egg-info/
|
|
10
|
+
*.egg
|
|
11
|
+
|
|
12
|
+
# Virtual environments
|
|
13
|
+
.venv/
|
|
14
|
+
venv/
|
|
15
|
+
env/
|
|
16
|
+
|
|
17
|
+
# uv
|
|
18
|
+
# (uv.lock IS committed; the environment is not)
|
|
19
|
+
|
|
20
|
+
# Test / coverage / type-check caches
|
|
21
|
+
.pytest_cache/
|
|
22
|
+
.ruff_cache/
|
|
23
|
+
.nox/
|
|
24
|
+
.tox/
|
|
25
|
+
.coverage
|
|
26
|
+
.coverage.*
|
|
27
|
+
coverage.xml
|
|
28
|
+
htmlcov/
|
|
29
|
+
.pyright/
|
|
30
|
+
.mypy_cache/
|
|
31
|
+
|
|
32
|
+
# Editors / OS
|
|
33
|
+
.idea/
|
|
34
|
+
.vscode/
|
|
35
|
+
.DS_Store
|
|
36
|
+
|
|
37
|
+
# Local AI-assistant configuration (kept out of the shared tree)
|
|
38
|
+
.claude/
|
|
39
|
+
.kiro/
|
|
40
|
+
CLAUDE.md
|
|
41
|
+
.mcp.json
|
|
42
|
+
|
|
43
|
+
# MkDocs documentation-site build output (generated; the sources under docs/site/ ARE committed)
|
|
44
|
+
docs/site/_site/
|
|
45
|
+
|
|
46
|
+
# Local backend store written by `just seed-fixture local` (generated, not source)
|
|
47
|
+
.calibrated-agent-evaluation/
|
|
48
|
+
# FastHTML session-signing secret, auto-created on first app start (a generated secret, never source)
|
|
49
|
+
.sesskey
|
|
50
|
+
.hypothesis/
|
|
51
|
+
|
|
52
|
+
# Inspect AI eval logs generated at runtime by the executor (the committed fixtures
|
|
53
|
+
# in fixtures/raw/inspect/*.eval are tracked explicitly via git add -f)
|
|
54
|
+
*.eval
|
|
55
|
+
!fixtures/raw/inspect/*.eval
|
|
56
|
+
|
|
57
|
+
.scratch/
|
|
58
|
+
|
|
59
|
+
# advisory lock files cassette writers hold while merging
|
|
60
|
+
.*.json.lock
|
|
61
|
+
checklist.md
|
|
@@ -0,0 +1,35 @@
|
|
|
1
|
+
# Changelog
|
|
2
|
+
|
|
3
|
+
All notable changes to this project are documented in this file.
|
|
4
|
+
|
|
5
|
+
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and this project
|
|
6
|
+
adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
|
|
7
|
+
|
|
8
|
+
## [0.1.0] - 2026-11-24
|
|
9
|
+
|
|
10
|
+
The first public release.
|
|
11
|
+
|
|
12
|
+
### Added
|
|
13
|
+
|
|
14
|
+
- A Python library and the `calibrated-agent-evaluation` CLI that take agent traces through one
|
|
15
|
+
pipeline: probe, define, generate, label, cluster, calibrate, bias, judge, run and improve.
|
|
16
|
+
- Trace ingestion from OpenTelemetry (OTLP JSON and protobuf, with the OpenInference and
|
|
17
|
+
OpenLLMetry conventions), Inspect AI logs, Strands agents, Langfuse, JSONL and Parquet datasets,
|
|
18
|
+
into a local SQLite/CSV store or an optional Langfuse backend.
|
|
19
|
+
- Judge calibration against human gold labels. A judge is certified only when the 95% lower bounds
|
|
20
|
+
on its true-positive and true-negative rates clear the gate, and each gold set is used once.
|
|
21
|
+
- Failure rates corrected for judge error (Rogan–Gladen) with an interval, a CI gate on each
|
|
22
|
+
failure mode's pass-rate lower bound, input-drift reports and multiple-comparison correction.
|
|
23
|
+
- A CALM bias battery for judges, Good–Turing saturation checks, failure-mode clustering and
|
|
24
|
+
sample-size planners.
|
|
25
|
+
- Before-and-after evaluation of a fix against a pre-registered target, with McNemar tests and a
|
|
26
|
+
Holm-corrected guard panel.
|
|
27
|
+
- A FastHTML review UI for human labeling, an inter-rater agreement gate, and pairwise preference
|
|
28
|
+
ranking for subjective qualities that never gates.
