diff --git a/llm_normalizer/data/autorun_generators/history.json b/llm_normalizer/data/autorun_generators/history.json index 6749a5a..83241ae 100644 --- a/llm_normalizer/data/autorun_generators/history.json +++ b/llm_normalizer/data/autorun_generators/history.json @@ -1,4 +1,69 @@ [ + { + "generation_id": "gen-ag06151952-5d8f12", + "created_at": "2026-06-15T19:52:40+00:00", + "mode": "saved_user_sessions", + "title": "AGENT | Phase107 mixed planner brain dogfood", + "count": 9, + "domain": "address_phase107_mixed_planner_brain_dogfood", + "questions": [ + "привет, ты на связи?", + "найди в 1С контрагента СВК", + "сколько получили по нему за 2020 год", + "а теперь сколько заплатили?", + "а какое нетто?", + "а по документам?", + "а по движениям?", + "Как ты оценишь деятельность компании ООО Альтернатива Плюс по данным 1С?", + "а чем капибара отличается от утки?" + ], + "generated_by": "codex_agent", + "saved_case_set_file": "assistant_autogen_saved_user_sessions_20260615195240_gen-ag06151952-5d8f12.json", + "context": { + "llm_provider": null, + "model": null, + "assistant_prompt_version": null, + "decomposition_prompt_version": null, + "prompt_fingerprint": null, + "autogen_personality_id": null, + "autogen_personality_prompt": null, + "source_session_id": null, + "saved_session_file": "assistant_saved_session_20260615195240_gen-ag06151952-5d8f12.json", + "saved_case_set_kind": "agent_semantic_scenario", + "agent_run": true, + "agent_focus": "mixed planner autonomy and counterparty evidence carryover", + "architecture_phase": "assistant-stage1 phase107", + "source_spec_file": "X:\\1C\\NDC_1C\\docs\\orchestration\\active_domain_contract.json", + "scenario_id": null, + "semantic_tags": [ + "alias_grounding", + "broad_business_evaluation", + "context_boundary", + "context_switch_from_counterparty_to_company", + "date_carryover", + "document_evidence", + "document_pivot", + "entity_resolution", + "followup_anchor", + "followup_reuse", + "grounded_summary", + "human_answer", + "incoming_value_flow", + "mcp_discovery_gate_sanity", + "meta_smalltalk", + "movement_evidence", + "net_value_flow", + "off_domain_living_chat", + "payout_switch", + "planner_catalog_alignment", + "stale_replay_forbidden", + "value_flow_pivot" + ], + "validation_status": "accepted_domain_pack_loop", + "validated_run_dir": "artifacts\\domain_runs\\phase107_mixed_planner_brain_dogfood_20260615_p11_loop", + "saved_after_validated_replay": true + } + }, { "generation_id": "gen-ag06151746-7d3a57", "created_at": "2026-06-15T17:46:11+00:00", diff --git a/llm_normalizer/data/autorun_generators/saved_sessions/assistant_saved_session_20260615195240_gen-ag06151952-5d8f12.json b/llm_normalizer/data/autorun_generators/saved_sessions/assistant_saved_session_20260615195240_gen-ag06151952-5d8f12.json new file mode 100644 index 0000000..9964e78 --- /dev/null +++ b/llm_normalizer/data/autorun_generators/saved_sessions/assistant_saved_session_20260615195240_gen-ag06151952-5d8f12.json @@ -0,0 +1,245 @@ +{ + "saved_at": "2026-06-15T19:52:40+00:00", + "generation_id": "gen-ag06151952-5d8f12", + "mode": "saved_user_sessions", + "title": "AGENT | Phase107 mixed planner brain dogfood", + "agent_run": true, + "questions": [ + "привет, ты на связи?", + "найди в 1С контрагента СВК", + "сколько получили по нему за 2020 год", + "а теперь сколько заплатили?", + "а какое нетто?", + "а по документам?", + "а по движениям?", + "Как ты оценишь деятельность компании ООО Альтернатива Плюс по данным 1С?", + "а чем капибара отличается от утки?" + ], + "metadata": { + "assistant_prompt_version": null, + "decomposition_prompt_version": null, + "prompt_fingerprint": null, + "agent_focus": "mixed planner autonomy and counterparty evidence carryover", + "architecture_phase": "assistant-stage1 phase107", + "source_spec_file": "X:\\1C\\NDC_1C\\docs\\orchestration\\active_domain_contract.json", + "scenario_id": null, + "semantic_tags": [ + "alias_grounding", + "broad_business_evaluation", + "context_boundary", + "context_switch_from_counterparty_to_company", + "date_carryover", + "document_evidence", + "document_pivot", + "entity_resolution", + "followup_anchor", + "followup_reuse", + "grounded_summary", + "human_answer", + "incoming_value_flow", + "mcp_discovery_gate_sanity", + "meta_smalltalk", + "movement_evidence", + "net_value_flow", + "off_domain_living_chat", + "payout_switch", + "planner_catalog_alignment", + "stale_replay_forbidden", + "value_flow_pivot" + ], + "validation_status": "accepted_domain_pack_loop", + "validated_run_dir": "artifacts\\domain_runs\\phase107_mixed_planner_brain_dogfood_20260615_p11_loop", + "saved_after_validated_replay": true, + "save_gate": { + "schema_version": "agent_semantic_save_gate_v1", + "validation_status": "accepted_domain_pack_loop", + "validated_run_dir": "artifacts\\domain_runs\\phase107_mixed_planner_brain_dogfood_20260615_p11_loop", + "final_status": "accepted", + "loop_id": "phase107_mixed_planner_brain_dogfood_20260615_p11_loop", + "target_score": 88, + "iterations_ran": 1, + "quality_score": 91, + "repair_target_count": 0, + "repair_target_severity_counts": { + "P0": 0, + "P1": 0, + "P2": 0 + }, + "accepted_gate": true, + "effective_runtime": { + "manifest_path": "artifacts\\domain_runs\\phase107_mixed_planner_brain_dogfood_20260615_p11_loop\\effective_runtime.json", + "runner": "domain_case_loop.run-pack-loop", + "git_sha": "7e70ec1d464548a9f422e35b717834211f8d7683", + "backend_url": "http://127.0.0.1:8787", + "mcp_proxy_url": "http://127.0.0.1:6003", + "llm_provider": "local", + "llm_model": "unsloth/qwen3-30b-a3b-instruct-2507", + "temperature": 0.8, + "max_output_tokens": 900, + "prompt_version": "normalizer_v2_0_2", + "prompt_source": "file", + "prompt_hash": "f36e36a5e491cd24511b380e0c0059cb01aea5e2af738fb9f01671cd291735c1", + "prompt_registry_status": "pass" + }, + "saved_after_validated_replay": true + } + }, + "source_session_id": null, + "session": { + "session_id": null, + "mode": "agent_semantic_run", + "items": [ + { + "message_id": "agent-user-001", + "role": "user", + "text": "привет, ты на связи?", + "created_at": "2026-06-15T19:52:40+00:00", + "reply_type": null, + "trace_id": null, + "debug": null + }, + { + "message_id": "agent-user-002", + "role": "user", + "text": "найди в 1С контрагента СВК", + "created_at": "2026-06-15T19:52:40+00:00", + "reply_type": null, + "trace_id": null, + "debug": null + }, + { + "message_id": "agent-user-003", + "role": "user", + "text": "сколько получили по нему за 2020 год", + "created_at": "2026-06-15T19:52:40+00:00", + "reply_type": null, + "trace_id": null, + "debug": null + }, + { + "message_id": "agent-user-004", + "role": "user", + "text": "а теперь сколько заплатили?", + "created_at": "2026-06-15T19:52:40+00:00", + "reply_type": null, + "trace_id": null, + "debug": null + }, + { + "message_id": "agent-user-005", + "role": "user", + "text": "а какое нетто?", + "created_at": "2026-06-15T19:52:40+00:00", + "reply_type": null, + "trace_id": null, + "debug": null + }, + { + "message_id": "agent-user-006", + "role": "user", + "text": "а по документам?", + "created_at": "2026-06-15T19:52:40+00:00", + "reply_type": null, + "trace_id": null, + "debug": null + }, + { + "message_id": "agent-user-007", + "role": "user", + "text": "а по движениям?", + "created_at": "2026-06-15T19:52:40+00:00", + "reply_type": null, + "trace_id": null, + "debug": null + }, + { + "message_id": "agent-user-008", + "role": "user", + "text": "Как ты оценишь деятельность компании ООО Альтернатива Плюс по данным 1С?", + "created_at": "2026-06-15T19:52:40+00:00", + "reply_type": null, + "trace_id": null, + "debug": null + }, + { + "message_id": "agent-user-009", + "role": "user", + "text": "а чем капибара отличается от утки?", + "created_at": "2026-06-15T19:52:40+00:00", + "reply_type": null, + "trace_id": null, + "debug": null + } + ], + "agent_run": true, + "metadata": { + "assistant_prompt_version": null, + "decomposition_prompt_version": null, + "prompt_fingerprint": null, + "agent_focus": "mixed planner autonomy and counterparty evidence carryover", + "architecture_phase": "assistant-stage1 phase107", + "source_spec_file": "X:\\1C\\NDC_1C\\docs\\orchestration\\active_domain_contract.json", + "scenario_id": null, + "semantic_tags": [ + "alias_grounding", + "broad_business_evaluation", + "context_boundary", + "context_switch_from_counterparty_to_company", + "date_carryover", + "document_evidence", + "document_pivot", + "entity_resolution", + "followup_anchor", + "followup_reuse", + "grounded_summary", + "human_answer", + "incoming_value_flow", + "mcp_discovery_gate_sanity", + "meta_smalltalk", + "movement_evidence", + "net_value_flow", + "off_domain_living_chat", + "payout_switch", + "planner_catalog_alignment", + "stale_replay_forbidden", + "value_flow_pivot" + ], + "validation_status": "accepted_domain_pack_loop", + "validated_run_dir": "artifacts\\domain_runs\\phase107_mixed_planner_brain_dogfood_20260615_p11_loop", + "saved_after_validated_replay": true, + "save_gate": { + "schema_version": "agent_semantic_save_gate_v1", + "validation_status": "accepted_domain_pack_loop", + "validated_run_dir": "artifacts\\domain_runs\\phase107_mixed_planner_brain_dogfood_20260615_p11_loop", + "final_status": "accepted", + "loop_id": "phase107_mixed_planner_brain_dogfood_20260615_p11_loop", + "target_score": 88, + "iterations_ran": 1, + "quality_score": 91, + "repair_target_count": 0, + "repair_target_severity_counts": { + "P0": 0, + "P1": 0, + "P2": 0 + }, + "accepted_gate": true, + "effective_runtime": { + "manifest_path": "artifacts\\domain_runs\\phase107_mixed_planner_brain_dogfood_20260615_p11_loop\\effective_runtime.json", + "runner": "domain_case_loop.run-pack-loop", + "git_sha": "7e70ec1d464548a9f422e35b717834211f8d7683", + "backend_url": "http://127.0.0.1:8787", + "mcp_proxy_url": "http://127.0.0.1:6003", + "llm_provider": "local", + "llm_model": "unsloth/qwen3-30b-a3b-instruct-2507", + "temperature": 0.8, + "max_output_tokens": 900, + "prompt_version": "normalizer_v2_0_2", + "prompt_source": "file", + "prompt_hash": "f36e36a5e491cd24511b380e0c0059cb01aea5e2af738fb9f01671cd291735c1", + "prompt_registry_status": "pass" + }, + "saved_after_validated_replay": true + } + } + } +} diff --git a/llm_normalizer/data/eval_cases/assistant_autogen_saved_user_sessions_20260615195240_gen-ag06151952-5d8f12.json b/llm_normalizer/data/eval_cases/assistant_autogen_saved_user_sessions_20260615195240_gen-ag06151952-5d8f12.json new file mode 100644 index 0000000..1a04433 --- /dev/null +++ b/llm_normalizer/data/eval_cases/assistant_autogen_saved_user_sessions_20260615195240_gen-ag06151952-5d8f12.json @@ -0,0 +1,52 @@ +{ + "suite_id": "assistant_saved_session_gen-ag06151952-5d8f12", + "suite_version": "0.1.0", + "schema_version": "assistant_saved_session_suite_v0_1", + "generated_at": "2026-06-15T19:52:40+00:00", + "generation_id": "gen-ag06151952-5d8f12", + "mode": "saved_user_sessions", + "title": "AGENT | Phase107 mixed planner brain dogfood", + "domain": "address_phase107_mixed_planner_brain_dogfood", + "scenario_count": 1, + "case_ids": [ + "SAVED-001" + ], + "cases": [ + { + "case_id": "SAVED-001", + "scenario_tag": "agent_saved_user_sessions", + "title": "AGENT | Phase107 mixed planner brain dogfood", + "question_type": "followup", + "broadness_level": "medium", + "turns": [ + { + "user_message": "привет, ты на связи?" + }, + { + "user_message": "найди в 1С контрагента СВК" + }, + { + "user_message": "сколько получили по нему за 2020 год" + }, + { + "user_message": "а теперь сколько заплатили?" + }, + { + "user_message": "а какое нетто?" + }, + { + "user_message": "а по документам?" + }, + { + "user_message": "а по движениям?" + }, + { + "user_message": "Как ты оценишь деятельность компании ООО Альтернатива Плюс по данным 1С?" + }, + { + "user_message": "а чем капибара отличается от утки?" + } + ] + } + ] +}