databricks-agentbricks 0.7.1__tar.gz → 0.9.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/PKG-INFO +247 -58
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/README.md +244 -57
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/pyproject.toml +4 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/app.py +38 -3
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/auth.py +6 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/deploy.py +8 -4
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/dev.py +30 -13
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/cli/endpoint.py +206 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/memory.py +20 -4
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/cli/permissions.py +185 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/sessions.py +4 -0
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/clients/app_probe.py +82 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/apps_client.py +49 -6
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/apps_user_auth_client.py +13 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/conversation_store_client.py +10 -4
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/databricks_cli.py +57 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/local_ports.py +14 -2
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/clients/terminal_relay.py +203 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/presentation/deploy.py +3 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/presentation/dev.py +35 -3
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/presentation/help.py +80 -7
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/presentation/loadtest.py +162 -0
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/presentation/permissions.py +25 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/deploy_service.py +99 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/deployment/deploy_state.py +2 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/dev_service.py +35 -12
- databricks_agentbricks-0.7.1/src/databricks_agentbricks/services/invoke_service.py → databricks_agentbricks-0.9.0/src/databricks_agentbricks/services/endpoint_service.py +94 -34
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/invoke/transport.py +11 -3
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/services/loadtest/__init__.py +1 -0
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/services/loadtest/_locustfile.py +305 -0
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/services/loadtest/plan.py +77 -0
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/services/loadtest/request.py +72 -0
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/services/loadtest/runner.py +165 -0
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/services/loadtest/streaming.py +175 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-bricks-workflow/SKILL.md +3 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-bricks-workflow/references/deployment.md +2 -2
- databricks_agentbricks-0.9.0/src/databricks_agentbricks/templates/agent-bricks-workflow/references/state-and-tools.md +95 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/AGENTKIT_CONTRACT.md +2 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/AGENTS.md +2 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/README.md +21 -6
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/agent.py +6 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/pyproject.toml +1 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/runtime/adapter.py +9 -4
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/AGENTKIT_CONTRACT.md +2 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/AGENTS.md +2 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/README.md +21 -5
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/agent/agent.py +4 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/pyproject.toml +1 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/pyproject.toml +1 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-openai/pyproject.toml +1 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/CHAT_APP.md +22 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/runtime/ui.py +221 -4
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/tests/test_demo_ui.py +103 -16
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/ui/app.js +664 -70
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/ui/index.html +82 -39
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/ui/styles.css +606 -55
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-openai/CHAT_APP.md +22 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-openai/runtime/ui.py +221 -4
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-openai/tests/test_demo_ui.py +100 -15
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-openai/ui/app.js +664 -70
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-openai/ui/index.html +82 -39
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-openai/ui/styles.css +606 -55
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/__init__.py +3 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/_api_client.py +77 -11
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/_pagination.py +6 -6
- databricks_agentbricks-0.9.0/src/databricks_agentkit/_permissions.py +44 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/langgraph/history.py +6 -2
- databricks_agentbricks-0.9.0/src/databricks_agentkit/langgraph/messages.py +168 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/memory_store.py +69 -7
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/model_upgrades/optimization.py +7 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/openai/__init__.py +4 -0
- databricks_agentbricks-0.9.0/src/databricks_agentkit/openai/models.py +27 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/ARCHITECTURE.md +29 -5
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/README.md +34 -7
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/app.py +99 -35
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/auth.py +4 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/durability/execution.py +2 -2
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/durability/lakebase_runtime_store.py +196 -74
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/durability/recovery.py +1 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/durability/store.py +6 -2
- databricks_agentbricks-0.9.0/src/databricks_agentkit/runtime/errors.py +247 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/execution.py +13 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/runtime.py +6 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/store.py +9 -1
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/tool_manifest.py +3 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/types.py +2 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/session_store.py +69 -7
- databricks_agentbricks-0.7.1/src/databricks_agentbricks/cli/endpoint.py +0 -81
- databricks_agentbricks-0.7.1/src/databricks_agentbricks/templates/agent-bricks-workflow/references/state-and-tools.md +0 -49
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/.gitignore +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/NOTICE +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/agent_hint.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/__main__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/doctor.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/experimental.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/init.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/mcp.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/models.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/pipeline.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/prompts.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/sandbox.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/telemetry.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/tools.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/cli/tracing.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/api_client_provider.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/legacy_runtime_store.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/local_tracing_client.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/managed_runtime_store.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/mlflow_trace_client.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/clients/tracing_client.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/errors.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/model_upgrade.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/presentation/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/presentation/endpoint.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/presentation/group.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/presentation/render.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/presentation/reporter.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/presentation/theme.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/presentation/tracing.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/projects/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/projects/agent_project.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/projects/app_manifest.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/projects/config.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/projects/resolver.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/projects/types.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/py.typed +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/reporting.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/deployment/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/deployment/app_provisioner.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/deployment/config.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/deployment/model_binding_provisioner.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/deployment/names.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/deployment/provisioners.