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 +1,652 @@
|
|
|
1
|
+
from __future__ import annotations
|
|
2
|
+
|
|
3
|
+
import hashlib
|
|
4
|
+
import json
|
|
5
|
+
import time
|
|
6
|
+
import urllib.error
|
|
7
|
+
import urllib.parse
|
|
8
|
+
import urllib.request
|
|
9
|
+
from dataclasses import dataclass
|
|
10
|
+
from typing import Any, Literal, Mapping
|
|
11
|
+
|
|
12
|
+
from fi.simulate.agent.definition import AgentDefinition
|
|
13
|
+
from fi.simulate.simulation.models import Scenario
|
|
14
|
+
|
|
15
|
+
from ._download import (
|
|
16
|
+
_config,
|
|
17
|
+
_field,
|
|
18
|
+
_headers,
|
|
19
|
+
_rows,
|
|
20
|
+
ScenarioDownloadError,
|
|
21
|
+
fetch_dataset_rows,
|
|
22
|
+
hydrate_platform_scenario,
|
|
23
|
+
validate_download,
|
|
24
|
+
)
|
|
25
|
+
from ._scan import DownloadRejected, scan_content
|
|
26
|
+
|
|
27
|
+
_AGENT_LIST_PATH = "/simulate/agent-definitions/"
|
|
28
|
+
_AGENT_CREATE_PATH = "/simulate/agent-definitions/create/"
|
|
29
|
+
_AGENT_VERSION_CREATE_PATH = "/simulate/agent-definitions/{agent_id}/versions/create/"
|
|
30
|
+
_SCENARIO_LIST_PATH = "/simulate/scenarios/"
|
|
31
|
+
_SCENARIO_CREATE_PATH = "/simulate/scenarios/create/"
|
|
32
|
+
_SCENARIO_DETAIL_PATH = "/simulate/scenarios/{scenario_id}/"
|
|
33
|
+
_TERMINAL_FAILURE_STATUSES = {"failed", "error", "cancelled"}
|
|
34
|
+
|
|
35
|
+
|
|
36
|
+
class ScenarioGenerationError(RuntimeError):
|
|
37
|
+
def __init__(
|
|
38
|
+
self,
|
|
39
|
+
message: str,
|
|
40
|
+
*,
|
|
41
|
+
scenario_id: str | None = None,
|
|
42
|
+
status: str | None = None,
|
|
43
|
+
retryable: bool = False,
|
|
44
|
+
) -> None:
|
|
45
|
+
super().__init__(message)
|
|
46
|
+
self.scenario_id = scenario_id
|
|
47
|
+
self.status = status
|
|
48
|
+
self.retryable = retryable
|
|
49
|
+
|
|
50
|
+
|
|
51
|
+
@dataclass(frozen=True)
|
|
52
|
+
class PlatformAgentReference:
|
|
53
|
+
agent_definition_id: str
|
|
54
|
+
agent_version_id: str
|
|
55
|
+
configuration_hash: str
|
|
56
|
+
reused: bool
|
|
57
|
+
|
|
58
|
+
|
|
59
|
+
@dataclass(frozen=True)
|
|
60
|
+
class PlatformScenarioRequest:
|
|
61
|
+
name: str
|
|
62
|
+
agent_definition: AgentDefinition | None = None
|
|
63
|
+
platform_agent_definition_id: str | None = None
|
|
64
|
+
platform_agent_version_id: str | None = None
|
|
65
|
+
description: str | None = None
|
|
66
|
+
custom_instruction: str | None = None
|
|
67
|
+
kind: Literal["graph"] = "graph"
|
|
68
|
+
no_of_rows: int = 10
|
|
69
|
+
poll_interval_seconds: float = 2.0
|
|
70
|
+
timeout_seconds: float = 900.0
|
|
71
|
+
|
|
72
|
+
def __post_init__(self) -> None:
|
|
73
|
+
if not self.name.strip():
|
|
74
|
+
raise ValueError("scenario name must be non-empty")
|
|
75
|
+
if self.kind != "graph":
|
|
76
|
+
raise ValueError("only graph scenario generation is supported")
|
|
77
|
+
if not 10 <= self.no_of_rows <= 20_000:
|
|
78
|
+
raise ValueError("no_of_rows must be between 10 and 20000")
|
|
79
|
+
if self.poll_interval_seconds <= 0:
|
|
80
|
+
raise ValueError("poll_interval_seconds must be greater than zero")
|
|
81
|
+
if self.timeout_seconds <= 0:
|
|
82
|
+
raise ValueError("timeout_seconds must be greater than zero")
|
|
83
|
+
