agent-learning-kit 0.1.0__py3-none-any.whl
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.
- agent_learning_kit-0.1.0.dist-info/METADATA +381 -0
- agent_learning_kit-0.1.0.dist-info/RECORD +642 -0
- agent_learning_kit-0.1.0.dist-info/WHEEL +4 -0
- agent_learning_kit-0.1.0.dist-info/entry_points.txt +5 -0
- agent_learning_kit-0.1.0.dist-info/licenses/LICENSE +173 -0
- agent_learning_kit-0.1.0.dist-info/licenses/NOTICE +7 -0
- fi/__init__.py +5 -0
- fi/alk/__init__.py +57 -0
- fi/alk/_facade.py +31 -0
- fi/alk/_module_alias.py +68 -0
- fi/alk/_paths.py +14 -0
- fi/alk/_schema.py +522 -0
- fi/alk/actions.py +727 -0
- fi/alk/bench/__init__.py +517 -0
- fi/alk/bench/_codeexec.py +213 -0
- fi/alk/bench/_coding.py +215 -0
- fi/alk/bench/_docker.py +237 -0
- fi/alk/bench/_grader.py +286 -0
- fi/alk/bench/_pull.py +212 -0
- fi/alk/bench/_voice.py +147 -0
- fi/alk/capabilities.py +627 -0
- fi/alk/cli.py +6396 -0
- fi/alk/config.py +130 -0
- fi/alk/cua_loop.py +562 -0
- fi/alk/evals.py +2351 -0
- fi/alk/extensions.py +163 -0
- fi/alk/harness/ARCHITECTURE.md +231 -0
- fi/alk/harness/DESIGN.md +246 -0
- fi/alk/harness/ENVIRONMENT_CONFORMANCE.md +127 -0
- fi/alk/harness/HOW-IT-WORKS.md +297 -0
- fi/alk/harness/IMPLEMENTATION_AND_VALIDATION_STATUS.md +229 -0
- fi/alk/harness/README.md +417 -0
- fi/alk/harness/__init__.py +77 -0
- fi/alk/harness/__main__.py +3 -0
- fi/alk/harness/amend.py +312 -0
- fi/alk/harness/artifacts.py +319 -0
- fi/alk/harness/authoring_entrypoint.py +189 -0
- fi/alk/harness/authoring_runtime_validation.py +267 -0
- fi/alk/harness/backends/README.md +43 -0
- fi/alk/harness/backends/__init__.py +122 -0
- fi/alk/harness/backends/base.py +241 -0
- fi/alk/harness/backends/claude.py +211 -0
- fi/alk/harness/backends/files.py +182 -0
- fi/alk/harness/backends/vertex_gemini.py +457 -0
- fi/alk/harness/background_noise.py +95 -0
- fi/alk/harness/build.py +385 -0
- fi/alk/harness/bundle.py +593 -0
- fi/alk/harness/bundle_author_v2.py +1831 -0
- fi/alk/harness/bundle_v2.py +719 -0
- fi/alk/harness/call_runner.py +1440 -0
- fi/alk/harness/callback_http_adapter.py +111 -0
- fi/alk/harness/catalogue.py +287 -0
- fi/alk/harness/chat.py +428 -0
- fi/alk/harness/chat_call_runner.py +506 -0
- fi/alk/harness/checks.py +136 -0
- fi/alk/harness/cli.py +1354 -0
- fi/alk/harness/config.py +338 -0
- fi/alk/harness/contract.py +718 -0
- fi/alk/harness/credentials.py +674 -0
- fi/alk/harness/data/persona_vocabulary.json +111 -0
- fi/alk/harness/environment.py +99 -0
- fi/alk/harness/environment_plan.py +168 -0
- fi/alk/harness/events.py +125 -0
- fi/alk/harness/executor.py +304 -0
- fi/alk/harness/folder.py +234 -0
- fi/alk/harness/generated_runtime.py +815 -0
- fi/alk/harness/github.py +72 -0
- fi/alk/harness/hosted_authoring_entrypoint.py +183 -0
- fi/alk/harness/hosted_entrypoint.py +2402 -0
- fi/alk/harness/hosted_scheduler.py +2218 -0
- fi/alk/harness/job.py +426 -0
- fi/alk/harness/judge.py +184 -0
- fi/alk/harness/livekit_source.py +50 -0
- fi/alk/harness/livekit_tool_trace_bootstrap.py +71 -0
- fi/alk/harness/observability.py +208 -0
- fi/alk/harness/outbound.py +3252 -0
- fi/alk/harness/packaging.py +515 -0
- fi/alk/harness/persona_guides.py +157 -0
- fi/alk/harness/platform.py +692 -0
- fi/alk/harness/process_preflight.py +764 -0
- fi/alk/harness/process_runtime.py +5670 -0
- fi/alk/harness/prove.py +425 -0
- fi/alk/harness/provider_import.py +703 -0
- fi/alk/harness/provider_lifecycle.py +392 -0
- fi/alk/harness/provision.py +2896 -0
- fi/alk/harness/reception.py +147 -0
- fi/alk/harness/retell_chat_call_runner.py +373 -0
- fi/alk/harness/run/__init__.py +296 -0
- fi/alk/harness/run/alk.py +184 -0
- fi/alk/harness/run/call.py +162 -0
- fi/alk/harness/run/conversation.py +264 -0
- fi/alk/harness/run/data/voices_by_language_and_gender.json +693 -0
- fi/alk/harness/run/evidence.py +195 -0
- fi/alk/harness/run/grade.py +598 -0
- fi/alk/harness/run/live.py +297 -0
- fi/alk/harness/run/models.py +56 -0
- fi/alk/harness/run/platform_evals.py +227 -0
- fi/alk/harness/run/sdk_voice.py +130 -0
- fi/alk/harness/run/simulation.py +1209 -0
- fi/alk/harness/run/stage.py +91 -0
- fi/alk/harness/run/targets.py +508 -0
- fi/alk/harness/run/tools.py +601 -0
- fi/alk/harness/run/voice.py +340 -0
- fi/alk/harness/runtime.py +172 -0
- fi/alk/harness/sandbox_server.py +2011 -0
- fi/alk/harness/sandbox_worker.py +44 -0
- fi/alk/harness/scenario.py +1048 -0
- fi/alk/harness/scenario_source.py +879 -0
- fi/alk/harness/scenario_tools.py +1143 -0
- fi/alk/harness/scenarios.py +915 -0
- fi/alk/harness/secrets.py +168 -0
- fi/alk/harness/service_catalog.py +97 -0
- fi/alk/harness/session.py +391 -0
- fi/alk/harness/sessions.py +372 -0
- fi/alk/harness/simulator.py +76 -0
- fi/alk/harness/simulator_voice.py +928 -0
- fi/alk/harness/skills/build-environment/SKILL.md +538 -0
- fi/alk/harness/skills/harness.md +131 -0
- fi/alk/harness/skills/kinds/chat.md +48 -0
- fi/alk/harness/skills/kinds/voice-voicemail.md +63 -0
- fi/alk/harness/skills/kinds/voice.md +59 -0
- fi/alk/harness/skills/plan-suite/SKILL.md +103 -0
- fi/alk/harness/skills/provision-environment/SKILL.md +136 -0
- fi/alk/harness/skills/run-scenarios/SKILL.md +112 -0
- fi/alk/harness/skills/understand-agent/SKILL.md +251 -0
- fi/alk/harness/skills/write-scenarios/SKILL.md +606 -0
- fi/alk/harness/skills/write-scenarios/references/refusals.md +28 -0
- fi/alk/harness/skills/write-scenarios/references/world-api.md +92 -0
- fi/alk/harness/source_data_invariants.py +444 -0
- fi/alk/harness/source_tool_evidence.py +79 -0
- fi/alk/harness/sources.py +253 -0
- fi/alk/harness/spend.py +140 -0
- fi/alk/harness/tool_trace_proxy.py +104 -0
