fast_unit_duration_allowlist.txt

# Backend unit-test node IDs grandfathered against BACKEND_FAST_UNIT_FAIL_SECONDS (default 0.12s locally).
#
# The duration guard measures per-test CPU time (call phase). CPU time is load-independent, so the
# guard is deterministic regardless of parallelism. The entries below are genuine unit tests whose
# measured CPU exceeds the failure threshold. The dominant cause is the file-isolation
# runner model: each test file runs in its own pytest process, so the first test of a file/class
# amortizes that process's module import (FastAPI app / router / database graph) into its call.
# This is structural, not a per-test regression. They STAY in the PR unit lane; this file only
# suppresses the duration guard for these baseline node IDs. NEW slow tests are still flagged.
#
# To shrink this list: (a) migrate the runner to a single pytest session (BACKEND_PYTEST_FILE_ISOLATION=0)
# so imports are paid once per worker instead of per file, or (b) raise BACKEND_FAST_UNIT_FAIL_SECONDS.
# Genuinely non-unit tests (real asyncio sleeps, network, stress, codebase greps, full-app wiring,
# per-test fresh module reload) are marked `slow`/`integration` and are NOT listed here.
# One full _proxy round trip (metering -> sanitize -> dispatch -> terminal telemetry) with an
# in-process capturing client; hermetic, but the end-to-end path costs ~0.15-0.19s CPU.
tests/unit/test_desktop_proxy.py::test_proxy_dispatches_server_paid_bodies_with_the_2048_cap
# Renders all eight first-party charts in dev and prod; still hermetic and required in the unit lane.
tests/unit/test_llm_gateway_deploy_contract.py::test_gateway_deploy_workflows_bind_identity_and_gate_serving_static_contract
tests/unit/test_rendered_deployment_contract.py::test_repository_charts_render_the_full_contract_when_helm_is_available
# Builds 10,001 in-memory records to prove universal export has no historical-data cap.
tests/unit/test_universal_memory_service.py::test_export_has_no_silent_record_cap
# Intentionally crosses the legacy 6,000-row boundary to prove cursor paging has no hidden cap.
tests/unit/test_universal_memory_list_cursor.py::test_cursor_pages_beyond_6000_historical_with_bounded_calls_no_cap
# Boots the developer memory HTTP surface to verify pending canonical rows remain observable.
tests/unit/test_memory_visibility_export_fixes.py::test_developer_list_requests_pending_processing
# Exhaustive compatibility and deployment regressions intentionally traverse large synthetic
# Firestore datasets or repository manifests while remaining hermetic unit tests.
tests/services/users/test_data_export.py::test_iter_user_data_export_paginates_complete_collections
# Exercises the complete FastAPI export route so memory preflight failure is proven to occur
# before streaming response headers; the full TestClient boundary costs ~0.17s CPU in isolation.
tests/routers/test_users.py::test_export_all_user_data_returns_500_before_streaming_headers_when_memory_preflight_fails
tests/unit/test_bounded_firestore_list_reads.py::test_knowledge_graph_route_exposes_truncation
tests/unit/test_memory_replace_policy.py::test_universal_reextract_failure_preserves_existing_memories
tests/unit/test_backend_runtime_env_validator.py::test_runtime_manifest_rejects_retired_per_user_memory_inventory
tests/unit/test_backend_runtime_env_validator.py::test_repo_parakeet_admission_deploy_contract_is_explicit
# Repository-manifest render/validation paths parse the full checked-in runtime contract while
# remaining hermetic; their CPU cost fluctuates narrowly around the unit guard threshold.
tests/unit/test_backend_runtime_env_validator.py::test_repo_rendered_cloud_run_artifact_matches_manifest
tests/unit/test_backend_runtime_env_validator.py::test_parakeet_selected_without_endpoint_is_rejected_for_rendered_cloud_run_state
tests/unit/test_backend_runtime_env_validator.py::test_repo_ilb_endpoints_use_http_scheme[prod]
tests/unit/test_backend_runtime_env_validator.py::test_parakeet_admission_deploy_contract_rejects_missing_or_invalid_values[PARAKEET_STREAM_CAPACITY-0-PARAKEET_STREAM_CAPACITY must be an integer >= 1]