|
|
29
|
+
- An opt-in Amazon Bedrock AgentCore plug-in: account checks, trace ingest, a Runtime executor,
|
|
30
|
+
evaluator judges, online evaluation, dataset publishing and A/B test audits.
|
|
31
|
+
- A Claude Code skills plugin with five skills that drive the CLI.
|
|
32
|
+
- `calibrated-agent-evaluation quickstart`, which runs a bundled sample corpus end to end, offline
|
|
33
|
+
and with no keys.
|
|
34
|
+
|
|
35
|
+
[0.1.0]: https://github.com/awslabs/calibrated-agent-evaluation/releases/tag/v0.1.0
|
|
@@ -0,0 +1,175 @@
|
|
|
1
|
+
|
|
2
|
+
Apache License
|
|
3
|
+
Version 2.0, January 2004
|
|
4
|
+
http://www.apache.org/licenses/
|
|
5
|
+
|
|
6
|
+
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
|
|
7
|
+
|
|
8
|
+
1. Definitions.
|
|
9
|
+
|
|
10
|
+
"License" shall mean the terms and conditions for use, reproduction,
|
|
11
|
+
and distribution as defined by Sections 1 through 9 of this document.
|
|
12
|
+
|
|
13
|
+
"Licensor" shall mean the copyright owner or entity authorized by
|
|
14
|
+
the copyright owner that is granting the License.
|
|
15
|
+
|
|
16
|
+
"Legal Entity" shall mean the union of the acting entity and all
|
|
17
|
+
other entities that control, are controlled by, or are under common
|
|
18
|
+
control with that entity. For the purposes of this definition,
|
|
19
|
+
"control" means (i) the power, direct or indirect, to cause the
|
|
20
|
+
direction or management of such entity, whether by contract or
|
|
21
|
+
otherwise, or (ii) ownership of fifty percent (50%) or more of the
|
|
22
|
+
outstanding shares, or (iii) beneficial ownership of such entity.
|
|
23
|
+
|
|
24
|
+
"You" (or "Your") shall mean an individual or Legal Entity
|
|
25
|
+
exercising permissions granted by this License.
|
|
26
|
+
|
|
27
|
+
"Source" form shall mean the preferred form for making modifications,
|
|
28
|
+
including but not limited to software source code, documentation
|
|
29
|
+
source, and configuration files.
|
|
30
|
+
|
|
31
|
+
"Object" form shall mean any form resulting from mechanical
|
|
32
|
+
transformation or translation of a Source form, including but
|
|
33
|
+
not limited to compiled object code, generated documentation,
|
|
34
|
+
and conversions to other media types.
|
|
35
|
+
|
|
36
|
+
"Work" shall mean the work of authorship, whether in Source or
|
|
37
|
+
Object form, made available under the License, as indicated by a
|
|
38
|
+
copyright notice that is included in or attached to the work
|
|
39
|
+
(an example is provided in the Appendix below).
|
|
40
|
+
|
|
41
|
+
"Derivative Works" shall mean any work, whether in Source or Object
|
|
42
|
+
form, that is based on (or derived from) the Work and for which the
|
|
43
|
+
editorial revisions, annotations, elaborations, or other modifications
|
|
44
|
+
represent, as a whole, an original work of authorship. For the purposes
|
|
45
|
+
of this License, Derivative Works shall not include works that remain
|
|
46
|
+
separable from, or merely link (or bind by name) to the interfaces of,
|
|
47
|
+
the Work and Derivative Works thereof.
|
|
48
|
+
|
|
49
|
+
"Contribution" shall mean any work of authorship, including
|
|
50
|
+
the original version of the Work and any modifications or additions
|
|
51
|
+
to that Work or Derivative Works thereof, that is intentionally
|
|
52
|
+
submitted to Licensor for inclusion in the Work by the copyright owner
|
|
53
|
+
or by an individual or Legal Entity authorized to submit on behalf of
|
|
54
|
+
the copyright owner. For the purposes of this definition, "submitted"
|
|
55
|
+
means any form of electronic, verbal, or written communication sent
|
|
56
|
+
to the Licensor or its representatives, including but not limited to
|
|
57
|
+
communication on electronic mailing lists, source code control systems,
|
|
58
|
+
and issue tracking systems that are managed by, or on behalf of, the
|
|
59
|
+
Licensor for the purpose of discussing and improving the Work, but
|
|
60
|
+
excluding communication that is conspicuously marked or otherwise
|
|
61
|
+
designated in writing by the copyright owner as "Not a Contribution."
|
|
62
|
+
|
|
63
|
+
"Contributor" shall mean Licensor and any individual or Legal Entity
|
|
64
|
+
on behalf of whom a Contribution has been received by Licensor and
|
|
65
|
+
subsequently incorporated within the Work.