- fi/alk/harness/tools.py +1018 -0
- fi/alk/harness/understand.py +169 -0
- fi/alk/harness/voicemail_audio.py +74 -0
- fi/alk/harness/world/__init__.py +33 -0
- fi/alk/harness/world/errors.py +68 -0
- fi/alk/harness/world/expectations.py +91 -0
- fi/alk/harness/world/handle.py +538 -0
- fi/alk/harness/world/kinds.py +196 -0
- fi/alk/harness/world/mutate.py +186 -0
- fi/alk/harness/world/probe.py +413 -0
- fi/alk/harness/world/provision.py +511 -0
- fi/alk/harness/world/provisioned.py +191 -0
- fi/alk/harness/world/runtime.py +616 -0
- fi/alk/harness/world/snapshot.py +288 -0
- fi/alk/harness/world/stores/__init__.py +305 -0
- fi/alk/harness/world/stores/container.py +215 -0
- fi/alk/harness/world/stores/inprocess.py +346 -0
- fi/alk/harness/world/stores/postgres.py +481 -0
- fi/alk/harness/world/stores/prove.py +202 -0
- fi/alk/harness/world/stores/sqlite.py +245 -0
- fi/alk/harness/world/stores/written.py +182 -0
- fi/alk/harness/world/tools.py +1516 -0
- fi/alk/harness/world/workspace.py +144 -0
- fi/alk/image_loop.py +453 -0
- fi/alk/image_perturb.py +241 -0
- fi/alk/improve.py +274 -0
- fi/alk/live/__init__.py +154 -0
- fi/alk/live/_attribution.py +184 -0
- fi/alk/live/_capture.py +264 -0
- fi/alk/live/_codec.py +391 -0
- fi/alk/live/_contract.py +134 -0
- fi/alk/live/_loopback.py +316 -0
- fi/alk/live/_perturb.py +449 -0
- fi/alk/live/_runner.py +386 -0
- fi/alk/live/_stats.py +561 -0
- fi/alk/live/_transcript.py +240 -0
- fi/alk/live/_workers/__init__.py +9 -0
- fi/alk/live/_workers/a2a_worker.py +316 -0
- fi/alk/live/_workers/langgraph_worker.py +217 -0
- fi/alk/live/_workers/livekit_worker.py +207 -0
- fi/alk/live/_workers/mcp_loopback_server.py +46 -0
- fi/alk/live/_workers/mcp_worker.py +158 -0
- fi/alk/live/_workers/pipecat_worker.py +189 -0
- fi/alk/live/a2a_lane.py +138 -0
- fi/alk/live/langgraph_lane.py +339 -0
- fi/alk/live/livekit_lane.py +376 -0
- fi/alk/live/mcp_lane.py +172 -0
- fi/alk/live/pipecat_lane.py +341 -0
- fi/alk/live/voice_redteam.py +494 -0
- fi/alk/loss.py +306 -0
- fi/alk/optimize.py +36260 -0
- fi/alk/practice/__init__.py +51 -0
- fi/alk/practice/_assess.py +103 -0
- fi/alk/practice/_budget.py +81 -0
- fi/alk/practice/_calibrate.py +69 -0
- fi/alk/practice/_capstone.py +86 -0
- fi/alk/practice/_contract.py +91 -0
- fi/alk/practice/_diagnose.py +79 -0
- fi/alk/practice/_drill.py +196 -0
- fi/alk/practice/_experiment.py +720 -0
- fi/alk/practice/_schedule.py +102 -0
- fi/alk/practice/_store.py +194 -0
- fi/alk/practice/_trainer.py +245 -0
- fi/alk/practice/_update.py +125 -0
- fi/alk/redteam.py +2621 -0
- fi/alk/rewardhack.py +237 -0
- fi/alk/simulate.py +10351 -0
- fi/alk/studio/__init__.py +82 -0
- fi/alk/studio/_bias.py +314 -0
- fi/alk/studio/_calibration.py +522 -0
- fi/alk/studio/_coverage.py +262 -0
- fi/alk/studio/_download.py +665 -0
- fi/alk/studio/_fidelity_attack.py +114 -0
- fi/alk/studio/_generate.py +652 -0
- fi/alk/studio/_library.py +370 -0
- fi/alk/studio/_scan.py +134 -0
- fi/alk/studio/_upgrade.py +42 -0
- fi/alk/studio/_vendor.py +172 -0
- fi/alk/suite.py +4200 -0
- fi/alk/tasks.py +828 -0
- fi/alk/telemetry/__init__.py +149 -0
- fi/alk/telemetry/_contract.py +141 -0
- fi/alk/telemetry/_emit.py +182 -0
- fi/alk/telemetry/_ledger.py +296 -0
- fi/alk/telemetry/_queue.py +127 -0
- fi/alk/telemetry/_row.py +294 -0
- fi/alk/telemetry/_run.py +233 -0
- fi/alk/telemetry/_sync.py +193 -0
- fi/alk/telemetry/_url.py +119 -0
- fi/alk/trinity.py +49397 -0
- fi/alk/voice_loop.py +174 -0
- fi/api/__init__.py +1 -0
- fi/api/auth.py +137 -0
- fi/api/types.py +29 -0
- fi/cli/__init__.py +9 -0
- fi/cli/assertions/__init__.py +25 -0
- fi/cli/assertions/conditions.py +76 -0
- fi/cli/assertions/evaluator.py +286 -0
- fi/cli/assertions/exit_codes.py +20 -0
- fi/cli/assertions/parser.py +131 -0
- fi/cli/assertions/reporter.py +194 -0
- fi/cli/commands/__init__.py +9 -0
- fi/cli/commands/config.py +165 -0
- fi/cli/commands/export.py +208 -0
- fi/cli/commands/init.py +112 -0
- fi/cli/commands/list_cmd.py +213 -0
- fi/cli/commands/run.py +486 -0
- fi/cli/commands/validate.py +173 -0
- fi/cli/commands/view.py +424 -0
- fi/cli/config/__init__.py +6 -0
- fi/cli/config/defaults.py +206 -0
- fi/cli/config/loader.py +155 -0
- fi/cli/config/schema.py +174 -0
- fi/cli/main.py +78 -0
- fi/cli/output/__init__.py +6 -0
- fi/cli/output/formatters.py +106 -0
- fi/cli/output/reporters.py +46 -0
- fi/cli/storage/__init__.py +5 -0
- fi/cli/storage/run_history.py +249 -0
- fi/cli/utils/__init__.py +5 -0
- fi/cli/utils/console.py +44 -0
- fi/evals/__init__.py +131 -0
- fi/evals/autoeval/__init__.py +137 -0
- fi/evals/autoeval/analyzer.py +211 -0
- fi/evals/autoeval/config.py +244 -0
- fi/evals/autoeval/export.py +213 -0
- fi/evals/autoeval/interactive.py +283 -0
- fi/evals/autoeval/pipeline.py +625 -0
- fi/evals/autoeval/prompts.py +139 -0
- fi/evals/autoeval/recommender.py +242 -0
- fi/evals/autoeval/rules.py +589 -0
- fi/evals/autoeval/templates.py +299 -0
- fi/evals/autoeval/types.py +232 -0
- fi/evals/core/__init__.py +16 -0
- fi/evals/core/cloud_registry.py +184 -0
- fi/evals/core/engines.py +368 -0
- fi/evals/core/evaluate.py +319 -0
- fi/evals/core/judge_prompt.py +90 -0
- fi/evals/core/prompt_generator.py +83 -0
- fi/evals/core/registry.py +57 -0
- fi/evals/core/result.py +55 -0
- fi/evals/evaluator.py +721 -0
- fi/evals/execution.py +168 -0
- fi/evals/feedback/__init__.py +32 -0
- fi/evals/feedback/calibrator.py +160 -0
- fi/evals/feedback/collector.py +214 -0
- fi/evals/feedback/hooks.py +81 -0
- fi/evals/feedback/retriever.py +128 -0
- fi/evals/feedback/store.py +272 -0
- fi/evals/feedback/types.py +99 -0
- fi/evals/framework/README.md +79 -0
