fix: OIDC multi-audience support and Swagger UI fix

- Accept tokens from multiple clients (core-api, tatlock-ui, tatlock)
- Fixed main.py to use oidc_audiences[0] for Swagger UI OAuth

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

.env.example

@@ -39,12 +39,9 @@ HOMEASSISTANT_URL=http://localhost:8123
 HOMEASSISTANT_TOKEN=your-long-lived-access-token
 
 # =============================================================================
-# AI Services
+# Search
 # =============================================================================
 
-# Ollama API
-OLLAMA_BASE_URL=http://localhost:11434
-
 # SearXNG (self-hosted search)
 SEARXNG_URL=http://localhost:8080
 

AGENTS.md

@@ -47,7 +47,7 @@ When changes are ready for deployment:
 5. **CI/CD triggers automatically**:
    - Gitea CI builds Docker image on new tag
    - Watchtower pulls and deploys to production
-   - Verify deployment: `curl http://192.168.86.149:8000/health`
+   - Verify deployment: `curl http://192.168.86.149:8083/health`
 
 ---
 

CHANGELOG.md

@@ -5,6 +5,108 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.10.6] - 2026-01-08
+
+### Fixed
+
+- **OIDC audience mismatch** - Accept tokens from multiple clients
+  - Changed `oidc_audience` (string) to `oidc_audiences` (list)
+  - Now accepts tokens with audience: `core-api`, `tatlock-ui`, or `tatlock`
+  - Fixes environment endpoint returning "default" user instead of authenticated username
+  - Fixed main.py to use `oidc_audiences[0]` for Swagger UI OAuth client
+
+### Changed
+
+- Documentation cleanup in README
+
+## [1.10.4] - 2026-01-07
+
+### Removed
+
+- **Ollama integration removed** - AI inference is no longer handled by this API
+  - Removed `src/models/ollama_client.py` and all Ollama-related configuration
+  - Removed `src/models/embeddings.py` and `src/models/embeddings_ollama.py`
+  - Removed model aliases and AI configuration from settings
+  - Health endpoints no longer check Ollama status
+  - Tests updated to reflect database-only health checks
+
+### Changed
+
+- Health check `/health/full` now only checks database connectivity
+- Diagnostics endpoint simplified (removed Ollama component info)
+
+## [1.10.3] - 2026-01-07
+
+### Fixed
+
+- **OIDC config not applied to domains module** - Both `src.auth.oidc` and `src.domains.auth.oidc` configs are now initialized
+  - Previously only `src.auth.oidc` was configured, leaving domains tools using hardcoded "local" user
+  - Environment endpoint now correctly uses authenticated user from OIDC token
+
+## [1.10.2] - 2026-01-07
+
+### Changed
+
+- **Enhanced environment endpoint logging** - Added detailed user claim logging for debugging
+  - Logs both `preferred_username` and `sub` claims when resolving user
+  - Distinguishes between authenticated and unauthenticated requests
+
+## [1.10.1] - 2026-01-06
+
+### Fixed
+
+- Fix OIDC import path in tools controller (`src.oidc.dependencies` → `src.auth.oidc`)
+  - Environment endpoint was returning `user: "local"` instead of authenticated username
+  - Caused queries to wrong Qdrant collection (`volatile_local` vs `volatile_{username}`)
+
+## [1.10.0] - 2026-01-06
+
+### Added
+
+- **Environment Data API** - Qdrant-backed endpoint for weather, forecast, and sun position data
+  - `GET /tools/environment` - Fetch environment data from user's volatile collection
+  - Weather: current temperature, conditions, humidity, wind speed
+  - Forecast: multi-day outlook with high/low temperatures
+  - Sun times: sunrise, sunset, daylight duration
+  - Air quality: AQI and quality level (when available)
+  - Data sourced from `volatile_{user}` Qdrant collection
+  - Uses `preferred_username` from OIDC, falls back to `default`
+- `qdrant-client` dependency for vector database access
+- `QdrantReadClient` wrapper for read-only collection queries
+- Comprehensive test suite for environment service parsing
+
+## [1.9.4] - 2026-01-04
+
+### Fixed
+
+- Fix SQLAlchemy async lazy loading error for new users in `/auth/sync`
+  - Initialize `user.roles = []` and `user.preferences` to avoid greenlet error
+  - Was causing "MissingGreenlet: greenlet_spawn has not been called" on new user creation
+
+## [1.9.3] - 2026-01-04
+
+### Fixed
+
+- Handle non-UUID `sub` claim in `/auth/sync` - Authentik JWT may return non-UUID subject identifiers
+  - Now derives deterministic UUID from sub string if direct parsing fails
+
+## [1.9.2] - 2026-01-04
+
+### Fixed
+
+- `/auth/users/me` endpoint now supports both NPM forward auth headers AND JWT Bearer tokens
+  - Added `get_current_user_or_forward_auth()` combined auth dependency
+  - Fixes web authentication where NPM passes `X-authentik-*` headers instead of JWT
+  - Mobile/native clients continue to use JWT Bearer tokens as before
+
+## [1.9.1] - 2026-01-03
+
+### Fixed
+
+- CORS configuration now uses explicit origins instead of `"*"`
+  - When `allow_credentials=True`, wildcard origins are rejected by browsers
+  - Added `home.schweitz.net`, `tatlock.schweitz.net`, and localhost origins
+
 ## [1.9.0] - 2026-01-03
 
 ### Added

README.md

@@ -19,8 +19,7 @@ Central API service providing infrastructure management, home automation, and ut
 
 ### Utilities
 - **DNS Lookup**: Query DNS records (A, AAAA, MX, TXT, CNAME, NS, SOA, PTR)
-- **Health Checks**: Comprehensive service health monitoring
-- **AI Metrics Proxy**: Forward metrics requests to Core-AI service
+- **Health Checks**: Service health monitoring with database connectivity status
 
 ## Architecture
 
@@ -35,10 +34,8 @@ src/
 ├── clients/
 │   ├── homeassistant_client.py    # Home Assistant REST client
 │   ├── npm_client.py              # Nginx Proxy Manager client
-│   ├── ollama_client.py           # Ollama LLM client
 │   └── portainer_client.py        # Portainer API client
 ├── controllers/
-│   ├── ai_controller.py           # AI metrics proxy
 │   ├── health_controller.py       # Health endpoints
 │   ├── housekeeping_controller.py # Home automation endpoints
 │   ├── infrastructure_controller.py # Infrastructure management
@@ -88,9 +85,6 @@ src/
 ### Tools (`/tools`)
 - `POST /tools/dns/lookup` - DNS record lookup
 
-### AI (`/ai`)
-- `GET /ai/metrics` - Proxy to Core-AI metrics
-
 ## Development
 