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/deployment/tool_access.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/deployment/tool_access_provisioner.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/invoke/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/invoke/auth.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/invoke/request.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/tracing/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/tracing/targets.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/services/tracing_service.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/skills.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-bricks-migrate/SKILL.md +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/.env.example +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/.gitignore +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/CLAUDE.md +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/mcps.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/tools/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/tools/sample_tool.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/tools/send_message.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/app.yaml +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/runtime/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/runtime/main.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/tests/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/tests/test_agent.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/tests/test_runtime.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-langgraph/tests/test_workspace.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/.env.example +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/.gitignore +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/CLAUDE.md +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/agent/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/agent/checkpoints.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/agent/mcps.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/agent/tools/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/agent/tools/sample_tool.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/agent/tools/send_message.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/app.yaml +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/runtime/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/runtime/adapter.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/runtime/main.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/tests/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/tests/test_agent.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/tests/test_checkpoints.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/tests/test_runtime.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/agent-openai/tests/test_workspace.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/.env.example +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/.gitignore +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/README.md +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/agent/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/agent/agent.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/app.yaml +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/runtime/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/runtime/main.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/tests/test_runtime.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-openai/.env.example +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-openai/.gitignore +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-openai/README.md +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-openai/agent/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-openai/agent/agent.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-openai/app.yaml +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-openai/runtime/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-openai/runtime/main.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/custom-agent-openai/tests/test_runtime.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/sandbox_mcp.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/sandbox_mcp_langgraph.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/runtime/main.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/ui/agentbricks-mark.png +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/ui/markdown-it-15.0.2.min.js +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-openai/runtime/main.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-openai/ui/agentbricks-mark.png +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentbricks/templates/ui/agent-openai/ui/markdown-it-15.0.2.min.js +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/client.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/langgraph/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/langgraph/genie.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/langgraph/mcp.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/langgraph/memory.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/langgraph/responses.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/langgraph/session_store.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/model_upgrades/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/models.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/openai/genie.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/openai/history.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/openai/mcp.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/openai/memory.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/openai/sessions.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/py.typed +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/cors.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/durability/__init__.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/durability/heartbeat.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/genie.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/mcp_auth.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/model_services.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/prompt_registry.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/session_store_client.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/tracing.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/runtime/workspace.py +0 -0
- {databricks_agentbricks-0.7.1 → databricks_agentbricks-0.9.0}/src/databricks_agentkit/timefmt.py +0 -0
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
Metadata-Version: 2.5
|
|
2
2
|
Name: databricks-agentbricks
|
|
3
|
-
Version: 0.
|
|
3
|
+
Version: 0.9.0
|
|
4
4
|
Summary: Agent Bricks CLI and AgentKit SDK for custom agents on Databricks
|
|
5
5
|
Author-email: Databricks <agent-feedback@databricks.com>
|
|
6
6
|
License-File: NOTICE
|
|
@@ -27,6 +27,8 @@ Requires-Dist: langchain>=1.0.0; extra == 'langgraph'
|
|
|
27
27
|
Requires-Dist: langgraph>=1.1.0; extra == 'langgraph'
|
|
28
28
|
Requires-Dist: mlflow>=3.10.1; extra == 'langgraph'
|
|
29
29
|
Requires-Dist: opentelemetry-exporter-otlp-proto-grpc>=1.25.0; extra == 'langgraph'
|
|
30
|
+
Provides-Extra: loadtest
|
|
31
|
+
Requires-Dist: locust>=2.31; extra == 'loadtest'
|
|
30
32
|
Provides-Extra: openai
|
|
31
33
|
Requires-Dist: databricks-agents>=1.9.3; extra == 'openai'
|
|
32
34
|
Requires-Dist: databricks-openai>=0.13.0; extra == 'openai'
|
|
@@ -206,6 +208,12 @@ agentbricks deploy # deploy to Databricks as agent-bricks-my-agent
|
|
|
206
208
|
`agentbricks dev` runs the agent locally on `http://localhost:8000`, wrapping the Databricks Apps
|
|
207
209
|
local runtime so local behavior matches a deployment.
|
|
208
210
|
|
|
211
|
+
Memory and session stores are off by default in `agentbricks dev`, even when a store is bound in
|
|
212
|
+
`agent.toml`: the chat app shows "Not connected" for memory, and conversation state stays in
|
|
213
|
+
process. They connect after `agentbricks deploy`. To use real stores locally, set
|
|
214
|
+
`AGENT_MEMORY_STORE` (the Store ID) and `AGENT_SESSION_STORE` (the store name) in the project's
|
|
215
|
+
`.env`.
|
|
216
|
+
|
|
209
217
|
`agentbricks init` records the deployment name in `agent.toml`, so `agentbricks deploy` from the
|
|
210
218
|
project directory deploys a Databricks App named `agent-bricks-my-agent`, provisions the
|
|
211
219
|
stores declared in `agent.toml`, and grants the app's service principal access to the stores and
|
|
@@ -232,10 +240,10 @@ agentbricks deployments wait agent-bricks-my-agent
|
|
|
232
240
|
```
|
|
233
241
|
|
|
234
242
|
`deployments wait` returns within 100 seconds (`--timeout`). Exit 0 means the agent is deployed and
|
|
235
|
-
|
|
236
|
-
of the deploy log; exit 75 means it is still deploying, so run the same command again.
|
|
237
|
-
with `sleep`. Progress lines with the elapsed time go to stderr, also under `-o json`,
|
|
238
|
-
is a single JSON document. A second `deploy` of the same app is refused while one is still running
|
|
243
|
+
answering requests, and it prints the URL; exit 1 means the deploy failed, and it prints the error
|
|
244
|
+
and the end of the deploy log; exit 75 means it is still deploying, so run the same command again.
|
|
245
|
+
Don't poll with `sleep`. Progress lines with the elapsed time go to stderr, also under `-o json`,
|
|
246
|
+
whose stdout is a single JSON document. A second `deploy` of the same app is refused while one is still running
|
|
239
247
|
from the project; the deploy's state and background log are in `.agentbricks/deploy-<app>.json` and
|
|
240
248
|
`.agentbricks/deploy-<app>.log`. If the App is busy with another operation, `deploy` retries for
|
|
241
249
|
about two minutes before failing.
|
|
@@ -374,9 +382,9 @@ writers to one path can overwrite each other. `add()`'s `write_mode` argument is
|
|
|
374
382
|
deprecated; use `put()` instead. `memory_store.list(actor_id=..., path=...)` returns the
|
|
375
383
|
memories at one exact path.