- fi/evals/framework/__init__.py +267 -0
- fi/evals/framework/backends/Dockerfile.eval-runner +33 -0
- fi/evals/framework/backends/__init__.py +99 -0
- fi/evals/framework/backends/_container.py +141 -0
- fi/evals/framework/backends/_utils.py +145 -0
- fi/evals/framework/backends/base.py +223 -0
- fi/evals/framework/backends/celery_backend.py +417 -0
- fi/evals/framework/backends/celery_worker.py +78 -0
- fi/evals/framework/backends/kubernetes_backend.py +665 -0
- fi/evals/framework/backends/ray_backend.py +521 -0
- fi/evals/framework/backends/temporal.py +350 -0
- fi/evals/framework/backends/temporal_worker.py +126 -0
- fi/evals/framework/backends/thread_pool.py +286 -0
- fi/evals/framework/context.py +258 -0
- fi/evals/framework/enrichment.py +306 -0
- fi/evals/framework/evals/__init__.py +68 -0
- fi/evals/framework/evals/agentic.py +399 -0
- fi/evals/framework/evals/builder.py +609 -0
- fi/evals/framework/evals/semantic.py +142 -0
- fi/evals/framework/evaluator.py +647 -0
- fi/evals/framework/evaluators/__init__.py +22 -0
- fi/evals/framework/evaluators/blocking.py +347 -0
- fi/evals/framework/evaluators/non_blocking.py +577 -0
- fi/evals/framework/propagation.py +421 -0
- fi/evals/framework/protocols.py +385 -0
- fi/evals/framework/registry.py +370 -0
- fi/evals/framework/resilience/__init__.py +150 -0
- fi/evals/framework/resilience/circuit_breaker.py +309 -0
- fi/evals/framework/resilience/degradation.py +355 -0
- fi/evals/framework/resilience/health.py +505 -0
- fi/evals/framework/resilience/rate_limiter.py +228 -0
- fi/evals/framework/resilience/retry.py +274 -0
- fi/evals/framework/resilience/types.py +288 -0
- fi/evals/framework/resilience/wrapper.py +433 -0
- fi/evals/framework/types.py +218 -0
- fi/evals/guardrails/README.md +915 -0
- fi/evals/guardrails/__init__.py +96 -0
- fi/evals/guardrails/backends/__init__.py +43 -0
- fi/evals/guardrails/backends/azure.py +361 -0
- fi/evals/guardrails/backends/base.py +88 -0
- fi/evals/guardrails/backends/generic_llm.py +163 -0
- fi/evals/guardrails/backends/granite.py +216 -0
- fi/evals/guardrails/backends/llamaguard.py +221 -0
- fi/evals/guardrails/backends/local_base.py +479 -0
- fi/evals/guardrails/backends/openai.py +365 -0
- fi/evals/guardrails/backends/qwen.py +170 -0
- fi/evals/guardrails/backends/shieldgemma.py +154 -0
- fi/evals/guardrails/backends/turing.py +235 -0
- fi/evals/guardrails/backends/vllm_client.py +321 -0
- fi/evals/guardrails/backends/wildguard.py +188 -0
- fi/evals/guardrails/base.py +888 -0
- fi/evals/guardrails/config.py +221 -0
- fi/evals/guardrails/discovery.py +243 -0
- fi/evals/guardrails/gateway.py +437 -0
- fi/evals/guardrails/registry.py +231 -0
- fi/evals/guardrails/scanners/__init__.py +127 -0
- fi/evals/guardrails/scanners/base.py +191 -0
- fi/evals/guardrails/scanners/code_injection.py +243 -0
- fi/evals/guardrails/scanners/eval_delegate.py +574 -0
- fi/evals/guardrails/scanners/invisible_chars.py +351 -0
- fi/evals/guardrails/scanners/jailbreak.py +412 -0
- fi/evals/guardrails/scanners/language.py +288 -0
- fi/evals/guardrails/scanners/pipeline.py +260 -0
- fi/evals/guardrails/scanners/regex.py +311 -0
- fi/evals/guardrails/scanners/secrets.py +274 -0
- fi/evals/guardrails/scanners/topics.py +649 -0
- fi/evals/guardrails/scanners/urls.py +341 -0
- fi/evals/guardrails/types.py +96 -0
- fi/evals/llm/__init__.py +3 -0
- fi/evals/llm/base_llm_provider.py +35 -0
- fi/evals/llm/providers/litellm.py +70 -0
- fi/evals/local/__init__.py +90 -0
- fi/evals/local/evaluator.py +690 -0
- fi/evals/local/execution_mode.py +121 -0
- fi/evals/local/llm.py +489 -0
- fi/evals/local/metrics/__init__.py +19 -0
- fi/evals/local/registry.py +360 -0
- fi/evals/manager.py +1018 -0
- fi/evals/manager_types.py +362 -0
- fi/evals/metrics/__init__.py +185 -0
- fi/evals/metrics/agents/__init__.py +74 -0
- fi/evals/metrics/agents/metrics.py +693 -0
- fi/evals/metrics/agents/report.py +36463 -0
- fi/evals/metrics/agents/types.py +160 -0
- fi/evals/metrics/base_llm_metric.py +111 -0
- fi/evals/metrics/base_metric.py +138 -0
- fi/evals/metrics/code_security/__init__.py +305 -0
- fi/evals/metrics/code_security/analyzer.py +985 -0
- fi/evals/metrics/code_security/benchmarks/__init__.py +73 -0
- fi/evals/metrics/code_security/benchmarks/builtin.py +750 -0
- fi/evals/metrics/code_security/benchmarks/loader.py +580 -0
- fi/evals/metrics/code_security/benchmarks/types.py +308 -0
- fi/evals/metrics/code_security/detectors/__init__.py +186 -0
- fi/evals/metrics/code_security/detectors/base.py +394 -0
- fi/evals/metrics/code_security/detectors/cryptography.py +345 -0
- fi/evals/metrics/code_security/detectors/injection.py +744 -0
- fi/evals/metrics/code_security/detectors/secrets.py +287 -0
- fi/evals/metrics/code_security/detectors/serialization.py +192 -0
- fi/evals/metrics/code_security/joint_metrics.py +588 -0
- fi/evals/metrics/code_security/judges/__init__.py +83 -0
- fi/evals/metrics/code_security/judges/base.py +238 -0
- fi/evals/metrics/code_security/judges/dual_judge.py +534 -0
- fi/evals/metrics/code_security/judges/llm_judge.py +301 -0
- fi/evals/metrics/code_security/judges/pattern_judge.py +515 -0
- fi/evals/metrics/code_security/metrics.py +388 -0
- fi/evals/metrics/code_security/modes/__init__.py +63 -0
- fi/evals/metrics/code_security/modes/adversarial.py +284 -0
- fi/evals/metrics/code_security/modes/autocomplete.py +198 -0
- fi/evals/metrics/code_security/modes/base.py +283 -0
- fi/evals/metrics/code_security/modes/instruct.py +253 -0
- fi/evals/metrics/code_security/modes/repair.py +230 -0
- fi/evals/metrics/code_security/reports/__init__.py +57 -0
- fi/evals/metrics/code_security/reports/generator.py +404 -0