 ### Requirements
@@ -103,10 +97,6 @@ src/
 # Install dependencies
 pip install -r requirements.txt
 
-# Copy credentials template
-cp src/credentials.example.py src/credentials.py
-# Edit src/credentials.py with your values
-
 # Run locally
 uvicorn src.main:app --reload --host 0.0.0.0 --port 8083
 ```
@@ -169,7 +159,6 @@ docker run -p 8083:8083 core-code:latest
 | `NPM_PASSWORD` | NPM admin password | - |
 | `HOMEASSISTANT_URL` | Home Assistant URL | `http://localhost:8123` |
 | `HOMEASSISTANT_TOKEN` | HA long-lived access token | - |
-| `OLLAMA_URL` | Ollama API URL | `http://localhost:11434` |
 | `OIDC_ENABLED` | Enable OIDC auth | `false` |
 | `OIDC_ISSUER` | OIDC issuer URL | - |
 | `OIDC_AUDIENCE` | OIDC audience | - |
@@ -183,9 +172,9 @@ Once deployed, access documentation at:
 
 ## Health Checks
 
-- **Basic**: `GET /health` - Returns status and Ollama connection
-- **Full**: `GET /health/full` - Returns all component statuses (503 if unhealthy)
-- **Diagnostics**: `GET /health/diagnostics` - Detailed service information
+- **Basic**: `GET /health` - Fast liveness check for container orchestration
+- **Full**: `GET /health/full` - Returns database status (503 if unhealthy)
+- **Diagnostics**: `GET /health/diagnostics` - Service info and configuration
 
 ## Security
 

npm-config.conf

@@ -0,0 +1,12 @@
+proxy_buffers 8 16k;
+proxy_buffer_size 32k;
+
+# CORS headers for Flutter web
+add_header Access-Control-Allow-Origin "https://home.schweitz.net" always;
+add_header Access-Control-Allow-Credentials true always;
+add_header Access-Control-Allow-Methods "GET, POST, PUT, DELETE, PATCH, OPTIONS" always;
+add_header Access-Control-Allow-Headers "Content-Type, Authorization" always;
+
+if ($request_method = OPTIONS) {
+    return 204;
+}

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "core-api"
-version = "1.9.0"
+version = "1.10.6"
 description = "Core Code API - Infrastructure management and tools API"
 readme = "README.md"
 requires-python = ">=3.12"

requirements.txt

@@ -32,3 +32,6 @@ cryptography>=44.0.1  # CVE-2024-12797
 sqlalchemy[asyncio]~=2.0.0
 asyncpg>=0.30.0
 alembic~=1.13.0
+
+# Vector Database
+qdrant-client>=1.9.0

src/auth/oidc.py

@@ -23,16 +23,16 @@ class OIDCConfig:
         # These will be set from environment variables in config.py
         self.enabled = False
         self.issuer = ""
-        self.audience = ""
+        self.audiences: list[str] = []
         self.jwks_uri = ""
 
-    def configure(self, enabled: bool, issuer: str, audience: str):
+    def configure(self, enabled: bool, issuer: str, audiences: list[str]):
         """Configure OIDC settings"""
         self.enabled = enabled
         self.issuer = issuer
-        self.audience = audience
+        self.audiences = audiences
         self.jwks_uri = f"{issuer.rstrip('/')}/jwks/"
-        logger.info(f"OIDC configured: enabled={enabled}, issuer={issuer}")
+        logger.info(f"OIDC configured: enabled={enabled}, issuer={issuer}, audiences={audiences}")
 
 
 # Global OIDC config instance
@@ -129,12 +129,12 @@ async def get_current_user(
             logger.warning(f"No matching key found for kid: {kid}")
             raise HTTPException(status_code=401, detail="Invalid token key")
 
-        # Verify and decode token
+        # Verify and decode token (accepts any of the configured audiences)
         payload = jwt.decode(
             token,
             rsa_key,
             algorithms=["RS256"],
-            audience=oidc_config.audience,
+            audience=oidc_config.audiences,
             issuer=oidc_config.issuer,
         )
 

src/config.py

@@ -45,35 +45,6 @@ class Settings(BaseSettings):
     # Logging
     log_level: str = "DEBUG"
 
-    # Ollama Configuration (for AI orchestration)
-    ollama_base_url: str  # Required - set OLLAMA_BASE_URL in .env
-    ollama_timeout: int = 300  # 5 minutes
-
-    # Model Configuration
-    default_model: str = "mistral-nemo-large:latest"
-    agent_model: str = "mistral-nemo-large:latest"  # Must support tool calling with ADK (~4GB VRAM)
-    code_models: str = "mistral-nemo-large:latest"
-    # Previous config (gemma3:12b used ~10GB VRAM)
-    # default_model: str = "gemma3:12b"
-    # agent_model: str = "gemma3:12b"
-
-    # System Prompt Variant (for A/B testing)
-    # Options: v1_verbose, v2_concise, v3_imperative, v4_minimal, v4_gemini_suggestion, v5_adk_optimized, v7_adk_best_practice, v8_holistic
-    system_prompt_variant: str = "v8_holistic"
-
-    # Agent Configuration
-    agent_fallback_enabled: bool = True
-
-    # Model Aliases (OpenAI → Local)
-    alias_gpt35: str = "gemma:7b"
-    alias_gpt4: str = "mistral:7b"
-    alias_gpt4_turbo: str = "mixtral:8x7b"
-    alias_gpt4_code: str = "codestral:latest"
-
-    # Memory Configuration
-    memory_tier1_max_turns: int = 10
-    memory_consolidation_threshold: int = 10
-
     # Qdrant Configuration
     qdrant_host: str = "qdrant"
     qdrant_port: int = 6333
@@ -81,11 +52,6 @@ class Settings(BaseSettings):
     qdrant_collection_documents: str = "core_api_documents"
     qdrant_collection_user_facts: str = "core_api_user_facts"
 
-    # Embeddings (using Ollama - no local models needed)
-    embedding_model: str = "nomic-embed-text"  # Ollama embedding model
-    embedding_dimension: int = 768  # nomic-embed-text dimension
-    embedding_batch_size: int = 32
-
     # Search Configuration
     search_provider: str = "searxng"
     searxng_url: str  # Required - set SEARXNG_URL in .env
@@ -118,7 +84,8 @@ class Settings(BaseSettings):
     # OIDC Authentication (Authentik)
     oidc_enabled: bool = False  # Set to True to require authentication
     oidc_issuer: str = "https://auth.schweitz.net/application/o/core-api/"
-    oidc_audience: str = "core-api"
+    # Accept tokens from multiple clients (core-api, tatlock-ui, tatlock)
+    oidc_audiences: list[str] = ["core-api", "tatlock-ui", "tatlock"]
 
     # Authentik API (for token validation and user management)
     # Must use domain name (not IP) when AUTHENTIK_COOKIE_DOMAIN is set
@@ -126,28 +93,6 @@ class Settings(BaseSettings):
     authentik_username: str = ""  # Admin username for API access (AUTHENTIK_USERNAME env var)
     authentik_password: str = ""  # Admin password for API access (AUTHENTIK_PASSWORD env var)
 
-    @property
-    def model_aliases(self) -> dict:
-        """Computed property for model aliases"""
-        return {
-            "gpt-3.5-turbo": self.alias_gpt35,
-            "gpt-4": self.alias_gpt4,
-            "gpt-4-turbo": self.alias_gpt4_turbo,
-            "gpt-4-code": self.alias_gpt4_code,
-        }
-
-    def get_lightweight_models(self) -> list[str]:
-        """Parse comma-separated lightweight models"""
-        return [m.strip().strip('"').strip("'") for m in self.lightweight_models.split(",") if m.strip()]
-
-    def get_heavy_models(self) -> list[str]:
-        """Parse comma-separated heavy models"""
-        return [m.strip().strip('"').strip("'") for m in self.heavy_models.split(",") if m.strip()]
-
-    def get_code_models(self) -> list[str]:
-        """Parse comma-separated code models"""
-        return [m.strip().strip('"').strip("'") for m in self.code_models.split(",") if m.strip()]
-
     class Config:
         env_file = ".env"
         case_sensitive = False

src/controllers/health_controller.py

@@ -9,11 +9,9 @@ from fastapi.responses import JSONResponse
 from src.controllers.base import BaseController
 from src.config import get_settings
 from src.logging_config import get_logger
-from src.models.ollama_client import get_ollama_client
 from src.db import get_database
 
 
-
 logger = get_logger(__name__)
 
 
@@ -73,63 +71,18 @@ class HealthController(BaseController):
 
         @router.get(
             "/health/full",
-            summary="Fast health check for Docker",
+            summary="Full health check with database",
         )
         async def full_health_check(response: Response):
             """
-            Fast health check for container orchestration (Docker/K8s).
+            Health check including database connectivity.
 
-            Checks component availability WITHOUT running expensive operations.
-            Returns 200 OK if all components are available, otherwise 503.
-
-            For detailed diagnostics, use /health/diagnostics instead.
+            Returns 200 OK if database is available, otherwise 503.
             """
             import time
             start_time = time.time()
 
-            # Check 1: Ollama connection + verify agent model is available
-            ollama_client = get_ollama_client()
-            ollama_healthy = False
-            ollama_error = None
-            model_available = False
-
-            try:
-                # Ping Ollama
-                ollama_healthy = await ollama_client.health_check()
-
-                # Verify the agent model is pulled and check what's currently loaded
-                models_info = {}
-                if ollama_healthy:
-                    try:
-                        models_response = await ollama_client.list_models()
-                        available_models = [m.get('name', '') for m in models_response.get('models', [])]
-                        model_available = settings.agent_model in available_models
-
-                        # Get info about currently loaded models (those with size in memory)
-                        loaded_models = [
-                            m.get('name', '') for m in models_response.get('models', [])
-                            if m.get('size', 0) > 0
-                        ]
-
-                        models_info = {
-                            "configured": settings.agent_model,
-                            "available": model_available,
-                            "total_in_ollama": len(available_models),
-                            "currently_loaded": loaded_models if loaded_models else ["none"]
-                        }
-
-                        if not model_available:
-                            ollama_error = f"Model '{settings.agent_model}' not found in Ollama. Available: {', '.join(available_models[:3])}"
-                            ollama_healthy = False
-                    except Exception as e:
-                        ollama_error = f"Could not list Ollama models: {str(e)}"
-                        ollama_healthy = False
-
-            except Exception as e:
-                ollama_error = str(e)
-                logger.warning(f"Ollama health check failed: {ollama_error}")
-
-            # Check 2: Database connection
+            # Check database connection
             database = get_database()
             db_healthy = False
             db_error = None
@@ -140,27 +93,17 @@ class HealthController(BaseController):
                 db_error = str(e)
                 logger.warning(f"Database health check failed: {db_error}")
 
-            is_healthy = ollama_healthy and db_healthy
-
             elapsed_ms = int((time.time() - start_time) * 1000)
-            status_code = 200 if is_healthy else 503
+            status_code = 200 if db_healthy else 503
             response.status_code = status_code
 
             return {
-                "status": "healthy" if is_healthy else "unhealthy",
+                "status": "healthy" if db_healthy else "unhealthy",
                 "status_code": status_code,
                 "response_time_ms": elapsed_ms,
                 "components": {
-                    "ollama": {
-                        "status": "✅ healthy" if ollama_healthy else "❌ unhealthy",
-                        "models": models_info if models_info else {
-                            "configured": settings.agent_model,
-                            "available": False
-                        },
-                        "error": ollama_error
-                    },
                     "database": {
-                        "status": "✅ healthy" if db_healthy else "❌ unhealthy",
+                        "status": "healthy" if db_healthy else "unhealthy",
                         "error": db_error
                     }
                 }
@@ -170,18 +113,13 @@ class HealthController(BaseController):
             "/health/diagnostics",
             summary="Detailed system diagnostics",
         )
-        async def diagnostics(deep_test: bool = False):
+        async def diagnostics():
             """
-            Comprehensive system diagnostics with detailed component information.
-
-            Query Parameters:
-            - deep_test: Set to true to actually test agent generation (slow, ~5-10s)
-
-            Returns detailed information about all system components.
+            System diagnostics with service information.
             """
             import time
-
             start_time = time.time()
+
             diagnostics = {
                 "timestamp": time.time(),
                 "service": {
@@ -189,30 +127,9 @@ class HealthController(BaseController):
                     "version": settings.app_version,
                     "purpose": "Infrastructure management and tools API"
                 },
-                "components": {}
-            }
-
-            # 1. Ollama Connection
-            ollama_client = get_ollama_client()
-            try:
-                ollama_healthy = await ollama_client.health_check()
-                diagnostics["components"]["ollama"] = {
-                    "status": "✅ connected",
-                    "url": settings.ollama_base_url,
-                    "timeout": settings.ollama_timeout,
-                    "default_model": settings.default_model
+                "configuration": {
+                    "cors_origins": settings.cors_origins[:2] if len(settings.cors_origins) > 2 else settings.cors_origins
                 }
-            except Exception as e:
-                diagnostics["components"]["ollama"] = {
-                    "status": "❌ error",
-                    "error": str(e)
-                }
-
-            # 2. Configuration
-            diagnostics["configuration"] = {
-                "agent_fallback_enabled": settings.agent_fallback_enabled,
-                "memory_tier1_max_turns": settings.memory_tier1_max_turns,
-                "cors_origins": settings.cors_origins[:2] if len(settings.cors_origins) > 2 else settings.cors_origins
             }
 
             elapsed_ms = int((time.time() - start_time) * 1000)

src/controllers/tools_controller.py

@@ -3,14 +3,20 @@ Tools Controller
 
 Provides utility tool endpoints including:
 - DNS lookups
+- Environment data (weather, forecast, sun times, air quality)
 """
-from fastapi import APIRouter, HTTPException, status
+from typing import Dict, Optional
+
+from fastapi import APIRouter, Depends, HTTPException, status
 
 from src.controllers.base import BaseController
 from src.logging_config import get_logger
 from src.dns.schemas import DNSLookupRequest, DNSLookupResponse
 from src.dns.service import DNSService
 from src.dns.exceptions import DNSQueryError
+from src.domains.tools.environment.schemas import EnvironmentResponse
+from src.domains.tools.environment.service import get_environment_service
+from src.auth.oidc import get_optional_user
 
 logger = get_logger(__name__)
 
@@ -26,6 +32,7 @@ class ToolsController(BaseController):
     def __init__(self):
         super().__init__(prefix="/tools", tags=["Tools"])