has_local = self.agent_definition is not None
|
|
84
|
+
has_platform = bool(self.platform_agent_definition_id)
|
|
85
|
+
if has_local == has_platform:
|
|
86
|
+
raise ValueError(
|
|
87
|
+
"provide exactly one of agent_definition or platform_agent_definition_id"
|
|
88
|
+
)
|
|
89
|
+
if self.platform_agent_version_id and not has_platform:
|
|
90
|
+
raise ValueError(
|
|
91
|
+
"platform_agent_version_id requires platform_agent_definition_id"
|
|
92
|
+
)
|
|
93
|
+
|
|
94
|
+
|
|
95
|
+
@dataclass(frozen=True)
|
|
96
|
+
class GeneratedScenario:
|
|
97
|
+
scenario: Scenario
|
|
98
|
+
platform_agent_definition_id: str
|
|
99
|
+
platform_agent_version_id: str | None
|
|
100
|
+
platform_scenario_id: str
|
|
101
|
+
platform_dataset_id: str
|
|
102
|
+
platform_status: str
|
|
103
|
+
polling_duration_seconds: float
|
|
104
|
+
checksum_sha256: str
|
|
105
|
+
|
|
106
|
+
|
|
107
|
+
def _request_json(
|
|
108
|
+
url: str,
|
|
109
|
+
headers: Mapping[str, str],
|
|
110
|
+
*,
|
|
111
|
+
method: str = "GET",
|
|
112
|
+
payload: Mapping[str, Any] | None = None,
|
|
113
|
+
timeout: float = 30.0,
|
|
114
|
+
) -> Any:
|
|
115
|
+
body = None
|
|
116
|
+
request_headers = dict(headers)
|
|
117
|
+
if payload is not None:
|
|
118
|
+
body = json.dumps(payload).encode("utf-8")
|
|
119
|
+
request_headers["Content-Type"] = "application/json"
|
|
120
|
+
request = urllib.request.Request(
|
|
121
|
+
url,
|
|
122
|
+
data=body,
|
|
123
|
+
headers=request_headers,
|
|
124
|
+
method=method,
|
|
125
|
+
)
|
|
126
|
+
try:
|
|
127
|
+
with urllib.request.urlopen(request, timeout=timeout) as response: # noqa: S310
|
|
128
|
+
return json.loads(response.read().decode("utf-8"))
|
|
129
|
+
except urllib.error.HTTPError as exc:
|
|
130
|
+
raise ScenarioGenerationError(
|
|
131
|
+
f"Future AGI API request failed with HTTP {exc.code}"
|
|
132
|
+
) from exc
|
|
133
|
+
except urllib.error.URLError as exc:
|
|
134
|
+
raise ScenarioGenerationError(
|
|
135
|
+
"Future AGI API request could not be completed",
|
|
136
|
+
retryable=True,
|
|
137
|
+
) from exc
|
|
138
|
+
|
|
139
|
+
|
|
140
|
+
def _required_config(config: Any | None) -> Any:
|
|
141
|
+
resolved = _config(config)
|
|
142
|
+
if not resolved.api_key or not resolved.secret_key:
|
|
143
|
+
raise ScenarioGenerationError(
|
|
144
|
+
"Future AGI API and secret keys are required for scenario generation"
|
|
145
|
+
)
|
|
146
|
+
return resolved
|
|
147
|
+
|
|
148
|
+
|
|
149
|
+
def _safe_livekit_url(value: object) -> str:
|
|
150
|
+
parsed = urllib.parse.urlsplit(str(value))
|
|
151
|
+
if parsed.scheme not in {"ws", "wss"} or not parsed.hostname:
|
|
152
|
+
raise ScenarioGenerationError("LiveKit agent URL must use ws:// or wss://")
|
|
153
|
+
if parsed.username or parsed.password or parsed.query or parsed.fragment:
|
|
154
|
+
raise ScenarioGenerationError(
|
|
155
|
+
"LiveKit agent URL must not contain credentials, query parameters, or fragments"
|
|
156
|
+
)
|
|
157
|
+
host = parsed.hostname
|
|
158
|
+
if parsed.port:
|
|
159
|
+
host = f"{host}:{parsed.port}"
|
|
160
|
+
return urllib.parse.urlunsplit((parsed.scheme, host, parsed.path, "", ""))
|
|
161
|
+
|
|
162
|
+
|
|
163
|
+