|
|
376
384
|
|
|
377
|
-
All `list()` methods return iterators that automatically consume server pages.
|
|
378
|
-
`page_size` and search `limit` values must be between 1 and 100
|
|
379
|
-
also auto-pages. `session.fork(...)` creates an independent copy, optionally through
|
|
385
|
+
All `list()` methods return iterators that automatically consume server pages. Store and content list
|
|
386
|
+
`page_size` and search `limit` values must be between 1 and 100; permission lists accept 1–1000.
|
|
387
|
+
`session.list_items()` also auto-pages. `session.fork(...)` creates an independent copy, optionally through
|
|
380
388
|
a specific item. Deleting a session with descendants requires
|
|
381
389
|
`session.delete(force=True)` to cascade the deletion.
|
|
382
390
|
|
|
@@ -401,9 +409,8 @@ agentbricks dev
|
|
|
401
409
|
agentbricks --profile <profile> deploy my-agent
|
|
402
410
|
```
|
|
403
411
|
|
|
404
|
-
Use the URL `agentbricks dev` prints (port 8000, or `--app-port`).
|
|
405
|
-
|
|
406
|
-
a few seconds while the app starts.
|
|
412
|
+
Use the URL `agentbricks dev` prints (port 8000, or `--app-port`). Agents on different ports can run
|
|
413
|
+
side by side. Requests can fail with HTTP 500 for a few seconds while the app starts.
|
|
407
414
|
|
|
408
415
|
Use `--framework openai` for OpenAI Agents. Templates keep agent code separate from the runtime
|
|
409
416
|
adapter and declare default Session and Memory Store bindings in `agent.toml`.
|
|
@@ -436,9 +443,10 @@ while its record is retained; using the ID for a different request returns `409`
|
|
|
436
443
|
|
|
437
444
|
`agentbricks dev` keeps execution state in process and loses it on restart. For projects with
|
|
438
445
|
`[agent].server = "agentbricks"`, `agentbricks deploy` provisions a persistent Runtime Store for requests,
|
|
439
|
-
status, events, and results.
|
|
440
|
-
|
|
441
|
-
|
|
446
|
+
status, events, and results. Queued work is reserved for the worker that accepted it, so other
|
|
447
|
+
replicas cannot start it through normal claims. Register `@app.recover` to restart interrupted
|
|
448
|
+
app-auth work or take over abandoned queued work after worker failures. Recovery is at-least-once,
|
|
449
|
+
so external side effects must be idempotent. Session and Memory Stores separately preserve agent state.
|
|
442
450
|
|
|
443
451
|
The managed path uses the internal Runtime Store API to create a dedicated database in the
|
|
444
452
|
workspace's shared Lakebase project and give the app SP ownership. The managed runtime initializes its schema and
|
|
@@ -450,8 +458,9 @@ The switch does not migrate existing deployments between backends. Managed clean
|
|
|
450
458
|
the app for retry. Direct app deletion bypasses managed store cleanup.
|
|
451
459
|
|
|
452
460
|
For a tool using `auth = "user"`, the Runtime Store still records token-free invocation state,
|
|
453
|
-
events, and results. The forwarded user credential remains
|
|
454
|
-
is never written to the Runtime Store. A replacement attempt
|
|
461
|
+
events, and results. The forwarded user credential remains in memory on the receiving worker while
|
|
462
|
+
queued and during the first attempt; it is never written to the Runtime Store. A replacement attempt
|
|
463
|
+
after failure recovery, including takeover before execution started, stops with
|
|
455
464
|
`MCP_USER_AUTH_RECOVERY_UNSUPPORTED` because the original request credential is no longer present.
|
|
456
465
|
|
|
457
466
|
Use `server = "custom"` to deploy your own HTTP server without provisioning a Runtime Store.
|
|
@@ -667,8 +676,9 @@ it in `recall` and `list_memories`. `memory_tools` resolves the store from the `
|
|
|
667
676
|
binding, carried to the runtime by the `AGENT_MEMORY_STORE` env var that `agentbricks deploy`
|
|
668
677
|
injects, and returns no tools when no store is set, so the agent runs unchanged.
|
|
669
678
|
That "no store set" path is also how it runs under `agentbricks dev`, which runs locally: memory is off
|
|
670
|
-
there (the store is provisioned and used only at deploy)
|
|
671
|
-
|
|
679
|
+
by default there (the store is provisioned and used only at deploy), even when a store is bound in
|
|
680
|
+
`agent.toml`. Setting `AGENT_MEMORY_STORE` in the project's `.env` uses that store locally. The
|
|
681
|
+
OpenAI Agents adapter exposes the same tools as `databricks_agentkit.openai.memory_tools(actor)`:
|
|
672
682
|
|
|
673
683
|
```python
|
|
674
684
|
from databricks_agentkit.langgraph import memory_tools
|
|
@@ -683,6 +693,60 @@ agent = create_agent(model=..., tools=[*your_tools, *memory_tools(actor)])
|
|
|
683
693
|
> `session_store.grant_permission(principal_id)` or `memory_store.grant_permission(principal_id)`;
|
|
684
694
|
> `agentbricks deploy` does this for the deployed app automatically.