         self.dns_service = DNSService()
+        self.environment_service = get_environment_service()
 
     def create_router(self) -> APIRouter:
         """Create and configure the router"""
@@ -95,6 +102,60 @@ class ToolsController(BaseController):
                     detail="An unexpected error occurred during DNS lookup"
                 )
 
+        @router.get(
+            "/environment",
+            response_model=EnvironmentResponse,
+            status_code=status.HTTP_200_OK,
+            summary="Get environment data",
+            description="""
+            Fetch current environment data including weather, forecast, sun times,
+            and optionally air quality.
+
+            Data is retrieved from the user's volatile Qdrant collection which is
+            populated by background data collectors.
+
+            **Data Sources:**
+            - Weather: Current temperature, conditions, humidity, wind
+            - Forecast: Multi-day weather outlook
+            - Sun Times: Sunrise, sunset, daylight duration
+            - Air Quality: AQI and pollutant levels (when available)
+
+            **Authentication:**
+            - Uses authenticated user's `preferred_username` if available
+            - Falls back to 'default' for unauthenticated requests
+            """
+        )
+        async def get_environment(
+            user: Optional[Dict] = Depends(get_optional_user),
+        ) -> EnvironmentResponse:
+            """
+            Get current environment data.
+
+            Args:
+                user: Optional authenticated user info
+
+            Returns:
+                Environment data with weather, forecast, sun times, and air quality
+            """
+            try:
+                # Determine user identifier
+                user_id = "default"
+                if user:
+                    logger.debug(f"User claims: {user}")
+                    user_id = user.get("preferred_username") or user.get("sub", "default")
+                    logger.info(f"Fetching environment data for user: {user_id} (preferred_username={user.get('preferred_username')}, sub={user.get('sub')})")
+                else:
+                    logger.info(f"Fetching environment data for user: {user_id} (no auth)")
+                result = await self.environment_service.get_current(user_id)
+                return result
+
+            except Exception as e:
+                logger.error(f"Error fetching environment data: {str(e)}", exc_info=True)
+                raise HTTPException(
+                    status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+                    detail="Failed to fetch environment data"
+                )
+
         return router
 
 

src/domains/auth/controller.py

@@ -20,7 +20,7 @@ from src.domains.auth.schemas import (
     ApiKeysListResponse,
 )
 from src.domains.auth.service import AuthService
-from src.domains.auth.oidc import get_current_user
+from src.domains.auth.oidc import get_current_user, get_current_user_or_forward_auth
 
 logger = get_logger(__name__)
 
@@ -322,22 +322,26 @@ class AuthController(BaseController):
             responses={
                 200: {"description": "User profile with roles and preferences"},
                 401: {"description": "Not authenticated"},
-                404: {"description": "User not found in database"},
             },
         )
         async def get_current_user_profile(
-            user_claims: dict = Depends(get_current_user),
+            user_claims: dict = Depends(get_current_user_or_forward_auth),
             session: AsyncSession = Depends(get_async_session),
         ) -> UserProfileResponse:
             """
             Get the current authenticated user's profile
 
             Returns the user's profile, roles, and preferences.
-            Requires authentication via Bearer token or API key.
+            Supports both:
+            - Bearer token (mobile/native clients)
+            - NPM forward auth headers (web clients via proxy)
+
+            For forward auth users, auto-creates the user in the database
+            if they don't exist yet (first login via web).
             """
             service = AuthService(session)
 
-            # Get authentik_id from claims (JWT 'sub' field)
+            # Get authentik_id from claims (JWT 'sub' field or forward auth 'uid')
             authentik_id_str = user_claims.get("sub")
             if not authentik_id_str or authentik_id_str == "local-user":
                 raise HTTPException(status_code=401, detail="Authentication required")
@@ -348,11 +352,27 @@ class AuthController(BaseController):
                 raise HTTPException(status_code=401, detail="Invalid user identifier")
 
             user = await service.get_user_by_authentik_id(authentik_id)
+
+            # If user not found and using forward auth, auto-create them
             if user is None:
-                raise HTTPException(
-                    status_code=404,
-                    detail="User not found - please sync via /auth/sync first",
-                )
+                auth_method = user_claims.get("auth_method")
+                if auth_method == "forward_auth":
+                    # Auto-sync user from forward auth headers
+                    logger.info(f"Auto-creating user from forward auth: {user_claims.get('email')}")
+                    user, is_new = await service.sync_user_from_claims(
+                        authentik_id=authentik_id,
+                        email=user_claims.get("email", ""),
+                        name=user_claims.get("name", user_claims.get("preferred_username", "")),
+                        groups=user_claims.get("groups", []),
+                    )
+                    await session.commit()
+                    await session.refresh(user, ["preferences", "roles"])
+                else:
+                    # JWT auth but user not in DB - they need to sync first
+                    raise HTTPException(
+                        status_code=404,
+                        detail="User not found - please sync via /auth/sync first",
+                    )
 
             return UserProfileResponse(
                 user=service.user_to_schema(user),

src/domains/auth/oidc.py

@@ -12,7 +12,6 @@ Permission Format: domain.category:action
 Examples:
 - control-room.general:admin - Full access to Control Room
 - media.general:viewer - View-only access to Media area
-- ai.ollama:user - User-level access to Ollama specifically (future)
 
 Action Hierarchy (higher implies lower):
 - admin > editor > user > viewer
@@ -65,16 +64,16 @@ class OIDCConfig:
         # These will be set from environment variables in config.py
         self.enabled = False
         self.issuer = ""
-        self.audience = ""
+        self.audiences: list[str] = []
         self.jwks_uri = ""
 
-    def configure(self, enabled: bool, issuer: str, audience: str):
+    def configure(self, enabled: bool, issuer: str, audiences: list[str]):
         """Configure OIDC settings"""
         self.enabled = enabled
         self.issuer = issuer
-        self.audience = audience
+        self.audiences = audiences
         self.jwks_uri = f"{issuer.rstrip('/')}/jwks/"
-        logger.info(f"OIDC configured: enabled={enabled}, issuer={issuer}")
+        logger.info(f"OIDC configured: enabled={enabled}, issuer={issuer}, audiences={audiences}")
 
 
 # Global OIDC config instance
@@ -178,12 +177,12 @@ async def get_current_user(
             logger.warning(f"No matching key found for kid: {kid}")
             raise HTTPException(status_code=401, detail="Invalid token key")
 
-        # Verify and decode token
+        # Verify and decode token (accepts any of the configured audiences)
         payload = jwt.decode(
             token,
             rsa_key,
             algorithms=["RS256"],
-            audience=oidc_config.audience,
+            audience=oidc_config.audiences,
             issuer=oidc_config.issuer,
         )
 
@@ -396,6 +395,56 @@ async def get_forward_auth_admin(
     return user
 
 
+async def get_current_user_or_forward_auth(
+    request: Request,
+    credentials: Optional[HTTPAuthorizationCredentials] = Security(security)
+) -> Dict:
+    """
+    Combined auth: Try forward auth headers first, then JWT Bearer token.
+
+    Supports both:
+    - Web clients via NPM forward auth (X-authentik-* headers from proxy)
+    - Mobile/native clients via OIDC JWT Bearer tokens
+
+    This is the preferred dependency for /auth/users/me and similar endpoints
+    that need to work with both web (cookie-based via NPM) and mobile (token-based).
+
+    Args:
+        request: FastAPI request object containing headers
+        credentials: HTTP Bearer token from Authorization header
+
+    Returns:
+        User claims dictionary with at minimum: sub, email, name, groups, auth_method
+
+    Raises:
+        HTTPException 401: If neither forward auth headers nor valid JWT provided
+    """
+    # 1. Try forward auth headers first (web via NPM)
+    username = request.headers.get("x-authentik-username")
+    email = request.headers.get("x-authentik-email")
+
+    if username and email:
+        # Forward auth headers present - use them
+        groups = request.headers.get("x-authentik-groups", "")
+        name = request.headers.get("x-authentik-name", username)
+        uid = request.headers.get("x-authentik-uid")
+
+        user_info = {
+            "sub": uid,  # Use authentik UID as subject (for user lookup)
+            "email": email,
+            "preferred_username": username,
+            "name": name,
+            "groups": [g.strip() for g in groups.split(",")] if groups else [],
+            "auth_method": "forward_auth"
+        }
+
+        logger.info(f"Authenticated via forward auth: {email}")
+        return user_info
+
+    # 2. Fall back to JWT Bearer token (mobile/native)
+    return await get_current_user(credentials)
+
+
 # =============================================================================
 # Permission-Based Access Control
 # =============================================================================
@@ -504,7 +553,7 @@ def _extract_permissions_from_groups(groups: List[str]) -> List[str]:
     Examples:
     - tatlock-control-room-general-admin -> control-room.general:admin
     - tatlock-media-viewer -> media.general:viewer (shorthand)
-    - tatlock-ai-ollama-user -> ai.ollama:user
+    - tatlock-tools-dns-user -> tools.dns:user
 
     Args:
         groups: List of Authentik group names

src/domains/auth/service.py

@@ -102,7 +102,13 @@ class AuthService:
         Returns:
             Tuple of (User, is_new_user)
         """
-        authentik_id = uuid.UUID(token_info.sub)
+        # Parse authentik_id - may be UUID or other format
+        try:
+            authentik_id = uuid.UUID(token_info.sub)
+        except ValueError:
+            # If sub is not a valid UUID, derive one deterministically
+            logger.warning(f"sub claim is not a UUID: {token_info.sub}, deriving UUID")
+            authentik_id = uuid.uuid5(uuid.NAMESPACE_OID, token_info.sub)
 
         # Try to find existing user
         stmt = (
@@ -124,11 +130,14 @@ class AuthService:
                 avatar_url=token_info.picture,
                 last_login=datetime.now(timezone.utc),
             )
+            # Initialize relationships to avoid lazy loading issues in async
+            user.roles = []
             self.session.add(user)
             await self.session.flush()  # Get the user ID
 
-            # Create default preferences
+            # Create default preferences and attach to user
             preferences = UserPreferences(user_id=user.id)
+            user.preferences = preferences
             self.session.add(preferences)
 
             logger.info(f"Created new user: {token_info.email}")
@@ -144,6 +153,74 @@ class AuthService:
         await self.session.flush()
         return user, is_new
 
+    async def sync_user_from_claims(
+        self,
+        authentik_id: uuid.UUID,
+        email: str,
+        name: str,
+        groups: list[str],
+        avatar_url: Optional[str] = None,
+    ) -> tuple[User, bool]:
+        """
+        Create or update user from forward auth claims (NPM X-authentik-* headers)
+
+        This is similar to sync_user() but works with raw claims instead of
+        TokenInfoSchema. Used for auto-syncing users on first web login via NPM.
+
+        Args:
+            authentik_id: The Authentik user UUID (from X-authentik-uid)
+            email: User email (from X-authentik-email)
+            name: User display name (from X-authentik-name)
+            groups: List of group names (from X-authentik-groups)
+            avatar_url: Optional avatar URL
+
+        Returns:
+            Tuple of (User, is_new_user)
+        """
+        # Try to find existing user
+        stmt = (
+            select(User)
+            .options(selectinload(User.roles), selectinload(User.preferences))
+            .where(User.authentik_id == authentik_id)
+        )
+        result = await self.session.execute(stmt)
+        user = result.scalar_one_or_none()
+
+        is_new = user is None
+
+        if is_new:
+            # Create new user
+            user = User(
+                authentik_id=authentik_id,
+                email=email,
+                name=name or email,
+                avatar_url=avatar_url,
+                last_login=datetime.now(timezone.utc),
+            )
+            self.session.add(user)
+            await self.session.flush()  # Get the user ID
+
+            # Create default preferences
+            preferences = UserPreferences(user_id=user.id)
+            self.session.add(preferences)
+
+            logger.info(f"Created new user from forward auth: {email}")
+        else:
+            # Update existing user
+            user.email = email
+            user.name = name or email
+            if avatar_url:
+                user.avatar_url = avatar_url
+            user.last_login = datetime.now(timezone.utc)
+
+            logger.info(f"Updated existing user from forward auth: {email}")
+
+        # Sync roles from groups
+        await self.sync_roles(user, groups)
+
+        await self.session.flush()
+        return user, is_new
+
     async def sync_roles(self, user: User, group_names: list[str]) -> list[Role]:
         """
         Synchronize user roles from Authentik groups via group_roles mapping

src/domains/health/controller.py

@@ -70,66 +70,18 @@ class HealthController(BaseController):
 
         @router.get(
             "/health/full",
-            summary="Fast health check for Docker",
+            summary="Full health check with database",
         )
         async def full_health_check(response: Response):
             """
-            Fast health check for container orchestration (Docker/K8s).
+            Health check including database connectivity.
 
-            Checks component availability WITHOUT running expensive operations.
-            Returns 200 OK if all components are available, otherwise 503.
-
-            For detailed diagnostics, use /health/diagnostics instead.
+            Returns 200 OK if database is available, otherwise 503.
             """
             import time
             start_time = time.time()
 
