Merge pull request #18 from vectornguyen76/develop

Develop triton inference server

.flake8

@@ -1,5 +1,5 @@
 [flake8]
-    ignore = E203, E266, E501, W503, F403, F401
+    ignore = F401
     max-line-length = 100
     max-complexity = 18
     select = B,C,E,F,W,T4,B9

.github/workflows/development_pipeline.yml

@@ -39,7 +39,7 @@ jobs:
         run: black --check ./image-search-engine/.
 
       - name: Run flake8
-        run: flake8 --ignore=E501,W503 ./image-search-engine
+        run: flake8 --ignore=E501,W503,F401 ./image-search-engine
 
       # - name: Run Pylint
       #   run: pylint ./image-search-engine/*.py
@@ -95,7 +95,7 @@ jobs:
         run: black --check ./text-search-engine/.
 
       - name: Run flake8
-        run: flake8 --ignore=E501,W503 ./text-search-engine
+        run: flake8 --ignore=E501,W503,F401 ./text-search-engine
 
       # - name: Run Pylint
       #   run: pylint ./image-search-engine/*.py

.gitignore

@@ -15,6 +15,9 @@ image-search-engine/assets/uploaded_images/*
 
 # Model
 *.pth
+*.pt
+*.onnx
+*.engine
 
 # Byte-compiled / optimized / DLL files
 __pycache__/

docker-compose-prod.yaml

@@ -0,0 +1,142 @@
+version: "3"
+services:
+  triton-server:
+    container_name: triton-server
+    image: nvcr.io/nvidia/tritonserver:23.09-py3
+    ports:
+      - 9000:8000
+      - 9001:8001
+      - 9002:8002
+    command: tritonserver --model-repository=/models
+    volumes:
+      - ./image-search-engine/model_repository:/models
+    deploy:
+      resources:
+        reservations:
+          devices:
+            - driver: nvidia
+              count: 1
+              capabilities: [gpu]
+
+  qdrant-vector-database:
+    container_name: qdrant-vector-database
+    image: qdrant/qdrant:v1.5.1
+    ports:
+      - 6333:6333
+      - 6334:6334
+    volumes:
+      - ./qdrant-vector-database:/qdrant/storage
+
+  image-search-engine:
+    container_name: image-search-container
+    image: vectornguyen76/image-search-engine
+    build:
+      context: ./image-search-engine
+      dockerfile: Dockerfile
+    environment:
+      - QDRANT_URL=http://qdrant-vector-database:6334
+      - TRITON_SERVER_URL=triton-server:8001
+    ports:
+      - 7000:7000
+    volumes:
+      - ./image-search-engine/logs:/app/logs
+    depends_on:
+      - qdrant-vector-database
+    profiles:
+      - dev.frontend
+      - prod
+
+  elasticsearch:
+    image: docker.elastic.co/elasticsearch/elasticsearch:7.10.0
+    container_name: elasticsearch
+    environment:
+      - node.name=elasticsearch
+      - cluster.name=es-docker-cluster
+      - discovery.type=single-node
+      - bootstrap.memory_lock=true
+      - "ES_JAVA_OPTS=-Xms1024m -Xmx1024m"
+    ulimits:
+      memlock:
+        soft: -1
+        hard: -1
+    ports:
+      - 9200:9200
+    healthcheck:
+      test: curl --fail http://localhost:9200/_cat/health || exit 1
+      interval: 10s
+      timeout: 1s
+      retries: 10
+    volumes:
+      - data-elastic-search:/usr/share/elasticsearch/data
+
+  kibana:
+    image: docker.elastic.co/kibana/kibana:7.10.0
+    container_name: kibana
+    ports:
+      - 5601:5601
+    environment:
+      ELASTICSEARCH_URL: http://elasticsearch:9200
+      ELASTICSEARCH_HOSTS: http://elasticsearch:9200
+    depends_on:
+      elasticsearch:
+        condition: service_healthy
+
+  text-search-engine:
+    container_name: text-search-container
+    image: vectornguyen76/text-search-engine
+    build:
+      context: ./text-search-engine
+      dockerfile: Dockerfile
+    ports:
+      - 8000:8000
+    environment:
+      ELASTICSEARCH_HOST: http://elasticsearch:9200
+    depends_on:
+      elasticsearch:
+        condition: service_healthy
+    profiles:
+      - dev.frontend
+      - prod
+
+  frontend_dev_service:
+    container_name: frontend_dev_container
+    image: vectornguyen76/frontend_dev_image
+    build:
+      context: ./frontend
+      dockerfile: Dockerfile.dev
+    ports:
+      - 3000:3000
+    profiles:
+      - dev.backend
+
+  frontend_service:
+    container_name: frontend_prod_container
+    image: vectornguyen76/frontend_prod_image
+    build:
+      context: ./frontend
+      dockerfile: Dockerfile
+    ports:
+      - 3000:3000
+    profiles:
+      - prod
+
+  nginx_service:
+    container_name: nginx_container
+    image: nginx:1.25.1-alpine
+    ports:
+      - 80:80
+    volumes:
+      - ./nginx-server/default.conf:/etc/nginx/conf.d/default.conf
+      - ./nginx-server/log:/var/log/nginx/
+    depends_on:
+      - frontend_service
+      - image-search-engine
+      - text-search-engine
+    profiles:
+      - prod
+
+volumes:
+  data-elastic-search:
+    driver: local
+  qdrant-vector-database:
+    driver: local