- fi/evals/metrics/code_security/reports/leaderboard.py +509 -0
- fi/evals/metrics/code_security/types.py +534 -0
- fi/evals/metrics/function_calling/__init__.py +34 -0
- fi/evals/metrics/function_calling/metrics.py +573 -0
- fi/evals/metrics/function_calling/types.py +87 -0
- fi/evals/metrics/hallucination/__init__.py +54 -0
- fi/evals/metrics/hallucination/detector.py +149 -0
- fi/evals/metrics/hallucination/metrics.py +390 -0
- fi/evals/metrics/hallucination/nli.py +253 -0
- fi/evals/metrics/hallucination/sentinel.py +106 -0
- fi/evals/metrics/hallucination/types.py +132 -0
- fi/evals/metrics/heuristics/aggregation_metrics.py +85 -0
- fi/evals/metrics/heuristics/json_metrics.py +87 -0
- fi/evals/metrics/heuristics/similarity_metrics.py +375 -0
- fi/evals/metrics/heuristics/string_metrics.py +391 -0
- fi/evals/metrics/llm_as_judges/__init__.py +17 -0
- fi/evals/metrics/llm_as_judges/custom_judge/metric.py +112 -0
- fi/evals/metrics/llm_as_judges/custom_judge/prompts.py +26 -0
- fi/evals/metrics/llm_as_judges/types.py +48 -0
- fi/evals/metrics/rag/__init__.py +111 -0
- fi/evals/metrics/rag/advanced/__init__.py +14 -0
- fi/evals/metrics/rag/advanced/multi_hop.py +283 -0
- fi/evals/metrics/rag/advanced/source_attribution.py +344 -0
- fi/evals/metrics/rag/generation/__init__.py +17 -0
- fi/evals/metrics/rag/generation/answer_relevancy.py +176 -0
- fi/evals/metrics/rag/generation/context_utilization.py +245 -0
- fi/evals/metrics/rag/generation/faithfulness.py +241 -0
- fi/evals/metrics/rag/generation/groundedness.py +131 -0
- fi/evals/metrics/rag/rag_score.py +277 -0
- fi/evals/metrics/rag/retrieval/__init__.py +20 -0
- fi/evals/metrics/rag/retrieval/context_entity_recall.py +124 -0
- fi/evals/metrics/rag/retrieval/context_precision.py +158 -0
- fi/evals/metrics/rag/retrieval/context_recall.py +106 -0
- fi/evals/metrics/rag/retrieval/noise_sensitivity.py +163 -0
- fi/evals/metrics/rag/retrieval/ranking.py +261 -0
- fi/evals/metrics/rag/types.py +100 -0
- fi/evals/metrics/rag/utils/__init__.py +62 -0
- fi/evals/metrics/rag/utils/claims.py +189 -0
- fi/evals/metrics/rag/utils/entities.py +244 -0
- fi/evals/metrics/rag/utils/nli.py +92 -0
- fi/evals/metrics/rag/utils/similarity.py +345 -0
- fi/evals/metrics/structured/__init__.py +114 -0
- fi/evals/metrics/structured/field_completeness.py +313 -0
- fi/evals/metrics/structured/hierarchy_score.py +366 -0
- fi/evals/metrics/structured/json_validation.py +190 -0
- fi/evals/metrics/structured/schema_compliance.py +280 -0
- fi/evals/metrics/structured/structured_output_score.py +298 -0
- fi/evals/metrics/structured/types.py +108 -0
- fi/evals/metrics/structured/validators/__init__.py +30 -0
- fi/evals/metrics/structured/validators/base.py +189 -0
- fi/evals/metrics/structured/validators/json_validator.py +196 -0
- fi/evals/metrics/structured/validators/pydantic_validator.py +178 -0
- fi/evals/metrics/structured/validators/yaml_validator.py +248 -0
- fi/evals/otel/__init__.py +266 -0
- fi/evals/otel/config.py +400 -0
- fi/evals/otel/conventions.py +463 -0
- fi/evals/otel/enrichment.py +371 -0
- fi/evals/otel/instrumentors/__init__.py +140 -0
- fi/evals/otel/instrumentors/anthropic.py +517 -0
- fi/evals/otel/instrumentors/base.py +382 -0
- fi/evals/otel/instrumentors/openai.py +673 -0
- fi/evals/otel/processors/__init__.py +36 -0
- fi/evals/otel/processors/base.py +473 -0
- fi/evals/otel/processors/cost.py +445 -0
- fi/evals/otel/processors/evaluation.py +559 -0
- fi/evals/otel/processors/llm.py +462 -0
- fi/evals/otel/tracer.py +506 -0
- fi/evals/otel/types.py +232 -0
- fi/evals/otel_utils.py +23 -0
- fi/evals/protect.py +671 -0
- fi/evals/protect_input_adapter.py +154 -0
- fi/evals/streaming/__init__.py +88 -0
- fi/evals/streaming/buffer.py +213 -0
- fi/evals/streaming/evaluator.py +551 -0
- fi/evals/streaming/policy.py +307 -0
- fi/evals/streaming/scorers.py +368 -0
- fi/evals/streaming/types.py +238 -0
- fi/evals/templates.py +472 -0
- fi/evals/types.py +156 -0
- fi/opt/__init__.py +221 -0
- fi/opt/_objective_scoring.py +85 -0
- fi/opt/base/__init__.py +11 -0
- fi/opt/base/base_generator.py +33 -0
- fi/opt/base/base_mapper.py +26 -0
- fi/opt/base/base_optimizer.py +45 -0
- fi/opt/base/evaluator.py +211 -0
- fi/opt/components.py +3095 -0
- fi/opt/datamappers/__init__.py +3 -0
- fi/opt/datamappers/basic_mapper.py +40 -0
- fi/opt/deployment.py +1021 -0
- fi/opt/evidence.py +4332 -0
- fi/opt/generators/__init__.py +3 -0
- fi/opt/generators/litellm.py +66 -0
- fi/opt/integrations/__init__.py +23 -0
- fi/opt/integrations/generative_suite.py +410 -0
- fi/opt/integrations/simulate.py +1313 -0
- fi/opt/mutations.py +771 -0
- fi/opt/observability.py +4639 -0
- fi/opt/optimizer_trace.py +889 -0
- fi/opt/optimizers/__init__.py +80 -0
- fi/opt/optimizers/agent.py +331 -0
- fi/opt/optimizers/agent_bandit.py +392 -0
- fi/opt/optimizers/agent_curriculum.py +635 -0
- fi/opt/optimizers/agent_evolution.py +894 -0
- fi/opt/optimizers/agent_feedback.py +1863 -0
- fi/opt/optimizers/agent_pareto.py +547 -0
- fi/opt/optimizers/agent_social_memory.py +1113 -0
- fi/opt/optimizers/agent_tpe.py +321 -0
- fi/opt/optimizers/bayesian_search.py +449 -0
- fi/opt/optimizers/council.py +2075 -0
- fi/opt/optimizers/futureagi_replay.py +799 -0
- fi/opt/optimizers/gepa.py +322 -0
- fi/opt/optimizers/metaprompt.py +243 -0
- fi/opt/optimizers/promptwizard.py +417 -0
- fi/opt/optimizers/protegi.py +329 -0
- fi/opt/optimizers/random_search.py +224 -0
- fi/opt/research.py +518 -0
- fi/opt/simulation.py +260 -0
- fi/opt/targets.py +232 -0
- fi/opt/types.py +66 -0
- fi/opt/utils/__init__.py +4 -0
- fi/opt/utils/early_stopping.py +266 -0
- fi/opt/utils/setup_logging.py +82 -0