-            # Import here to avoid circular imports
-            from src.models.ollama_client import get_ollama_client
-
-            # Check 1: Ollama connection + verify agent model is available
-            ollama_client = get_ollama_client()
-            ollama_healthy = False
-            ollama_error = None
-            model_available = False
-
-            try:
-                # Ping Ollama
-                ollama_healthy = await ollama_client.health_check()
-
-                # Verify the agent model is pulled and check what's currently loaded
-                models_info = {}
-                if ollama_healthy:
-                    try:
-                        models_response = await ollama_client.list_models()
-                        available_models = [m.get('name', '') for m in models_response.get('models', [])]
-                        model_available = settings.agent_model in available_models
-
-                        # Get info about currently loaded models (those with size in memory)
-                        loaded_models = [
-                            m.get('name', '') for m in models_response.get('models', [])
-                            if m.get('size', 0) > 0
-                        ]
-
-                        models_info = {
-                            "configured": settings.agent_model,
-                            "available": model_available,
-                            "total_in_ollama": len(available_models),
-                            "currently_loaded": loaded_models if loaded_models else ["none"]
-                        }
-
-                        if not model_available:
-                            ollama_error = f"Model '{settings.agent_model}' not found in Ollama. Available: {', '.join(available_models[:3])}"
-                            ollama_healthy = False
-                    except Exception as e:
-                        ollama_error = f"Could not list Ollama models: {str(e)}"
-                        ollama_healthy = False
-
-            except Exception as e:
-                ollama_error = str(e)
-                logger.warning(f"Ollama health check failed: {ollama_error}")
-
-            # Check 2: Database connection
+            # Check database connection
             database = get_database()
             db_healthy = False
             db_error = None
@@ -140,25 +92,15 @@ class HealthController(BaseController):
                 db_error = str(e)
                 logger.warning(f"Database health check failed: {db_error}")
 
-            is_healthy = ollama_healthy and db_healthy
-
             elapsed_ms = int((time.time() - start_time) * 1000)
-            status_code = 200 if is_healthy else 503
+            status_code = 200 if db_healthy else 503
             response.status_code = status_code
 
             return {
-                "status": "healthy" if is_healthy else "unhealthy",
+                "status": "healthy" if db_healthy else "unhealthy",
                 "status_code": status_code,
                 "response_time_ms": elapsed_ms,
                 "components": {
-                    "ollama": {
-                        "status": "healthy" if ollama_healthy else "unhealthy",
-                        "models": models_info if models_info else {
-                            "configured": settings.agent_model,
-                            "available": False
-                        },
-                        "error": ollama_error
-                    },
                     "database": {
                         "status": "healthy" if db_healthy else "unhealthy",
                         "error": db_error
@@ -170,19 +112,13 @@ class HealthController(BaseController):
             "/health/diagnostics",
             summary="Detailed system diagnostics",
         )
-        async def diagnostics(deep_test: bool = False):
+        async def diagnostics():
             """
-            Comprehensive system diagnostics with detailed component information.
-
-            Query Parameters:
-            - deep_test: Set to true to actually test agent generation (slow, ~5-10s)
-
-            Returns detailed information about all system components.
+            System diagnostics with service information.
             """
             import time
-            from src.models.ollama_client import get_ollama_client
-
             start_time = time.time()
+
             diagnostics = {
                 "timestamp": time.time(),
                 "service": {
@@ -190,30 +126,9 @@ class HealthController(BaseController):
                     "version": settings.app_version,
                     "purpose": "Infrastructure management and tools API"
                 },
-                "components": {}
-            }
-
-            # 1. Ollama Connection
-            ollama_client = get_ollama_client()
-            try:
-                ollama_healthy = await ollama_client.health_check()
-                diagnostics["components"]["ollama"] = {
-                    "status": "connected",
-                    "url": settings.ollama_base_url,
-                    "timeout": settings.ollama_timeout,
-                    "default_model": settings.default_model
+                "configuration": {
+                    "cors_origins": settings.cors_origins[:2] if len(settings.cors_origins) > 2 else settings.cors_origins
                 }
-            except Exception as e:
-                diagnostics["components"]["ollama"] = {
-                    "status": "error",
-                    "error": str(e)
-                }
-
-            # 2. Configuration
-            diagnostics["configuration"] = {
-                "agent_fallback_enabled": settings.agent_fallback_enabled,
-                "memory_tier1_max_turns": settings.memory_tier1_max_turns,
-                "cors_origins": settings.cors_origins[:2] if len(settings.cors_origins) > 2 else settings.cors_origins
             }
 
             elapsed_ms = int((time.time() - start_time) * 1000)

src/domains/tools/controller.py

@@ -4,8 +4,10 @@ Tools Controller
 Provides utility tool endpoints including:
 - DNS lookups
 - System stats
+- Environment data (weather, forecast, sun times)
 """
-from fastapi import APIRouter, HTTPException, status
+from typing import Dict, Optional
+from fastapi import APIRouter, HTTPException, status, Depends
 
 from src.shared.base import BaseController
 from src.shared.logging import get_logger
@@ -14,6 +16,9 @@ from src.domains.tools.dns.service import DNSService
 from src.domains.tools.dns.exceptions import DNSQueryError
 from src.domains.tools.system.schemas import SystemStatsResponse
 from src.domains.tools.system.service import SystemStatsService
+from src.domains.tools.environment.schemas import EnvironmentResponse
+from src.domains.tools.environment.service import EnvironmentService
+from src.domains.auth.oidc import get_optional_user
 
 logger = get_logger(__name__)
 
@@ -25,12 +30,14 @@ class ToolsController(BaseController):
     Provides endpoints for:
     - DNS lookups
     - System stats
+    - Environment data (weather, forecast, sun times)
     """
 
     def __init__(self):
         super().__init__(prefix="/tools", tags=["Tools"])
         self.dns_service = DNSService()
         self.system_stats_service = SystemStatsService()
+        self.environment_service = EnvironmentService()
 
     def create_router(self) -> APIRouter:
         """Create and configure the router"""
@@ -146,6 +153,63 @@ class ToolsController(BaseController):
                     detail=f"Failed to collect system stats: {str(e)}"
                 )
 
+        @router.get(
+            "/environment",
+            response_model=EnvironmentResponse,
+            status_code=status.HTTP_200_OK,
+            summary="Get environment data",
+            description="""
+            Get current environment data including weather, forecast, and sun times.
+
+            Fetches data from the Qdrant volatile collection for the authenticated user.
+            Falls back to 'default' user if not authenticated.
+
+            **Data Returned:**
+            - **Weather:** Current temperature, conditions, humidity, wind
+            - **Forecast:** Multi-day weather outlook
+            - **Sun Times:** Sunrise, sunset, daylight duration
+            - **Air Quality:** AQI and pollutant levels (if available)
+
+            **Data Source:** Qdrant volatile_{user} collection
+
+            **Use Cases:**
+            - Dashboard environment widgets
+            - Home automation context
+            - Weather-based automations
+            """
+        )
+        async def get_environment(
+            user: Optional[Dict] = Depends(get_optional_user),
+        ) -> EnvironmentResponse:
+            """
+            Get current environment data
+
+            Args:
+                user: Optional authenticated user from OIDC
+
+            Returns:
+                Environment data including weather, forecast, sun times
+
+            Raises:
+                HTTPException: 500 for processing errors
+            """
+            try:
+                # Get user identifier from OIDC claims, fallback to 'default'
+                user_id = "default"
+                if user:
+                    user_id = user.get("preferred_username") or user.get("sub", "default")
+
+                logger.info(f"Fetching environment data for user: {user_id}")
+                result = await self.environment_service.get_current(user_id)
+                return result
+
+            except Exception as e:
+                logger.error(f"Failed to get environment data: {str(e)}", exc_info=True)
+                raise HTTPException(
+                    status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+                    detail=f"Failed to fetch environment data: {str(e)}"
+                )
+
         return router
 
 

src/domains/tools/environment/__init__.py

@@ -0,0 +1,23 @@
+"""
+Environment data module for Tools domain.
+
+Provides access to weather, forecast, sun times, and air quality data
+from the Qdrant volatile collection.
+"""
+from src.domains.tools.environment.schemas import (
+    WeatherData,
+    ForecastDay,
+    SunTimesData,
+    AirQualityData,
+    EnvironmentResponse,
+)
+from src.domains.tools.environment.service import EnvironmentService
+
+__all__ = [
+    "WeatherData",
+    "ForecastDay",
+    "SunTimesData",
+    "AirQualityData",
+    "EnvironmentResponse",
+    "EnvironmentService",
+]

src/domains/tools/environment/schemas.py

@@ -0,0 +1,185 @@
+"""
+Environment data schemas for Tools domain.
+
+Provides Pydantic models for weather, forecast, sun times, and air quality data
+retrieved from the Qdrant volatile collection.
+"""
+from datetime import datetime
+from typing import Optional, List, Any
+from pydantic import Field
+
+from src.shared.base import BaseSchema
+
+
+class WeatherData(BaseSchema):
+    """Current weather conditions."""
+
+    temperature: Optional[float] = Field(
+        None,
+        description="Current temperature in Celsius"
+    )
+    feels_like: Optional[float] = Field(
+        None,
+        description="Feels-like temperature in Celsius"
+    )
+    conditions: Optional[str] = Field(
+        None,
+        description="Weather conditions description (e.g., 'Partly Cloudy')"
+    )
+    humidity: Optional[int] = Field(
+        None,
+        ge=0,
+        le=100,
+        description="Humidity percentage"
+    )
+    wind_speed: Optional[float] = Field(
+        None,
+        description="Wind speed in km/h"
+    )
+    wind_direction: Optional[str] = Field(
+        None,
+        description="Wind direction (e.g., 'NW')"
+    )
+    pressure: Optional[float] = Field(
+        None,
+        description="Atmospheric pressure in hPa"
+    )
+    visibility: Optional[float] = Field(
+        None,
+        description="Visibility in km"
+    )
+    uv_index: Optional[float] = Field(
+        None,
+        description="UV index"
+    )
+    location: Optional[str] = Field(
+        None,
+        description="Location name"
+    )
+    icon: Optional[str] = Field(
+        None,
+        description="Weather icon code or URL"
+    )
+
+
+class ForecastDay(BaseSchema):
+    """Single day forecast data."""
+
+    date: str = Field(
+        ...,
+        description="Date string (e.g., '2025-01-07')"
+    )
+    high: Optional[float] = Field(
+        None,
+        description="High temperature in Celsius"
+    )
+    low: Optional[float] = Field(
+        None,
+        description="Low temperature in Celsius"
+    )
+    conditions: Optional[str] = Field(
+        None,
+        description="Weather conditions description"
+    )
+    precipitation_chance: Optional[int] = Field(
+        None,
+        ge=0,
+        le=100,
+        description="Chance of precipitation percentage"
+    )
+    icon: Optional[str] = Field(
+        None,
+        description="Weather icon code or URL"
+    )
+
+
+class SunTimesData(BaseSchema):
+    """Sunrise and sunset times."""
+
+    sunrise: Optional[datetime] = Field(
+        None,
+        description="Sunrise time"
+    )
+    sunset: Optional[datetime] = Field(
+        None,
+        description="Sunset time"
+    )
+    daylight_minutes: Optional[int] = Field(
+        None,
+        description="Total daylight duration in minutes"
+    )
+    solar_noon: Optional[datetime] = Field(
+        None,
+        description="Solar noon time"
+    )
+    dawn: Optional[datetime] = Field(
+        None,
+        description="Civil dawn time"
+    )
+    dusk: Optional[datetime] = Field(
+        None,
+        description="Civil dusk time"
+    )
+
+
+class AirQualityData(BaseSchema):
+    """Air quality information."""
+
+    aqi: Optional[int] = Field(
+        None,
+        ge=0,
+        description="Air Quality Index"
+    )
+    quality: Optional[str] = Field(
+        None,
+        description="Quality category (Good, Moderate, Unhealthy, etc.)"
+    )
+    pm25: Optional[float] = Field(
+        None,
+        description="PM2.5 concentration in microg/m3"
+    )
+    pm10: Optional[float] = Field(
+        None,
+        description="PM10 concentration in microg/m3"
+    )
+    o3: Optional[float] = Field(
+        None,
+        description="Ozone concentration in ppb"
+    )
+    no2: Optional[float] = Field(
+        None,
+        description="Nitrogen dioxide concentration in ppb"
+    )
+    location: Optional[str] = Field(
+        None,
+        description="Location name"
+    )
+
+
+class EnvironmentResponse(BaseSchema):
+    """Combined environment data response."""
+
+    weather: Optional[WeatherData] = Field(
+        None,
+        description="Current weather conditions"
+    )
+    forecast: Optional[List[ForecastDay]] = Field(
+        None,
+        description="Multi-day weather forecast"
+    )
+    sun_times: Optional[SunTimesData] = Field(
+        None,
+        description="Sunrise/sunset times"
+    )
+    air_quality: Optional[AirQualityData] = Field(
+        None,
+        description="Air quality data (None if not available)"
+    )
+    updated_at: datetime = Field(
+        default_factory=datetime.utcnow,
+        description="Timestamp when data was fetched"
+    )
+    user: Optional[str] = Field(
+        None,
+        description="User identifier used for data lookup"