|
|
685
695
|
|
|
696
|
+
### Managing store permissions
|
|
697
|
+
|
|
698
|
+
Both store types support `permissions get`, `list`, `grant`, and `revoke`:
|
|
699
|
+
|
|
700
|
+
```sh
|
|
701
|
+
# Inspect your own effective access, or list explicit grants as the creator/admin.
|
|
702
|
+
agentbricks sessions stores permissions get support-agent-sessions
|
|
703
|
+
agentbricks memory stores permissions list memory-stores/<id> --all
|
|
704
|
+
|
|
705
|
+
# Grant a user READ or an app service principal WRITE.
|
|
706
|
+
agentbricks sessions stores permissions grant support-agent-sessions \
|
|
707
|
+
--user teammate@example.com --permission READ
|
|
708
|
+
agentbricks memory stores permissions grant memory-stores/<id> \
|
|
709
|
+
--service-principal <application-id> --permission WRITE
|
|
710
|
+
|
|
711
|
+
# Remove both READ and WRITE grants. --yes skips confirmation.
|
|
712
|
+
agentbricks memory stores permissions revoke memory-stores/<id> \
|
|
713
|
+
--service-principal <application-id> --yes
|
|
714
|
+
```
|
|
715
|
+
|
|
716
|
+
Session stores accept their configured name or `session-stores/<name>`. Memory stores accept an ID
|
|
717
|
+
or `memory-stores/<id>`; permission Get also accepts the display-name lookup supported by store Get
|
|
718
|
+
and rejects ambiguous matches. The global `--profile` and `--output json` options go before
|
|
719
|
+
`sessions` or `memory`. List follows the existing CLI pagination convention: one JSON/piped page
|
|
720
|
+
with `next_page_token`, interactive text pages, or every page with `--all`.
|
|
721
|
+
|
|
722
|
+
Get always returns the authenticated caller's effective permission, including
|
|
723
|
+
`PERMISSION_LEVEL_UNSPECIFIED` for no access (shown as "No access" in text). List reports direct
|
|
724
|
+
grants and excludes implicit ownership and inherited access. Grant/List/Revoke require the store
|
|
725
|
+
creator or a workspace admin, even if the caller has WRITE. Service principals use application/client
|
|
726
|
+
IDs, not numeric SCIM IDs. WRITE includes READ; granting READ preserves existing WRITE. Revoke
|
|
727
|
+
removes both levels of direct grants, is idempotent, and cannot revoke ownership or inherited access.
|
|
728
|
+
A later deploy can restore the app's grants. These APIs require workspace preview/feature enablement.
|
|
729
|
+
|
|
730
|
+
The Python SDK exposes the same operations on both store objects:
|
|
731
|
+
|
|
732
|
+
```python
|
|
733
|
+
# Existing service-principal grants retain the WRITE default.
|
|
734
|
+
session_store.grant_permission("<application-id>")
|
|
735
|
+
memory_store.grant_permission("teammate@example.com", principal_type="USER", permission="READ")
|
|
736
|
+
|
|
737
|
+
permission = session_store.get_permission()
|
|
738
|
+
print(permission.principal.name, permission.principal.principal_type, permission.permission)
|
|
739
|
+
for grant in memory_store.list_permissions(page_size=100):
|
|
740
|
+
print(grant.principal.name, grant.permission)
|
|
741
|
+
|
|
742
|
+
memory_store.revoke_permission("teammate@example.com", principal_type="USER")
|
|
743
|
+
```
|
|
744
|
+
|
|
745
|
+
`get_permission()` returns a `StorePermission`; `list_permissions()` returns an iterator of those
|
|
746
|
+
values and automatically follows every page. Each contains a `Principal` with `name` and
|
|
747
|
+
`principal_type`, plus a `permission` value. Grant/Revoke default to `SERVICE_PRINCIPAL`;
|
|
748
|
+
pass `principal_type="USER"` for user emails.
|
|
749
|
+
|
|
686
750
|
### How sessions and memory are keyed
|
|
687
751
|
|
|
688
752
|
In an agent generated by `agentbricks init`, each invocation picks its session and its memory
|
|
@@ -697,8 +761,8 @@ trusted application context. The chat app sends the signed-in user; HTTP clients
|
|
|
697
761
|
`agentbricks endpoint invoke`, have to send it themselves. With a request-user (`auth = "user"`) tool
|
|
698
762
|
bound, the runtime also namespaces a supplied `actor` to the signed-in user.
|
|
699
763
|
|
|
700
|
-
Memory is off under `agentbricks dev`, so try this against a deployed agent: two
|
|
701
|
-
actor.
|
|
764
|
+
Memory is off by default under `agentbricks dev`, so try this against a deployed agent: two
|
|
765
|
+
conversations, one actor.
|
|
702
766
|
|
|
703
767
|
```sh
|
|
704
768
|
ask() { agentbricks --profile <profile> endpoint invoke agent-bricks-my-agent --path /api/invocations \
|
|
@@ -707,6 +771,37 @@ ask "Remember that I prefer answers as bullet points." # conversation 1 saves a
|
|
|
707
771
|
ask "How do I like my answers formatted?" # conversation 2 recalls it
|
|
708
772
|
```
|
|
709
773
|
|
|
774
|
+
### Seeding and using memory from the CLI
|
|
775
|
+
|
|
776
|
+
There are two ways to get a memory store behind a deployed agent. Both end with
|
|
777
|
+
`agentbricks deploy`, which creates the store if it doesn't exist and grants the app access to it.
|
|
778
|
+
|
|
779
|
+
**Start with an empty store.** Bind the store name, then deploy:
|
|
780
|
+
|
|
781
|
+
```sh
|
|
782
|
+
agentbricks memory bind support-agent-memory
|
|
783
|
+
agentbricks deploy my-agent
|
|
784
|
+
```
|
|
785
|
+
|
|
786
|
+
**Start with memories already in the store.** Create the store, add entries, then bind and deploy:
|
|
787
|
+
|
|
788
|
+
```sh
|
|
789
|
+
agentbricks memory stores create --name support-agent-memory # prints a Store ID
|
|
790
|
+
agentbricks memory entries create --store <Store ID> \
|
|
791
|
+
--actor-id you@example.com \
|
|
792
|
+
--path /preferences/style.md \
|
|
793
|
+
--content "Prefers Python and concise answers with code examples."