|
|
66
|
+
|
|
67
|
+
2. Grant of Copyright License. Subject to the terms and conditions of
|
|
68
|
+
this License, each Contributor hereby grants to You a perpetual,
|
|
69
|
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
|
70
|
+
copyright license to reproduce, prepare Derivative Works of,
|
|
71
|
+
publicly display, publicly perform, sublicense, and distribute the
|
|
72
|
+
Work and such Derivative Works in Source or Object form.
|
|
73
|
+
|
|
74
|
+
3. Grant of Patent License. Subject to the terms and conditions of
|
|
75
|
+
this License, each Contributor hereby grants to You a perpetual,
|
|
76
|
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
|
77
|
+
(except as stated in this section) patent license to make, have made,
|
|
78
|
+
use, offer to sell, sell, import, and otherwise transfer the Work,
|
|
79
|
+
where such license applies only to those patent claims licensable
|
|
80
|
+
by such Contributor that are necessarily infringed by their
|
|
81
|
+
Contribution(s) alone or by combination of their Contribution(s)
|
|
82
|
+
with the Work to which such Contribution(s) was submitted. If You
|
|
83
|
+
institute patent litigation against any entity (including a
|
|
84
|
+
cross-claim or counterclaim in a lawsuit) alleging that the Work
|
|
85
|
+
or a Contribution incorporated within the Work constitutes direct
|
|
86
|
+
or contributory patent infringement, then any patent licenses
|
|
87
|
+
granted to You under this License for that Work shall terminate
|
|
88
|
+
as of the date such litigation is filed.
|
|
89
|
+
|
|
90
|
+
4. Redistribution. You may reproduce and distribute copies of the
|
|
91
|
+
Work or Derivative Works thereof in any medium, with or without
|
|
92
|
+
modifications, and in Source or Object form, provided that You
|
|
93
|
+
meet the following conditions:
|
|
94
|
+
|
|
95
|
+
(a) You must give any other recipients of the Work or
|
|
96
|
+
Derivative Works a copy of this License; and
|
|
97
|
+
|
|
98
|
+
(b) You must cause any modified files to carry prominent notices
|
|
99
|
+
stating that You changed the files; and
|
|
100
|
+
|
|
101
|
+
(c) You must retain, in the Source form of any Derivative Works
|
|
102
|
+
that You distribute, all copyright, patent, trademark, and
|
|
103
|
+
attribution notices from the Source form of the Work,
|
|
104
|
+
excluding those notices that do not pertain to any part of
|
|
105
|
+
the Derivative Works; and
|
|
106
|
+
|
|
107
|
+
(d) If the Work includes a "NOTICE" text file as part of its
|
|
108
|
+
distribution, then any Derivative Works that You distribute must
|
|
109
|
+
include a readable copy of the attribution notices contained
|
|
110
|
+
within such NOTICE file, excluding those notices that do not
|
|
111
|
+
pertain to any part of the Derivative Works, in at least one
|
|
112
|
+
of the following places: within a NOTICE text file distributed
|
|
113
|
+
as part of the Derivative Works; within the Source form or
|
|
114
|
+
documentation, if provided along with the Derivative Works; or,
|
|
115
|
+
within a display generated by the Derivative Works, if and
|
|
116
|
+
wherever such third-party notices normally appear. The contents
|
|
117
|
+
of the NOTICE file are for informational purposes only and
|
|
118
|
+
do not modify the License. You may add Your own attribution
|
|
119
|
+
notices within Derivative Works that You distribute, alongside
|
|
120
|
+
or as an addendum to the NOTICE text from the Work, provided
|
|
121
|
+
that such additional attribution notices cannot be construed
|
|
122
|
+
as modifying the License.
|
|
123
|
+
|
|
124
|
+
You may add Your own copyright statement to Your modifications and
|
|
125
|
+
may provide additional or different license terms and conditions
|
|
126
|
+
for use, reproduction, or distribution of Your modifications, or
|
|
127
|
+
for any such Derivative Works as a whole, provided Your use,
|
|
128
|
+
reproduction, and distribution of the Work otherwise complies with
|
|
129
|
+
the conditions stated in this License.
|
|
130
|
+
|
|
131
|
+
5. Submission of Contributions. Unless You explicitly state otherwise,
|
|
132
|
+
any Contribution intentionally submitted for inclusion in the Work
|
|
133
|
+
by You to the Licensor shall be under the terms and conditions of
|
|
134
|
+
this License, without any additional terms or conditions.