def _agent_payload(agent_definition: AgentDefinition) -> tuple[dict[str, Any], str]:
|
|
164
|
+
transport = agent_definition.transport
|
|
165
|
+
transport_kind = transport.kind if transport else "webrtc"
|
|
166
|
+
inbound = transport_kind != "sip_inbound"
|
|
167
|
+
description = agent_definition.system_prompt
|
|
168
|
+
scan = scan_content({"description": description})
|
|
169
|
+
if scan["status"] == "flagged":
|
|
170
|
+
raise ScenarioGenerationError(
|
|
171
|
+
"agent description was rejected by the local content and secret scan"
|
|
172
|
+
)
|
|
173
|
+
|
|
174
|
+
safe_configuration: dict[str, Any] = {
|
|
175
|
+
"name": agent_definition.name,
|
|
176
|
+
"description": description,
|
|
177
|
+
"transport": transport_kind,
|
|
178
|
+
"inbound": inbound,
|
|
179
|
+
"model": agent_definition.llm.model,
|
|
180
|
+
"model_provider": agent_definition.llm.provider,
|
|
181
|
+
"language": agent_definition.stt.language,
|
|
182
|
+
}
|
|
183
|
+
identity_configuration: dict[str, Any] = {"transport": transport_kind}
|
|
184
|
+
payload: dict[str, Any] = {
|
|
185
|
+
"agent_type": "voice",
|
|
186
|
+
"commit_message": "Created by Agent Learning Kit for scenario generation",
|
|
187
|
+
"description": description,
|
|
188
|
+
"inbound": inbound,
|
|
189
|
+
"language": agent_definition.stt.language,
|
|
190
|
+
"languages": (
|
|
191
|
+
[agent_definition.stt.language] if agent_definition.stt.language else None
|
|
192
|
+
),
|
|
193
|
+
"model": agent_definition.llm.model,
|
|
194
|
+
"model_details": {
|
|
195
|
+
"provider": agent_definition.llm.provider,
|
|
196
|
+
"temperature": agent_definition.llm.temperature,
|
|
197
|
+
},
|
|
198
|
+
}
|
|
199
|
+
target = agent_definition.target
|
|
200
|
+
if target is not None:
|
|
201
|
+
target_id = (
|
|
202
|
+
target.assistant_id if target.provider == "vapi" else target.agent_id
|
|
203
|
+
)
|
|
204
|
+
target_url = (
|
|
205
|
+
target.api_base_url if target.provider == "vapi" else target.api_url
|
|
206
|
+
)
|
|
207
|
+
payload.update(
|
|
208
|
+
{
|
|
209
|
+
"provider": target.provider,
|
|
210
|
+
"assistant_id": target_id,
|
|
211
|
+
}
|
|
212
|
+
)
|
|
213
|
+
safe_configuration.update(
|
|
214
|
+
{
|
|
215
|
+
"provider": target.provider,
|
|
216
|
+
"assistant_id": target_id,
|
|
217
|
+
"provider_api_url": str(target_url),
|
|
218
|
+
}
|
|
219
|
+
)
|
|
220
|
+
identity_configuration.update(
|
|
221
|
+
{"provider": target.provider, "assistant_id": target_id}
|
|
222
|
+
)
|
|
223
|
+
elif transport_kind in {"sip_outbound", "sip_inbound"}:
|
|
224
|
+
if transport is None or not transport.sip_call_to:
|
|
225
|
+
raise ScenarioGenerationError(
|
|
226
|
+
"SIP platform agent creation requires a target contact number"
|
|
227
|
+
)
|
|
228
|
+
payload.update(
|
|
229
|
+
{
|
|
230
|
+
"provider": "others",
|
|
231
|
+
"contact_number": transport.sip_call_to,
|
|
232
|
+
}
|
|
233
|
+
)
|
|
234
|
+
safe_configuration["contact_number"] = transport.sip_call_to
|
|
235
|
+
identity_configuration.update(
|
|
236
|
+
{"provider": "others", "contact_number": transport.sip_call_to}
|
|
237
|
+
)
|
|
238
|
+
else:
|
|
239
|
+
if agent_definition.url is None:
|
|
240
|
+
raise ScenarioGenerationError(
|
|
241
|
+