- fi/simulate/__init__.py +540 -0
- fi/simulate/_hashing.py +35 -0
- fi/simulate/_logging.py +10 -0
- fi/simulate/adapters.py +87 -0
- fi/simulate/agent/__init__.py +120 -0
- fi/simulate/agent/browser.py +658 -0
- fi/simulate/agent/definition.py +587 -0
- fi/simulate/agent/frameworks.py +3528 -0
- fi/simulate/agent/generic.py +8286 -0
- fi/simulate/agent/import_probe.py +227 -0
- fi/simulate/agent/memory.py +905 -0
- fi/simulate/agent/mocks.py +101 -0
- fi/simulate/agent/multi_agent.py +361 -0
- fi/simulate/agent/orchestration.py +903 -0
- fi/simulate/agent/realtime.py +665 -0
- fi/simulate/agent/wrapper.py +99 -0
- fi/simulate/agent/wrappers/__init__.py +18 -0
- fi/simulate/agent/wrappers/anthropic.py +62 -0
- fi/simulate/agent/wrappers/gemini.py +65 -0
- fi/simulate/agent/wrappers/http.py +404 -0
- fi/simulate/agent/wrappers/langchain.py +80 -0
- fi/simulate/agent/wrappers/openai.py +75 -0
- fi/simulate/agent/wrappers/websocket.py +326 -0
- fi/simulate/artifacts/__init__.py +11 -0
- fi/simulate/artifacts/manifest.py +62 -0
- fi/simulate/cli.py +20560 -0
- fi/simulate/endpoints/__init__.py +45 -0
- fi/simulate/endpoints/_http_actor.py +73 -0
- fi/simulate/endpoints/actor_sources.py +243 -0
- fi/simulate/endpoints/base.py +107 -0
- fi/simulate/endpoints/builtins.py +10 -0
- fi/simulate/endpoints/callable.py +95 -0
- fi/simulate/endpoints/http.py +76 -0
- fi/simulate/endpoints/livekit.py +138 -0
- fi/simulate/endpoints/originators.py +132 -0
- fi/simulate/endpoints/profiles.py +348 -0
- fi/simulate/endpoints/retell.py +633 -0
- fi/simulate/endpoints/vapi.py +205 -0
- fi/simulate/endpoints/websocket.py +76 -0
- fi/simulate/environment.py +33026 -0
- fi/simulate/environments/__init__.py +11 -0
- fi/simulate/environments/base.py +73 -0
- fi/simulate/environments/chat.py +697 -0
- fi/simulate/environments/voice.py +212 -0
- fi/simulate/evaluation/__init__.py +4 -0
- fi/simulate/evaluation/ai_eval.py +227 -0
- fi/simulate/evidence/__init__.py +35 -0
- fi/simulate/evidence/base.py +59 -0
- fi/simulate/evidence/caller_observed.py +50 -0
- fi/simulate/evidence/livekit_instrumentation.py +51 -0
- fi/simulate/evidence/livekit_room.py +50 -0
- fi/simulate/evidence/otel.py +49 -0
- fi/simulate/evidence/providers/__init__.py +24 -0
- fi/simulate/evidence/providers/base.py +61 -0
- fi/simulate/evidence/providers/retell.py +376 -0
- fi/simulate/evidence/providers/vapi.py +426 -0
- fi/simulate/hosted/__init__.py +32 -0
- fi/simulate/hosted/child_entrypoint.py +306 -0
- fi/simulate/hosted/job.py +150 -0
- fi/simulate/hosted/targets.py +53 -0
- fi/simulate/instrumentation/__init__.py +5 -0
- fi/simulate/instrumentation/livekit/__init__.py +122 -0
- fi/simulate/manifest.py +1033 -0
- fi/simulate/matrix_cli.py +165 -0
- fi/simulate/realtime/__init__.py +40 -0
- fi/simulate/realtime/events.py +107 -0
- fi/simulate/realtime/media.py +61 -0
- fi/simulate/realtime/session.py +91 -0
- fi/simulate/recording/__init__.py +5 -0
- fi/simulate/recording/room_recorder.py +326 -0
- fi/simulate/registry.py +185 -0
- fi/simulate/results/__init__.py +9 -0
- fi/simulate/results/base.py +18 -0
- fi/simulate/results/filesystem.py +71 -0
- fi/simulate/results/futureagi.py +1340 -0
- fi/simulate/runtime/__init__.py +85 -0
- fi/simulate/runtime/capabilities.py +40 -0
- fi/simulate/runtime/events.py +63 -0
- fi/simulate/runtime/failures.py +25 -0
- fi/simulate/runtime/ids.py +34 -0
- fi/simulate/runtime/plan.py +70 -0
- fi/simulate/runtime/planner.py +102 -0
- fi/simulate/runtime/report.py +174 -0
- fi/simulate/runtime/run.py +75 -0
- fi/simulate/runtime/runner.py +333 -0
- fi/simulate/runtime/spec.py +186 -0
- fi/simulate/simulation/__init__.py +30 -0
- fi/simulate/simulation/behavior_policy.py +425 -0
- fi/simulate/simulation/bridge/__init__.py +9 -0
- fi/simulate/simulation/bridge/audio.py +29 -0
- fi/simulate/simulation/bridge/connector.py +46 -0
- fi/simulate/simulation/bridge/livekit.py +252 -0
- fi/simulate/simulation/bridge/retell.py +188 -0
- fi/simulate/simulation/bridge/vapi.py +177 -0
- fi/simulate/simulation/contract.py +419 -0
- fi/simulate/simulation/engines/__init__.py +12 -0
- fi/simulate/simulation/engines/base.py +21 -0
- fi/simulate/simulation/engines/cloud.py +517 -0
- fi/simulate/simulation/engines/livekit.py +4167 -0
- fi/simulate/simulation/engines/local_text.py +89 -0
- fi/simulate/simulation/fidelity.py +374 -0
- fi/simulate/simulation/gemini_tts_stream.py +110 -0
- fi/simulate/simulation/generator.py +91 -0
- fi/simulate/simulation/goal_machine.py +185 -0
- fi/simulate/simulation/livekit_models.py +467 -0
- fi/simulate/simulation/matrix.py +170 -0
- fi/simulate/simulation/models.py +279 -0
- fi/simulate/simulation/runner.py +153 -0
- fi/simulate/simulation/synthetic.py +880 -0
- fi/simulate/simulation/voice_prompt.py +502 -0
- fi/simulate/simulator/__init__.py +55 -0
- fi/simulate/simulator/builtins.py +53 -0
- fi/simulate/suite.py +1288 -0
- fi/simulate/utils/routes.py +164 -0
- fi/simulate/voice.py +225 -0
- fi/simulate/voice_cli.py +182 -0
- fi/utils/__init__.py +1 -0
- fi/utils/constants.py +14 -0
- fi/utils/errors.py +200 -0
- fi/utils/executor.py +26 -0
- fi/utils/routes.py +119 -0
- fi/utils/utils.py +17 -0
|
@@ -0,0 +1,426 @@
|
|
|
1
|
+
"""Vapi post-call evidence adapter.
|
|
2
|
+
|
|
3
|
+
Polls ``GET https://api.vapi.ai/call/{call_id}`` after the SDK's phone
|
|
4
|
+
leg finishes. Downloads the provider recording, extracts transcript,
|
|
5
|
+
tool calls, analysis, cost, and latency, and returns those as an
|
|
6
|
+
``EvidenceSourceSummary`` plus ``ArtifactManifestEntry`` rows. The
|
|
7
|
+
adapter never imports from ``futureagi/`` and reads credentials from
|
|
8
|
+
env only.