|
|
135
|
+
Notwithstanding the above, nothing herein shall supersede or modify
|
|
136
|
+
the terms of any separate license agreement you may have executed
|
|
137
|
+
with Licensor regarding such Contributions.
|
|
138
|
+
|
|
139
|
+
6. Trademarks. This License does not grant permission to use the trade
|
|
140
|
+
names, trademarks, service marks, or product names of the Licensor,
|
|
141
|
+
except as required for reasonable and customary use in describing the
|
|
142
|
+
origin of the Work and reproducing the content of the NOTICE file.
|
|
143
|
+
|
|
144
|
+
7. Disclaimer of Warranty. Unless required by applicable law or
|
|
145
|
+
agreed to in writing, Licensor provides the Work (and each
|
|
146
|
+
Contributor provides its Contributions) on an "AS IS" BASIS,
|
|
147
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
|
148
|
+
implied, including, without limitation, any warranties or conditions
|
|
149
|
+
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
|
|
150
|
+
PARTICULAR PURPOSE. You are solely responsible for determining the
|
|
151
|
+
appropriateness of using or redistributing the Work and assume any
|
|
152
|
+
risks associated with Your exercise of permissions under this License.
|
|
153
|
+
|
|
154
|
+
8. Limitation of Liability. In no event and under no legal theory,
|
|
155
|
+
whether in tort (including negligence), contract, or otherwise,
|
|
156
|
+
unless required by applicable law (such as deliberate and grossly
|
|
157
|
+
negligent acts) or agreed to in writing, shall any Contributor be
|
|
158
|
+
liable to You for damages, including any direct, indirect, special,
|
|
159
|
+
incidental, or consequential damages of any character arising as a
|
|
160
|
+
result of this License or out of the use or inability to use the
|
|
161
|
+
Work (including but not limited to damages for loss of goodwill,
|
|
162
|
+
work stoppage, computer failure or malfunction, or any and all
|
|
163
|
+
other commercial damages or losses), even if such Contributor
|
|
164
|
+
has been advised of the possibility of such damages.
|
|
165
|
+
|
|
166
|
+
9. Accepting Warranty or Additional Liability. While redistributing
|
|
167
|
+
the Work or Derivative Works thereof, You may choose to offer,
|
|
168
|
+
and charge a fee for, acceptance of support, warranty, indemnity,
|
|
169
|
+
or other liability obligations and/or rights consistent with this
|
|
170
|
+
License. However, in accepting such obligations, You may act only
|
|
171
|
+
on Your own behalf and on Your sole responsibility, not on behalf
|
|
172
|
+
of any other Contributor, and only if You agree to indemnify,
|
|
173
|
+
defend, and hold each Contributor harmless for any liability
|
|
174
|
+
incurred by, or claims asserted against, such Contributor by reason
|
|
175
|
+
of your accepting any such warranty or additional liability.
|
|
@@ -0,0 +1,237 @@
|
|
|
1
|
+
Metadata-Version: 2.5
|
|
2
|
+
Name: calibrated-agent-evaluation
|
|
3
|
+
Version: 0.1.0
|
|
4
|
+
Summary: Error analysis and judge calibration for AI agents: certify an LLM judge against human labels, then gate CI on failure rates corrected for its error.
|
|
5
|
+
Project-URL: Homepage, https://github.com/awslabs/calibrated-agent-evaluation
|
|
6
|
+
Project-URL: Documentation, https://awslabs.github.io/calibrated-agent-evaluation/
|
|
7
|
+
Project-URL: Repository, https://github.com/awslabs/calibrated-agent-evaluation
|
|
8
|
+
Project-URL: Issues, https://github.com/awslabs/calibrated-agent-evaluation/issues
|
|
9
|
+
Project-URL: Changelog, https://github.com/awslabs/calibrated-agent-evaluation/blob/main/CHANGELOG.md
|
|
10
|
+
Author: Amazon Web Services