# Same group: each renders the desktop-backend compose contract end to end. Measured 0.10-0.13s
# CPU on an idle host, i.e. already above the 0.1s target before any contention.
tests/unit/test_backend_runtime_env_validator.py::test_desktop_backend_compose_requires_vertex_pt_env[dev]
tests/unit/test_backend_runtime_env_validator.py::test_desktop_backend_compose_requires_vertex_pt_env[prod]
tests/unit/test_backend_runtime_env_validator.py::test_desktop_backend_vertex_pt_omission_fails_loud
tests/unit/test_render_backend_runtime_env.py::test_selected_job_renders_only_shared_network_and_named_job_outputs
tests/unit/test_render_backend_runtime_env.py::test_render_prod_gateway_callers_inject_verified_endpoint
tests/unit/test_render_backend_runtime_env.py::test_render_prod_emits_memory_maintenance_job_cron_on
tests/unit/test_render_backend_runtime_env.py::test_render_prod_requires_vpc_env_vars_before_job_outputs
tests/unit/test_render_backend_runtime_env.py::test_selected_job_rejects_unknown_name_without_emitting_partial_output
# MEMORY_ENABLED product-flag tests parse the composed runtime manifest.
tests/unit/test_memory_enabled_product_flag.py::test_dev_overlay_pins_memory_enabled_on_and_prod_stays_off
tests/unit/test_memory_enabled_product_flag.py::test_memory_enabled_on_does_not_require_maintenance
tests/unit/test_backend_runtime_env_validator.py::test_repo_ilb_endpoints_use_http_scheme[dev]
tests/unit/test_backend_runtime_env_validator.py::test_scheduler_runtime_surfaces_declare_orphan_stale_setting[prod]
tests/unit/test_backend_runtime_env_validator.py::test_scheduler_runtime_surfaces_declare_orphan_stale_setting[dev]
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_auto_dev_workflow_is_listed_and_targets_job
tests/unit/test_backend_runtime_env_validator.py::test_parakeet_admission_deploy_contract_rejects_missing_or_invalid_values[PARAKEET_STREAM_CAPACITY-None-missing PARAKEET_STREAM_CAPACITY]
tests/unit/test_backend_runtime_env_validator.py::test_parakeet_admission_deploy_contract_rejects_missing_or_invalid_values[PARAKEET_STREAM_ALLOCATION_PERCENT-101-PARAKEET_STREAM_ALLOCATION_PERCENT must be an integer from 0 through 100]
tests/unit/test_backend_runtime_env_validator.py::test_prod_account_deletion_dispatch_contract_rejects_missing_or_inline_profile
tests/unit/test_backend_runtime_env_validator.py::test_prod_listen_finalization_contract_requires_the_dedicated_worker_bindings
tests/unit/test_backend_runtime_env_validator.py::test_listen_finalization_contract_requires_maps_key_on_backend_sync
tests/unit/test_backend_runtime_env_validator.py::test_prod_cloud_run_secret_bindings_exclude_stale_optional_secrets
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_allows_write_without_job_cron
tests/unit/test_backend_runtime_env_validator.py::test_repo_prod_rendered_cloud_run_state_matches_manifest
tests/unit/test_backend_runtime_env_validator.py::test_dev_cloud_run_pusher_contract_rejects_legacy_and_non_listener_bindings
tests/unit/test_backend_runtime_env_validator.py::test_dev_cloud_run_pusher_contract_rejects_job_binding
tests/unit/test_backend_runtime_env_validator.py::test_manual_firestore_readiness_contract_allows_one_staged_workflow_control_checkout
tests/unit/test_backend_runtime_env_validator.py::test_automatic_firestore_readiness_contract_requires_current_main_then_admitted_sha
tests/unit/test_backend_runtime_env_validator.py::test_automatic_firestore_readiness_contract_requires_readiness_admitted_sha_for_deploy
tests/unit/test_backend_runtime_env_validator.py::test_firestore_readiness_contract_requires_isolated_job_dependency[gcp_backend.yml]
tests/unit/test_backend_runtime_env_validator.py::test_firestore_readiness_contract_requires_isolated_job_dependency[gcp_backend_auto_dev.yml]
tests/unit/test_backend_runtime_env_validator.py::test_firestore_readiness_contract_requires_validation_before_artifact_upload[gcp_backend.yml]
tests/unit/test_backend_runtime_env_validator.py::test_firestore_readiness_contract_requires_validation_before_artifact_upload[gcp_backend_auto_dev.yml]
tests/unit/test_backend_runtime_env_validator.py::test_firestore_readiness_contract_rejects_backend_deployment_credentials[gcp_backend.yml]
tests/unit/test_backend_runtime_env_validator.py::test_firestore_readiness_contract_rejects_backend_deployment_credentials[gcp_backend_auto_dev.yml]
tests/unit/test_backend_runtime_env_validator.py::test_workflow_validation_uses_immutable_workflow_root_with_admitted_runtime_manifest
tests/unit/test_runtime_image_contracts.py::test_registered_runtime_image_sources_are_closed
tests/unit/test_runtime_image_contracts.py::test_pusher_contract_rejects_omitted_shared_package
tests/unit/test_runtime_image_contracts.py::test_pusher_dependency_probe_includes_jsonschema
tests/unit/test_runtime_image_contracts.py::test_modal_contract_rejects_omitted_shared_package