|
|
9
|
+
"""
|
|
10
|
+
|
|
11
|
+
from __future__ import annotations
|
|
12
|
+
|
|
13
|
+
import asyncio
|
|
14
|
+
import logging
|
|
15
|
+
import os
|
|
16
|
+
import uuid
|
|
17
|
+
from datetime import timedelta, timezone
|
|
18
|
+
from typing import Any
|
|
19
|
+
|
|
20
|
+
import httpx
|
|
21
|
+
|
|
22
|
+
from fi.simulate.agent.definition import ProviderEvidenceConfig
|
|
23
|
+
from fi.simulate.artifacts.manifest import ArtifactManifestEntry
|
|
24
|
+
from fi.simulate.evidence.base import (
|
|
25
|
+
EvidenceCapabilities,
|
|
26
|
+
EvidenceClass,
|
|
27
|
+
EvidenceSourceSummary,
|
|
28
|
+
)
|
|
29
|
+
|
|
30
|
+
from .base import (
|
|
31
|
+
EvidenceContext,
|
|
32
|
+
ProviderConfigError,
|
|
33
|
+
ProviderFetchResult,
|
|
34
|
+
checksum_bytes,
|
|
35
|
+
coerce_json,
|
|
36
|
+
redact_phone,
|
|
37
|
+
)
|
|
38
|
+
|
|
39
|
+
_VAPI_API_BASE = "https://api.vapi.ai"
|
|
40
|
+
_TERMINAL_STATUSES = {"ended", "failed", "cancelled"}
|
|
41
|
+
_ADAPTER = "vapi"
|
|
42
|
+
_RECORDING_ENDPOINTS = {
|
|
43
|
+
"combined": "mono-recording",
|
|
44
|
+
"assistant": "assistant-recording",
|
|
45
|
+
"customer": "customer-recording",
|
|
46
|
+
"stereo": "stereo-recording",
|
|
47
|
+
}
|
|
48
|
+
logger = logging.getLogger(__name__)
|
|
49
|
+
|
|
50
|
+
|
|
51
|
+
class VapiEvidenceSource:
|
|
52
|
+
capabilities = EvidenceCapabilities(
|
|
53
|
+
transcript=True,
|
|
54
|
+
audio=True,
|
|
55
|
+
tool_calls=True,
|
|
56
|
+
tool_results=True,
|
|
57
|
+
usage=True,
|
|
58
|
+
internal_latency=True,
|
|
59
|
+
configuration_snapshot=False,
|
|
60
|
+
)
|
|
61
|
+
|
|
62
|
+
def __init__(
|
|
63
|
+
self,
|
|
64
|
+
config: ProviderEvidenceConfig,
|
|
65
|
+
*,
|
|
66
|
+
api_key: str | None = None,
|
|
67
|
+
api_base_url: str | None = None,
|
|
68
|
+
client: httpx.AsyncClient | None = None,
|
|
69
|
+
) -> None:
|
|
70
|
+
if config.provider != "vapi":
|
|
71
|
+
raise ProviderConfigError(
|
|
72
|
+
f"vapi adapter requires provider='vapi', got {config.provider!r}"
|
|
73
|
+
)
|
|
74
|
+
self._config = config
|
|
75
|
+
self._api_key = api_key or os.environ.get("VAPI_API_KEY")
|
|
76
|
+
if not self._api_key:
|
|
77
|
+
raise ProviderConfigError("VAPI_API_KEY is required for the Vapi adapter")
|
|
78
|
+
self._client = client or httpx.AsyncClient(
|
|
79
|
+
base_url=api_base_url or _VAPI_API_BASE,
|
|
80
|
+
headers={"Authorization": f"Bearer {self._api_key}"},
|
|
81
|
+
timeout=httpx.Timeout(30.0, connect=10.0),
|
|
82
|
+
)
|
|
83
|
+
self._owns_client = client is None
|
|
84
|
+
self._context: EvidenceContext | None = None
|
|
85
|
+
self._source_id = f"vapi:{uuid.uuid4().hex[:12]}"
|
|
86
|
+
|
|
87
|
+
async def connect(self, context: EvidenceContext) -> None:
|
|
88
|
+
self._context = context
|
|
89
|
+
|
|
90
|
+
async def fetch_final(self) -> ProviderFetchResult:
|
|
91
|
+
if self._context is None:
|
|
92
|
+
raise RuntimeError("vapi_adapter_not_connected")
|
|
93
|
+
context = self._context
|
|
94
|
+
call_id = context.call_id_hint
|
|
95
|
+
try:
|
|
96
|
+
if not call_id and self._config.call_id_source == "polling_window":
|
|
97
|
+
call_id = await self._locate_call_id()
|
|
98
|
+
if not call_id:
|
|
99
|
+
return self._unavailable("vapi_call_id_missing")
|
|
100
|
+
call_payload = await self._poll_call(call_id)
|
|
101
|
+
except httpx.HTTPError as exc:
|
|
102
|
+
return self._unavailable(
|
|
103
|
+
"vapi_fetch_failed",
|
|
104
|
+
error=type(exc).__name__,
|
|
105
|
+
)
|
|
106
|
+
artifacts = await self._download_recordings(call_payload, call_id)
|
|
107
|
+
summary = self._summarize(call_payload, call_id, artifacts)
|
|
108
|
+
return ProviderFetchResult(summary=summary, artifacts=artifacts)
|
|
109
|
+
|
|
110
|
+
async def close(self) -> None:
|
|
111
|
+
if self._owns_client:
|
|
112
|
+
await self._client.aclose()
|
|
113
|
+
|
|
114
|
+
async def _locate_call_id(self) -> str | None:
|
|
115
|
+
assert self._context is not None
|
|
116
|
+
window = self._config.polling_window_seconds
|
|
117
|
+
if not window:
|
|
118
|
+
return None
|
|
119
|
+
started = self._context.started_at.astimezone(timezone.utc)
|
|
120
|
+
params = {
|
|
121
|
+
"limit": 100,
|
|
122
|
+
"createdAtGt": (started - timedelta(seconds=window)).isoformat(),
|
|
123
|
+
"createdAtLt": (started + timedelta(seconds=window)).isoformat(),
|
|
124
|
+
}
|
|
125
|
+
deadline = (
|
|
126
|
+
asyncio.get_running_loop().time() + self._config.poll_deadline_seconds
|
|
127
|
+
)
|
|
128
|
+
while True:
|
|
129
|
+
response = await self._client.get("/call", params=params)
|
|
130
|
+
response.raise_for_status()
|
|
131
|
+
candidates = _select_vapi_calls(response.json())
|
|
132
|
+
matches = [
|
|
133
|
+
call
|
|
134
|
+
for call in candidates
|
|
135
|
+
if _matches_call_numbers(
|
|
136
|
+
call,
|
|
137
|
+
caller=self._context.caller_phone,
|
|
138
|
+
callee=self._context.callee_phone,
|
|
139
|
+
)
|
|
140
|
+
]
|
|
141
|
+
if matches:
|
|
142
|
+
matches.sort(
|
|
143
|
+
key=lambda call: str(
|
|
144
|
+
call.get("startedAt") or call.get("createdAt") or ""
|
|
145
|
+
),
|
|
146
|
+
reverse=True,
|
|
147
|
+
)
|
|
148
|
+
call_id = matches[0].get("id")
|
|
149
|
+
if call_id:
|
|
150
|
+
return str(call_id)
|
|
151
|
+
if asyncio.get_running_loop().time() >= deadline:
|
|
152
|
+
return None
|
|
153
|
+
await asyncio.sleep(self._config.poll_interval_seconds)
|
|
154
|
+
|
|
155
|
+
async def _poll_call(self, call_id: str) -> dict[str, Any]:
|
|
156
|
+
deadline = (
|
|
157
|
+
asyncio.get_running_loop().time() + self._config.poll_deadline_seconds
|
|
158
|
+
)
|
|
159
|
+
while True:
|
|
160
|
+
response = await self._client.get(f"/call/{call_id}")
|
|
161
|
+
response.raise_for_status()
|
|
162
|
+
payload = response.json()
|
|
163
|
+
status = str(payload.get("status") or "").lower()
|
|
164
|
+
if status in _TERMINAL_STATUSES:
|
|
165
|