+    )

src/domains/tools/environment/service.py

@@ -0,0 +1,246 @@
+"""
+Environment data service for Tools domain.
+
+Fetches weather, forecast, sun times, and air quality data from
+the Qdrant volatile collection.
+"""
+from datetime import datetime
+from typing import Optional, Dict, Any, List
+
+from src.shared.logging import get_logger
+from src.shared.clients.qdrant_client import get_qdrant_client
+from src.domains.tools.environment.schemas import (
+    WeatherData,
+    ForecastDay,
+    SunTimesData,
+    AirQualityData,
+    EnvironmentResponse,
+)
+
+logger = get_logger(__name__)
+
+
+class EnvironmentService:
+    """
+    Service for fetching environment data from Qdrant volatile collection.
+
+    Retrieves weather, forecast, sun times, and optionally air quality
+    data for a specific user.
+    """
+
+    def __init__(self):
+        """Initialize environment service with Qdrant client."""
+        self.qdrant = get_qdrant_client()
+
+    def _parse_weather(self, raw_data: Optional[Dict[str, Any]]) -> Optional[WeatherData]:
+        """
+        Parse raw weather data into WeatherData schema.
+
+        Handles various field naming conventions that might come from
+        different weather APIs.
+        """
+        if not raw_data:
+            return None
+
+        try:
+            return WeatherData(
+                temperature=raw_data.get("temperature") or raw_data.get("temp"),
+                feels_like=raw_data.get("feels_like") or raw_data.get("feelslike"),
+                conditions=raw_data.get("conditions") or raw_data.get("weather") or raw_data.get("description"),
+                humidity=raw_data.get("humidity"),
+                wind_speed=raw_data.get("wind_speed") or raw_data.get("windspeed") or raw_data.get("wind"),
+                wind_direction=raw_data.get("wind_direction") or raw_data.get("wind_dir"),
+                pressure=raw_data.get("pressure"),
+                visibility=raw_data.get("visibility"),
+                uv_index=raw_data.get("uv_index") or raw_data.get("uv"),
+                location=raw_data.get("location") or raw_data.get("city"),
+                icon=raw_data.get("icon") or raw_data.get("icon_url"),
+            )
+        except Exception as e:
+            logger.warning(f"Failed to parse weather data: {e}")
+            return None
+
+    def _parse_forecast(self, raw_data: Any) -> Optional[List[ForecastDay]]:
+        """
+        Parse raw forecast data into list of ForecastDay schemas.
+
+        Handles both list format and dict with nested list.
+        """
+        if not raw_data:
+            return None
+
+        try:
+            # Normalize to list
+            forecast_list = raw_data
+            if isinstance(raw_data, dict):
+                forecast_list = raw_data.get("days") or raw_data.get("forecast") or []
+
+            if not isinstance(forecast_list, list):
+                return None
+
+            days = []
+            for day in forecast_list:
+                if isinstance(day, dict):
+                    days.append(ForecastDay(
+                        date=day.get("date", ""),
+                        high=day.get("high") or day.get("maxtemp") or day.get("temp_max"),
+                        low=day.get("low") or day.get("mintemp") or day.get("temp_min"),
+                        conditions=day.get("conditions") or day.get("weather") or day.get("description"),
+                        precipitation_chance=day.get("precipitation_chance") or day.get("pop") or day.get("precip"),
+                        icon=day.get("icon"),
+                    ))
+
+            return days if days else None
+
+        except Exception as e:
+            logger.warning(f"Failed to parse forecast data: {e}")
+            return None
+
+    def _parse_sun_times(self, raw_data: Optional[Dict[str, Any]]) -> Optional[SunTimesData]:
+        """
+        Parse raw sun times data into SunTimesData schema.
+
+        Handles datetime strings and calculates daylight minutes if not provided.
+        """
+        if not raw_data:
+            return None
+
+        try:
+            sunrise = raw_data.get("sunrise")
+            sunset = raw_data.get("sunset")
+
+            # Parse datetime strings if needed
+            if isinstance(sunrise, str):
+                sunrise = datetime.fromisoformat(sunrise.replace("Z", "+00:00"))
+            if isinstance(sunset, str):
+                sunset = datetime.fromisoformat(sunset.replace("Z", "+00:00"))
+
+            # Calculate daylight minutes if not provided
+            daylight_minutes = raw_data.get("daylight_minutes") or raw_data.get("daylight")
+            if daylight_minutes is None and sunrise and sunset:
+                daylight_minutes = int((sunset - sunrise).total_seconds() / 60)
+
+            # Parse optional fields
+            solar_noon = raw_data.get("solar_noon")
+            if isinstance(solar_noon, str):
+                solar_noon = datetime.fromisoformat(solar_noon.replace("Z", "+00:00"))
+
+            dawn = raw_data.get("dawn") or raw_data.get("civil_dawn")
+            if isinstance(dawn, str):
+                dawn = datetime.fromisoformat(dawn.replace("Z", "+00:00"))
+
+            dusk = raw_data.get("dusk") or raw_data.get("civil_dusk")
+            if isinstance(dusk, str):
+                dusk = datetime.fromisoformat(dusk.replace("Z", "+00:00"))
+
+            return SunTimesData(
+                sunrise=sunrise,
+                sunset=sunset,
+                daylight_minutes=daylight_minutes,
+                solar_noon=solar_noon,
+                dawn=dawn,
+                dusk=dusk,
+            )
+
+        except Exception as e:
+            logger.warning(f"Failed to parse sun times data: {e}")
+            return None
+
+    def _parse_air_quality(self, raw_data: Any) -> Optional[AirQualityData]:
+        """
+        Parse raw air quality data into AirQualityData schema.
+
+        Handles both dict format and simple integer AQI value.
+        """
+        if raw_data is None:
+            return None
+
+        try:
+            # Handle simple integer AQI
+            if isinstance(raw_data, (int, float)):
+                aqi = int(raw_data)
+                return AirQualityData(
+                    aqi=aqi,
+                    quality=self._aqi_to_quality(aqi),
+                )
+
+            if not isinstance(raw_data, dict):
+                return None
+
+            aqi = raw_data.get("aqi") or raw_data.get("index")
+            if isinstance(aqi, (int, float)):
+                aqi = int(aqi)
+
+            return AirQualityData(
+                aqi=aqi,
+                quality=raw_data.get("quality") or (self._aqi_to_quality(aqi) if aqi else None),
+                pm25=raw_data.get("pm25") or raw_data.get("pm2_5"),
+                pm10=raw_data.get("pm10"),
+                o3=raw_data.get("o3") or raw_data.get("ozone"),
+                no2=raw_data.get("no2"),
+                location=raw_data.get("location"),
+            )
+
+        except Exception as e:
+            logger.warning(f"Failed to parse air quality data: {e}")
+            return None
+
+    def _aqi_to_quality(self, aqi: int) -> str:
+        """Convert AQI value to quality category string."""
+        if aqi <= 50:
+            return "Good"
+        elif aqi <= 100:
+            return "Moderate"
+        elif aqi <= 150:
+            return "Unhealthy for Sensitive Groups"
+        elif aqi <= 200:
+            return "Unhealthy"
+        elif aqi <= 300:
+            return "Very Unhealthy"
+        else:
+            return "Hazardous"
+
+    async def get_current(self, user: str = "default") -> EnvironmentResponse:
+        """
+        Get current environment data for a user.
+
+        Fetches weather, forecast, sun times, and air quality from
+        the user's volatile collection.
+
+        Args:
+            user: User identifier (default: 'default')
+
+        Returns:
+            EnvironmentResponse with all available data
+        """
+        logger.info(f"Fetching environment data for user: {user}")
+
+        # Get raw data from Qdrant
+        raw_data = await self.qdrant.get_environment_data(user)
+
+        # Parse each data type
+        weather = self._parse_weather(raw_data.get("weather"))
+        forecast = self._parse_forecast(raw_data.get("forecast"))
+        sun_times = self._parse_sun_times(raw_data.get("sun_times"))
+        air_quality = self._parse_air_quality(raw_data.get("air_quality"))
+
+        return EnvironmentResponse(
+            weather=weather,
+            forecast=forecast,
+            sun_times=sun_times,
+            air_quality=air_quality,
+            updated_at=datetime.utcnow(),
+            user=user,
+        )
+
+
+# Singleton instance
+_environment_service: Optional[EnvironmentService] = None
+
+
+def get_environment_service() -> EnvironmentService:
+    """Get or create singleton environment service instance."""
+    global _environment_service
+    if _environment_service is None:
+        _environment_service = EnvironmentService()
+    return _environment_service

src/main.py

@@ -10,7 +10,6 @@ from src.shared.config import get_settings
 from src.shared.logging import setup_logging, get_logger
 from src.shared.database import get_database
 from src.shared.security import initialize_oidc
-from src.models.ollama_client import get_ollama_client, close_ollama_client
 
 # Import domain controllers
 from src.domains.health import health_controller
@@ -42,17 +41,8 @@ async def lifespan(app: FastAPI):
     logger.info(f"Starting {settings.app_name} v{settings.app_version}")
     logger.info(f"Debug mode: {settings.debug}")
     logger.info(f"Log level: {settings.log_level}")
-    logger.info(f"Ollama URL: {settings.ollama_base_url}")
     logger.info("=" * 60)
 
-    # Check Ollama connectivity
-    ollama_client = get_ollama_client()
-    ollama_healthy = await ollama_client.health_check()
-    if ollama_healthy:
-        logger.info("Ollama connection successful")
-    else:
-        logger.warning("Ollama connection failed - AI features may not work")
-
     # Check database connectivity
     database = get_database()
     db_healthy = await database.health_check()
@@ -68,7 +58,6 @@ async def lifespan(app: FastAPI):
 
     # Shutdown
     logger.info("Shutting down application")
-    await close_ollama_client()
     await database.close()
 
 
@@ -94,7 +83,7 @@ See `/docs` for the full API reference.
     lifespan=lifespan,
     debug=settings.debug,
     swagger_ui_init_oauth={
-        "clientId": settings.oidc_audience,
+        "clientId": settings.oidc_audiences[0] if settings.oidc_audiences else "core-api",
         "usePkceWithAuthorizationCodeGrant": True,
     } if settings.oidc_enabled else None
 )

src/models/embeddings.py

@@ -1,128 +0,0 @@
-"""
-Embedding model client for text vectorization
-
-Uses sentence-transformers for generating embeddings.
-"""
-import logging
-from typing import List, Optional
-from sentence_transformers import SentenceTransformer
-from src.config import get_settings
-
-logger = logging.getLogger(__name__)
-settings = get_settings()
-
-
-class EmbeddingClient:
-    """Client for generating text embeddings"""
-
-    def __init__(self, model_name: Optional[str] = None):
-        """
-        Initialize embedding client
-
-        Args:
-            model_name: Optional model name, defaults to config
-        """
-        self.model_name = model_name or settings.embedding_model
-        self.dimension = settings.embedding_dimension
-        self._model: Optional[SentenceTransformer] = None
-        logger.info(f"Initializing EmbeddingClient with model: {self.model_name}")
-
-    def _load_model(self) -> SentenceTransformer:
-        """
-        Lazy load the embedding model
-
-        Returns:
-            Loaded SentenceTransformer model
-        """
-        if self._model is None:
-            logger.info(f"Loading embedding model: {self.model_name}")
-            self._model = SentenceTransformer(self.model_name)
-            logger.info(f"Model loaded successfully. Embedding dimension: {self.dimension}")
-        return self._model
-
-    def embed_text(self, text: str) -> List[float]:
-        """
-        Generate embedding for a single text
-
-        Args:
-            text: Input text to embed
-
-        Returns:
-            List of floats representing the embedding vector
-        """
-        model = self._load_model()
-        embedding = model.encode(text, convert_to_numpy=True)
-        return embedding.tolist()
-
-    def embed_batch(self, texts: List[str]) -> List[List[float]]:
-        """
-        Generate embeddings for multiple texts
-
-        Args:
-            texts: List of input texts
-
-        Returns:
-            List of embedding vectors
-        """
-        model = self._load_model()
-        embeddings = model.encode(
-            texts,
-            batch_size=settings.embedding_batch_size,
-            convert_to_numpy=True,
-            show_progress_bar=False
-        )
-        return embeddings.tolist()
-
-    def get_dimension(self) -> int:
-        """
-        Get embedding dimension
-
-        Returns:
-            Embedding vector dimension
-        """
-        return self.dimension
-
-
-# Global instance
-_embedding_client: Optional[EmbeddingClient] = None
-
-
-def get_embedding_client() -> EmbeddingClient:
-    """
-    Get or create global embedding client instance
-
-    Returns:
-        EmbeddingClient instance
-    """
-    global _embedding_client
-    if _embedding_client is None:
-        _embedding_client = EmbeddingClient()
-    return _embedding_client
-
-
-async def embed_text_async(text: str) -> List[float]:
-    """
-    Async wrapper for embedding text
-
-    Args:
-        text: Input text
-
-    Returns:
-        Embedding vector
-    """
-    client = get_embedding_client()
-    return client.embed_text(text)
-
-
-async def embed_batch_async(texts: List[str]) -> List[List[float]]:
-    """
-    Async wrapper for batch embedding
-
-    Args:
-        texts: List of input texts
-
-    Returns:
-        List of embedding vectors
-    """
-    client = get_embedding_client()
-    return client.embed_batch(texts)

src/models/embeddings_ollama.py

@@ -1,136 +0,0 @@
-"""
-Ollama-based embedding client for text vectorization
-
-Uses Ollama's embedding API instead of local sentence-transformers.
-This eliminates the need for PyTorch and heavy ML dependencies.
-"""
-import logging
-import httpx
-from typing import List, Optional
-from src.config import get_settings
-
-logger = logging.getLogger(__name__)
-settings = get_settings()
-
-
-class OllamaEmbeddingClient:
-    """Client for generating text embeddings using Ollama"""
-
-    def __init__(
-        self,
-        model_name: Optional[str] = None,
-        base_url: Optional[str] = None,
-        timeout: int = 30
-    ):
-        """
-        Initialize Ollama embedding client
-
-        Args:
-            model_name: Embedding model name (default: nomic-embed-text)
-            base_url: Ollama base URL (default from settings)
-            timeout: Request timeout in seconds
-        """
-        self.model_name = model_name or settings.embedding_model
-        self.base_url = (base_url or settings.ollama_base_url).rstrip("/")
