ohmu/third/server/openwebui/docker-compose.yml.bak-embed-20260913-190443

version: "3.9"
# Tailscale ühendus:
# sudo tailscale serve --bg --https 3000 http://127.0.0.1:3000
# sudo tailscale serve --https=3000 off
# sudo tailscale serve status

# ═══════════════════════════════════════════════════════════════
# OPEN WEBUI - Professionaalne LLM liides
# ═══════════════════════════════════════════════════════════════
services:
  open-webui:
    image: ghcr.io/open-webui/open-webui:main
    container_name: open-webui
    mem_limit: 8g
    restart: always
    ports:
      # Kui kasutad tailscale'i siis
      #- "127.0.0.1:3000:8080"
      - "100.87.1.24:3010:8080"   # ainult tailnet; avalikult läbi prod (openwebui.odamus.com)
    #env_file:
      #- /data/docker/open-webui/.env
    volumes:
      - data:/app/backend/data
      - /data/docker/open-webui/scripts:/scripts
      - /data/docker/open-webui/patches/qdrant_multitenancy.py:/app/backend/open_webui/retrieval/vector/dbs/qdrant_multitenancy.py
      #- /data/docker/open-webui/patches/searxng.py:/app/backend/open_webui/retrieval/web/searxng.py
      - /backups/docker/open-webui:/backups
      - /etc/localtime:/etc/localtime
      #- /data/docker/open-webui/.env:/app/backend/open_webui/.env
      #- /data/docker/open-webui/patches:/patches:ro
    #entrypoint: ["/bin/sh", "-c", "python /patches/patch_ollama_embeddings.py; python /patches/patch_qdrant.py && exec /app/backend/start.sh"]
    environment:
       # --- PÕHISED ---
      - WEBUI_SECRET_KEY=ai_webui_secret_2026_change_me_to_random_string
      - TZ=Europe/Tallinn
      - LANG=C.UTF-8
      - CHAT_STREAM_RESPONSE_CHUNK_MAX_BUFFER_SIZE=52428800
      - ENABLE_SIGNUP=false   # esimesel korral true, hiljem false
      - OPENAI_API_BASE_URLS=http://100.87.1.24:8070/v1
      - OPENAI_API_KEYS=sk-dummy-key-not-required

      - OLLAMA_API_BASE_URL=http://100.87.1.24:11434
      - RAG_EMBEDDING_OPENAI_API_BASE_URL=http://100.87.1.24:11434

      # --- RAG ja EMBEDDINGUD ---
      - OLLAMA_BASE_URL=http://100.87.1.24:11434
      - RAG_EMBEDDING_ENGINE=ollama
      - RAG_EMBEDDING_MODEL=nomic-embed-text
      - USE_OLLAMA_DOCKER=false
      # Chunking (märgid, mitte sõnad)
      #- RAG_TEXT_SPLITTER=recursive
      #- RAG_CHUNK_SIZE=1000
      #- RAG_CHUNK_OVERLAP=100
      #- RAG_CHUNK_MIN_SIZE=100
      # Hybrid search + reranker
      #- RAG_HYBRID_SEARCH_ENABLED=true
      #- RAG_TOP_K=20
      # Full context režiim maha (kasuta fokuseeritud RAG-i)
      - RAG_FULL_CONTEXT_MODE=false

      # TEKSTI TÜKELDAMINE – need on kõige olulisemad suurte failide jaoks
      - TEXT_SPLITTER=recursive
      # kasuta rekursiivset tükeldajat (toetab keelepiire)
      - CHUNK_SIZE=1500
      # vähenda 500 tähemärgile, et vältida kontekstiületust
      - CHUNK_OVERLAP=150
      # kattuvus tükkide vahel
      - CHUNK_MIN_SIZE_TARGET=300
      # kui vaja, aga pole kohustuslik
      - HYBRID_SEARCH_ENABLED=true
      - TOP_K=20