|
|
794
|
+
agentbricks memory bind support-agent-memory
|
|
795
|
+
agentbricks deploy my-agent
|
|
796
|
+
```
|
|
797
|
+
|
|
798
|
+
`--path` is the entry's key inside the store, written like a file path; it isn't a file on your
|
|
799
|
+
machine. `--actor-id` is the user the memory belongs to. In the deployed chat app, the actor is the
|
|
800
|
+
signed-in user's email, so seed entries under your own email to see them recalled. `--store` takes
|
|
801
|
+
the Store ID that `memory stores create` (or `memory stores list`) prints, while `bind` takes the
|
|
802
|
+
store name. Check what you seeded with
|
|
803
|
+
`agentbricks memory entries list --store <Store ID> --actor-id you@example.com`.
|
|
804
|
+
|
|
710
805
|
### Declaring and provisioning stores
|
|
711
806
|
|
|
712
807
|
For a deployed agent, `agent.toml` declares which stores it uses and `agentbricks deploy` provisions them —
|
|
@@ -728,11 +823,18 @@ agentbricks memory bind support-agent-memory
|
|
|
728
823
|
agentbricks deploy my-agent
|
|
729
824
|
```
|
|
730
825
|
|
|
826
|
+
A store that already exists is reused as is: `deploy` doesn't create a second one or change who else
|
|
827
|
+
can access it. Its only permission change is a WRITE grant for the app's service principal, which
|
|
828
|
+
needs you to be the store's creator or a workspace admin; if you aren't, the deploy still succeeds
|
|
829
|
+
and reports the missing grant. In a shared workspace, give teammates access with
|
|
830
|
+
`agentbricks memory stores permissions grant` (or `agentbricks sessions stores permissions grant`);
|
|
831
|
+
see [Managing store permissions](#managing-store-permissions).
|
|
832
|
+
|
|
731
833
|
Memory and session stores are independent resources: deleting one never affects the other.
|
|
732
834
|
For projects with `agent.toml`, unbinding either store removes its declaration; the next deploy
|
|
733
835
|
removes the corresponding `AGENT_MEMORY_STORE` or `AGENT_SESSION_STORE` entry from `app.yaml`.
|
|
734
836
|
Unbinding does not delete the store or revoke an existing grant to the App's service principal.
|
|
735
|
-
|
|
837
|
+
Use the permission commands below as the store creator or a workspace admin to revoke access.
|
|
736
838
|
|
|
737
839
|
## Model and prompt upgrades (experimental)
|
|
738
840
|
|
|
@@ -748,7 +850,7 @@ and declares the Prompt Registry prompts it loads so the search can rewrite them
|
|
|
748
850
|
|
|
749
851
|
```sh
|
|
750
852
|
agentbricks experimental models bind main.my_agent.router_llm --role router --default system.ai.claude-haiku-4-5
|
|
751
|
-
agentbricks experimental models bind main.my_agent.writer_llm --role writer --default system.ai.claude-sonnet-
|
|
853
|
+
agentbricks experimental models bind main.my_agent.writer_llm --role writer --default system.ai.claude-sonnet-5-5
|
|
752
854
|
agentbricks experimental prompts bind main.my_agent.writer
|
|
753
855
|
agentbricks deploy my-agent # creates the services and grants the app access to them and the prompt
|
|
754
856
|
agentbricks experimental models upgrade -c router=claude-haiku-4-5,gpt-5-4-nano -c writer=claude-haiku-4-5 \
|
|
@@ -790,7 +892,7 @@ result = optimize_prompts_and_models(
|
|
|
790
892
|
prompt_uris=["prompts:/main.my_agent.router@production"],
|
|
791
893
|
gateway_endpoints={
|
|
792
894
|
"main.my_agent.router_llm": ["claude-haiku-4-5", "gpt-5-4-mini"],
|
|
793
|
-
"main.my_agent.writer_llm": ["claude-sonnet-
|
|
895
|
+
"main.my_agent.writer_llm": ["claude-sonnet-5-5", "claude-haiku-4-5"],
|
|
794
896
|
},
|
|
795
897
|
scorers=scorers, # MLflow scorers, or (inputs, expectations, answer) -> float
|
|
796
898
|
max_metric_calls=200,
|
|
@@ -867,6 +969,9 @@ agentbricks [-p <profile>] [-o text|json]
|
|
|
867
969
|
list | get | logs | start | stop | delete
|
|
868
970
|
endpoint
|
|
869
971
|
invoke [APP] --path PATH [--url URL] [--json JSON] [--sse]
|
|
972
|
+
loadtest [APP] --path PATH [--url URL] [--requests FILE] [--sse]
|
|
973
|
+
[--concurrency N] [--max-duration SECONDS] [--max-requests N] [--no-warmup]
|
|
974
|
+
[--report-dir PATH] [--save-responses FILE]
|
|
870
975
|
```
|
|
871
976
|
|
|
872
977
|
## Bring an existing agent
|
|
@@ -967,6 +1072,96 @@ as the session id - put the session id at the top level of the `--json` body for
|
|
|
967
1072
|
also works with a direct App URL and with the generated runtime on localhost. OAuth and session headers are managed by
|
|
968
1073
|
the runtime; arbitrary custom request headers are intentionally not exposed by this command.
|
|
969
1074
|
|
|
1075
|
+
## Load test HTTP endpoints
|
|
1076
|
+
|
|
1077
|
+
`agentbricks endpoint loadtest` repeats HTTP requests with Locust. The load generator runs on the
|
|
1078
|
+
machine where you execute the command; the target can be a deployed Databricks App, localhost, or
|
|
1079
|
+
another HTTP server. Each measured request is one HTTP exchange, including the complete response
|
|
1080
|
+
body or SSE stream. A `202 Accepted` response measures submission, not eventual background work.