-        self.timeout = timeout
-        self.dimension = settings.embedding_dimension
-
-        logger.info(f"Initializing OllamaEmbeddingClient with model: {self.model_name}")
-        logger.info(f"Ollama URL: {self.base_url}")
-
-    async def embed_text(self, text: str) -> List[float]:
-        """
-        Generate embedding for a single text using Ollama
-
-        Args:
-            text: Input text to embed
-
-        Returns:
-            List of floats representing the embedding vector
-        """
-        try:
-            async with httpx.AsyncClient(timeout=self.timeout) as client:
-                response = await client.post(
-                    f"{self.base_url}/api/embeddings",
-                    json={
-                        "model": self.model_name,
-                        "prompt": text
-                    }
-                )
-                response.raise_for_status()
-                result = response.json()
-                return result["embedding"]
-
-        except Exception as e:
-            logger.error(f"Error generating embedding via Ollama: {e}")
-            raise
-
-    async def embed_batch(self, texts: List[str]) -> List[List[float]]:
-        """
-        Generate embeddings for multiple texts
-
-        Args:
-            texts: List of input texts
-
-        Returns:
-            List of embedding vectors
-        """
-        embeddings = []
-        for text in texts:
-            embedding = await self.embed_text(text)
-            embeddings.append(embedding)
-        return embeddings
-
-    def get_dimension(self) -> int:
-        """
-        Get embedding dimension
-
-        Returns:
-            Embedding vector dimension
-        """
-        return self.dimension
-
-
-# Global instance
-_embedding_client: Optional[OllamaEmbeddingClient] = None
-
-
-def get_embedding_client() -> OllamaEmbeddingClient:
-    """
-    Get or create global Ollama embedding client instance
-
-    Returns:
-        OllamaEmbeddingClient instance
-    """
-    global _embedding_client
-    if _embedding_client is None:
-        _embedding_client = OllamaEmbeddingClient()
-    return _embedding_client
-
-
-async def embed_text_async(text: str) -> List[float]:
-    """
-    Async wrapper for embedding text
-
-    Args:
-        text: Input text
-
-    Returns:
-        Embedding vector
-    """
-    client = get_embedding_client()
-    return await client.embed_text(text)
-
-
-async def embed_batch_async(texts: List[str]) -> List[List[float]]:
-    """
-    Async wrapper for batch embedding
-
-    Args:
-        texts: List of input texts
-
-    Returns:
-        List of embedding vectors
-    """
-    client = get_embedding_client()
-    return await client.embed_batch(texts)

src/models/ollama_client.py

@@ -1,223 +0,0 @@
-"""
-Ollama client for model inference.
-Handles both streaming and non-streaming requests.
-"""
-
-import httpx
-import json
-import logging
-from typing import AsyncIterator, Dict, Any, Optional
-from src.config import get_settings
-
-logger = logging.getLogger(__name__)
-settings = get_settings()
-
-
-class OllamaClient:
-    """Client for interacting with Ollama API."""
-
-    def __init__(self):
-        self.base_url = settings.ollama_base_url
-        self.timeout = settings.ollama_timeout
-        self.client = httpx.AsyncClient(timeout=self.timeout)
-        logger.info(f"Initialized Ollama client: {self.base_url}")
-
-    async def close(self):
-        """Close the HTTP client."""
-        await self.client.aclose()
-
-    def resolve_model(self, model_name: str) -> str:
-        """
-        Resolve model alias to actual Ollama model.
-
-        Args:
-            model_name: Requested model name (e.g., "gpt-3.5-turbo")
-
-        Returns:
-            Actual Ollama model name (e.g., "gemma:7b")
-        """
-        resolved = settings.model_aliases.get(model_name, model_name)
-        if resolved != model_name:
-            logger.info(f"Model resolution: {model_name} → {resolved}")
-        return resolved
-
-    async def generate_non_streaming(
-        self,
-        model: str,
-        prompt: str,
-        temperature: float = 0.7,
-        max_tokens: Optional[int] = None
-    ) -> Dict[str, Any]:
-        """
-        Generate non-streaming response from Ollama using chat endpoint.
-
-        Args:
-            model: Model name
-            prompt: User prompt
-            temperature: Sampling temperature
-            max_tokens: Maximum tokens to generate
-
-        Returns:
-            Dict with 'response' and 'tokens' keys
-        """
-        actual_model = self.resolve_model(model)
-
-        payload = {
-            "model": actual_model,
-            "messages": [
-                {"role": "user", "content": prompt}
-            ],
-            "stream": False,
-            "options": {
-                "temperature": temperature,
-            }
-        }
-
-        if max_tokens:
-            payload["options"]["num_predict"] = max_tokens
-
-        logger.debug(f"Ollama request to {actual_model}")
-
-        try:
-            response = await self.client.post(
-                f"{self.base_url}/api/chat",
-                json=payload
-            )
-            response.raise_for_status()
-            result = response.json()
-
-            return {
-                "response": result.get("message", {}).get("content", ""),
-                "tokens": {
-                    "prompt": result.get("prompt_eval_count", 0),
-                    "completion": result.get("eval_count", 0),
-                    "total": result.get("prompt_eval_count", 0) + result.get("eval_count", 0)
-                }
-            }
-
-        except httpx.HTTPError as e:
-            logger.error(f"Ollama request failed: {e}")
-            raise
-
-    async def generate_streaming(
-        self,
-        model: str,
-        prompt: str,
-        temperature: float = 0.7,
-        max_tokens: Optional[int] = None
-    ) -> AsyncIterator[str]:
-        """
-        Generate streaming response from Ollama using chat endpoint.
-
-        Args:
-            model: Model name
-            prompt: User prompt
-            temperature: Sampling temperature
-            max_tokens: Maximum tokens to generate
-
-        Yields:
-            Token strings
-        """
-        actual_model = self.resolve_model(model)
-
-        payload = {
-            "model": actual_model,
-            "messages": [
-                {"role": "user", "content": prompt}
-            ],
-            "stream": True,
-            "options": {
-                "temperature": temperature,
-            }
-        }
-
-        if max_tokens:
-            payload["options"]["num_predict"] = max_tokens
-
-        logger.debug(f"Ollama streaming request to {actual_model}")
-
-        try:
-            async with self.client.stream(
-                "POST",
-                f"{self.base_url}/api/chat",
-                json=payload
-            ) as response:
-                response.raise_for_status()
-
-                async for line in response.aiter_lines():
-                    if not line:
-                        continue
-
-                    try:
-                        chunk = json.loads(line)
-                        if "message" in chunk:
-                            content = chunk["message"].get("content", "")
-                            if content:
-                                yield content
-
-                        # Check if done
-                        if chunk.get("done", False):
-                            break
-
-                    except json.JSONDecodeError:
-                        logger.warning(f"Failed to parse JSON: {line}")
-                        continue
-
-        except httpx.HTTPError as e:
-            logger.error(f"Ollama streaming request failed: {e}")
-            raise
-
-    async def health_check(self) -> bool:
-        """
-        Check if Ollama is healthy.
-
-        Returns:
-            True if healthy, False otherwise
-        """
-        try:
-            response = await self.client.get(
-                f"{self.base_url}/api/tags",
-                timeout=5.0
-            )
-            return response.status_code == 200
-        except Exception as e:
-            logger.error(f"Ollama health check failed: {e}")
-            return False
-
-    async def list_models(self) -> Dict[str, Any]:
-        """
-        List all available models in Ollama.
-
-        Returns:
-            Dict with 'models' key containing list of model info
-        """
-        try:
-            response = await self.client.get(
-                f"{self.base_url}/api/tags",
-                timeout=5.0
-            )
-            response.raise_for_status()
-            return response.json()
-        except Exception as e:
-            logger.error(f"Failed to list Ollama models: {e}")
-            raise
-
-
-# Global client instance
-_ollama_client: Optional[OllamaClient] = None
-
-
-def get_ollama_client() -> OllamaClient:
-    """Get or create the global Ollama client instance."""
-    global _ollama_client
-    if _ollama_client is None:
-        _ollama_client = OllamaClient()
-    return _ollama_client
-
-
-async def close_ollama_client():
-    """Close the global Ollama client."""
-    global _ollama_client
-    if _ollama_client is not None:
-        await _ollama_client.close()
-        _ollama_client = None

src/shared/clients/__init__.py

@@ -7,6 +7,7 @@ from src.shared.clients.portainer_client import PortainerClient, get_portainer_c
 from src.shared.clients.npm_client import NPMClient, get_npm_client
 from src.shared.clients.homeassistant_client import HomeAssistantClient, get_homeassistant_client
 from src.shared.clients.authentik_client import AuthentikClient, get_authentik_client
+from src.shared.clients.qdrant_client import QdrantReadClient, get_qdrant_client
 
 __all__ = [
     "PortainerClient",
@@ -17,4 +18,6 @@ __all__ = [
     "get_homeassistant_client",
     "AuthentikClient",
     "get_authentik_client",
+    "QdrantReadClient",
+    "get_qdrant_client",
 ]

src/shared/clients/qdrant_client.py

@@ -0,0 +1,258 @@
+"""
+Qdrant Vector Database Client (Read-Only)
+
+Provides read-only access to Qdrant collections for querying volatile data.
+Used to fetch weather, forecast, and sun times from the volatile_{user} collection.
+"""
+import time
+from typing import List, Dict, Any, Optional
+from qdrant_client import QdrantClient
+from qdrant_client.models import Filter, FieldCondition, MatchValue, Range
+
+from src.shared.logging import get_logger
+from src.shared.config import get_settings
+
+logger = get_logger(__name__)
+settings = get_settings()
+
+
+class QdrantReadClient:
+    """
+    Read-only Qdrant client for accessing volatile data.
+
+    Connects to Qdrant and provides methods to query collections
+    with filtering by namespace and TTL expiry.
+    """
+
+    VOLATILE_COLLECTION_PREFIX = "volatile_"
+
+    def __init__(
+        self,
+        host: Optional[str] = None,
+        port: Optional[int] = None,
+    ):
+        """
+        Initialize Qdrant read client.
+
+        Args:
+            host: Qdrant server host (default from settings)
+            port: Qdrant server port (default from settings)
+        """
+        self.host = host or settings.qdrant_host
+        self.port = port or settings.qdrant_port
+        self._client: Optional[QdrantClient] = None
+
+        logger.info(f"Initialized QdrantReadClient: {self.host}:{self.port}")
+
+    @property
+    def client(self) -> QdrantClient:
+        """Lazy-load Qdrant client connection."""
+        if self._client is None:
+            self._client = QdrantClient(
+                host=self.host,
+                port=self.port,
+            )
+        return self._client
+
+    def _get_volatile_collection(self, user: str) -> str:
+        """Get volatile collection name for user."""
+        return f"{self.VOLATILE_COLLECTION_PREFIX}{user}"
+
+    def _current_timestamp_ms(self) -> int:
+        """Get current timestamp in milliseconds."""
+        return int(time.time() * 1000)
+
+    async def collection_exists(self, collection_name: str) -> bool:
+        """
+        Check if a collection exists.
+
+        Args:
+            collection_name: Name of collection to check
+
+        Returns:
+            True if collection exists
+        """
+        try:
+            collections = self.client.get_collections()
+            existing = [c.name for c in collections.collections]
+            return collection_name in existing
+        except Exception as e:
+            logger.error(f"Error checking collection existence: {e}")
+            return False
+
+    async def get_by_namespace(
+        self,
+        user: str,
+        namespace: str,
+        include_expired: bool = False
+    ) -> List[Dict[str, Any]]:
+        """
+        Get all records for a specific namespace from user's volatile collection.
+
+        Args:
+            user: User identifier (e.g., 'jpmschweitzer' or 'default')
+            namespace: Namespace to filter (e.g., 'weather', 'forecast', 'sun')
+            include_expired: Whether to include expired records (default False)
+
+        Returns:
+            List of records with payload data
+        """
+        collection_name = self._get_volatile_collection(user)
+
+        if not await self.collection_exists(collection_name):
+            logger.debug(f"Collection {collection_name} does not exist")
+            return []
+
+        # Build filter conditions
+        conditions = [
+            FieldCondition(
+                key="namespace",
+                match=MatchValue(value=namespace)
+            )
+        ]
+
+        # Add TTL expiry filter unless including expired
+        if not include_expired:
+            now_ms = self._current_timestamp_ms()
+            conditions.append(
+                FieldCondition(
+                    key="ttl_expiry",
+                    range=Range(gt=now_ms)
+                )
+            )
+
+        query_filter = Filter(must=conditions)
+
+        try:
+            # Scroll through matching records
+            points, _ = self.client.scroll(
+                collection_name=collection_name,
+                scroll_filter=query_filter,
+                limit=100,
+                with_payload=True,
+                with_vectors=False
+            )
+
+            results = []
+            for point in points:
+                payload = dict(point.payload) if point.payload else {}
+                results.append({
+                    "id": str(point.id),
+                    "namespace": payload.get("namespace"),
+                    "key": payload.get("key"),
+                    "raw_data": payload.get("raw_data", {}),
+                    "source": payload.get("source"),
+                    "ttl_expiry": payload.get("ttl_expiry"),
+                    "updated_at": payload.get("updated_at"),
+                })
+
+            logger.debug(
+                f"Found {len(results)} records in {collection_name}/{namespace}"
+            )
+            return results
+
+        except Exception as e:
+            logger.error(f"Error fetching from {collection_name}/{namespace}: {e}")
+            return []
+
+    async def get_environment_data(