"LiveKit target creation requires an AgentDefinition url"
|
|
242
|
+
)
|
|
243
|
+
livekit_url = _safe_livekit_url(agent_definition.url)
|
|
244
|
+
livekit_agent_name = agent_definition.agent_name or agent_definition.name
|
|
245
|
+
payload.update(
|
|
246
|
+
{
|
|
247
|
+
"provider": "livekit",
|
|
248
|
+
"livekit_url": livekit_url,
|
|
249
|
+
"livekit_agent_name": livekit_agent_name,
|
|
250
|
+
}
|
|
251
|
+
)
|
|
252
|
+
safe_configuration.update(
|
|
253
|
+
{
|
|
254
|
+
"livekit_url": livekit_url,
|
|
255
|
+
"livekit_agent_name": livekit_agent_name,
|
|
256
|
+
}
|
|
257
|
+
)
|
|
258
|
+
identity_configuration.update(
|
|
259
|
+
{
|
|
260
|
+
"provider": "livekit",
|
|
261
|
+
"livekit_url": livekit_url,
|
|
262
|
+
"livekit_agent_name": livekit_agent_name,
|
|
263
|
+
}
|
|
264
|
+
)
|
|
265
|
+
|
|
266
|
+
configuration_hash = hashlib.sha256(
|
|
267
|
+
json.dumps(
|
|
268
|
+
safe_configuration,
|
|
269
|
+
sort_keys=True,
|
|
270
|
+
separators=(",", ":"),
|
|
271
|
+
default=str,
|
|
272
|
+
).encode("utf-8")
|
|
273
|
+
).hexdigest()
|
|
274
|
+
identity_hash = hashlib.sha256(
|
|
275
|
+
json.dumps(
|
|
276
|
+
identity_configuration,
|
|
277
|
+
sort_keys=True,
|
|
278
|
+
separators=(",", ":"),
|
|
279
|
+
default=str,
|
|
280
|
+
).encode("utf-8")
|
|
281
|
+
).hexdigest()
|
|
282
|
+
base_name = "-".join(agent_definition.name.strip().split()) or "agent"
|
|
283
|
+
payload["agent_name"] = f"{base_name[:220]}-alk-{identity_hash[:12]}"
|
|
284
|
+
payload["commit_message"] = _configuration_commit(configuration_hash)
|
|
285
|
+
return {
|
|
286
|
+
key: value for key, value in payload.items() if value is not None
|
|
287
|
+
}, configuration_hash
|
|
288
|
+
|
|
289
|
+
|
|
290
|
+
def _configuration_commit(configuration_hash: str) -> str:
|
|
291
|
+
return f"Agent Learning Kit configuration {configuration_hash}"
|
|
292
|
+
|
|
293
|
+
|
|
294
|
+
def _active_version(detail: Mapping[str, Any]) -> Mapping[str, Any] | None:
|
|
295
|
+
active = detail.get("active_version")
|
|
296
|
+
if isinstance(active, Mapping) and active.get("id"):
|
|
297
|
+
return active
|
|
298
|
+
versions = detail.get("versions")
|
|
299
|
+
if isinstance(versions, list):
|
|
300
|
+
return next(
|
|
301
|
+
(
|
|
302
|
+
version
|
|
303
|
+
for version in versions
|
|
304
|
+
if isinstance(version, Mapping) and version.get("id")
|
|
305
|
+
),
|
|
306
|
+
None,
|
|
307
|
+
)
|
|
308
|
+
return None
|
|
309
|
+
|
|
310
|
+
|
|
311
|
+
def _active_version_id(detail: Mapping[str, Any]) -> str | None:
|
|
312
|
+
active = _active_version(detail)
|
|
313
|
+
if active is not None:
|
|
314
|
+
return str(active["id"])
|
|
315
|
+
latest = detail.get("latest_version_id")
|
|
316
|
+
return str(latest) if latest else None
|
|
317
|
+
|
|
318
|
+
|
|
319
|
+
def _logical_agent_match(
|
|
320
|
+
item: Mapping[str, Any],
|
|
321
|
+
payload: Mapping[str, Any],
|
|
322
|
+
) -> bool:
|
|
323
|
+
if str(_field(item, "agent_name") or "") == str(payload["agent_name"]):
|
|
324
|
+
return True
|
|
325
|
+
provider = str(payload.get("provider") or "")
|
|
326
|
+
assistant_id = payload.get("assistant_id")
|
|
327
|
+
if assistant_id:
|
|
328
|
+
return bool(
|
|
329
|
+
str(_field(item, "provider") or "") == provider
