kincses/service-finder
1
0
Fork 0
Code Issues 33 Pull Requests Packages Projects 2 Releases Wiki Activity
Files
ba8b6579ef99b239224684bb49d00bb4428083e6
service-finder/.roo/history_gemini.md

9.4 KiB
Raw Blame History

🧠 Roo Context File - Service Finder Master Book 2.0.1

Generated: 2026-03-26
Auditor: Projekt Manager Gemini
Purpose: System reality snapshot for Roo AI memory & future development context

📊 CURRENT SYSTEM REALITY (70+ DB Tables, Multi-Schema)

🗄️ Database Schema Overview (19 Schemas, 127+ Tables)

Schema Table Count Status
audit 5 Active (Audit logs)
data 0 Not in list (maybe missing)
finance 6 Active (Triple Wallet, Ledger)
fleet 6 Active (Fleet management)
gamification 11 Active (XP, Badges, Leaderboard)
identity 7 Active (Person/User dual model)
marketplace 12 Active (Services, Providers)
public 4 System tables
system 15 Active (Parameters, Translations)
tiger 34 PostGIS extension tables
topology 2 PostGIS extension tables
vehicle 25 Active (Assets, Definitions, History)
TOTAL 127+ Established multi-schema architecture

✅ 100% WORKING COMPONENTS (Verified)

  1. Authentication & Token Handling - JWT with dual entity (Person/User)
  2. Basic Routing - FastAPI operational on port 8000
  3. Database Connection Pool - Async SQLAlchemy 2.0+ with PostgreSQL
  4. Docker Stack - 30+ containers running (API, Frontend, DB, Redis, MinIO, etc.)
  5. Robot Ecosystem - 10+ specialized workers active (Discovery, Hunter, Enricher, Validator, Auditor)
  6. Multi-Schema Isolation - DDD principles enforced (identity, finance, vehicle domains separated)

🔄 ESTABLISHED WORKFLOWS

  • Vehicle Creation: 2-step Draft → Active process (frontend calls /api/v1/vehicles/register, backend uses /api/v1/assets/vehicles)
  • MDM Deduplication: Merge only when make, technical_code, and engine_capacity match
  • Triple Wallet Economy: Local/EUR/Token balances with audit trail
  • Robot Quota Management: .quota_dvla.json tracking for API rate limits

⚠️ KNOWN GAPS & BROKEN PIPES

🔴 CRITICAL (Blocking Production)

  1. API Endpoint Mismatch

    • Frontend AddVehicle.vue calls /api/v1/vehicles/register (404 Not Found)
    • Correct endpoint is /api/v1/assets/vehicles (requires authentication)
    • Impact: Vehicle creation fails for users

  2. Missing 2-Step Vehicle Creation Backend Logic

    • Documented 2-step process (Draft → Active) not fully implemented
    • AssetService.create_or_claim_vehicle() handles VIN conflicts but draft logic unclear

  3. Authentication Endpoint 404

    • /api/v1/auth/me returns 404 (should be /api/v1/users/me or similar)
    • Impact: Frontend cannot validate user session

🟡 HIGH PRIORITY (Functional but Incomplete)

  1. Mocked/Broken Catalog Endpoints

    • /api/v1/catalog/brands - 404 (likely moved to /api/v1/vehicles/search/brands)
    • /api/v1/vehicles/search/brands - 404 (needs implementation)

  2. Frontend/Backend API Version Drift

    • Frontend uses hardcoded IP 192.168.100.43:8000 instead of env variable
    • Multiple API calls expecting different response formats

  3. Missing Historical Data (occurrence_date)

    • Epic requirement: "Historical Data (múltbéli költségek, szervizek) bevezetése"
    • occurrence_date field not consistently implemented across cost tables

🟢 MEDIUM PRIORITY (Architectural Debt)

  1. Inconsistent Error Handling

    • Some endpoints return plain text errors, others JSON
    • No standardized error schema

  2. Missing AnalyticsService (TCO/km)

    • Flotta Analytics (Total Cost of Ownership per km) not implemented
    • Required for fleet manager dashboards

  3. Robot-0-GB Discovery CSV Path

    • Robot expects /mnt/nas/app_data/uk_mot_data.csv but file existence not verified

🏛️ ARCHITECTURAL RULES (Must Preserve)

🚫 STRICT PROHIBITIONS

  1. No Yellow Text on White Backgrounds - Accessibility violation
  2. Never Hardcode API Keys - Use config.py + .env only
  3. No Direct Database Drops - Use Alembic migrations only
  4. No Sync Blocking Calls - All I/O must be async in FastAPI endpoints
  5. No Schema Mixing - Finance data stays in finance schema, vehicle in vehicle, etc.