+        self,
+        user: str
+    ) -> Dict[str, Any]:
+        """
+        Get all environment data (weather, forecast, sun times) for a user.
+
+        Convenience method that fetches all environment-related namespaces
+        in a single call.
+
+        Args:
+            user: User identifier
+
+        Returns:
+            Dict with 'weather', 'forecast', 'sun_times', 'air_quality' keys
+            (each may be None if no data found)
+        """
+        result = {
+            "weather": None,
+            "forecast": None,
+            "sun_times": None,
+            "air_quality": None,
+        }
+
+        # Fetch weather data
+        weather_records = await self.get_by_namespace(user, "weather")
+        if weather_records:
+            # Get the first/most recent weather record
+            result["weather"] = weather_records[0].get("raw_data")
+            # Check if air quality is embedded in weather data
+            if result["weather"]:
+                aqi = result["weather"].get("aqi") or result["weather"].get("air_quality")
+                if aqi:
+                    result["air_quality"] = aqi if isinstance(aqi, dict) else {"aqi": aqi}
+
+        # Fetch forecast data
+        forecast_records = await self.get_by_namespace(user, "forecast")
+        if forecast_records:
+            # Forecast might be a single record with list or multiple records
+            first_record = forecast_records[0].get("raw_data")
+            if isinstance(first_record, list):
+                result["forecast"] = first_record
+            elif isinstance(first_record, dict):
+                # Could be a dict with 'days' or 'forecast' key
+                result["forecast"] = first_record.get(
+                    "days",
+                    first_record.get("forecast", [first_record])
+                )
+
+        # Fetch sun times data
+        sun_records = await self.get_by_namespace(user, "sun")
+        if sun_records:
+            result["sun_times"] = sun_records[0].get("raw_data")
+
+        # Check for separate air quality namespace if not embedded
+        if result["air_quality"] is None:
+            aq_records = await self.get_by_namespace(user, "air_quality")
+            if aq_records:
+                result["air_quality"] = aq_records[0].get("raw_data")
+
+        return result
+
+    async def health_check(self) -> Dict[str, Any]:
+        """
+        Check Qdrant connectivity.
+
+        Returns:
+            Dict with connection status and info
+        """
+        try:
+            collections = self.client.get_collections()
+            volatile_collections = [
+                c.name for c in collections.collections
+                if c.name.startswith(self.VOLATILE_COLLECTION_PREFIX)
+            ]
+            return {
+                "status": "healthy",
+                "connected": True,
+                "host": f"{self.host}:{self.port}",
+                "volatile_collections": volatile_collections,
+            }
+        except Exception as e:
+            logger.error(f"Qdrant health check failed: {e}")
+            return {
+                "status": "unhealthy",
+                "connected": False,
+                "host": f"{self.host}:{self.port}",
+                "error": str(e),
+            }
+
+
+# Singleton instance for reuse
+_qdrant_client: Optional[QdrantReadClient] = None
+
+
+def get_qdrant_client() -> QdrantReadClient:
+    """Get or create singleton Qdrant client instance."""
+    global _qdrant_client
+    if _qdrant_client is None:
+        _qdrant_client = QdrantReadClient()
+    return _qdrant_client

src/shared/config.py

@@ -36,8 +36,15 @@ class Settings(BaseSettings):
     host: str = "0.0.0.0"
     port: int = 8083
 
-    # CORS
-    cors_origins: list[str] = ["*"]
+    # CORS - Note: When cors_credentials is True, cannot use "*" for origins
+    # Set CORS_ORIGINS env var to override (comma-separated list)
+    cors_origins: list[str] = [
+        "https://home.schweitz.net",
+        "https://tatlock.schweitz.net",
+        "http://localhost:8080",
+        "http://localhost:3000",
+        "http://127.0.0.1:8080",
+    ]
     cors_credentials: bool = True
     cors_methods: list[str] = ["*"]
     cors_headers: list[str] = ["*"]
@@ -45,31 +52,6 @@ class Settings(BaseSettings):
     # Logging
     log_level: str = "DEBUG"
 
-    # Ollama Configuration (for AI orchestration)
-    ollama_base_url: str  # Required - set OLLAMA_BASE_URL in .env
-    ollama_timeout: int = 300  # 5 minutes
-
-    # Model Configuration
-    default_model: str = "mistral-nemo-large:latest"
-    agent_model: str = "mistral-nemo-large:latest"
-    code_models: str = "mistral-nemo-large:latest"
-
-    # System Prompt Variant (for A/B testing)
-    system_prompt_variant: str = "v8_holistic"
-
-    # Agent Configuration
-    agent_fallback_enabled: bool = True
-
-    # Model Aliases (OpenAI → Local)
-    alias_gpt35: str = "gemma:7b"
-    alias_gpt4: str = "mistral:7b"
-    alias_gpt4_turbo: str = "mixtral:8x7b"
-    alias_gpt4_code: str = "codestral:latest"
-
-    # Memory Configuration
-    memory_tier1_max_turns: int = 10
-    memory_consolidation_threshold: int = 10
-
     # Qdrant Configuration
     qdrant_host: str = "qdrant"
     qdrant_port: int = 6333
@@ -77,11 +59,6 @@ class Settings(BaseSettings):
     qdrant_collection_documents: str = "core_api_documents"
     qdrant_collection_user_facts: str = "core_api_user_facts"
 
-    # Embeddings (using Ollama)
-    embedding_model: str = "nomic-embed-text"
-    embedding_dimension: int = 768
-    embedding_batch_size: int = 32
-
     # Search Configuration
     search_provider: str = "searxng"
     searxng_url: str  # Required - set SEARXNG_URL in .env
@@ -114,27 +91,14 @@ class Settings(BaseSettings):
     # OIDC Authentication (Authentik)
     oidc_enabled: bool = False
     oidc_issuer: str = "https://auth.schweitz.net/application/o/core-api/"
-    oidc_audience: str = "core-api"
+    # Accept tokens from multiple clients (core-api, tatlock-ui, tatlock)
+    oidc_audiences: list[str] = ["core-api", "tatlock-ui", "tatlock"]
 
     # Authentik API (for token validation and user management)
     authentik_url: str = "https://auth.schweitz.net"
     authentik_username: str = ""
     authentik_password: str = ""
 
-    @property
-    def model_aliases(self) -> dict:
-        """Computed property for model aliases"""
-        return {
-            "gpt-3.5-turbo": self.alias_gpt35,
-            "gpt-4": self.alias_gpt4,
-            "gpt-4-turbo": self.alias_gpt4_turbo,
-            "gpt-4-code": self.alias_gpt4_code,
-        }
-
-    def get_code_models(self) -> list[str]:
-        """Parse comma-separated code models"""
-        return [m.strip().strip('"').strip("'") for m in self.code_models.split(",") if m.strip()]
-
     class Config:
         env_file = ".env"
         case_sensitive = False

src/shared/security.py

@@ -17,12 +17,20 @@ def initialize_oidc(settings: Settings) -> None:
         settings: Application settings containing OIDC configuration
     """
     # Import here to avoid circular imports
-    from src.auth.oidc import oidc_config
+    # Configure BOTH oidc modules (src.auth and src.domains.auth)
+    from src.auth.oidc import oidc_config as auth_oidc_config
+    from src.domains.auth.oidc import oidc_config as domains_oidc_config
 
-    oidc_config.configure(
+    auth_oidc_config.configure(
         enabled=settings.oidc_enabled,
         issuer=settings.oidc_issuer,
-        audience=settings.oidc_audience
+        audiences=settings.oidc_audiences
+    )
+
+    domains_oidc_config.configure(
+        enabled=settings.oidc_enabled,
+        issuer=settings.oidc_issuer,
+        audiences=settings.oidc_audiences
     )
 
     if settings.oidc_enabled:

tests/test_environment.py

@@ -0,0 +1,159 @@
+"""Tests for environment service and schemas."""
+import pytest
+from datetime import datetime
+
+
+class TestEnvironmentService:
+    """Test EnvironmentService methods."""
+
+    def test_aqi_to_quality_good(self):
+        """AQI 0-50 should return Good."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        assert service._aqi_to_quality(0) == "Good"
+        assert service._aqi_to_quality(25) == "Good"
+        assert service._aqi_to_quality(50) == "Good"
+
+    def test_aqi_to_quality_moderate(self):
+        """AQI 51-100 should return Moderate."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        assert service._aqi_to_quality(51) == "Moderate"
+        assert service._aqi_to_quality(75) == "Moderate"
+        assert service._aqi_to_quality(100) == "Moderate"
+
+    def test_aqi_to_quality_unhealthy_sensitive(self):
+        """AQI 101-150 should return Unhealthy for Sensitive Groups."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        assert service._aqi_to_quality(101) == "Unhealthy for Sensitive Groups"
+        assert service._aqi_to_quality(150) == "Unhealthy for Sensitive Groups"
+
+    def test_aqi_to_quality_unhealthy(self):
+        """AQI 151-200 should return Unhealthy."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        assert service._aqi_to_quality(151) == "Unhealthy"
+        assert service._aqi_to_quality(200) == "Unhealthy"
+
+    def test_aqi_to_quality_very_unhealthy(self):
+        """AQI 201-300 should return Very Unhealthy."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        assert service._aqi_to_quality(201) == "Very Unhealthy"
+        assert service._aqi_to_quality(300) == "Very Unhealthy"
+
+    def test_aqi_to_quality_hazardous(self):
+        """AQI >300 should return Hazardous."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        assert service._aqi_to_quality(301) == "Hazardous"
+        assert service._aqi_to_quality(500) == "Hazardous"
+
+    def test_parse_weather_with_valid_data(self):
+        """Parse weather should return WeatherData for valid input."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        raw = {
+            "temperature": 15.5,
+            "conditions": "Cloudy",
+            "humidity": 72,
+            "location": "Rotterdam",
+        }
+        result = service._parse_weather(raw)
+
+        assert result is not None
+        assert result.temperature == 15.5
+        assert result.conditions == "Cloudy"
+        assert result.humidity == 72
+
+    def test_parse_weather_with_none(self):
+        """Parse weather should return None for None input."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        result = service._parse_weather(None)
+        assert result is None
+
+    def test_parse_forecast_with_list(self):
+        """Parse forecast should handle list format."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        raw = [
+            {"date": "2026-01-06", "high": 12, "low": 5, "conditions": "Cloudy"},
+            {"date": "2026-01-07", "high": 14, "low": 6, "conditions": "Sunny"},
+        ]
+        result = service._parse_forecast(raw)
+
+        assert result is not None
+        assert len(result) == 2
+        assert result[0].date == "2026-01-06"
+        assert result[0].high == 12
+
+    def test_parse_forecast_with_dict(self):
+        """Parse forecast should handle dict with days key."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        raw = {
+            "days": [
+                {"date": "2026-01-06", "high": 12, "low": 5, "conditions": "Cloudy"},
+            ]
+        }
+        result = service._parse_forecast(raw)
+
+        assert result is not None
+        assert len(result) == 1
+
+    def test_parse_sun_times_with_strings(self):
+        """Parse sun times should handle ISO datetime strings."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        raw = {
+            "sunrise": "2026-01-06T08:45:00",
+            "sunset": "2026-01-06T16:50:00",
+        }
+        result = service._parse_sun_times(raw)
+
+        assert result is not None
+        assert result.sunrise.hour == 8
+        assert result.sunrise.minute == 45
+        assert result.sunset.hour == 16
+        assert result.daylight_minutes == 485
+
+    def test_parse_air_quality_with_int(self):
+        """Parse air quality should handle simple integer AQI."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        result = service._parse_air_quality(42)
+
+        assert result is not None
+        assert result.aqi == 42
+        assert result.quality == "Good"
+
+    def test_parse_air_quality_with_dict(self):
+        """Parse air quality should handle dict format."""
+        from src.domains.tools.environment.service import EnvironmentService
+        service = EnvironmentService.__new__(EnvironmentService)
+
+        raw = {
+            "aqi": 75,
+            "pm25": 8.5,
+            "pm10": 15,
+        }
+        result = service._parse_air_quality(raw)
+
+        assert result is not None
+        assert result.aqi == 75
+        assert result.quality == "Moderate"
+        assert result.pm25 == 8.5

tests/test_health.py

@@ -80,24 +80,24 @@ class TestFullHealthCheck:
     """Test /health/full endpoint."""
 
     @patch("src.shared.database.Database.health_check")
-    @patch("src.models.ollama_client.get_ollama_client")
-    def test_full_health_returns_503_when_unhealthy(self, mock_get_ollama, mock_db_health, client):
-        """Full health should return 503 when Ollama unhealthy."""
-        mock_client = AsyncMock()
-        mock_client.health_check.return_value = False
-        mock_get_ollama.return_value = mock_client
+    def test_full_health_returns_200_when_healthy(self, mock_db_health, client):
+        """Full health should return 200 when database is healthy."""
         mock_db_health.return_value = True
 
+        response = client.get("/health/full")
+        assert response.status_code == 200
+
+    @patch("src.shared.database.Database.health_check")
+    def test_full_health_returns_503_when_unhealthy(self, mock_db_health, client):
+        """Full health should return 503 when database is unhealthy."""
+        mock_db_health.return_value = False
+
         response = client.get("/health/full")
         assert response.status_code == 503
 
     @patch("src.shared.database.Database.health_check")
-    @patch("src.models.ollama_client.get_ollama_client")
-    def test_full_health_returns_components_status(self, mock_get_ollama, mock_db_health, client):
+    def test_full_health_returns_components_status(self, mock_db_health, client):
         """Full health should return component status."""
-        mock_client = AsyncMock()
-        mock_client.health_check.return_value = False
-        mock_get_ollama.return_value = mock_client
         mock_db_health.return_value = True
 
         response = client.get("/health/full")
@@ -105,63 +105,32 @@ class TestFullHealthCheck:
 
         assert "status" in data
         assert "components" in data
-        assert "ollama" in data["components"]
         assert "database" in data["components"]
         assert "response_time_ms" in data
 
     @patch("src.shared.database.Database.health_check")
-    @patch("src.models.ollama_client.get_ollama_client")
-    def test_full_health_handles_list_models_error(self, mock_get_ollama, mock_db_health, client):
-        """Full health should handle list_models errors."""
-        mock_client = AsyncMock()