|
|
1081
|
+
|
|
1082
|
+
Install the optional dependency and create a JSON array of complete request bodies. A one-element
|
|
1083
|
+
array repeats one body; multiple bodies cycle in order. Omit `--requests` for a bodyless request.
|
|
1084
|
+
`endpoint invoke` retains `--json` for a single request; load testing uses only `--requests`.
|
|
1085
|
+
|
|
1086
|
+
```sh
|
|
1087
|
+
pip install 'databricks-agentbricks[loadtest]'
|
|
1088
|
+
cat > requests.json <<'JSON'
|
|
1089
|
+
[
|
|
1090
|
+
{
|
|
1091
|
+
"id": "{{request_id}}",
|
|
1092
|
+
"session_id": "{{user_id}}",
|
|
1093
|
+
"input": [{"role": "user", "content": "Hello"}],
|
|
1094
|
+
"stream": true
|
|
1095
|
+
}
|
|
1096
|
+
]
|
|
1097
|
+
JSON
|
|
1098
|
+
agentbricks --profile <profile> endpoint loadtest agent-bricks-my-agent \
|
|
1099
|
+
--path /api/invocations --requests requests.json --sse \
|
|
1100
|
+
--concurrency 5 --max-duration 60 --max-requests 200 --report-dir results/loadtest
|
|
1101
|
+
```
|
|
1102
|
+
|
|
1103
|
+
`--concurrency` sets simultaneous users (default: 1). Each sends another request after its previous
|
|
1104
|
+
request finishes. `--max-duration` is in seconds (default: 60); `--max-requests` optionally caps measured
|
|
1105
|
+
requests. Whichever limit is reached first stops dispatch, then in-flight requests drain for up to
|
|
1106
|
+
`--timeout` seconds. Ctrl-C also drains and preserves a partial report. Failed or incomplete
|
|
1107
|
+
requests make the command exit nonzero; cancellation exits 130 or 143 for SIGINT or SIGTERM.
|
|
1108
|
+
|
|
1109
|
+
By default, one real **warm-up** request runs before measurement using the first body. It can incur
|
|
1110
|
+
model/tool costs, mutate state, and warm the endpoint. Its request and latency, including cold-start
|
|
1111
|
+
latency, are excluded from the cap, report counts, and timings. A failed warm-up stops the test.
|
|
1112
|
+
Use `--no-warmup` to send only measured requests, including any cold-start latency. Reports state
|
|
1113
|
+
whether the excluded warm-up ran. Set a suitable `--timeout` for the complete response.
|
|
1114
|
+
|
|
1115
|
+
Request bodies support two identity lifetimes:
|
|
1116
|
+
|
|
1117
|
+
| Placeholder | Lifetime | Example use |
|
|
1118
|
+
| --- | --- | --- |
|
|
1119
|
+
| `{{request_id}}` | Fresh UUID per request, including the excluded warm-up | Agent invocation/idempotency ID |
|
|
1120
|
+
| `{{user_id}}` | Stable UUID per simulated user; warm-up uses its own identity | Agent session ID |
|
|
1121
|
+
| `{{uuid}}` | Alias of the current `{{request_id}}` | Existing request datasets |
|
|
1122
|
+
|
|
1123
|
+
Durable agents need a fresh invocation `id` for every request; a fixed ID can replay cached output.
|
|
1124
|
+
Use `{{request_id}}` for `id` and `{{user_id}}` for a session that should persist across one user's
|
|
1125
|
+
requests. By default, each App user's `X-Routing-Key` matches its generated user identity; an explicit
|
|
1126
|
+
`--routing-key` pins all users to that routing key. Routing and session contents are separate: include
|
|
1127
|
+
`session_id` in the body when the server requires it. Tests run closed-loop, without a think-time pause.
|
|
1128
|
+
|
|
1129
|
+
The final summary includes counts, throughput including drain, successful-only and all-completed
|
|
1130
|
+
latency percentiles, answer-text TTFT, token usage, coverage, incomplete attempts, and generator CPU
|
|
1131
|
+
warnings. With `--output json` before `endpoint`, stdout contains one Agent Bricks summary object;
|
|
1132
|
+
progress stays on stderr. This is the same object saved as `summary.json`.
|
|
1133
|
+
Locust's rounded histogram buckets are used for percentiles; all-completed timings include failures.
|
|
1134
|
+
These bounded tests measure the selected workload and do not establish sustainable capacity or answer quality.
|
|
1135
|
+
|
|
1136
|
+
`--report-dir` writes **all HTML, CSV, and JSON formats together**: `index.html`, native `locust.html`
|
|
1137
|
+
charts, `locust_*.csv` statistics/history/errors, and `summary.json`. There is no format selector.
|
|
1138
|
+
Use a new or empty directory. Reports contain metrics and the target host, without raw request or
|
|
1139
|
+
answer content or credentials. Review the target host before sharing a report.
|
|
1140
|
+
|
|
1141
|
+
For SSE responses, latency includes stream completion. Answer-text TTFT recognizes supported
|
|
1142
|
+
DurableAgentServer, OpenAI Chat Completions, and Responses SSE events, ignoring metadata and tool
|
|
1143
|
+
events. Missing text/token usage is unknown, with coverage reported; token counts are never inferred.
|
|
1144
|
+
Runtime model-message usage is summed across distinct messages; repeated message IDs update that
|
|
1145
|
+
message's total. Chat Completions and Responses usage snapshots replace earlier totals.
|
|
1146
|
+
Recognized error events and truncated supported streams fail the request. This command does not
|
|
1147
|
+
perform background polling, reconnect, or stream replay.