|
|
11
|
+
License-Expression: Apache-2.0
|
|
12
|
+
License-File: LICENSE
|
|
13
|
+
License-File: NOTICE
|
|
14
|
+
License-File: THIRD-PARTY-LICENSES
|
|
15
|
+
Keywords: agents,ai,error-analysis,evals,evaluation,inspect-ai,langfuse,llm
|
|
16
|
+
Classifier: Development Status :: 3 - Alpha
|
|
17
|
+
Classifier: Environment :: Console
|
|
18
|
+
Classifier: Intended Audience :: Developers
|
|
19
|
+
Classifier: License :: OSI Approved :: Apache Software License
|
|
20
|
+
Classifier: Programming Language :: Python :: 3
|
|
21
|
+
Classifier: Programming Language :: Python :: 3.11
|
|
22
|
+
Classifier: Programming Language :: Python :: 3.12
|
|
23
|
+
Classifier: Programming Language :: Python :: 3.13
|
|
24
|
+
Classifier: Programming Language :: Python :: 3.14
|
|
25
|
+
Classifier: Topic :: Software Development :: Testing
|
|
26
|
+
Classifier: Typing :: Typed
|
|
27
|
+
Requires-Python: >=3.11.5
|
|
28
|
+
Requires-Dist: pydantic<3,>=2.8
|
|
29
|
+
Provides-Extra: agentcore
|
|
30
|
+
Requires-Dist: bedrock-agentcore<2,>=1.24.2; extra == 'agentcore'
|
|
31
|
+
Requires-Dist: boto3<2,>=1.43.103; extra == 'agentcore'
|
|
32
|
+
Requires-Dist: botocore<2,>=1.43.103; extra == 'agentcore'
|
|
33
|
+
Requires-Dist: jinja2<4,>=3.1.6; extra == 'agentcore'
|
|
34
|
+
Requires-Dist: strands-agents-evals<1.7,>=1.6.0; extra == 'agentcore'
|
|
35
|
+
Requires-Dist: strands-agents<2,>=1.48; extra == 'agentcore'
|
|
36
|
+
Requires-Dist: urllib3!=2.2.0,<3,>=1.26.5; extra == 'agentcore'
|
|
37
|
+
Provides-Extra: bedrock
|
|
38
|
+
Requires-Dist: boto3<2,>=1.40; extra == 'bedrock'
|
|
39
|
+
Requires-Dist: inspect-ai<0.4,>=0.3.242; extra == 'bedrock'
|
|
40
|
+
Provides-Extra: datasets
|
|
41
|
+
Requires-Dist: pyarrow<25,>=17; extra == 'datasets'
|
|
42
|
+
Provides-Extra: hf
|
|
43
|
+
Requires-Dist: huggingface-hub<3,>=0.24; extra == 'hf'
|
|
44
|
+
Provides-Extra: inspect
|
|
45
|
+
Requires-Dist: inspect-ai<0.4,>=0.3.242; extra == 'inspect'
|
|
46
|
+
Provides-Extra: langfuse
|
|
47
|
+
Requires-Dist: langfuse<4,>=3.15; extra == 'langfuse'
|
|
48
|
+
Provides-Extra: otel
|
|
49
|
+
Requires-Dist: opentelemetry-proto<2,>=1.43; extra == 'otel'
|
|
50
|
+
Provides-Extra: probe
|
|
51
|
+
Requires-Dist: inspect-ai<0.3.288,>=0.3.242; extra == 'probe'
|
|
52
|
+
Requires-Dist: inspect-evals<0.15,>=0.14.3; extra == 'probe'
|
|
53
|
+
Provides-Extra: probe-ws
|
|
54
|
+
Requires-Dist: websockets<17,>=16; extra == 'probe-ws'
|
|
55
|
+
Provides-Extra: sampling
|
|
56
|
+
Requires-Dist: allpairspy<3,>=2.5; extra == 'sampling'
|
|
57
|
+
Provides-Extra: scorers
|
|
58
|
+
Requires-Dist: autoevals<1,>=0.3.0; extra == 'scorers'
|
|
59
|
+
Provides-Extra: stats
|
|
60
|
+
Requires-Dist: numpy<3,>=2.4; extra == 'stats'
|
|
61
|
+
Requires-Dist: scipy<2,>=1.15; extra == 'stats'
|
|
62
|
+
Requires-Dist: statsmodels<1,>=0.14; extra == 'stats'
|
|
63
|
+
Provides-Extra: strands
|
|
64
|
+
Requires-Dist: opentelemetry-exporter-otlp-proto-common<2,>=1.43; extra == 'strands'
|
|
65
|
+
Requires-Dist: opentelemetry-sdk<2,>=1.43; extra == 'strands'
|
|
66
|
+
Requires-Dist: strands-agents<2,>=1.48; extra == 'strands'
|
|
67
|
+
Provides-Extra: synth
|
|
68
|
+
Requires-Dist: strands-agents-evals<1.7,>=1.6.0; extra == 'synth'
|
|
69
|
+
Provides-Extra: ui
|
|
70
|
+
Requires-Dist: python-fasthtml<0.15,>=0.14; extra == 'ui'
|
|
71
|
+
Requires-Dist: starlette<2,>=1.3; extra == 'ui'
|
|
72
|
+
Requires-Dist: uvicorn<1,>=0.49; extra == 'ui'
|
|
73
|
+
Description-Content-Type: text/markdown
|
|
74
|
+
|
|
75
|
+
<p align="center">
|
|
76
|
+
<picture>
|
|
77
|
+