+
return payload
|
|
166
|
+
if asyncio.get_running_loop().time() >= deadline:
|
|
167
|
+
return payload
|
|
168
|
+
await asyncio.sleep(self._config.poll_interval_seconds)
|
|
169
|
+
|
|
170
|
+
async def _download_recordings(
|
|
171
|
+
self,
|
|
172
|
+
payload: dict[str, Any],
|
|
173
|
+
call_id: str,
|
|
174
|
+
) -> list[ArtifactManifestEntry]:
|
|
175
|
+
assert self._context is not None
|
|
176
|
+
artifact_dir = self._context.case_directory / "provider"
|
|
177
|
+
artifact_dir.mkdir(parents=True, exist_ok=True)
|
|
178
|
+
urls = _extract_vapi_recording_urls(payload)
|
|
179
|
+
entries: list[ArtifactManifestEntry] = []
|
|
180
|
+
for label, url in urls.items():
|
|
181
|
+
if not url:
|
|
182
|
+
continue
|
|
183
|
+
try:
|
|
184
|
+
data = await self._get_recording(call_id, label)
|
|
185
|
+
except httpx.HTTPError as exc:
|
|
186
|
+
logger.warning(
|
|
187
|
+
"vapi recording download failed",
|
|
188
|
+
extra={"label": label, "error": type(exc).__name__},
|
|
189
|
+
)
|
|
190
|
+
continue
|
|
191
|
+
filename = f"vapi_{call_id}_{label}.wav"
|
|
192
|
+
path = artifact_dir / filename
|
|
193
|
+
path.write_bytes(data)
|
|
194
|
+
entries.append(
|
|
195
|
+
ArtifactManifestEntry(
|
|
196
|
+
artifact_id=f"{self._source_id}:{label}",
|
|
197
|
+
test_case_id=self._context.test_case_id,
|
|
198
|
+
type="audio",
|
|
199
|
+
path=str(path),
|
|
200
|
+
checksum=checksum_bytes(data),
|
|
201
|
+
size_bytes=len(data),
|
|
202
|
+
mime_type="audio/wav",
|
|
203
|
+
codec="pcm",
|
|
204
|
+
evidence_class=EvidenceClass.PROVIDER_REPORTED,
|
|
205
|
+
evidence_source_id=self._source_id,
|
|
206
|
+
leg_id=label,
|
|
207
|
+
metadata={"provider": "vapi", "recording_label": label},
|
|
208
|
+
)
|
|
209
|
+
)
|
|
210
|
+
return entries
|
|
211
|
+
|
|
212
|
+
async def _get_recording(self, call_id: str, label: str) -> bytes:
|
|
213
|
+
endpoint = _RECORDING_ENDPOINTS[label]
|
|
214
|
+
response = await self._client.get(
|
|
215
|
+
f"/call/{call_id}/{endpoint}",
|
|
216
|
+
follow_redirects=True,
|
|
217
|
+
)
|
|
218
|
+
response.raise_for_status()
|
|
219
|
+
return response.content
|
|
220
|
+
|
|
221
|
+
def _summarize(
|
|
222
|
+
self,
|
|
223
|
+
payload: dict[str, Any],
|
|
224
|
+
call_id: str,
|
|
225
|
+
artifacts: list[ArtifactManifestEntry],
|
|
226
|
+
) -> EvidenceSourceSummary:
|
|
227
|
+
assert self._context is not None
|
|
228
|
+
artifact = payload.get("artifact") or {}
|
|
229
|
+
performance = (
|
|
230
|
+
artifact.get("performanceMetrics")
|
|
231
|
+
or payload.get("performanceMetrics")
|
|
232
|
+
or {}
|
|
233
|
+
)
|
|
234
|
+
transcript_messages = artifact.get("messages") or payload.get("messages") or []
|
|
235
|
+
tool_calls = _extract_tool_calls(transcript_messages)
|
|
236
|
+
tool_results = _extract_tool_results(transcript_messages)
|
|
237
|
+
cost_summary = _cost_summary(payload)
|
|
238
|
+
metadata: dict[str, Any] = {
|
|
239
|
+
"provider": "vapi",
|
|
240
|
+
"call_id": call_id,
|
|
241
|
+
"status": payload.get("status"),
|
|
242
|
+
"ended_reason": payload.get("endedReason"),
|
|
243
|
+
"started_at": payload.get("startedAt"),
|
|
244
|
+
"ended_at": payload.get("endedAt"),
|
|
245
|
+
"tool_call_count": len(tool_calls),
|
|
246
|
+
"tool_calls": tool_calls or None,
|
|
247
|
+
"tool_result_count": len(tool_results),
|
|
248
|
+
"tool_results": tool_results or None,
|
|
249
|
+
"message_count": len(transcript_messages),
|
|
250
|
+
"latency": coerce_json(performance) or None,
|
|
251
|
+
"cost": cost_summary,
|
|
252
|
+
"analysis_summary": coerce_json(
|
|
253
|
+
(payload.get("analysis") or {}).get("summary")
|
|
254
|
+
),
|
|
255
|
+
"recording_labels": [entry.leg_id for entry in artifacts if entry.leg_id],
|
|
256
|
+
}
|
|
257
|
+
if self._context.caller_phone:
|
|
258
|
+
metadata["caller_phone"] = redact_phone(self._context.caller_phone)
|
|
259
|
+
if self._context.termination_source:
|
|
260
|
+
metadata["termination_source"] = self._context.termination_source
|
|
261
|
+
if payload.get("endedReason") == "call-deleted":
|
|
262
|
+
metadata["ended_reason_interpretation"] = "sdk_originator_teardown"
|
|
263
|
+
return EvidenceSourceSummary(
|
|
264
|
+
source_id=self._source_id,
|
|
265
|
+
adapter=_ADAPTER,
|
|
266
|
+
evidence_class=EvidenceClass.PROVIDER_REPORTED,
|
|
267
|
+
capabilities=self.capabilities,
|
|
268
|
+
available=str(payload.get("status") or "").lower() in _TERMINAL_STATUSES,
|
|
269
|
+
redactions=["auth", "phone_e164"],
|
|
270
|
+
metadata={k: v for k, v in metadata.items() if v is not None},
|
|
271
|
+
)
|
|
272
|
+
|
|
273
|
+
def _unavailable(self, code: str, **details: Any) -> ProviderFetchResult:
|
|
274
|
+
summary = EvidenceSourceSummary(
|
|
275
|
+
source_id=self._source_id,
|
|
276
|
+
adapter=_ADAPTER,
|
|
277
|
+
evidence_class=EvidenceClass.PROVIDER_REPORTED,
|
|
278
|
+
capabilities=self.capabilities,
|
|
279
|
+
available=False,
|
|
280
|
+
redactions=["auth", "phone_e164"],
|
|
281
|
+
metadata={"provider": "vapi", "reason": code, **coerce_json(details)},
|
|
282
|
+
)
|
|
283
|
+
return ProviderFetchResult(summary=summary, artifacts=[])
|
|
284
|
+
|
|
285
|
+
|
|
286
|
+
def _select_vapi_calls(payload: Any) -> list[dict[str, Any]]:
|
|
287
|
+
if isinstance(payload, list):
|
|
288
|
+
return [item for item in payload if isinstance(item, dict)]
|
|
289
|
+
if isinstance(payload, dict):
|
|
290
|
+
for key in ("calls", "results", "items", "data"):
|
|
291
|
+
value = payload.get(key)
|
|
292
|
+
if isinstance(value, list):
|
|
293
|
+
return [item for item in value if isinstance(item, dict)]
|
|
294
|
+
return []
|
|
295
|
+
|
|
296
|
+
|
|
297
|
+
def _matches_call_numbers(
|
|
298
|
+
payload: dict[str, Any],
|
|
299
|
+
*,
|
|
300
|
+
caller: str | None,
|
|
301
|
+
callee: str | None,
|
|
302
|
+
) -> bool:
|
|
303
|
+
if caller:
|
|
304
|
+
