diff --git a/app/api/routes/scout_setup.py b/app/api/routes/scout_setup.py index 7ff4e74..36f8717 100644 --- a/app/api/routes/scout_setup.py +++ b/app/api/routes/scout_setup.py @@ -1,4 +1,4 @@ -"""Chatbot Journey — WS-based guided conversation to build an AgentConfig. +"""Chatbot Journey — WS-based guided conversation to build an ScoutConfig. The journey is driven entirely through WebSocket frames (no REST endpoints). The device WS handler dispatches ``journey_start`` and ``journey_message`` @@ -13,7 +13,7 @@ Journey flow: 3. FE sends ``journey_message`` frames for each user reply. 4. Server appends the user message, calls the LLM (which may read files via tools), and sends back a ``journey_reply``. - 5. After 3-5 turns the LLM wraps up by emitting an ``AgentConfig`` JSON + 5. After 3-5 turns the LLM wraps up by emitting an ``ScoutConfig`` JSON block delimited by ``AGENT_CONFIG_START`` / ``AGENT_CONFIG_END``. 6. Server parses and validates the JSON with Pydantic, sends ``journey_reply`` with ``done=True`` and the serialised config. @@ -34,7 +34,7 @@ from langchain_core.messages import AIMessage, HumanMessage, SystemMessage, Tool from app.agents.filesystem_agent import make_directory_tools from app.core.langfuse_client import compile_prompt, extract_usage, get_langfuse, get_prompt_or_fallback, langfuse_context from app.core.llm import get_agent_llm, model_for_agent -from app.schemas import AgentConfig +from app.schemas import ScoutConfig logger = logging.getLogger(__name__) @@ -42,7 +42,7 @@ logger = logging.getLogger(__name__) _SESSION_TTL_SECONDS: int = 1800 # 30 minutes -# Sentinel strings used to delimit the LLM-produced AgentConfig JSON. +# Sentinel strings used to delimit the LLM-produced ScoutConfig JSON. _CONFIG_START = "AGENT_CONFIG_START" _CONFIG_END = "AGENT_CONFIG_END" @@ -92,7 +92,7 @@ def get_journey_session(session_id: str, user_id: str) -> JourneySession | None: _JOURNEY_SYSTEM_PROMPT = """\ You are a friendly assistant helping a freelancer configure a data-extraction agent. Your job is to understand what files the user has in their directory and produce a -structured AgentConfig JSON that the extraction agent will use as its instruction set. +structured ScoutConfig JSON that the extraction agent will use as its instruction set. You have access to file-system tools to explore the user's directory: - list_directory: see folder structure and file names @@ -122,7 +122,7 @@ Cover these topics based on what you discovered: 4. Date extraction (e.g. "by Friday" → dueDate) 5. Exclusion rules (e.g. skip newsletters, skip files with no project match) -### Step 4 — Produce the AgentConfig JSON +### Step 4 — Produce the ScoutConfig JSON Once you are ≥ 90% confident, output the final config between these exact markers (each on its own line): @@ -168,7 +168,7 @@ def _build_system_prompt( ) -> tuple[str, Any]: """Return ``(compiled_system_prompt, langfuse_prompt_obj_or_None)``.""" existing_section = ( - "\nThe user already has the following AgentConfig — refine it based on their answers:\n" + "\nThe user already has the following ScoutConfig — refine it based on their answers:\n" f"```json\n{existing_config}\n```\n" if existing_config else "" @@ -189,11 +189,11 @@ def _build_system_prompt( return compiled, prompt_obj -# ── AgentConfig extraction ──────────────────────────────────────────────── +# ── ScoutConfig extraction ──────────────────────────────────────────────── def _extract_agent_config(text: str) -> str | None: - """Return validated AgentConfig JSON string from between markers, or None. + """Return validated ScoutConfig JSON string from between markers, or None. Parses the JSON with Pydantic to ensure it conforms to the schema before returning. Returns None if markers are absent or JSON is invalid. @@ -206,10 +206,10 @@ def _extract_agent_config(text: str) -> str | None: if not raw: return None try: - parsed = AgentConfig.model_validate_json(raw) + parsed = ScoutConfig.model_validate_json(raw) return parsed.model_dump_json() except Exception as exc: - logger.warning("agent_setup: failed to parse AgentConfig JSON: %s", exc) + logger.warning("agent_setup: failed to parse ScoutConfig JSON: %s", exc) return None @@ -475,7 +475,7 @@ async def handle_journey_message( if turns >= _MAX_TURNS: nudge_content = ( "[System: You have enough information. Please generate the final " - f"AgentConfig JSON now, wrapped in {_CONFIG_START} / {_CONFIG_END} markers.]" + f"ScoutConfig JSON now, wrapped in {_CONFIG_START} / {_CONFIG_END} markers.]" ) session.history.append({"role": "user", "content": nudge_content}) diff --git a/app/api/routes/scouts.py b/app/api/routes/scouts.py index 9d4bbb0..973a0d5 100644 --- a/app/api/routes/scouts.py +++ b/app/api/routes/scouts.py @@ -30,11 +30,11 @@ from app.core.note_summarizer import generate_note_summary from app.db import get_session from app.models import ScoutRunLog, LocalScoutConfig from app.schemas import ( - AgentCatalogItem, - AgentCreationCheckRequest, - AgentCreationCheckResponse, - AgentRunLogResponse, - AgentTriggerRequest, + ScoutCatalogItem, + ScoutCreationCheckRequest, + ScoutCreationCheckResponse, + ScoutRunLogResponse, + ScoutTriggerRequest, UserProfile, ) @@ -70,8 +70,8 @@ def _to_data_types(values: list[str]) -> list[str]: return result -def _to_run_log_response(log: ScoutRunLog) -> AgentRunLogResponse: - return AgentRunLogResponse( +def _to_run_log_response(log: ScoutRunLog) -> ScoutRunLogResponse: + return ScoutRunLogResponse( id=log.id, agent_id=log.scout_id, agent_type=log.scout_type, # type: ignore[arg-type] @@ -124,28 +124,28 @@ async def _enforce_run_frequency( # ── Catalog ─────────────────────────────────────────────────────────── -@router.get("/catalog", response_model=list[AgentCatalogItem]) +@router.get("/catalog", response_model=list[ScoutCatalogItem]) async def get_agent_catalog( current_user: UserProfile = Depends(get_current_user), -) -> list[AgentCatalogItem]: +) -> list[ScoutCatalogItem]: """Return the static list of available agent types and their descriptions.""" return [ - AgentCatalogItem( + ScoutCatalogItem( type="local_directory", name="Local Directory Monitor", description="Watches local directories, extracts data from files using AI", ), - AgentCatalogItem( + ScoutCatalogItem( type="gmail", name="Gmail Connector", description="Scans Gmail inbox, extracts tasks/notes from emails", ), - AgentCatalogItem( + ScoutCatalogItem( type="teams", name="Microsoft Teams Connector", description="Monitors Teams messages, extracts action items", ), - AgentCatalogItem( + ScoutCatalogItem( type="outlook", name="Outlook Connector", description="Scans Outlook inbox, extracts tasks/notes", @@ -153,11 +153,11 @@ async def get_agent_catalog( ] -@router.post("/can-create", response_model=AgentCreationCheckResponse) +@router.post("/can-create", response_model=ScoutCreationCheckResponse) async def can_create_agent( - body: AgentCreationCheckRequest, + body: ScoutCreationCheckRequest, current_user: UserProfile = Depends(get_current_user), -) -> AgentCreationCheckResponse: +) -> ScoutCreationCheckResponse: """Check if the user can create one more agent based on billing tier. Since configuration is client-owned, the Electron app sends its current @@ -165,7 +165,7 @@ async def can_create_agent( """ limit: int = FEATURES.get(current_user.tier, FEATURES["free"])["batch_active"] allowed = limit == -1 or body.active_agents < limit - return AgentCreationCheckResponse( + return ScoutCreationCheckResponse( allowed=allowed, tier=current_user.tier, active_agents=body.active_agents, @@ -173,12 +173,12 @@ async def can_create_agent( ) -@router.post("/trigger", response_model=AgentRunLogResponse, status_code=status.HTTP_202_ACCEPTED) +@router.post("/trigger", response_model=ScoutRunLogResponse, status_code=status.HTTP_202_ACCEPTED) async def trigger_agent_run( - body: AgentTriggerRequest, + body: ScoutTriggerRequest, current_user: UserProfile = Depends(get_current_user), db: AsyncSession = Depends(get_session), -) -> AgentRunLogResponse: +) -> ScoutRunLogResponse: """Trigger a local agent run using client-provided configuration.""" _enforce_agent_limit(current_user.tier, body.active_agents) await _enforce_run_frequency(current_user.tier, current_user.id, db) diff --git a/app/schemas/__init__.py b/app/schemas/__init__.py index ca45c19..9350c97 100644 --- a/app/schemas/__init__.py +++ b/app/schemas/__init__.py @@ -207,10 +207,10 @@ class WsStreamEnd(BaseModel): mutations: list[dict[str, Any]] | None = None -# ── Agent Config V2 ─────────────────────────────────────────────────── +# ── Scout Config V2 ─────────────────────────────────────────────────── -class ContentTypeConfig(BaseModel): +class ScoutContentTypeConfig(BaseModel): """Per-type extraction config produced by the journey chatbot.""" id: str @@ -220,34 +220,34 @@ class ContentTypeConfig(BaseModel): extraction_prompt: str -class AgentConfig(BaseModel): - """Structured agent configuration (replaces freeform prompt_template).""" +class ScoutConfig(BaseModel): + """Structured scout configuration (replaces freeform prompt_template).""" - content_types: list[ContentTypeConfig] = [] + content_types: list[ScoutContentTypeConfig] = [] global_rules: list[str] = [] data_types: list[str] = [] -# ── Agent Catalog ───────────────────────────────────────────────────── +# ── Scout Catalog ───────────────────────────────────────────────────── -class AgentCatalogItem(BaseModel): +class ScoutCatalogItem(BaseModel): type: str name: str description: str -class AgentCreationCheckRequest(BaseModel): +class ScoutCreationCheckRequest(BaseModel): active_agents: int = Field(ge=0, default=0) -class AgentCreationCheckResponse(BaseModel): +class ScoutCreationCheckResponse(BaseModel): allowed: bool tier: BillingTier active_agents: int limit: int -class AgentTriggerRequest(BaseModel): +class ScoutTriggerRequest(BaseModel): directory: str = Field(min_length=1) device_id: str = Field(default="") agent_id: str | None = None # FE stable agent ID (electron-store UUID) @@ -259,9 +259,9 @@ class AgentTriggerRequest(BaseModel): last_run_at: int | None = None # epoch ms from FE — enables incremental scanning -# ── Agent Run Log ───────────────────────────────────────────────────── +# ── Scout Run Log ───────────────────────────────────────────────────── -class AgentRunLogResponse(BaseModel): +class ScoutRunLogResponse(BaseModel): id: str agent_id: str agent_type: Literal["local", "cloud"] diff --git a/tests/test_journey_v2.py b/tests/test_journey_v2.py index 6076de9..bf3af00 100644 --- a/tests/test_journey_v2.py +++ b/tests/test_journey_v2.py @@ -1,6 +1,6 @@ """Tests for Local Agent V2 journey setup (Step 4). -Covers the chatbot journey that produces a structured AgentConfig JSON +Covers the chatbot journey that produces a structured ScoutConfig JSON instead of a freeform prompt_template string. Unit tests (no LLM) @@ -16,7 +16,7 @@ Eval test (real LLM + Langfuse scoring) ---------------------------------------- 4.1 Journey start explores directory → first reply contains a question -Cases 4.2–4.5 (multi-turn conversations producing a full AgentConfig) are +Cases 4.2–4.5 (multi-turn conversations producing a full ScoutConfig) are non-deterministic and tested manually — results tracked in Langfuse. Run: @@ -48,7 +48,7 @@ from app.api.routes.scout_setup import ( ) from app.core.langfuse_client import get_langfuse from app.core.ws_context import clear_client_executor, set_client_executor -from app.schemas import AgentConfig +from app.schemas import ScoutConfig from tests.conftest import TEST_USER_IDS # ── Constants ───────────────────────────────────────────────────────────── @@ -179,7 +179,7 @@ def _evaluate_case(case: dict, reply: dict) -> tuple[float, str]: def test_4_6a_extract_valid_json(): """_extract_agent_config: valid JSON between markers → returns serialised config.""" - config = AgentConfig( + config = ScoutConfig( content_types=[], global_rules=["No project = no entity"], data_types=["tasks"], @@ -187,7 +187,7 @@ def test_4_6a_extract_valid_json(): text = f"Some preamble\n{_CONFIG_START}\n{config.model_dump_json()}\n{_CONFIG_END}\nTrailing" result = _extract_agent_config(text) assert result is not None - parsed = AgentConfig.model_validate_json(result) + parsed = ScoutConfig.model_validate_json(result) assert parsed.global_rules == ["No project = no entity"]