Merge pull request #1157 from shiyu22/image-search-src

Update image search with latest Milvus and Towhee

README.md

@@ -97,7 +97,7 @@ And more solutions you can refer to the [**Examples**](https://github.com/towhee
      </td>
       <td>
         <p>- <a href="https://github.com/towhee-io/examples/tree/main/image/reverse_image_search">Jupyter notebook</a></p>
-        <p>- <a href="solutions/image/reverse_image_search/quick_deploy">Quick deploy</a></p>
+        <p>- <a href="solutions/image/reverse_image_search">Quick deploy</a></p>
      </td>
       <td>
         <p>- <a href="https://zhuanlan.zhihu.com/p/517360724">10 行代码搞定以图搜图</a></p>
@@ -125,7 +125,7 @@ And more solutions you can refer to the [**Examples**](https://github.com/towhee
      </td>
       <td>
         <p>-<a href="https://github.com/towhee-io/examples/tree/main/nlp/question_answering"> Jupyter notebook</a></p>
-        <p>-<a href="solutions/nlp/question_answering_system/quick_deploy"> Quick deploy</a></p>
+        <p>-<a href="solutions/nlp/question_answering_system"> Quick deploy</a></p>
      </td>
       <td>
           <p>-<a href="https://mp.weixin.qq.com/s/BZp4CMv2yuVb0oEyuDKNkw">快速搭建对话机器人，就用这一招！</a></p>
@@ -238,7 +238,6 @@ And more solutions you can refer to the [**Examples**](https://github.com/towhee
       <td>- <a href="https://www.bilibili.com/video/BV1dD4y1D7zS">中文</a></td>
    </tr>
 </table>
-
 ### :clapper: Live Demo
 
 We have built [online demos](https://milvus.io/milvus-demos/) for reverse image search, chatbot and molecular search that everyone can have fun with.

solutions/image/reverse_image_search/README.md

@@ -4,7 +4,7 @@ This demo uses [towhee](https://github.com/towhee-io/towhee) image embedding ope
 
 The system architecture is as below:
 
-<img src="pic/demo.jpg" width = "450" height = "600" alt="system_arch" align=center />
+<img src="pic/workflow.png" height = "500" alt="arch" align=center />
 
 
 ## Data Source
@@ -21,7 +21,8 @@ Download location: https://drive.google.com/file/d/1n_370-5Stk4t0uDV1QqvYkcvyV8r
 
 ### Requirements
 
-- [Milvus 2.0](https://milvus.io/docs/v2.0.0/install_standalone-docker.md)
+- [Milvus 2.2.10](https://milvus.io/)
+- [Towhee](https://towhee.io/)
 - [MySQL](https://hub.docker.com/r/mysql/mysql-server)
 - [Python3](https://www.python.org/downloads/)
 - [Docker](https://docs.docker.com/engine/install/)
@@ -33,11 +34,10 @@ The reverse image search system requires Milvus, MySQL, WebServer and WebClient
 
 - Modify docker-compose.yaml to map your data directory to the docker container of WebServer
 ```bash
-$ git clone https://github.com/milvus-io/bootcamp.git
-$ cd solutions/image/reverse_image_search/quick_deploy
+$ wget https://raw.githubusercontent.com/milvus-io/bootcamp/master/solutions/image/reverse_image_search/docker-compose.yaml
 $ vim docker-compose.yaml
 ```
-> Change line 73: `./data:/data` --> `your_data_path:/data`
+**Then to change line 75:** `./data:/data` --> `your_data_path:/data`
 
 - Create containers & start servers with docker-compose.yaml
 ```bash
@@ -59,28 +59,33 @@ Creating img-search-webserver ... done
 And show all containers with `docker ps`, and you can use `docker logs img-search-webserver` to get the logs of **server** container.
 