# Walk every registered import smoke's first-party closure through the AST; hermetic but CPU-bound.
tests/unit/test_runtime_image_contracts.py::test_registered_import_smokes_declare_their_import_time_environment
tests/unit/test_runtime_image_contracts.py::test_import_smoke_reaching_encryption_without_its_secret_is_rejected
tests/unit/test_app_client_schema_inventory.py::test_inventory_ignores_static_base_url_field_routes
tests/unit/test_app_client_schema_inventory.py::test_inventory_normalizes_interpolated_dart_route_segments
tests/unit/test_app_client_schema_inventory.py::test_inventory_openapi_route_response_decode_migration_complete
tests/unit/test_app_client_schema_inventory.py::test_inventory_scopes_record_return_type_route_functions
tests/unit/test_app_client_schema_inventory.py::test_inventory_separates_generated_backed_adapters_from_raw_manual_dtos
tests/unit/test_async_http_infrastructure.py::TestWebhookClientConfig::test_webhook_client_read_timeout_is_30s
tests/unit/test_audio_merge_tasks.py::TestEnqueueAudioMergeJob::test_task_named_by_conversation_and_file
tests/unit/test_backend_runtime_env_validator.py::test_cloud_run_state_rejects_old_secret_versions
tests/unit/test_backend_runtime_env_validator.py::test_cloud_run_state_reports_missing_gateway_url
tests/unit/test_backend_runtime_env_validator.py::test_cloud_run_workflow_validation_uses_custom_manifest_for_runtime_env_outputs
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_passes_for_repo_manifest
# Each parameter renders and validates the complete runtime manifest after injecting the three
# forbidden daily-sweep bindings; this is the same hermetic manifest path as the entries above.
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_rejects_daily_sweep_and_posthog_bindings[dev]
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_rejects_daily_sweep_and_posthog_bindings[prod]
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_rejects_empty_surface_allowlist
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_rejects_mismatched_surface_allowlist
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_rejects_missing_job
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_rejects_notifications_job_maintenance_config
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_rejects_read_mode_without_job_cron
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_rejects_request_path_cron
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_rejects_wrong_dev_capacity_value
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_requires_gateway_luna_bindings_when_enabled
tests/unit/test_backend_runtime_env_validator.py::test_memory_maintenance_job_contract_rejects_pinned_gateway_endpoint_when_enabled
tests/unit/test_backend_runtime_env_validator.py::test_parakeet_selected_without_endpoint_is_rejected_for_all_cloud_run_validation_modes
tests/unit/test_backend_runtime_env_validator.py::test_repo_prod_manifest_rejects_noncanonical_model_order_for_every_surface
tests/unit/test_backend_runtime_env_validator.py::test_repo_cloud_run_workflows_match_manifest
tests/unit/test_backend_runtime_env_validator.py::test_repo_gke_values_match_manifest
tests/unit/test_backend_runtime_env_validator.py::test_repo_prod_cloud_run_workflows_match_manifest
tests/unit/test_backend_runtime_env_validator.py::test_repo_prod_gke_values_match_manifest
tests/unit/test_backend_runtime_env_validator.py::test_repo_rendered_cloud_run_matches_manifest
tests/unit/test_verify_llm_gateway_serving.py::test_gateway_promotion_intent_tracks_runtime_and_helm_listener_surfaces