<source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/awslabs/calibrated-agent-evaluation/v0.1.0/docs/site/content/assets/readme/banner-dark.svg">
|
|
78
|
+
<img alt="Calibrated Agent Evaluation" src="https://raw.githubusercontent.com/awslabs/calibrated-agent-evaluation/v0.1.0/docs/site/content/assets/readme/banner-light.svg" width="100%">
|
|
79
|
+
</picture>
|
|
80
|
+
</p>
|
|
81
|
+
|
|
82
|
+
<h3 align="center">Find out how often your AI agent fails, with a number you can defend.</h3>
|
|
83
|
+
|
|
84
|
+
<p align="center">
|
|
85
|
+
<a href="https://pypi.org/project/calibrated-agent-evaluation/"><img alt="PyPI" src="https://img.shields.io/pypi/v/calibrated-agent-evaluation"></a>
|
|
86
|
+
<a href="https://pypi.org/project/calibrated-agent-evaluation/"><img alt="Python versions" src="https://img.shields.io/pypi/pyversions/calibrated-agent-evaluation"></a>
|
|
87
|
+
<a href="https://github.com/awslabs/calibrated-agent-evaluation/actions/workflows/ci.yml"><img alt="CI" src="https://img.shields.io/github/actions/workflow/status/awslabs/calibrated-agent-evaluation/ci.yml?branch=main&label=CI"></a>
|
|
88
|
+
<a href="https://github.com/awslabs/calibrated-agent-evaluation/blob/v0.1.0/LICENSE"><img alt="License: Apache-2.0" src="https://img.shields.io/github/license/awslabs/calibrated-agent-evaluation"></a>
|
|
89
|
+
</p>
|
|
90
|
+
|
|
91
|
+
<p align="center">
|
|
92
|
+
<a href="https://awslabs.github.io/calibrated-agent-evaluation/">Documentation</a> ·
|
|
93
|
+
<a href="https://awslabs.github.io/calibrated-agent-evaluation/getting-started/quickstart/">Quickstart</a> ·
|
|
94
|
+
<a href="https://awslabs.github.io/calibrated-agent-evaluation/guides/lifecycle/">Lifecycle guide</a> ·
|
|
95
|
+
<a href="https://awslabs.github.io/calibrated-agent-evaluation/cli/">CLI reference</a> ·
|
|
96
|
+
<a href="https://github.com/awslabs/calibrated-agent-evaluation/blob/v0.1.0/CONTRIBUTING.md">Contributing</a>
|
|
97
|
+
</p>
|
|
98
|
+
|
|
99
|
+
---
|
|
100
|
+
|
|
101
|
+
An LLM judge reports that 12% of your agent's runs fail. How wrong is the judge, and how wide is
|
|
102
|
+
that 12% once you count its mistakes? Calibrated Agent Evaluation answers both questions. It takes
|
|
103
|
+
your agent's traces through error analysis, certifies a judge against human labels, corrects the
|
|
104
|
+
measured failure rate for the judge's own error, and gates CI on the interval's lower bound.
|
|
105
|
+
|
|
106
|
+
It grades what the agent did (a normalized trace), so a custom loop, an LLM with MCP tools, a
|
|
107
|
+
Strands agent and a multi-agent crew all go through the same pipeline. It is a library and CLI you
|
|
108
|
+
run yourself, offline on a laptop or inside a corporate network. There is no hosted service.
|
|
109
|
+
|
|
110
|
+
## What it checks that a raw judge score skips
|
|
111
|
+
|
|
112
|
+
- **Judge certification.** A judge is certified only when the 95% lower bounds on both its
|
|
113
|
+
true-positive and true-negative rates reach 0.80, measured on gold labels that are used once.
|
|
114
|
+
- **Corrected failure rates.** The judge's measured rate is corrected for its error
|
|
115
|
+
(Rogan–Gladen), with an interval that carries the uncertainty of the human gold and the judge.
|
|
116
|
+
The correction is refused when the judge is too close to chance.
|
|
117
|
+
- **Enough labels.** Good–Turing coverage tells you when labeling has stopped finding new failure
|
|
118
|
+
modes, and a planner sizes the gold set and the CI sample before you start.
|
|
119
|
+
- **Fixes that held.** A fix is accepted against a target registered before the change, with
|
|
120
|
+
McNemar tests and a Holm-corrected panel of guard metrics.