-        mock_client.health_check.return_value = True
-        mock_client.list_models.side_effect = Exception("Connection error")
-        mock_get_ollama.return_value = mock_client
-        mock_db_health.return_value = True
+    def test_full_health_handles_database_error(self, mock_db_health, client):
+        """Full health should handle database errors gracefully."""
+        mock_db_health.side_effect = Exception("Connection error")
 
         response = client.get("/health/full")
         data = response.json()
 
-        # Should report error in component status
-        assert "ollama" in data["components"]
-
-    @patch("src.shared.database.Database.health_check")
-    @patch("src.models.ollama_client.get_ollama_client")
-    def test_full_health_handles_health_check_exception(self, mock_get_ollama, mock_db_health, client):
-        """Full health should handle health check exceptions gracefully."""
-        mock_client = AsyncMock()
-        # Return False instead of raising exception to test unhealthy path
-        mock_client.health_check.return_value = False
-        mock_get_ollama.return_value = mock_client
-        mock_db_health.return_value = True
-
-        response = client.get("/health/full")
-        # Should return 503 for unhealthy
         assert response.status_code == 503
-        data = response.json()
         assert data["status"] == "unhealthy"
+        assert "error" in data["components"]["database"]
 
 
 class TestDiagnosticsEndpoint:
     """Test /health/diagnostics endpoint."""
 
-    @patch("src.models.ollama_client.get_ollama_client")
-    def test_diagnostics_returns_200(self, mock_get_ollama, client):
+    def test_diagnostics_returns_200(self, client):
         """Diagnostics should return 200."""
-        mock_client = AsyncMock()
-        mock_client.health_check.return_value = True
-        mock_get_ollama.return_value = mock_client
-
         response = client.get("/health/diagnostics")
         assert response.status_code == 200
 
-    @patch("src.models.ollama_client.get_ollama_client")
-    def test_diagnostics_returns_service_info(self, mock_get_ollama, client):
+    def test_diagnostics_returns_service_info(self, client):
         """Diagnostics should return service information."""
-        mock_client = AsyncMock()
-        mock_client.health_check.return_value = True
-        mock_get_ollama.return_value = mock_client
-
         response = client.get("/health/diagnostics")
         data = response.json()
 
@@ -169,54 +138,17 @@ class TestDiagnosticsEndpoint:
         assert "name" in data["service"]
         assert "version" in data["service"]
 
-    @patch("src.models.ollama_client.get_ollama_client")
-    def test_diagnostics_returns_components(self, mock_get_ollama, client):
-        """Diagnostics should return component details."""
-        mock_client = AsyncMock()
-        mock_client.health_check.return_value = True
-        mock_get_ollama.return_value = mock_client
-
-        response = client.get("/health/diagnostics")
-        data = response.json()
-
-        assert "components" in data
-        assert "ollama" in data["components"]
-
-    @patch("src.models.ollama_client.get_ollama_client")
-    def test_diagnostics_returns_configuration(self, mock_get_ollama, client):
+    def test_diagnostics_returns_configuration(self, client):
         """Diagnostics should return configuration info."""
-        mock_client = AsyncMock()
-        mock_client.health_check.return_value = True
-        mock_get_ollama.return_value = mock_client
-
         response = client.get("/health/diagnostics")
         data = response.json()
 
         assert "configuration" in data
 
-    @patch("src.models.ollama_client.get_ollama_client")
-    def test_diagnostics_returns_response_time(self, mock_get_ollama, client):
+    def test_diagnostics_returns_response_time(self, client):
         """Diagnostics should return response time."""
-        mock_client = AsyncMock()
-        mock_client.health_check.return_value = True
-        mock_get_ollama.return_value = mock_client
-
         response = client.get("/health/diagnostics")
         data = response.json()
 
         assert "response_time_ms" in data
         assert isinstance(data["response_time_ms"], int)
-
-    @patch("src.models.ollama_client.get_ollama_client")
-    def test_diagnostics_handles_ollama_error(self, mock_get_ollama, client):
-        """Diagnostics should handle Ollama connection errors."""
-        mock_client = AsyncMock()
-        mock_client.health_check.side_effect = Exception("Connection refused")
-        mock_get_ollama.return_value = mock_client
-
-        response = client.get("/health/diagnostics")
-        data = response.json()
-
-        # Should still return 200 with error info
-        assert response.status_code == 200
-        assert "error" in data["components"]["ollama"]

tests/test_ollama_client.py

@@ -1,307 +0,0 @@
-"""Tests for Ollama client."""
-import pytest
-from unittest.mock import patch, AsyncMock, MagicMock
-import json
-
-from src.models.ollama_client import OllamaClient, get_ollama_client, close_ollama_client
-
-
-class TestOllamaClientInit:
-    """Test OllamaClient initialization."""
-
-    @patch("src.models.ollama_client.settings")
-    def test_uses_settings_defaults(self, mock_settings):
-        """Client should use settings for defaults."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 60
-
-        client = OllamaClient()
-
-        assert client.base_url == "http://ollama:11434"
-        assert client.timeout == 60
-
-    @patch("src.models.ollama_client.settings")
-    def test_creates_http_client(self, mock_settings):
-        """Client should create httpx AsyncClient."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-
-        client = OllamaClient()
-
-        assert client.client is not None
-
-
-class TestOllamaClientClose:
-    """Test client close functionality."""
-
-    @pytest.mark.asyncio
-    @patch("src.models.ollama_client.settings")
-    async def test_close_closes_client(self, mock_settings):
-        """close should close the HTTP client."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-
-        client = OllamaClient()
-
-        with patch.object(client.client, "aclose", new_callable=AsyncMock) as mock_close:
-            await client.close()
-            mock_close.assert_called_once()
-
-
-class TestOllamaClientResolveModel:
-    """Test model resolution."""
-
-    @patch("src.models.ollama_client.settings")
-    def test_resolves_aliased_model(self, mock_settings):
-        """resolve_model should map alias to actual model."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-        mock_settings.model_aliases = {"gpt-3.5-turbo": "gemma:7b"}
-
-        client = OllamaClient()
-        result = client.resolve_model("gpt-3.5-turbo")
-
-        assert result == "gemma:7b"
-
-    @patch("src.models.ollama_client.settings")
-    def test_returns_original_if_no_alias(self, mock_settings):
-        """resolve_model should return original if no alias found."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-        mock_settings.model_aliases = {}
-
-        client = OllamaClient()
-        result = client.resolve_model("llama2")
-
-        assert result == "llama2"
-
-
-class TestOllamaClientHealthCheck:
-    """Test health check functionality."""
-
-    @pytest.mark.asyncio
-    @patch("src.models.ollama_client.settings")
-    async def test_health_check_returns_true_on_200(self, mock_settings):
-        """Health check should return True when Ollama responds 200."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-
-        client = OllamaClient()
-
-        mock_response = MagicMock()
-        mock_response.status_code = 200
-
-        with patch.object(client.client, "get", new_callable=AsyncMock) as mock_get:
-            mock_get.return_value = mock_response
-            result = await client.health_check()
-
-            assert result is True
-            mock_get.assert_called_once()
-
-    @pytest.mark.asyncio
-    @patch("src.models.ollama_client.settings")
-    async def test_health_check_returns_false_on_error(self, mock_settings):
-        """Health check should return False on connection error."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-
-        client = OllamaClient()
-
-        with patch.object(client.client, "get", new_callable=AsyncMock) as mock_get:
-            mock_get.side_effect = Exception("Connection refused")
-            result = await client.health_check()
-
-            assert result is False
-
-    @pytest.mark.asyncio
-    @patch("src.models.ollama_client.settings")
-    async def test_health_check_returns_false_on_non_200(self, mock_settings):
-        """Health check should return False on non-200 status."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-
-        client = OllamaClient()
-
-        mock_response = MagicMock()
-        mock_response.status_code = 500
-
-        with patch.object(client.client, "get", new_callable=AsyncMock) as mock_get:
-            mock_get.return_value = mock_response
-            result = await client.health_check()
-
-            assert result is False
-
-
-class TestOllamaClientListModels:
-    """Test list models functionality."""
-
-    @pytest.mark.asyncio
-    @patch("src.models.ollama_client.settings")
-    async def test_list_models_returns_dict(self, mock_settings):
-        """list_models should return dict with models."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-
-        client = OllamaClient()
-
-        models_data = {
-            "models": [
-                {"name": "llama2", "size": 1000000},
-                {"name": "gemma:7b", "size": 2000000}
-            ]
-        }
-
-        mock_response = MagicMock()
-        mock_response.status_code = 200
-        mock_response.json.return_value = models_data
-        mock_response.raise_for_status = MagicMock()
-
-        with patch.object(client.client, "get", new_callable=AsyncMock) as mock_get:
-            mock_get.return_value = mock_response
-            result = await client.list_models()
-
-            assert result == models_data
-            assert len(result["models"]) == 2
-
-    @pytest.mark.asyncio
-    @patch("src.models.ollama_client.settings")
-    async def test_list_models_raises_on_error(self, mock_settings):
-        """list_models should raise on error."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-
-        client = OllamaClient()
-
-        with patch.object(client.client, "get", new_callable=AsyncMock) as mock_get:
-            mock_get.side_effect = Exception("Connection error")
-
-            with pytest.raises(Exception):
-                await client.list_models()
-
-
-class TestOllamaClientGenerateNonStreaming:
-    """Test non-streaming generation."""
-
-    @pytest.mark.asyncio
-    @patch("src.models.ollama_client.settings")
-    async def test_generate_non_streaming_returns_response(self, mock_settings):
-        """generate_non_streaming should return response dict."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-        mock_settings.model_aliases = {}
-
-        client = OllamaClient()
-
-        response_data = {
-            "message": {"content": "Hello! How can I help?"},
-            "prompt_eval_count": 10,
-            "eval_count": 20
-        }
-
-        mock_response = MagicMock()
-        mock_response.status_code = 200
-        mock_response.json.return_value = response_data
-        mock_response.raise_for_status = MagicMock()
-
-        with patch.object(client.client, "post", new_callable=AsyncMock) as mock_post:
-            mock_post.return_value = mock_response
-            result = await client.generate_non_streaming("llama2", "Hello")
-
-            assert result["response"] == "Hello! How can I help?"
-            assert result["tokens"]["prompt"] == 10
-            assert result["tokens"]["completion"] == 20
-            assert result["tokens"]["total"] == 30
-
-    @pytest.mark.asyncio
-    @patch("src.models.ollama_client.settings")
-    async def test_generate_non_streaming_includes_max_tokens(self, mock_settings):
-        """generate_non_streaming should include max_tokens in payload."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-        mock_settings.model_aliases = {}
-
-        client = OllamaClient()
-
-        mock_response = MagicMock()
-        mock_response.status_code = 200
-        mock_response.json.return_value = {"message": {"content": "Hi"}}
-        mock_response.raise_for_status = MagicMock()
-
-        with patch.object(client.client, "post", new_callable=AsyncMock) as mock_post:
-            mock_post.return_value = mock_response
-            await client.generate_non_streaming("llama2", "Hello", max_tokens=100)
-
-            call_args = mock_post.call_args
-            assert call_args[1]["json"]["options"]["num_predict"] == 100
-
-
-class TestOllamaClientGenerateStreaming:
-    """Test streaming generation."""
-
-    @pytest.mark.asyncio
-    @patch("src.models.ollama_client.settings")
-    async def test_generate_streaming_yields_content(self, mock_settings):
-        """generate_streaming should yield content chunks."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-        mock_settings.model_aliases = {}
-
-        client = OllamaClient()
-
-        # Create mock streaming response
-        async def mock_aiter_lines():
-            yield json.dumps({"message": {"content": "Hello"}})
-            yield json.dumps({"message": {"content": " world"}})
-            yield json.dumps({"done": True})
-
-        mock_response = MagicMock()
-        mock_response.status_code = 200
-        mock_response.raise_for_status = MagicMock()
-        mock_response.aiter_lines = mock_aiter_lines
-
-        mock_stream_context = AsyncMock()
-        mock_stream_context.__aenter__.return_value = mock_response
-        mock_stream_context.__aexit__.return_value = None
-
-        with patch.object(client.client, "stream", return_value=mock_stream_context):
-            chunks = []
-            async for chunk in client.generate_streaming("llama2", "Hi"):
-                chunks.append(chunk)
-
-            assert "Hello" in chunks
-            assert " world" in chunks
-
-
-class TestOllamaClientSingleton:
-    """Test singleton pattern."""
-
-    @patch("src.models.ollama_client.settings")
-    def test_get_ollama_client_returns_same_instance(self, mock_settings):
-        """get_ollama_client should return singleton."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-
-        import src.models.ollama_client as module
-        module._ollama_client = None
-
-        client1 = get_ollama_client()
-        client2 = get_ollama_client()
-
-        assert client1 is client2
-
-    @pytest.mark.asyncio
-    @patch("src.models.ollama_client.settings")
-    async def test_close_ollama_client_clears_singleton(self, mock_settings):
-        """close_ollama_client should clear the singleton."""
-        mock_settings.ollama_base_url = "http://ollama:11434"
-        mock_settings.ollama_timeout = 30
-
-        import src.models.ollama_client as module
-        module._ollama_client = None
-
-        client = get_ollama_client()
-
-        with patch.object(client.client, "aclose", new_callable=AsyncMock):
-            await close_ollama_client()
-
-        assert module._ollama_client is None