tests/unit/test_byok_security_enrollment.py::TestBYOKActivationValidation::test_valid_activation_persists_fingerprints
tests/unit/test_verify_pusher_config_references.py::test_dev_pusher_contract_requires_typesense_host_secret_clear
tests/unit/test_verify_pusher_config_references.py::test_rendered_dev_pusher_direct_bindings_match_source_contract
tests/unit/test_byok_security_credentials.py::TestCacheKeyHashing::test_hash_key_returns_sha256
tests/unit/test_byok_security_quota.py::TestChatQuotaBYOKBypass::test_enforce_chat_quota_bypasses_for_byok_with_openai_key
tests/unit/test_byok_security_context.py::TestContextVarSafety::test_default_is_none
tests/unit/test_byok_llm_logging.py::test_classify_byok_llm_error_authentication
tests/unit/test_canonical_short_term_maintenance_cron.py::test_cohort_runner_noops_when_cron_disabled
tests/unit/test_canonical_short_term_maintenance_cron.py::test_cohort_runner_uses_real_maintenance_with_fake_firestore
tests/unit/test_canonical_short_term_maintenance_cron.py::test_daily_cadence_after_first_promotion_run
tests/unit/test_canonical_short_term_maintenance_cron.py::test_first_cron_tick_does_not_mass_promote_below_batch_threshold
tests/unit/test_canonical_short_term_maintenance_cron.py::test_first_cron_tick_promotes_at_batch_threshold
tests/unit/test_chat_generate_reply_stateless.py::test_generate_reply_does_not_write_any_turn_to_chat_history
tests/unit/test_chat_quota_counting_router.py::test_v2_messages_quota_exceeded_reply_does_not_record_quota_question
tests/unit/test_chat_quota_counting_router.py::test_v2_messages_records_failure_when_the_production_stream_errors
tests/unit/test_chat_quota_counting_router.py::test_v2_messages_records_quota_question_after_human_message_persisted
tests/unit/test_chat_quota_counting_router.py::test_v2_voice_messages_records_quota_question_from_visible_message_chunk
tests/unit/test_chat_quota_counting_router.py::test_v2_voice_messages_without_visible_message_does_not_record_quota_question
tests/unit/test_chat_stream_error_fallback.py::test_build_stream_error_reply_persists_ai_message
tests/unit/test_chat_stream_error_fallback.py::test_emit_stream_error_fallback_reports_exhausted_when_persist_fails
tests/unit/test_chat_stream_error_fallback.py::test_v2_messages_does_not_double_emit_canned_after_typed_stream_error
tests/unit/test_chat_stream_error_fallback.py::test_v2_messages_emits_canned_done_after_error_without_staged_answer
tests/unit/test_chat_stream_error_fallback.py::test_v2_messages_emits_fallback_done_frame_on_pipeline_error
tests/unit/test_chat_stream_error_fallback.py::test_v2_messages_keeps_persisted_id_when_app_usage_recording_fails
tests/unit/test_chat_stream_error_fallback.py::test_v2_messages_normal_answer_still_emits_single_done_frame
tests/unit/test_chat_stream_error_fallback.py::test_voice_stream_emits_fallback_done_frame_on_pipeline_error
tests/unit/test_chat_stream_error_fallback.py::test_voice_stream_emits_fallback_when_pipeline_yields_no_answer
tests/unit/test_client_device_provenance.py::test_record_client_device_upserts_and_throttles
tests/unit/test_conversation_search_date_validation.py::test_bad_start_date_returns_400_not_500
tests/unit/test_conversation_model_split.py::TestPhase4RuntimeBehavior::test_trends_extractor_signature_callable
tests/unit/test_daily_notification_timezone_selection.py::test_sub_hour_offset_timezones_are_included_at_target_hour
tests/unit/test_desktop_message_quota_router.py::test_desktop_ai_message_does_not_record_quota