|
|
330
|
+
and str(_field(item, "assistant_id") or "") == str(assistant_id)
|
|
331
|
+
)
|
|
332
|
+
contact_number = payload.get("contact_number")
|
|
333
|
+
return bool(
|
|
334
|
+
contact_number
|
|
335
|
+
and str(_field(item, "provider") or "") == provider
|
|
336
|
+
and str(_field(item, "contact_number") or "") == str(contact_number)
|
|
337
|
+
)
|
|
338
|
+
|
|
339
|
+
|
|
340
|
+
def ensure_platform_agent(
|
|
341
|
+
agent_definition: AgentDefinition,
|
|
342
|
+
*,
|
|
343
|
+
config: Any | None = None,
|
|
344
|
+
) -> PlatformAgentReference:
|
|
345
|
+
cfg = _required_config(config)
|
|
346
|
+
headers = _headers(cfg)
|
|
347
|
+
base = str(cfg.api_url).rstrip("/")
|
|
348
|
+
payload, configuration_hash = _agent_payload(agent_definition)
|
|
349
|
+
search = payload.get("assistant_id") or agent_definition.name
|
|
350
|
+
query = urllib.parse.urlencode(
|
|
351
|
+
{"search": search, "agent_type": "voice", "limit": 100}
|
|
352
|
+
)
|
|
353
|
+
listing = _request_json(f"{base}{_AGENT_LIST_PATH}?{query}", headers)
|
|
354
|
+
existing = next(
|
|
355
|
+
(
|
|
356
|
+
item
|
|
357
|
+
for item in _rows(listing)
|
|
358
|
+
if isinstance(item, Mapping) and _logical_agent_match(item, payload)
|
|
359
|
+
),
|
|
360
|
+
None,
|
|
361
|
+
)
|
|
362
|
+
if existing is not None:
|
|
363
|
+
agent_id = str(_field(existing, "id") or "")
|
|
364
|
+
if not agent_id:
|
|
365
|
+
raise ScenarioGenerationError(
|
|
366
|
+
"reused platform Agent Definition has no agent id"
|
|
367
|
+
)
|
|
368
|
+
detail = _request_json(f"{base}{_AGENT_LIST_PATH}{agent_id}/", headers)
|
|
369
|
+
active = _active_version(detail) if isinstance(detail, Mapping) else None
|
|
370
|
+
if active is None:
|
|
371
|
+
raise ScenarioGenerationError(
|
|
372
|
+
"reused platform Agent Definition has no active version"
|
|
373
|
+
)
|
|
374
|
+
version_id = str(active["id"])
|
|
375
|
+
if str(active.get("commit_message") or "") == _configuration_commit(
|
|
376
|
+
configuration_hash
|
|
377
|
+
):
|
|
378
|
+
return PlatformAgentReference(
|
|
379
|
+
agent_definition_id=agent_id,
|
|
380
|
+
agent_version_id=version_id,
|
|
381
|
+
configuration_hash=configuration_hash,
|
|
382
|
+
reused=True,
|
|
383
|
+
)
|
|
384
|
+
created_version = _request_json(
|
|
385
|
+
f"{base}{_AGENT_VERSION_CREATE_PATH.format(agent_id=agent_id)}",
|
|
386
|
+
headers,
|
|
387
|
+
method="POST",
|
|
388
|
+
payload=payload,
|
|
389
|
+
)
|
|
390
|
+
version = (
|
|
391
|
+
created_version.get("version")
|
|
392
|
+
if isinstance(created_version, Mapping)
|
|
393
|
+
else None
|
|
394
|
+
)
|
|
395
|
+
version_id = (
|
|
396
|
+
str(_field(version, "id") or "")
|
|
397
|
+
if isinstance(version, Mapping)
|
|
398
|
+
else ""
|
|
399
|
+
)
|
|
400
|
+
if not version_id:
|
|
401
|
+
raise ScenarioGenerationError(
|
|
402
|
+
"platform Agent Version response did not include a version id"
|
|
403
|
+
)
|
|
404
|
+
return PlatformAgentReference(
|
|
405
|
+
agent_definition_id=agent_id,
|
|
406
|
+
agent_version_id=version_id,
|
|
407
|
+
configuration_hash=configuration_hash,
|
|
408
|
+
reused=False,
|
|
409
|
+
)
|
|
410
|
+
|
|
411
|
+
created = _request_json(
|
|
412
|
+
f"{base}{_AGENT_CREATE_PATH}",
|
