ollamonym/docker-compose.yml at main · H-Ismael/ollamonym · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
services:
  # FastAPI service for anonymization
  api:
    build: .
    container_name: pii-anonymizer-api
    ports:
      - "8000:8000"
    environment:
      - TEMPLATES_DIR=/app/templates
      - CUSTOM_TEMPLATES_DIR=/app/data/templates
      - PSEUDONYM_SECRET=${PSEUDONYM_SECRET:-your-secret-key-change-me}
      - DEFAULT_TEMPLATE_ID=default-pii-v1
      - OLLAMA_BASE_URL=http://ollama:11434
      - OLLAMA_FALLBACK_URLS=http://host.docker.internal:11434
      - LLM_MODEL=llama3.1:8b-instruct-q4_K_M
      - LLM_CONCURRENCY=2
      - LLM_NUM_PREDICT=512
      - LLM_TEMPERATURE=0
      - OLLAMA_KEEP_ALIVE=30m
      - CHUNKING_ENABLED=true
      - CHUNK_CHAR_TARGET=8000
      - CHUNK_OVERLAP_CHARS=200
      - CHUNK_MAX_PARALLEL=2
      - RULE_PREEXTRACT_ENABLED=true
      - LOG_SENSITIVE=false
      - TOKEN_ID_LEN=6
    volumes:
      - ./templates:/app/templates:ro
      - ./data/templates:/app/data/templates
    extra_hosts:
      - "host.docker.internal:host-gateway"
    depends_on:
      ollama:
        condition: service_healthy
    networks:
      - anonymizer
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 10s
      timeout: 3s
      retries: 5
      start_period: 30s

  # Ollama service for LLM inference
  ollama:
    image: ollama/ollama:latest
    container_name: pii-anonymizer-ollama
    ports:
      - "11434:11434"
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
    environment:
      - OLLAMA_NUM_GPU=1
      - OLLAMA_MAX_LOADED_MODELS=1
      - OLLAMA_NUM_PARALLEL=2
      - NVIDIA_VISIBLE_DEVICES=all
      - NVIDIA_DRIVER_CAPABILITIES=compute,utility
    volumes:
      - ollama_models:/root/.ollama
    networks:
      - anonymizer
    healthcheck:
      test: ["CMD", "ollama", "list"]
      interval: 10s
      timeout: 3s
      retries: 5
      start_period: 30s
    # Optional: pre-pull manually once:
    # docker exec pii-anonymizer-ollama ollama pull llama3.1:8b-instruct-q4_K_M
    # docker exec pii-anonymizer-ollama ollama pull qwen2.5:7b-instruct-q4_K_M
    # docker exec pii-anonymizer-ollama ollama pull qwen2.5:14b-instruct-q4_K_M

networks:
  anonymizer:
    driver: bridge

volumes:
  ollama_models:
    driver: local