# File-isolation import of routers.chat + multipart/STT stubs amortizes into
# whichever test in this file runs first; CPU sits on the 0.30s fail line.
tests/unit/test_desktop_transcribe.py
tests/unit/test_desktop_transcribe.py::TestDurationBudgetEnforcement::test_multipart_budget_exhausted_429
tests/unit/test_desktop_transcribe.py::TestDurationBudgetEnforcement::test_octet_stream_budget_exhausted_429
tests/unit/test_desktop_transcribe.py::TestMultipartBudgetAggregation::test_multipart_multi_file_budget_sums_durations
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_accepts_boundary_sample_rate_48000
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_accepts_boundary_sample_rate_8000
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_connects_and_receives_segments
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_dg_connection_failure_closes_1011
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_finalize_flushes_and_finalizes
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_rejects_channels_above_2
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_rejects_invalid_channels
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_rejects_sample_rate_above_48000
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_rejects_unsupported_codec
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_rejects_zero_sample_rate
tests/unit/test_desktop_transcribe.py::TestTranscribeStreamWebSocket::test_ws_stereo_flush_accounts_for_channels
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeBoundary::test_octet_stream_accepts_boundary_sample_rate_48000
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeBoundary::test_octet_stream_accepts_boundary_sample_rate_8000
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeBoundary::test_octet_stream_channels_above_2_rejected
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeBoundary::test_octet_stream_sample_rate_above_48000_rejected
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeEndpoint::test_octet_stream_bad_channels_422
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeEndpoint::test_octet_stream_bad_sample_rate_422
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeEndpoint::test_octet_stream_channels_zero_422
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeEndpoint::test_octet_stream_empty_body_400
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeEndpoint::test_octet_stream_sample_rate_zero_400
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeEndpoint::test_multipart_browser_container_preserves_extension_for_prerecorded_stt[mp4-audio/mp4]
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeEndpoint::test_openapi_declares_typed_transcription_failures
tests/unit/test_desktop_transcribe.py::TestVoiceMessageTranscribeEndpoint::test_octet_stream_returns_transcript
tests/unit/test_desktop_transcribe.py::TestVoiceMessagesBudgetHappyPath::test_voice_messages_budget_consumed_on_success
tests/unit/test_dev_api_folder_filters.py::TestDevGetFolders::test_dev_get_folders_returns_folder_list
tests/unit/test_dev_api_lock_bypass.py::TestDevApiMemoryLockEnforcement::test_delete_memory_allows_unlocked
tests/unit/test_developer_memory_adapter.py::test_developer_create_route_checks_split_brain_guard_before_legacy_write
tests/unit/test_developer_memory_adapter.py::test_developer_delete_route_checks_split_brain_guard_before_reads_and_legacy_delete
tests/unit/test_dg_usage_batch.py::TestDgUsageBatchingBehavior::test_batched_60s_single_redis_write
tests/unit/test_dg_usage_batch.py::TestDgUsageBatchingBehavior::test_large_accumulation_no_overflow
tests/unit/test_dg_usage_batch.py::TestRedisOpsReduction::test_reduction_factor
tests/unit/test_diarizer_embedding_decoder_bypass.py::test_embedding_endpoint_passes_preloaded_audio_to_inference
tests/unit/test_experience_summary_date_grounding.py::test_summarize_experience_falls_back_to_utc_on_invalid_timezone
tests/unit/test_fair_use_classifier.py::TestPrepareConversationSummaries::test_extracts_metadata_correctly
tests/unit/test_google_calendar_auth_recovery.py::test_refresh_google_token_marks_missing_refresh_token_reauth_required
tests/unit/test_high_priority_usage_tracking.py::TestExternalIntegrationsTracking::test_get_conversation_summary_sets_context
tests/unit/test_high_priority_usage_tracking.py::TestKnowledgeGraphTracking::test_extract_knowledge_sets_context
tests/unit/test_llm_gateway_client_config.py::test_perplexity_tool_uses_legacy_provider_when_gateway_feature_mode_off
tests/unit/test_llm_gateway_client_config.py::test_gateway_langchain_client_uses_internal_gateway_base_url_and_auth
tests/unit/test_llm_gateway_dependencies.py::test_provider_registry_is_cached_and_closed
tests/unit/test_llm_gateway_executor.py::test_byok_missing_key_and_unsupported_provider_fail_without_fallback_or_lkg
tests/unit/test_llm_gateway_executor.py::test_executor_does_not_retry_terminal_provider_failures[invalid_config-GatewayInvalidRouteConfigError]
tests/unit/test_llm_gateway_executor.py::test_executor_does_not_retry_terminal_provider_failures[byok_rate_limit-GatewayCredentialFailureError]
tests/unit/test_llm_gateway_executor.py::test_raw_byok_key_is_not_in_provider_error_repr_or_dump
tests/unit/test_llm_gateway_openai_compatible.py::test_chat_completions_invalid_json_records_pre_route_rejection
tests/unit/test_llm_gateway_openai_compatible.py::test_chat_completions_rejects_unknown_auto_lane
tests/unit/test_llm_gateway_openai_provider.py::test_openai_compatible_provider_closes_owned_http_client