|
413
|
+
headers,
|
|
414
|
+
method="POST",
|
|
415
|
+
payload=payload,
|
|
416
|
+
)
|
|
417
|
+
agent = created.get("agent") if isinstance(created, Mapping) else None
|
|
418
|
+
agent_id = str(_field(agent, "id") or "") if isinstance(agent, Mapping) else ""
|
|
419
|
+
if not agent_id:
|
|
420
|
+
raise ScenarioGenerationError(
|
|
421
|
+
"platform Agent Definition response did not include an agent id"
|
|
422
|
+
)
|
|
423
|
+
detail = _request_json(f"{base}{_AGENT_LIST_PATH}{agent_id}/", headers)
|
|
424
|
+
version_id = _active_version_id(detail) if isinstance(detail, Mapping) else None
|
|
425
|
+
if not version_id:
|
|
426
|
+
raise ScenarioGenerationError(
|
|
427
|
+
"created platform Agent Definition has no active version"
|
|
428
|
+
)
|
|
429
|
+
return PlatformAgentReference(
|
|
430
|
+
agent_definition_id=agent_id,
|
|
431
|
+
agent_version_id=version_id,
|
|
432
|
+
configuration_hash=configuration_hash,
|
|
433
|
+
reused=False,
|
|
434
|
+
)
|
|
435
|
+
|
|
436
|
+
|
|
437
|
+
def _extract_scenario(payload: Any) -> Mapping[str, Any]:
|
|
438
|
+
if not isinstance(payload, Mapping):
|
|
439
|
+
return {}
|
|
440
|
+
scenario = payload.get("scenario")
|
|
441
|
+
if isinstance(scenario, Mapping):
|
|
442
|
+
return scenario
|
|
443
|
+
result = payload.get("result")
|
|
444
|
+
if isinstance(result, Mapping):
|
|
445
|
+
nested = result.get("scenario")
|
|
446
|
+
return nested if isinstance(nested, Mapping) else result
|
|
447
|
+
return payload
|
|
448
|
+
|
|
449
|
+
|
|
450
|
+
def _completed_scenario(
|
|
451
|
+
detail: Mapping[str, Any],
|
|
452
|
+
*,
|
|
453
|
+
agent_definition_id: str,
|
|
454
|
+
agent_version_id: str | None,
|
|
455
|
+
polling_duration_seconds: float,
|
|
456
|
+
base: str,
|
|
457
|
+
headers: Mapping[str, str],
|
|
458
|
+
) -> GeneratedScenario:
|
|
459
|
+
scenario_id = str(_field(detail, "id") or "")
|
|
460
|
+
dataset_id = str(_field(detail, "dataset_id") or _field(detail, "dataset") or "")
|
|
461
|
+
if not dataset_id:
|
|
462
|
+
raise ScenarioGenerationError(
|
|
463
|
+
"completed platform Scenario has no dataset id",
|
|
464
|
+
scenario_id=scenario_id,
|
|
465
|
+
status=str(_field(detail, "status") or "Completed"),
|
|
466
|
+
)
|
|
467
|
+
try:
|
|
468
|
+
rows = fetch_dataset_rows(base, headers, dataset_id)
|
|
469
|
+
except (
|
|
470
|
+
urllib.error.HTTPError,
|
|
471
|
+
urllib.error.URLError,
|
|
472
|
+
ScenarioDownloadError,
|
|
473
|
+
) as exc:
|
|
474
|
+
raise ScenarioGenerationError(
|
|
475
|
+
"completed platform Scenario dataset could not be retrieved",
|
|
476
|
+
scenario_id=scenario_id,
|
|
477
|
+
status=str(_field(detail, "status") or "Completed"),
|
|
478
|
+
retryable=True,
|
|
479
|
+
) from exc
|
|
480
|
+
artifact = {
|
|
481
|
+
"id": scenario_id,
|
|
482
|
+
"updated_at": _field(detail, "updated_at"),
|
|
483
|
+
"scenario": dict(detail),
|
|
484
|
+
"rows": rows,
|
|
485
|
+
}
|
|
486
|
+
try:
|
|
487
|
+
pin = validate_download(
|
|
488
|
+
artifact,
|
|
489
|
+
source=urllib.parse.urlsplit(base).netloc or base,
|
|
490
|
+
)
|
|
491
|
+
except DownloadRejected as exc:
|
|
492
|
+
raise ScenarioGenerationError(
|
|
493
|
+
"generated platform Scenario was rejected by the local content scan",