|
|
121
|
+
- **Crews.** In a multi-agent run, the span that caused a failure is kept apart from the span
|
|
122
|
+
where it showed.
|
|
123
|
+
|
|
124
|
+
It does not label traces for you (a human writes the gold) and it never gates on subjective
|
|
125
|
+
qualities such as tone. `calibrated-agent-evaluation prefer` ranks those instead.
|
|
126
|
+
|
|
127
|
+
## Install
|
|
128
|
+
|
|
129
|
+
```bash
|
|
130
|
+
pip install calibrated-agent-evaluation
|
|
131
|
+
```
|
|
132
|
+
|
|
133
|
+
Or `uv add calibrated-agent-evaluation`, or `uv tool install calibrated-agent-evaluation` for the
|
|
134
|
+
CLI alone. Python 3.11.5 or later. The base install depends only on pydantic. Everything that needs
|
|
135
|
+
a third-party package is an opt-in extra:
|
|
136
|
+
|
|
137
|
+
```bash
|
|
138
|
+
pip install "calibrated-agent-evaluation[ui,stats,langfuse]"
|
|
139
|
+
```
|
|
140
|
+
|
|
141
|
+
| Extra | Adds |
|
|
142
|
+
| ----------- | --------------------------------------------------------------------- |
|
|
143
|
+
| `ui` | the trace-review web app for human labeling |
|
|
144
|
+
| `stats` | judge calibration, corrected rates and failure-mode clustering |
|
|
145
|
+
| `langfuse` | Langfuse as the trace store |
|
|
146
|
+
| `inspect` | Inspect AI logs, agents and judge models |
|
|
147
|
+
| `strands` | running and capturing Strands agents |
|
|
148
|
+
| `bedrock` | Amazon Bedrock input generation and embeddings |
|
|
149
|
+
| `agentcore` | the Amazon Bedrock AgentCore plug-in |
|
|
150
|
+
|
|
151
|
+
`calibrated-agent-evaluation doctor` lists what is installed. The full set is in the
|
|
152
|
+
[installation guide](https://awslabs.github.io/calibrated-agent-evaluation/getting-started/installation/).
|
|
153
|
+
|
|
154
|
+
## Quickstart
|
|
155
|
+
|
|
156
|
+
The bundled sample corpus runs end to end in a temporary directory, with no keys or network:
|
|
157
|
+
|
|
158
|
+
```console
|
|
159
|
+
$ calibrated-agent-evaluation quickstart
|
|
160
|
+
Seeded 6 agent traces + 9 gold labels into an offline local store
|
|
161
|
+
|
|
162
|
+
First-pass result: 5 of 6 traces have a labeled failure.
|
|
163
|
+
Failure modes found:
|
|
164
|
+
- trusted_broken_policy_tool: 2
|
|
165
|
+
- dropped_material_fact: 1
|
|
166
|
+
- misread_amount: 1
|
|
167
|
+
- vacancy_threshold_margin_mishandled: 1
|
|
168
|
+
```
|
|
169
|
+
|
|
170
|
+
`calibrated-agent-evaluation quickstart --certify` goes further. It seeds 848 traces and writes a
|
|
171
|
+
judge kit with recorded model replies, then prints the commands that certify the judge, run its
|
|
172
|
+
bias battery and gate CI, all offline. Then bring your own traces from
|
|
173
|
+
OpenTelemetry (OpenInference or OpenLLMetry), Inspect AI, Strands, Langfuse, JSONL or Parquet:
|
|
174
|
+
[choose your integration path](https://awslabs.github.io/calibrated-agent-evaluation/integrate/).
|
|
175
|
+
|
|
176
|
+
## How it works
|
|
177
|
+
|
|
178
|
+
<p align="center">
|
|
179
|
+
<picture>
|
|
180
|
+
<source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/awslabs/calibrated-agent-evaluation/v0.1.0/docs/site/content/assets/readme/pipeline-dark.svg">
|
|
181
|
+
<img alt="The pipeline in order: probe, define, generate, label and cluster for error analysis; calibrate and bias as the two trust gates; judge and run for scoring; improve to test a fix" src="https://raw.githubusercontent.com/awslabs/calibrated-agent-evaluation/v0.1.0/docs/site/content/assets/readme/pipeline-light.svg" width="100%">
|
|
182
|
+
</picture>
|
|
183
|
+
</p>
|
|
184
|
+
|
|
185
|
+
Each stage is one subcommand, run in this order. Error analysis comes first: you look at real
|
|
186
|
+
traces, label them, and name the failure modes before anything is scored. A judge then has to pass
|
|
187
|
+
both trust gates before `judge` will run it, and `run` fails the build when a failure mode's
|
|
188
|
+
pass-rate 95% lower bound falls below the threshold. The
|
|
189
|
+
[stage map](https://awslabs.github.io/calibrated-agent-evaluation/concepts/stage-map/) covers each
|
|
190
|
+
stage, and
|
|
191
|
+
[how the numbers work](https://awslabs.github.io/calibrated-agent-evaluation/concepts/how-the-numbers-work/)
|
|
192
|
+
explains the statistics, with every method cited.