tests/unit/test_lock_bypass_fixes.py::TestConversationToolFiltering::test_get_conversations_tool_filters_locked
tests/unit/test_lock_bypass_fixes.py::TestMcpSseMemoryLockEnforcement::test_mcp_sse_edit_memory_allows_unlocked
tests/unit/test_lock_bypass_fixes.py::TestMcpSseMemoryLockEnforcement::test_mcp_sse_edit_memory_rejects_locked
tests/unit/test_lock_bypass_fixes.py::TestSyncAudioLockEnforcement::test_precache_rejects_locked
tests/unit/test_lock_bypass_fixes.py::TestUsersLockEnforcement::test_followup_question_rejects_locked
tests/unit/test_memories_validation.py::TestExtractedMemoryValidation::test_legacy_llm_category_maps_to_current_category
tests/unit/test_memory_mutation_contract.py::test_edit_memory_accepts_canonical_json_body
tests/unit/test_product_memory_read_service.py::test_fetch_default_product_memory_search_reads_authoritative_items_and_filters_default_visibility
tests/unit/test_modulate_stt.py::TestModulatePrerecorded::test_basic_transcription
tests/unit/test_modulate_stt.py::TestUtteranceResults::test_partial_superseded_by_utterance
tests/unit/test_optional_audio_codecs.py::test_sync_import_defers_missing_native_opus
tests/unit/test_parakeet_batch_routing.py::TestSetGpuWorker::test_set_and_use
tests/unit/test_parakeet_builtin_embedding.py::TestGetEmbedding::test_falls_back_to_http_when_builtin_unavailable
tests/unit/test_parakeet_endpoints.py::TestDurationGuardHTTP413::test_v1_guard_disabled_when_zero
tests/unit/test_parakeet_nim.py::TestNimTranscribe::test_nim_calls_correct_endpoint
tests/unit/test_parakeet_nim.py::TestNimTranscribe::test_nim_text_only_fallback
tests/unit/test_parakeet_prerecorded.py::TestProviderClass::test_provider_transcribe_url_delegates
tests/unit/test_parakeet_stream_session.py::TestStreamSessionFeed::test_silence_produces_no_segments
tests/unit/test_parakeet_stream_session.py::TestStreamSessionRNNTStreaming::test_streaming_utterance_does_not_call_batch_transcribe_when_text_not_ready
tests/unit/test_payment_promotion_codes.py::test_checkout_invalid_promo_returns_400
tests/unit/test_payment_promotion_codes.py::test_release_attached_schedules_releases_only_matching_active
tests/unit/test_payment_promotion_codes.py::test_release_attached_schedules_noop_without_customer
tests/unit/test_people_conversations_500s.py::TestGetPeopleDocIdInjection::test_get_people_injects_doc_id
tests/unit/test_product_capability_synthetics.py::test_output_uses_required_status_vocabulary_and_json_formatting
tests/unit/test_prompt_cache_integration.py::test_core_tools_has_26_tools
tests/unit/test_prompt_cache_integration.py::test_llm_agent_model_kwargs_via_real_instantiation
tests/unit/test_pusher_docker_context.py::test_pusher_image_layout_includes_backend_config_imports
tests/unit/test_pusher_ghost_connections.py::TestTranscribeSupervisor::test_transcribe_supervisor_session_in_try_finally
tests/unit/test_rate_limit_json_failopen.py::test_partial_cache_entry_fails_open
tests/unit/test_rate_limiting.py::TestRouterWiring::test_developer_rate_limit_failures_log_without_request
tests/unit/test_short_audio_embedding.py::TestGetWavDuration::test_different_sample_rates
tests/unit/test_speaker_id_pipeline.py::TestAudioRingBuffer::test_write_overwrites_oldest_when_capacity_exceeded
# First test to build a live SpeakerMatcher amortizes importing routers.listen.speakers
# (FastAPI router graph) into its call; later tests in this file reuse the cached import.
tests/unit/test_speaker_match.py::test_short_clips_are_pooled_before_a_live_decision
tests/unit/test_speaker_sample.py::test_verify_and_transcribe_sample_transcription_failure
tests/unit/test_speaker_sample_migration.py::test_get_migration_lock_reuses_same_key
tests/unit/test_streaming_deepgram_backoff.py::test_gated_socket_death_reason_delegates_to_safe_socket
tests/unit/test_sync_cloud_tasks.py::TestVerifyCloudTasksOidc::test_env_unset_fails_closed
tests/unit/test_sync_fair_use_gate.py::TestCreateConversationLockPropagation::test_create_conversation_default_unlocked
tests/unit/test_sync_fair_use_gate.py::TestHardRestrictionRetryAfter::test_returns_none_when_restrict_until_is_missing_or_expired
tests/unit/test_sync_fair_use_gate.py::TestSpeechDurationComputation::test_duration_from_raw_vad_segments
tests/unit/test_sync_silent_failure.py::TestProcessSegmentReal::test_success_adds_to_new_memories
tests/unit/test_sync_transcription_prefs.py::TestProcessSegmentPreferences::test_private_cloud_sync_flag_passed_to_new_conversation
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_decode_generic_exception_marks_failed
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_decode_http_exception_marks_failed
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_dg_budget_exhausted_marks_failed
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_dg_budget_not_exhausted_continues
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_empty_decode_completes_zero_segments
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_partial_segment_failure_completes
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_target_conversation_id_forwarded
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_vad_errors_mark_failed
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_zero_segments_after_vad_completes
tests/unit/test_sync_v2.py::TestV2EndpointExecution::test_get_poll_404_expired_job
tests/unit/test_sync_v2.py::TestV2EndpointExecution::test_get_poll_403_wrong_owner
tests/unit/test_sync_v2.py::TestV2EndpointExecution::test_get_poll_completed_includes_result