|
|
494
|
+
scenario_id=scenario_id,
|
|
495
|
+
status=str(_field(detail, "status") or "Completed"),
|
|
496
|
+
) from exc
|
|
497
|
+
scenario = hydrate_platform_scenario(detail, rows, pin=pin)
|
|
498
|
+
return GeneratedScenario(
|
|
499
|
+
scenario=scenario,
|
|
500
|
+
platform_agent_definition_id=agent_definition_id,
|
|
501
|
+
platform_agent_version_id=agent_version_id,
|
|
502
|
+
platform_scenario_id=scenario_id,
|
|
503
|
+
platform_dataset_id=dataset_id,
|
|
504
|
+
platform_status=str(_field(detail, "status") or "Completed"),
|
|
505
|
+
polling_duration_seconds=polling_duration_seconds,
|
|
506
|
+
checksum_sha256=str(pin["checksum_sha256"]),
|
|
507
|
+
)
|
|
508
|
+
|
|
509
|
+
|
|
510
|
+
def fetch_scenario(
|
|
511
|
+
scenario_id: str,
|
|
512
|
+
*,
|
|
513
|
+
platform_agent_definition_id: str = "",
|
|
514
|
+
platform_agent_version_id: str | None = None,
|
|
515
|
+
poll_interval_seconds: float = 2.0,
|
|
516
|
+
timeout_seconds: float = 900.0,
|
|
517
|
+
config: Any | None = None,
|
|
518
|
+
) -> GeneratedScenario:
|
|
519
|
+
if not scenario_id.strip():
|
|
520
|
+
raise ValueError("scenario_id must be non-empty")
|
|
521
|
+
if poll_interval_seconds <= 0 or timeout_seconds <= 0:
|
|
522
|
+
raise ValueError("poll interval and timeout must be greater than zero")
|
|
523
|
+
cfg = _required_config(config)
|
|
524
|
+
headers = _headers(cfg)
|
|
525
|
+
base = str(cfg.api_url).rstrip("/")
|
|
526
|
+
started = time.monotonic()
|
|
527
|
+
while True:
|
|
528
|
+
raw = _request_json(
|
|
529
|
+
f"{base}{_SCENARIO_DETAIL_PATH.format(scenario_id=scenario_id)}",
|
|
530
|
+
headers,
|
|
531
|
+
)
|
|
532
|
+
detail = _extract_scenario(raw)
|
|
533
|
+
status = str(_field(detail, "status") or "").strip()
|
|
534
|
+
normalized_status = status.lower()
|
|
535
|
+
elapsed = time.monotonic() - started
|
|
536
|
+
if normalized_status == "completed":
|
|
537
|
+
return _completed_scenario(
|
|
538
|
+
detail,
|
|
539
|
+
agent_definition_id=platform_agent_definition_id,
|
|
540
|
+
agent_version_id=platform_agent_version_id,
|
|
541
|
+
polling_duration_seconds=elapsed,
|
|
542
|
+
base=base,
|
|
543
|
+
headers=headers,
|
|
544
|
+
)
|
|
545
|
+
if normalized_status in _TERMINAL_FAILURE_STATUSES:
|
|
546
|
+
raise ScenarioGenerationError(
|
|
547
|
+
f"platform Scenario ended with status {status}",
|
|
548
|
+
scenario_id=scenario_id,
|
|
549
|
+
status=status,
|
|
550
|
+
)
|
|
551
|
+
if elapsed >= timeout_seconds:
|
|
552
|
+
raise ScenarioGenerationError(
|
|
553
|
+
"platform Scenario generation timed out; resume with fetch_scenario",
|
|
554
|
+
scenario_id=scenario_id,
|
|
555
|
+
status=status or "Processing",
|
|
556
|
+
retryable=True,
|
|
557
|
+
)
|
|
558
|
+
time.sleep(min(poll_interval_seconds, timeout_seconds - elapsed))
|
|
559
|
+
|
|
560
|
+
|
|
561
|
+
def generate_scenario(
|
|
562
|
+
request: PlatformScenarioRequest,
|
|
563
|
+
*,
|
|
564
|
+
config: Any | None = None,
|
|
565
|
+
) -> GeneratedScenario:
|
|
566
|
+
cfg = _required_config(config)
|
|
567
|
+
if request.agent_definition is not None:
|
|
568
|
+
agent = ensure_platform_agent(request.agent_definition, config=cfg)
|
|
569
|
+
agent_definition_id = agent.agent_definition_id