|
|
193
|
+
|
|
194
|
+
## Skills plugin
|
|
195
|
+
|
|
196
|
+
`skills/` is a Claude Code and Kiro plugin that helps a coding agent drive the pipeline: it drafts
|
|
197
|
+
the taxonomy, the judge prompt or the review session, then hands off to the CLI command that
|
|
198
|
+
enforces the gate. Load it for one session with `claude --plugin-dir <path-to-checkout>`, or
|
|
199
|
+
install it from inside Claude Code:
|
|
200
|
+
|
|
201
|
+
```text
|
|
202
|
+
/plugin marketplace add <path-to-checkout>
|
|
203
|
+
/plugin install calibrated-agent-evaluation@calibrated-agent-evaluation
|
|
204
|
+
```
|
|
205
|
+
|
|
206
|
+
For Kiro, see [`skills/README.md`](https://github.com/awslabs/calibrated-agent-evaluation/blob/v0.1.0/skills/README.md).
|
|
207
|
+
|
|
208
|
+
## Documentation
|
|
209
|
+
|
|
210
|
+
- [Why calibrate](https://awslabs.github.io/calibrated-agent-evaluation/why/) and
|
|
211
|
+
[how it compares with alternatives](https://awslabs.github.io/calibrated-agent-evaluation/why/alternatives/)
|
|
212
|
+
- [The lifecycle guide](https://awslabs.github.io/calibrated-agent-evaluation/guides/lifecycle/),
|
|
213
|
+
from defining behavior to gating in production
|
|
214
|
+
- [Calibrate a judge](https://awslabs.github.io/calibrated-agent-evaluation/guides/calibrate-a-judge/)
|
|
215
|
+
and [gate a pull request in CI](https://awslabs.github.io/calibrated-agent-evaluation/guides/gate-in-ci/)
|
|
216
|
+
- [Methods and citations](https://awslabs.github.io/calibrated-agent-evaluation/concepts/methodology/)
|
|
217
|
+
- [Python API](https://awslabs.github.io/calibrated-agent-evaluation/reference/) and
|
|
218
|
+
[CLI](https://awslabs.github.io/calibrated-agent-evaluation/cli/) reference
|
|
219
|
+
|
|
220
|
+
The project is at 0.1.0. Before 1.0, a minor release may change the API
|
|
221
|
+
([versioning policy](https://awslabs.github.io/calibrated-agent-evaluation/versioning/)).
|
|
222
|
+
|
|
223
|
+
## Contributing
|
|
224
|
+
|
|
225
|
+
[CONTRIBUTING.md](https://github.com/awslabs/calibrated-agent-evaluation/blob/v0.1.0/CONTRIBUTING.md) covers the development setup, the code layout and the checks a
|
|
226
|
+
change must pass (`just check`). Coding agents should read [AGENTS.md](https://github.com/awslabs/calibrated-agent-evaluation/blob/v0.1.0/AGENTS.md).
|
|
227
|
+
|
|
228
|
+
## Security
|
|
229
|
+
|
|
230
|
+
See [CONTRIBUTING](https://github.com/awslabs/calibrated-agent-evaluation/blob/v0.1.0/CONTRIBUTING.md#security-issue-notifications) for how to report a security
|
|
231
|
+
issue, and [SECURITY.md](https://github.com/awslabs/calibrated-agent-evaluation/blob/v0.1.0/SECURITY.md) for the security policy and the security model.
|
|
232
|
+
|
|
233
|
+
## License
|
|
234
|
+
|
|
235
|
+
This project is licensed under the Apache-2.0 License. See [LICENSE](https://github.com/awslabs/calibrated-agent-evaluation/blob/v0.1.0/LICENSE) and
|
|
236
|
+
[NOTICE](https://github.com/awslabs/calibrated-agent-evaluation/blob/v0.1.0/NOTICE). Vendored code and third-party fixture data are listed with their copyright
|
|
237
|
+
notices and full license texts in [THIRD-PARTY-LICENSES](https://github.com/awslabs/calibrated-agent-evaluation/blob/v0.1.0/THIRD-PARTY-LICENSES).
|