# Full-file isolation amortizes sync router/pipeline imports and Firestore fakes into these calls.
tests/unit/test_sync_v2.py::TestV2EndpointExecution::test_get_poll_failed_includes_error
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_fresh_soft_caps_use_the_current_subscription_plan[unlimited]
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_retry_hydration_excludes_durably_fenced_conversation_from_audio_finalization
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_inline_renew_errors_cancel_before_capacity_admission_without_releasing_retry_material
tests/unit/test_sync_v2.py::TestAsyncCoordinatorBehavioral::test_content_completion_failure_cannot_publish_completed_job
tests/unit/test_tools_rest_memory_runtime_adapter.py::test_tools_rest_get_memories_text_requests_legacy_safe_memory_decision
tests/unit/test_tools_router.py::TestRouterEnvelope::test_ok_normal
tests/unit/test_transcript_segment.py::test_forward_merge_on_short_incomplete_last_sentence
tests/unit/test_translation_optimization.py::test_confident_language_detection_handles_short_fillers_and_real_text
tests/unit/test_upstream_boundary.py::TestExtractionSeamFanOut::test_fan_out_invokes_memory_action_item_and_goal_paths_separately
tests/unit/test_user_speaker_embedding.py::TestSetUserSpeakerEmbedding::test_stores_embedding_on_user_document
tests/unit/test_user_speaker_embedding.py::TestSpeechProfileEmbeddingExtraction::test_extraction_called_after_upload
tests/unit/test_user_speaker_embedding.py::TestSpeechProfileEmbeddingExtraction::test_extraction_failure_does_not_block_upload
tests/unit/test_v3_real_router_default_off_f4.py::test_production_default_dependency_is_disabled_and_preserves_legacy_get_behavior
tests/unit/test_v3_real_router_default_off_f4.py::test_router_logs_are_sanitized_and_mutating_route_dependencies_unchanged
tests/unit/test_workstream_router_contract.py::test_openapi_exposes_intent_and_thread_resources_without_manual_workstream_create
tests/unit/test_vad_gate.py::TestVADStreamingGate::test_silence_produces_empty_output
tests/unit/test_vertex_ai_system_role.py::test_apps_router_uses_langchain_messages_not_raw_dicts
tests/unit/test_vertex_ai_system_role.py::test_generate_app_from_prompt_uses_langchain_messages
tests/unit/test_verify_backend_release_vector.py::test_legacy_binding_migration_is_idempotent_after_removal
tests/unit/test_verify_backend_release_vector.py::test_dev_deploy_migrates_only_exact_legacy_google_client_id_secrets_without_traffic
tests/unit/test_verify_backend_release_vector.py::test_legacy_binding_migration_strips_prod_legacy_bindings
tests/unit/test_voice_duration_limiter.py::TestPCMDurationComputation::test_16khz_mono_1s
tests/unit/test_workflow_contracts.py::test_workflow_contracts_static_check_accepts_current_allowlist
tests/unit/test_workflow_contracts.py::test_workflow_contract_sources_select_adjacent_tests
tests/unit/test_ws_b_short_term_lifecycle.py::test_promoted_item_readable_and_idempotent_on_second_run
tests/unit/test_ws_b_short_term_lifecycle.py::test_promotion_fires_on_batch_threshold_via_apply
tests/unit/test_ws_b_short_term_lifecycle.py::test_required_promotion_merges_multiple_sources_in_same_run
tests/unit/test_ws_c_backfill.py::test_archive_hidden_long_term_visible_in_canonical_read
tests/unit/test_ws_l_surface_routing.py::TestMcpSseLegacySearchParity::test_legacy_sse_search_uses_vector_order_and_full_firestore_docs
tests/unit/test_sync_cloud_tasks.py::test_sync_post_enqueue_cleanup_does_not_unstage