|
|
570
|
+
agent_version_id: str | None = agent.agent_version_id
|
|
571
|
+
else:
|
|
572
|
+
agent_definition_id = str(request.platform_agent_definition_id)
|
|
573
|
+
agent_version_id = request.platform_agent_version_id
|
|
574
|
+
|
|
575
|
+
payload: dict[str, Any] = {
|
|
576
|
+
"name": request.name.strip(),
|
|
577
|
+
"kind": request.kind,
|
|
578
|
+
"generate_graph": True,
|
|
579
|
+
"agent_definition_id": agent_definition_id,
|
|
580
|
+
"no_of_rows": request.no_of_rows,
|
|
581
|
+
}
|
|
582
|
+
if agent_version_id:
|
|
583
|
+
payload["agent_definition_version_id"] = agent_version_id
|
|
584
|
+
if request.description is not None:
|
|
585
|
+
payload["description"] = request.description
|
|
586
|
+
if request.custom_instruction is not None:
|
|
587
|
+
payload["custom_instruction"] = request.custom_instruction
|
|
588
|
+
if scan_content(payload)["status"] == "flagged":
|
|
589
|
+
raise ScenarioGenerationError(
|
|
590
|
+
"scenario generation request was rejected by the local content and secret scan"
|
|
591
|
+
)
|
|
592
|
+
|
|
593
|
+
headers = _headers(cfg)
|
|
594
|
+
base = str(cfg.api_url).rstrip("/")
|
|
595
|
+
query = urllib.parse.urlencode(
|
|
596
|
+
{
|
|
597
|
+
"search": request.name.strip(),
|
|
598
|
+
"agent_definition_id": agent_definition_id,
|
|
599
|
+
"limit": 100,
|
|
600
|
+
}
|
|
601
|
+
)
|
|
602
|
+
listing = _request_json(f"{base}{_SCENARIO_LIST_PATH}?{query}", headers)
|
|
603
|
+
existing = next(
|
|
604
|
+
(
|
|
605
|
+
item
|
|
606
|
+
for item in _rows(listing)
|
|
607
|
+
if str(_field(item, "name") or "") == request.name.strip()
|
|
608
|
+
),
|
|
609
|
+
None,
|
|
610
|
+
)
|
|
611
|
+
if existing is not None:
|
|
612
|
+
scenario_id = str(_field(existing, "id") or "")
|
|
613
|
+
if scenario_id:
|
|
614
|
+
return fetch_scenario(
|
|
615
|
+
scenario_id,
|
|
616
|
+
platform_agent_definition_id=agent_definition_id,
|
|
617
|
+
platform_agent_version_id=agent_version_id,
|
|
618
|
+
poll_interval_seconds=request.poll_interval_seconds,
|
|
619
|
+
timeout_seconds=request.timeout_seconds,
|
|
620
|
+
config=cfg,
|
|
621
|
+
)
|
|
622
|
+
created = _request_json(
|
|
623
|
+
f"{base}{_SCENARIO_CREATE_PATH}",
|
|
624
|
+
headers,
|
|
625
|
+
method="POST",
|
|
626
|
+
payload=payload,
|
|
627
|
+
)
|
|
628
|
+
scenario = _extract_scenario(created)
|
|
629
|
+
scenario_id = str(_field(scenario, "id") or "")
|
|
630
|
+
if not scenario_id:
|
|
631
|
+
raise ScenarioGenerationError(
|
|
632
|
+
"platform Scenario response did not include a scenario id"
|
|
633
|
+
)
|
|
634
|
+
return fetch_scenario(
|
|
635
|
+
scenario_id,
|
|
636
|
+
platform_agent_definition_id=agent_definition_id,
|
|
637
|
+
platform_agent_version_id=agent_version_id,
|
|
638
|
+
poll_interval_seconds=request.poll_interval_seconds,
|
|
639
|
+
timeout_seconds=request.timeout_seconds,
|
|
640
|
+
config=cfg,
|
|
641
|
+
)
|
|
642
|
+
|
|
643
|
+
|
|
644
|
+
__all__ = [
|
|
645
|
+
"GeneratedScenario",
|
|
646
|
+
"PlatformAgentReference",
|
|
647
|
+
"PlatformScenarioRequest",
|
|
648
|
+
"ScenarioGenerationError",
|
|
649
|
+
"ensure_platform_agent",
|
|
650
|
+
"fetch_scenario",
|
|
651
|
+
"generate_scenario",
|
|
652
|
+
]
|