✅ MANDATORY PATTERNS

  1. Vehicle Creation is 2-Step

    • Step 1: Draft (VIN optional, basic info)
    • Step 2: Active (technical enrichment, digital twin creation)
    • XP reward only after Step 2 completion

  2. Deduplication Logic

    • Merge ONLY when make, technical_code, AND engine_capacity match
    • Generate N/A and UNKNOWN fallback codes for SQL constraint compatibility

  3. Robot Quota Enforcement

    • All external API calls must respect DVLA_DAILY_LIMIT from .env
    • Log usage in .quota_dvla.json with timestamp

  4. Triple Wallet Transactions

    • Every financial movement must create audit trail in finance.ledger
    • Balance checks before deductions

🔧 TECHNICAL CONSTRAINTS

  1. Docker Compose V2 Only - Use docker compose (space), not docker-compose
  2. Roo-Helper Container for Scripts - All Python operations via docker exec roo-helper
  3. Test Database Isolation - Unit tests must use service_finder_test or SQLite in-memory
  4. Logging Standard - Use logging.getLogger(__name__), no print() in production

📈 SYSTEM HEALTH ASSESSMENT (60% → 70% Complete)

✅ STRENGTHS

  • Robust Database Foundation: 127+ tables across 19 schemas, well-normalized
  • Active Robot Fleet: 10+ specialized workers running continuously
  • Containerized Infrastructure: Full Docker stack with monitoring
  • Domain-Driven Design: Clear separation of concerns (identity, finance, vehicle, marketplace)

⚠️ WEAKNESSES

  • Frontend/Backend Integration: API mismatches causing 404 errors
  • Documentation/Code Drift: Master Book 2.0 docs don't match actual endpoints
  • Incomplete User Flows: Vehicle creation, authentication need fixing
  • Limited Testing: Unit test coverage unknown, integration tests sparse

🎯 IMMEDIATE NEXT STEPS (PHASE 2)

  1. Fix API Endpoint Mismatches - Align frontend calls with backend routes
  2. Implement 2-Step Vehicle Creation - Complete draft/active workflow
  3. Add Historical Data Fields - occurrence_date across cost tables
  4. Build AnalyticsService - TCO/km calculations for fleet managers
  5. Create Integration Test Suite - Verify end-to-end user journeys

🔗 KEY DEPENDENCIES & RISKS

🔄 INTERNAL DEPENDENCIES

  • PostgreSQL 15+ with PostGIS extension (spatial data)
  • Redis for caching and session management
  • MinIO for document/evidence storage
  • Ollama (local AI) + Gemini/Groq (fallback) for OCR/AI

🌐 EXTERNAL DEPENDENCIES

  • DVLA VES API (UK vehicle data) - Rate limited, requires API key
  • RDW API (Dutch vehicle data) - Public but rate limited
  • OpenStreetMap (service location data) - No API key required
  • SendGrid (email) - API key required

⚠️ RISK FACTORS

  • Single Point of Failure: Shared PostgreSQL instance
  • API Rate Limits: DVLA (1000/day), RDW (unknown)
  • Data Volume: UK MOT CSV (~10M records) requires efficient processing
  • Complexity: 30+ Docker containers increase orchestration complexity

📝 AUDIT METHODOLOGY

  1. Codebase Analysis: Read backend endpoints, frontend views/stores
  2. Database Inspection: Schema enumeration via PostgreSQL queries
  3. API Testing: Direct endpoint calls from roo-helper container
  4. Documentation Comparison: Master Book 2.0 vs actual implementation
  5. Rule Extraction: From .roo/rules/ and code patterns

NEXT ACTION: Create Gitea issues for each identified gap, prioritize by criticality, begin implementation with Fast Coder mode.

🏗️ RULE ARCHITECTURE OPTIMIZATION (2026-03-27)

Multi-Mode Schema Consolidation

  • Updated 00-global.md: Added mandatory directives: Docker Compose V2, color scheme (#1e3a8a), DB verification with sync_engine.py, ticket verification with gitea_manager.py, mandatory 2-step vehicle flow (Draft → Active).
  • Merged 02-architecture.md into architect.md: Enhanced architect rules with DDD, schema separation, project directory map, and SQL error handling.
  • Merged 04-debug-protocol.md into fast-coder.md: Added debug protocol steps and rapid API wiring guidelines (Pydantic validation, frontend integration).
  • Path verification: Updated all references to use docker compose exec roo-helper (consistent with Docker Compose V2).

Environment Cleanup Plan

Identified redundant test folders and orphaned files for archiving (rename to .old and move to archive):

  • backend/app/test_outside/ → archive/test_outside_old/
  • backend/app/tests_internal/ → archive/tests_internal_old/
  • Various .bak and .old files scattered across the codebase (to be moved to archive).

Mode Boundary Clarification

Each mode now has clear boundaries to prevent hallucination overlap:

  • Architect: Focus on DDD, schema design, system integrity, and Kanban management.
  • Fast Coder: Focus on surgical coding, API wiring, Pydantic validation, and frontend integration.
  • Auditor/Debugger: Focus on verification, logging, and systematic debugging.

Next Steps

  • Execute archiving of identified redundant files.
  • Validate that all rule paths are correct and functional.
  • Ensure each mode's custom instructions reflect the updated rule sets.

This file will be updated after each major system change. Maintain as single source of truth for Roo AI context.

Reference in New Issue View Git Blame Copy Permalink