|
|
1148
|
+
|
|
1149
|
+
To save responses, add `--save-responses responses.json`. This exports **every successful measured
|
|
1150
|
+
request**, without sampling; it excludes warm-up, failed requests, and incomplete responses.
|
|
1151
|
+
The new file has mode 0600 and must be outside `--report-dir`, in an existing parent directory.
|
|
1152
|
+
It contains raw request inputs, outputs, and explicit trace IDs returned by an already instrumented
|
|
1153
|
+
agent (`X-MLflow-Trace-ID` or top-level `trace_id`). Keep it private and out of shared report bundles.
|
|
1154
|
+
The export is streamed to disk rather than retaining all results in memory; size grows with the load.
|
|
1155
|
+
|
|
1156
|
+
```json
|
|
1157
|
+
{"schema_version":1,"case_count":1,"observations":[{"case_index":0,"inputs":{},"outputs":"...","trace_id":null}]}
|
|
1158
|
+
```
|
|
1159
|
+
|
|
1160
|
+
`case_index` is zero-based. Supported SSE outputs contain accumulated answer text; JSON outputs use
|
|
1161
|
+
its `output` value when present, otherwise the response body. Bodyless inputs are `null`.
|
|
1162
|
+
Use native MLflow APIs separately for evaluation; load testing does not run scorers, fetch traces,
|
|
1163
|
+
or enable server tracing.
|
|
1164
|
+
|
|
970
1165
|
## Command help
|
|
971
1166
|
|
|
972
1167
|
Use the conventional help flag at any command level. Every command's help includes runnable
|
|
@@ -996,6 +1191,9 @@ agentbricks tools remove mcp system.ai.web_search
|
|
|
996
1191
|
agentbricks tools list
|
|
997
1192
|
```
|
|
998
1193
|
|
|
1194
|
+
`agentbricks tools list` shows integrations you can add. To see the tools an agent already has, open
|
|
1195
|
+
**Tools** under Capabilities in the chat app.
|
|
1196
|
+
|
|
999
1197
|
### Managed tool identity and migration
|
|
1000
1198
|
|
|
1001
1199
|
`agentbricks tools add mcp`, `agentbricks tools add sandbox`, `agentbricks tools add uc-function`,
|
|
@@ -1091,11 +1289,16 @@ generated request-aware agent function.
|
|
|
1091
1289
|
|
|
1092
1290
|
Request-user invocations use the same synchronous, streaming, background, status, event-replay,
|
|
1093
1291
|
and idempotency APIs as app-auth invocations. The Runtime Store records only token-free request
|
|
1094
|
-
state, events, and results. The forwarded credential stays process-local
|
|
1095
|
-
attempt and closes when that attempt completes, fails, or is cancelled.
|
|
1292
|
+
state, events, and results. The forwarded credential stays process-local while queued and during
|
|
1293
|
+
the first attempt, and closes when that attempt completes, fails, or is cancelled. The store
|
|
1294
|
+
enforces affinity to the accepting worker. A replacement attempt after
|
|
1096
1295
|
failure recovery stops with `MCP_USER_AUTH_RECOVERY_UNSUPPORTED` because no user credential is
|
|
1097
1296
|
available; neither the invoke nor recovery handler runs for that attempt.
|
|
1098
1297
|
|
|
1298
|
+
Event reconnection does not resume execution. OBO approval/resume remains unsupported
|
|
1299
|
+
(`MCP_USER_AUTH_HITL_UNSUPPORTED`). Upgrade all replicas before relying on affinity; see the
|
|
1300
|
+
[runtime guide](src/databricks_agentkit/runtime/README.md#deployed-execution) for details.
|
|
1301
|
+
|
|
1099
1302
|
Before deploying user-auth tools from an older project, migrate its request handler and framework
|
|
1100
1303
|
adapter to the current request-auth-aware `DurableAgentServer` template, then explicitly choose `user` or
|
|
1101
1304
|
`app` on **every** managed MCP, sandbox, or Genie entry. Changing `agent.toml` alone does not
|
|
@@ -1114,7 +1317,7 @@ Deploy derives Apps user scopes from explicit `auth = "user"` bindings and union
|
|
|
1114
1317
|
| UC function | `mcp.functions` |
|
|
1115
1318
|
| Sandbox with a Volume downscope | `ai-gateway`, `files` |
|
|
1116
1319
|
| First-class Genie One or Genie Agent | `genie` |
|
|
1117
|
-
| Sandbox with token injection | `ai-gateway`, `workspace.workspace` |
|
|
1320
|
+
| Sandbox with token injection | `ai-gateway`, `workspace.workspace`, `files`, `sql`, `catalog.catalogs:read`, `catalog.schemas:read`, `catalog.tables:read` |
|
|
1118
1321
|
| Sandbox with token injection disabled | `ai-gateway` |
|
|
1119
1322
|
|
|
1120
1323
|
For example, bind Genie tools in a current project with `server = "agentbricks"`:
|
|
@@ -1142,15 +1345,20 @@ The `system.ai.dbsql` managed MCP additionally requests the Apps `sql` user scop
|
|
|
1142
1345
|
API consent, not `sql:restricted-query`; read-only enforcement remains the service policy plus the
|
|
1143
1346
|
requesting user's Unity Catalog grants. DBSQL does not use Databricks Connect.
|
|
1144
1347
|
|
|
1145
|
-
When a user-auth sandbox has `databricks_access_token_included = true`,
|
|
1146
|
-
|
|
1147
|
-
|
|
1148
|
-
|
|
1149
|
-
Unity Catalog
|
|
1150
|
-
|
|
1151
|
-
|
|
1152
|
-
|
|
1153
|
-
the
|
|
1348
|
+
When a user-auth sandbox has `databricks_access_token_included = true`, the injected credential
|
|
1349
|
+
carries only the App's user API scopes, so the sandbox requests `workspace.workspace` (workspace
|
|
1350
|
+
files), `files` (Volume files through the Files API), `sql` (SQL warehouses through the Statement
|
|
1351
|
+
Execution API), and `catalog.catalogs:read`, `catalog.schemas:read`, and `catalog.tables:read`
|
|
1352
|
+
(Unity Catalog metadata). Apps has no user API scope for Databricks Connect, clusters, jobs, or
|
|
1353
|
+
secrets, so sandbox code cannot call those APIs as the request user; use `auth = "app"` for them.