docker-compose.yaml

@@ -1,55 +1,22 @@
 version: "3"
 services:
-  triton:
-    image: nvcr.io/nvidia/tritonserver:22.02-py3
-    ipc: host
+  triton-server:
+    container_name: triton-server
+    image: nvcr.io/nvidia/tritonserver:23.09-py3
     ports:
       - 9000:8000
       - 9001:8001
       - 9002:8002
     command: tritonserver --model-repository=/models
     volumes:
       - ./image-search-engine/model_repository:/models
-    # deploy:
-    #   resources:
-    #     reservations:
-    #       devices:
-    #         - capabilities: [gpu]
-
-  elasticsearch:
-    image: docker.elastic.co/elasticsearch/elasticsearch:7.10.0
-    container_name: elasticsearch
-    environment:
-      - node.name=elasticsearch
-      - cluster.name=es-docker-cluster
-      - discovery.type=single-node
-      - bootstrap.memory_lock=true
-      - "ES_JAVA_OPTS=-Xms1024m -Xmx1024m"
-    ulimits:
-      memlock:
-        soft: -1
-        hard: -1
-    ports:
-      - 9200:9200
-    healthcheck:
-      test: curl --fail http://localhost:9200/_cat/health || exit 1
-      interval: 10s
-      timeout: 1s
-      retries: 10
-    volumes:
-      - data-elastic-search:/usr/share/elasticsearch/data
-
-  kibana:
-    image: docker.elastic.co/kibana/kibana:7.10.0
-    container_name: kibana
-    ports:
-      - 5601:5601
-    environment:
-      ELASTICSEARCH_URL: http://elasticsearch:9200
-      ELASTICSEARCH_HOSTS: http://elasticsearch:9200
-    depends_on:
-      elasticsearch:
-        condition: service_healthy
+    deploy:
+      resources:
+        reservations:
+          devices:
+            - driver: nvidia
+              count: 1
+              capabilities: [gpu]
 
   qdrant-vector-database:
     container_name: qdrant-vector-database
@@ -67,73 +34,38 @@ services:
       context: ./image-search-engine
       dockerfile: Dockerfile
     environment:
-      - QDRANT_HOST=qdrant-vector-database
+      - QDRANT_URL=http://qdrant-vector-database:6334
+      - TRITON_SERVER_URL=triton-server:8001
     ports:
       - 7000:7000
     volumes:
       - ./image-search-engine/logs:/app/logs
     depends_on:
       - qdrant-vector-database
-    profiles:
-      - dev.frontend
-      - prod
 
-  text-search-engine:
-    container_name: text-search-container
-    image: vectornguyen76/text-search-engine
-    build:
-      context: ./text-search-engine
-      dockerfile: Dockerfile
-    ports:
-      - 8000:8000
-    environment:
-      ELASTICSEARCH_HOST: http://elasticsearch:9200
-    depends_on:
-      elasticsearch:
-        condition: service_healthy
-    profiles:
-      - dev.frontend
-      - prod
-
-  frontend_dev_service:
-    container_name: frontend_dev_container
-    image: vectornguyen76/frontend_dev_image
-    build:
-      context: ./frontend
-      dockerfile: Dockerfile.dev
-    ports:
-      - 3000:3000
-    profiles:
-      - dev.backend
-
-  frontend_service:
-    container_name: frontend_prod_container
-    image: vectornguyen76/frontend_prod_image
-    build:
-      context: ./frontend
-      dockerfile: Dockerfile
-    ports:
-      - 3000:3000
-    profiles:
-      - prod
-
-  nginx_service:
-    container_name: nginx_container
-    image: nginx:1.25.1-alpine
-    ports:
-      - 80:80
-    volumes:
-      - ./nginx-server/default.conf:/etc/nginx/conf.d/default.conf
-      - ./nginx-server/log:/var/log/nginx/
-    depends_on:
-      - frontend_service
-      - image-search-engine
-      - text-search-engine
-    profiles:
-      - prod
+  # image-search-engine-gpu:
+  #   container_name: image-search-container-gpu
+  #   image: vectornguyen76/image-search-engine-gpu
+  #   build:
+  #     context: ./image-search-engine
+  #     dockerfile: Dockerfile.gpu
+  #   environment:
+  #     - QDRANT_URL=http://qdrant-vector-database:6334
+  #     - TRITON_SERVER_URL=triton-server:8001
+  #   ports:
+  #     - 7000:7000
+  #   volumes:
+  #     - ./image-search-engine/logs:/app/logs
+  #   depends_on:
+  #     - qdrant-vector-database
+  #   deploy:
+  #     resources:
+  #       reservations:
+  #         devices:
+  #           - driver: nvidia
+  #             count: 1
+  #             capabilities: [gpu]
 
 volumes:
-  data-elastic-search:
-    driver: local
   qdrant-vector-database:
     driver: local

image-search-engine/.dockerignore

@@ -1 +1,4 @@
+data
+locust
 faiss_store
+model_repository

image-search-engine/Dockerfile

@@ -4,6 +4,9 @@ FROM python:3.9-slim
 # Set the working directory inside the container to /app
 WORKDIR /app
 
+# Install for OpenCV
+RUN apt-get update && apt-get install ffmpeg libsm6 libxext6  -y
+
 # Copy the requirements.txt file from the host into the container's /app directory
 COPY requirements.txt /app
 
@@ -31,9 +34,6 @@ RUN python -c 'from torchvision.models import efficientnet_b3, EfficientNet_B3_W
 # Copy all files from the host into the container's /app directory
 COPY . /app
 
-# Ingest data to Faiss
-RUN python faiss_ingest.py
-
 # Expose port 7000 to the host machine
 EXPOSE 7000