# File-isolation amortizes router/STT/VAD imports across different node IDs per run.
tests/unit/test_sync_cloud_tasks.py
tests/unit/test_fallback_observability.py
tests/unit/test_task_recommendations.py::test_no_production_branch_consumes_debug_summary_or_reason_codes
tests/unit/test_utils_import_boundary.py::test_utils_modules_do_not_import_routers
tests/unit/test_listen_session_bootstrap.py::test_project_listen_connect_decisions_pins_vocabulary_and_translation
tests/unit/test_memory_replace_policy.py::test_legacy_reextract_failure_preserves_existing_memories[extractor_side_effect0]
tests/unit/test_memory_replace_policy.py::test_legacy_reextract_failure_preserves_existing_memories[extractor_side_effect1]
tests/unit/test_inv_mem_1_guard.py::TestInvMemSourceRatchet::test_no_non_canonical_product_tier_literals_in_memory_paths
tests/unit/test_inv_mem_1_guard.py::TestInvMemSourceRatchet::test_default_read_paths_do_not_include_archive_without_explicit_mode
tests/unit/test_task_intelligence_contract_freeze.py
tests/unit/test_lock_bypass_fixes.py
tests/unit/test_byok_security_enrollment.py::TestCacheRouting::test_cached_openai_chat_no_raw_key_in_cache
# Behavior test drives the real process_segment target-attach fallback (#10033/#10119 review ask);
# it is the sole heavy test in its file, so it amortizes the sync pipeline import into its call phase.
tests/unit/test_sync_merge_target_selection.py::test_sync_target_attach_fallback_to_closest
tests/unit/test_bounded_firestore_list_reads.py::test_legacy_get_memories_no_first_page_5000_force
# File-isolation amortizes llm_gateway app/provider imports into the first oracle call.
tests/unit/test_replay_llm_gateway_fake_upstream_oracle.py::test_loopback_oracle_bypasses_ambient_proxy_configuration
# File-isolation amortizes the merge_conversations import (models/database graph) into the first call.
tests/unit/test_merge_audio_chunk_copy_failure.py::test_merge_propagates_custom_stt_marker_from_any_source
tests/unit/test_agent_proxy_async_boundaries.py::test_persistent_lease_heartbeat_failure_fails_closed_before_ttl_expires
# First test amortizes the file-isolated utils.llm.clients import into its call phase.
tests/unit/test_prompt_cache_wire_payload.py::test_get_llm_forwards_explicit_cache_options_and_production_cache_key
# Suppressed-prefix expansion round walks 500 historical stubs with status batches;
# ~0.5s CPU before and after the list-read budget threading (structural, unchanged test).
tests/unit/test_memory_offset_read_cost.py::test_status_lookups_are_not_repeated_across_expansion_rounds
# 20k-row large-offset fixture intentionally crosses the legacy 10k boundary to charge
# server-side offset() work; the in-memory doc list construction is hermetic but structural.
tests/unit/test_list_read_budget_contract.py::test_conversations_large_offsets_are_served_and_charged[20000-19000-1000]
# First test in the file-isolated memories router suite amortizes the module import.
tests/unit/test_memories_archive_and_read_contracts.py::test_get_memories_forwards_include_archive
# Full listen-runtime bootstrap (STT selection, fair-use, onboarding admission)
# sits exactly at the 0.30s CPU budget under a saturated pre-push fanout.
tests/unit/test_listen_runtime_regressions.py::test_bootstrap_forces_single_language_before_selecting_stt_for_onboarding
tests/unit/test_llm_gateway_validator.py::test_accepts_matching_output_limit_aliases
tests/unit/test_desktop_proxy.py::test_proxy_reports_upstream_unavailable_as_retryable
tests/unit/test_chat_file_upload_unsupported.py::test_heic_photo_is_rejected_as_bad_request[/v2/files]
tests/unit/test_chat_file_upload_unsupported.py::test_heic_photo_is_rejected_as_bad_request[/v1/files]
tests/unit/test_llm_gateway_coverage_guardrails.py::test_every_model_config_feature_has_inventory_and_gateway_lane
tests/unit/test_llm_gateway_coverage_guardrails.py::test_resolver_feature_literals_are_configured_and_dynamic_sites_are_reviewed
tests/unit/test_llm_gateway_embeddings_route.py::test_embeddings_success_returns_openai_shape_and_records_accounting
tests/unit/test_llm_gateway_vertex_provider.py::test_gateway_registry_uses_native_vertex_for_gemini
# File-level: red-proof AST walks of production sources exceed 0.30s CPU per
# test under saturated pre-push once this file is in the changed set.
tests/unit/test_client_processing_trust_boundary.py
# First route test in its file: amortizes the routers.users FastAPI/pydantic import into its call
# phase. Structural (the file-isolation runner model), not a per-test regression.
tests/unit/test_daily_summary_generation.py::test_create_route_happy_path