|
|
1354
|
+
An existing App deployed before these defaults needs `--allow-user-scope-update` once. A sandbox
|
|
1355
|
+
binding with a Volume downscope also requests the Apps `files` user scope.
|
|
1356
|
+
OAuth consent does not grant data access: the requesting user still needs the corresponding
|
|
1357
|
+
Unity Catalog and warehouse privileges, and the sandbox downscope remains authoritative. A sandbox
|
|
1358
|
+
binding with token injection disabled requests none of these credential scopes; its other
|
|
1359
|
+
resource-derived scopes still apply. Databricks Apps rejects the legacy bare `workspace` scope, so
|
|
1360
|
+
Agent Bricks requests `workspace.workspace`. These scopes are requested only for `auth = "user"`;
|
|
1361
|
+
`auth = "app"` uses the App service principal's permissions instead.
|
|
1154
1362
|
|
|
1155
1363
|
Review the target App's scopes and coordinate with its other owners before allowing the update. Once
|
|
1156
1364
|
those scopes are present, later deploys do not need the flag. The CLI preserves unrelated scopes,
|
|
@@ -1220,16 +1428,8 @@ with TOML inspection. Replace `agentbricks mcp list [--schema catalog.schema]` w
|
|
|
1220
1428
|
`agentbricks tools list --kind mcp [--schema catalog.schema]`; the former command is removed. Use
|
|
1221
1429
|
`agentbricks tools list --help` for the new discovery contract.
|
|
1222
1430
|
|
|
1223
|
-
|
|
1224
|
-
|
|
1225
|
-
|
|
1226
|
-
```sh
|
|
1227
|
-
AGENTBRICKS_E2E_PROFILE=<profile> .venv-functional/bin/pytest tests/e2e/tool_discovery_test.py -v
|
|
1228
|
-
```
|
|
1229
|
-
|
|
1230
|
-
The live checks compare default and MCP-filtered discovery with the compatibility service list.
|
|
1231
|
-
Set `AGENTBRICKS_E2E_SCHEMA=catalog.schema` to exercise an additional schema. The installed CLI's local
|
|
1232
|
-
add/review/remove flows and all updated help pages are covered by `tests/functional/cli_smoke_test.py`.
|
|
1431
|
+
The local add/review/remove flows and all updated help pages are covered by the unit tests
|
|
1432
|
+
(`tools_test.py`, `tools_discovery_test.py`, `cli_test.py`).
|
|
1233
1433
|
|
|
1234
1434
|
In managed-server templates, custom Python tools are code-first. Write them with the framework's native
|
|
1235
1435
|
decorator in `agent/tools/`: LangGraph uses `@tool`, while OpenAI Agents uses `@function_tool`. The
|
|
@@ -1358,18 +1558,6 @@ The effective caller needs access to the data, the SQL warehouse, and the select
|
|
|
1358
1558
|
where applicable. Agent Bricks CLI does not grant permissions or promise a service-principal fallback when
|
|
1359
1559
|
caller credentials lack access. An offline add succeeding does not establish runtime access.
|
|
1360
1560
|
|
|
1361
|
-
The opt-in live tests exercise both frameworks against the configured workspace and an existing
|
|
1362
|
-
Genie space. From `integrations/agentbricks`, with both framework extras installed:
|
|
1363
|
-
|
|
1364
|
-
```sh
|
|
1365
|
-
DATABRICKS_CONFIG_PROFILE=my-workspace RUN_AGENTBRICKS_GENIE_TESTS=1 \
|
|
1366
|
-
AGENTBRICKS_GENIE_SPACE_ID=SPACE_ID \
|
|
1367
|
-
uv run pytest tests/integration_tests/genie_tools_test.py
|
|
1368
|
-
```
|
|
1369
|
-
|
|
1370
|
-
By default they ask for the row count of `samples.nyctaxi.trips`. Set `AGENTBRICKS_GENIE_QUESTION` for
|
|
1371
|
-
another dataset and `AGENTBRICKS_GENIE_EXPECTED_VALUE` to assert a known result cell.
|
|
1372
|
-
|
|
1373
1561
|
## Initialize the chat app demo
|
|
1374
1562
|
|
|
1375
1563
|
The chat app is a LangGraph-specific init overlay, not a command that mutates an existing project.
|
|
@@ -1399,7 +1587,8 @@ agentbricks --profile <profile> deploy agent-bricks-agent-demo --source .
|
|
|
1399
1587
|
(`bind` declares the store name in `agent.toml`; `agentbricks deploy` creates any declared-but-missing
|
|
1400
1588
|
store and grants the app's service principal access to it. The memory store id flows to the runtime
|
|
1401
1589
|
via the `AGENT_MEMORY_STORE` env var that `deploy` injects; `agentbricks dev` runs locally with memory off
|
|
1402
|
-
and does not inject it
|
|
1590
|
+
and does not inject it, so the chat app shows "Not connected" for memory until you deploy. The id is
|
|
1591
|
+
not persisted in `agent.toml`.)
|
|
1403
1592
|
|
|
1404
1593
|
The chat UI generates a stable application session UUID in browser local storage, sends it as the
|
|
1405
1594
|
invocation's top-level `session_id`, sends the signed-in user as `input.actor`, and creates a fresh
|