customer = payload.get("customer") or {}
|
|
305
|
+
if _normalized_phone(customer.get("number")) != _normalized_phone(caller):
|
|
306
|
+
return False
|
|
307
|
+
if callee:
|
|
308
|
+
phone_number = payload.get("phoneNumber") or {}
|
|
309
|
+
target_number = (
|
|
310
|
+
phone_number.get("number")
|
|
311
|
+
or payload.get("phoneNumberNumber")
|
|
312
|
+
or payload.get("toNumber")
|
|
313
|
+
)
|
|
314
|
+
if target_number and _normalized_phone(target_number) != _normalized_phone(
|
|
315
|
+
callee
|
|
316
|
+
):
|
|
317
|
+
return False
|
|
318
|
+
return True
|
|
319
|
+
|
|
320
|
+
|
|
321
|
+
def _normalized_phone(value: Any) -> str:
|
|
322
|
+
return "".join(character for character in str(value or "") if character.isdigit())
|
|
323
|
+
|
|
324
|
+
|
|
325
|
+
def _extract_vapi_recording_urls(payload: dict[str, Any]) -> dict[str, str | None]:
|
|
326
|
+
artifact = payload.get("artifact") or {}
|
|
327
|
+
recording = artifact.get("recording") or payload.get("recording") or {}
|
|
328
|
+
mono = recording.get("mono") if isinstance(recording, dict) else {}
|
|
329
|
+
urls: dict[str, str | None] = {
|
|
330
|
+
"combined": (
|
|
331
|
+
(mono or {}).get("combinedUrl") if isinstance(mono, dict) else None
|
|
332
|
+
)
|
|
333
|
+
or (recording if isinstance(recording, str) else None)
|
|
334
|
+
or artifact.get("recordingUrl")
|
|
335
|
+
or payload.get("recordingUrl"),
|
|
336
|
+
"assistant": (mono or {}).get("assistantUrl")
|
|
337
|
+
if isinstance(mono, dict)
|
|
338
|
+
else None,
|
|
339
|
+
"customer": (mono or {}).get("customerUrl") if isinstance(mono, dict) else None,
|
|
340
|
+
"stereo": (recording.get("stereoUrl") if isinstance(recording, dict) else None)
|
|
341
|
+
or artifact.get("stereoRecordingUrl")
|
|
342
|
+
or payload.get("stereoRecordingUrl"),
|
|
343
|
+
}
|
|
344
|
+
return urls
|
|
345
|
+
|
|
346
|
+
|
|
347
|
+
def _extract_tool_calls(messages: list[Any]) -> list[dict[str, Any]]:
|
|
348
|
+
calls: list[dict[str, Any]] = []
|
|
349
|
+
for entry in messages:
|
|
350
|
+
if not isinstance(entry, dict):
|
|
351
|
+
continue
|
|
352
|
+
tool_calls = (
|
|
353
|
+
entry.get("toolCalls")
|
|
354
|
+
or entry.get("tool_calls")
|
|
355
|
+
or entry.get("toolCallList")
|
|
356
|
+
or []
|
|
357
|
+
)
|
|
358
|
+
if not tool_calls and isinstance(entry.get("toolWithToolCallList"), list):
|
|
359
|
+
tool_calls = [
|
|
360
|
+
wrapped.get("toolCall")
|
|
361
|
+
for wrapped in entry["toolWithToolCallList"]
|
|
362
|
+
if isinstance(wrapped, dict)
|
|
363
|
+
and isinstance(wrapped.get("toolCall"), dict)
|
|
364
|
+
]
|
|
365
|
+
if not isinstance(tool_calls, list):
|
|
366
|
+
continue
|
|
367
|
+
for call in tool_calls:
|
|
368
|
+
if isinstance(call, dict):
|
|
369
|
+
function = call.get("function") or {}
|
|
370
|
+
calls.append(
|
|
371
|
+
{
|
|
372
|
+
key: value
|
|
373
|
+
for key, value in {
|
|
374
|
+
"id": call.get("id") or call.get("toolCallId"),
|
|
375
|
+
"name": function.get("name") or call.get("name"),
|
|
376
|
+
"arguments": function.get("arguments")
|
|
377
|
+
or call.get("parameters")
|
|
378
|
+
or call.get("arguments"),
|
|
379
|
+
}.items()
|
|
380
|
+
if value is not None
|
|
381
|
+
}
|
|
382
|
+
)
|
|
383
|
+
return calls
|
|
384
|
+
|
|
385
|
+
|
|
386
|
+
def _extract_tool_results(messages: list[Any]) -> list[dict[str, Any]]:
|
|
387
|
+
results: list[dict[str, Any]] = []
|
|
388
|
+
for entry in messages:
|
|
389
|
+
if not isinstance(entry, dict):
|
|
390
|
+
continue
|
|
391
|
+
candidates = (
|
|
392
|
+
entry.get("toolCallResultList")
|
|
393
|
+
or entry.get("tool_call_results")
|
|
394
|
+
or entry.get("results")
|
|
395
|
+
or []
|
|
396
|
+
)
|
|
397
|
+
if not candidates and entry.get("role") == "tool":
|
|
398
|
+
candidates = [entry]
|
|
399
|
+
if not isinstance(candidates, list):
|
|
400
|
+
continue
|
|
401
|
+
for result in candidates:
|
|
402
|
+
if not isinstance(result, dict):
|
|
403
|
+
continue
|
|
404
|
+
results.append(
|
|
405
|
+
{
|
|
406
|
+
key: value
|
|
407
|
+
for key, value in {
|
|
408
|
+
"tool_call_id": result.get("toolCallId")
|
|
409
|
+
or result.get("tool_call_id"),
|
|
410
|
+
"name": result.get("name"),
|
|
411
|
+
"result": result.get("result"),
|
|
412
|
+
"content": result.get("content") or result.get("message"),
|
|
413
|
+
"error": result.get("error"),
|
|
414
|
+
}.items()
|
|
415
|
+
if value is not None
|
|
416
|
+
}
|
|
417
|
+
)
|
|
418
|
+
return results
|
|
419
|
+
|
|
420
|
+
|
|
421
|
+
def _cost_summary(payload: dict[str, Any]) -> dict[str, Any] | None:
|
|
422
|
+
total = payload.get("cost")
|
|
423
|
+
breakdown = payload.get("costBreakdown")
|
|
424
|
+
if total is None and not breakdown:
|
|
425
|
+
return None
|
|
426
|
+
return {"total": total, "breakdown": coerce_json(breakdown)}
|
|
@@ -0,0 +1,32 @@
|
|
|
1
|
+
"""Hosted-runner surface: the platform-triggered execution path for the SDK.
|
|
2
|
+
|
|
3
|
+
The platform dispatches a ``StartRunnerJob`` to a ``simulation-runner`` worker,
|
|
4
|
+
which spawns ``child_entrypoint`` to run the released SDK and submit results
|
|
5
|
+
through the existing ingestion API. See the plan doc §9.
|
|
6
|
+
"""
|
|
7
|
+
|
|
8
|
+
from __future__ import annotations
|
|
9
|
+
|
|
10
|
+
from .job import (
|
|
11
|
+
RUNNER_JOB_SCHEMA_VERSION,
|
|
12
|
+
HostedRunnerPort,
|
|
13
|
+
ResultSinkConfig,
|
|
14
|
+
RunnerJobHandle,
|
|
15
|
+
RunnerJobPhase,
|
|
16
|
+
RunnerJobStatus,
|
|
17
|
+
RunnerMode,
|
|
18
|
+
RunnerReconcileResult,
|
|
19
|
+
StartRunnerJob,
|
|
20
|
+
)
|
|
21
|
+
|
|
22
|
+
__all__ = [
|
|
23
|
+
"RUNNER_JOB_SCHEMA_VERSION",
|
|
24
|
+
"HostedRunnerPort",
|
|
25
|
+
"ResultSinkConfig",
|
|
26
|
+
"RunnerJobHandle",
|
|
27
|
+
"RunnerJobPhase",
|
|
28
|
+
"RunnerJobStatus",
|
|
29
|
+
"RunnerMode",
|
|
30
|
+
"RunnerReconcileResult",
|
|
31
|
+
"StartRunnerJob",
|
|
32
|
+
]
|