      - RAG_WEB_SEARCH_RESULT_COUNT=3
      - WEB_LOADER_ENGINE=default
      - CONCURRENT_REQUESTS=20   # kui pole juba

      # Failide üleslaadimise piirangud
      - UPLOAD_MAX_SIZE=6291456
      # 6 MB baitides
      - ALLOWED_FILE_EXTENSIONS=md,pdf,txt,docx
      # luba ka markdown

      # --- QDRANT --- (AINULT NEED!)
      - VECTOR_DB=qdrant
      - QDRANT_URI=http://qdrant:6333
      #- VECTOR_DB=postgres
      #- VECTOR_DB=weaviate
      #- WEAVIATE_HOST=http://weaviate:8080

      # --- SEARXNG (veebiotsing) ---
      - ENABLE_RAG_WEB_SEARCH=true
      - BYPASS_RETRIEVAL_ACCESS_CONTROL=true
      - BYPASS_WEB_SEARCH_EMBEDDING_AND_RETRIEVAL=true
      - RAG_WEB_SEARCH_ENGINE=searxng
      - SEARXNG_QUERY_URL=http://100.87.1.24:8888/search?q=<query>&format=json&categories=general,news,science
      #- SEARXNG_QUERY_URL=http://searxng:8080/search?q=<query>&format=json&categories=general,news,science
      #- SEARXNG_QUERY_URL=http://localhost:8080/search?q=<query>&format=json&categories=general,news,science
      - DATABASE_URL=postgresql://ai_user:ai_secure_pass_2026@postgres:5432/openwebui_test_db

      # --- MCP TOOL SERVER ÜHENDUSED ---
      - 'TOOL_SERVER_CONNECTIONS=[{"url":"http://mcp-filesystem-sse:3100/mcp","path":"","type":"mcp","auth_type":"none","key":"","headers":{},"config":{"enable":true,"access_grants":[]},"info":{"id":"filesystem","name":"Filesystem","description":"YOU CAN AND MUST USE THIS TOOL to create directories, write files, and read files on the server. Creates parent directories automatically. Use write_file to save code files. Allowed paths: /projects/data, /data, /scripts, /backups"}},{"url":"http://mcp-terminal:3110/mcp","path":"","type":"mcp","auth_type":"none","key":"","headers":{},"config":{"enable":true,"access_grants":[]},"info":{"id":"terminal","name":"Terminal","description":"YOU CAN AND MUST USE THIS TOOL to execute shell commands (mkdir, npm, npx, node, ng, python3, ls, cat). Also provides write_file and create_directory tools for saving files. This has WORKING WRITE ACCESS to /projects/data"}}]'
    networks:
      - ai_network
      - prod_net
    ulimits:
      nofile:
        soft: 65536
        hard: 65536
    extra_hosts:
      - "host.docker.internal:host-gateway"
    healthcheck:
      test: ["CMD-SHELL", "curl -f http://localhost:8080/health || exit 1"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 60s
  mcp-terminal:
    build: ./mcp-servers
    container_name: mcp-terminal
    entrypoint: ["/entrypoint-terminal.sh"]
    restart: unless-stopped
    ports:
      - "127.0.0.1:3110:3110"
    environment:
      - TZ=Europe/Tallinn
      - LANG=C.UTF-8
    volumes:
      - /data/mcp-filesystem:/projects/data
      - /etc/localtime:/etc/localtime
      - /var/run/docker.sock:/var/run/docker.sock
    networks:
      - ai_network

  tika:
    image: apache/tika:latest-full     # "full" sisaldab Tesseract OCR-i (soovitatav)
    container_name: tika
    ports:
      - "9998:9998"                     # Vajadusel väljapoole eksponeerimiseks
    restart: unless-stopped
    environment:
      - TZ=Europe/Tallinn
      - LANG=C.UTF-8
    volumes:
      - /etc/localtime:/etc/localtime
    networks:
      - ai_network

volumes:
  data:
    name: 'open-webui_data'
    driver: local

networks:
  ai_network:
    external: true
    name: main_ai_network
  prod_net:
    external: true