 ```bash
-CONTAINER ID   IMAGE                                         COMMAND                  CREATED              STATUS                             PORTS                               NAMES
-25b4c8e13590   milvusbootcamp/img-search-server:towhee       "/bin/sh -c 'python3…"   59 seconds ago       Up 49 seconds                      0.0.0.0:5000->5000/tcp              img-search-webserver
-ae9a9a783952   milvusdb/milvus:v2.0.0-rc8-20211104-d1f4106   "/tini -- milvus run…"   59 seconds ago       Up 58 seconds                      0.0.0.0:19530->19530/tcp            milvus-standalone
-7e88bdf66d96   minio/minio:RELEASE.2020-12-03T00-03-10Z      "/usr/bin/docker-ent…"   About a minute ago   Up 59 seconds (healthy)            9000/tcp                            milvus-minio
-4a3ea5fff0f9   mysql:5.7                                     "docker-entrypoint.s…"   About a minute ago   Up 59 seconds                      0.0.0.0:3306->3306/tcp, 33060/tcp   img-search-mysql
-f3c7440d5dc4   milvusbootcamp/img-search-client:1.0          "/bin/bash -c '/usr/…"   About a minute ago   Up 59 seconds (health: starting)   0.0.0.0:8001->80/tcp                img-search-webclient
-cc6b473d905d   quay.io/coreos/etcd:v3.5.0                    "etcd -advertise-cli…"   About a minute ago   Up 59 seconds                      2379-2380/tcp                       milvus-etcd
+CONTAINER ID   IMAGE                                      COMMAND                  CREATED          STATUS                      PORTS                                                                                      NAMES
+f24ddb15a529   milvusbootcamp/img-search-server:2.2.10    "/bin/sh -c 'python3…"   35 minutes ago   Up 34 minutes               0.0.0.0:5000->5000/tcp, :::5000->5000/tcp                                                  img-search-webserver
+d26b5cf21822   milvusdb/milvus:v2.2.10                    "/tini -- milvus run…"   35 minutes ago   Up 35 minutes               0.0.0.0:9091->9091/tcp, :::9091->9091/tcp, 0.0.0.0:19530->19530/tcp, :::19530->19530/tcp   milvus-standalone
+1e9254f56006   minio/minio:RELEASE.2023-03-20T20-16-18Z   "/usr/bin/docker-ent…"   35 minutes ago   Up 35 minutes (healthy)     9000/tcp                                                                                   milvus-minio
+445cef3fe474   milvusbootcamp/img-search-client:2.2.10    "/docker-entrypoint.…"   35 minutes ago   Up 35 minutes (unhealthy)   0.0.0.0:8001->80/tcp, :::8001->80/tcp                                                      img-search-webclient
+d9bbab39f325   mysql:5.7                                  "docker-entrypoint.s…"   35 minutes ago   Up 35 minutes               33060/tcp, 0.0.0.0:3306->3306/tcp, :::3306->3306/tcp                                       img-search-mysql
+807c2ace0b28   quay.io/coreos/etcd:v3.5.5                 "etcd -advertise-cli…"   35 minutes ago   Up 35 minutes               2379-2380/tcp                                                                              milvus-etcd
 ```
 
 ## Option 2: Deploy with source code
 
-We recommend using Docker Compose to deploy the reverse image search system. However, you also can run from source code, you need to manually start [Milvus](https://milvus.io/docs/v2.0.0/install_standalone-docker.md) and [Mysql](https://dev.mysql.com/doc/mysql-installation-excerpt/5.7/en/docker-mysql-getting-started.html). Next show you how to run the API server and Client.
+We recommend using Docker Compose to deploy the reverse image search system. However, you also can run from source code, you need to manually start [Milvus](https://milvus.io) and [Mysql](https://dev.mysql.com/doc/mysql-installation-excerpt/5.7/en/docker-mysql-getting-started.html). Next show you how to run the API server and Client.
 
 ### 1. Start Milvus & Mysql
 
 First, you need to start Milvus & Mysql servers.
 
-Refer [Milvus Standalone](https://milvus.io/docs/v2.0.0/install_standalone-docker.md) for how to install Milvus. Please note the Milvus version should match pymilvus version in [config.py](./server/src/config.py).
+Refer [Milvus Standalone](https://milvus.io/docs/install_standalone-docker.md) for how to install Milvus.
+
+```bash
+$ wget https://github.com/milvus-io/milvus/releases/download/v2.2.10/milvus-standalone-docker-compose.yml -O docker-compose.yml
+$ sudo docker-compose up -d
+```
 
 There are several ways to start Mysql. One option is using docker to create a container:
 ```bash
-$ docker run -p 3306:3306 -e MYSQL_ROOT_PASSWORD=123456 -d --name qa_mysql mysql:5.7
+$ docker run -p 3306:3306 -e MYSQL_ROOT_PASSWORD=123456 -d --name image_search_mysql mysql:5.7
 ```
 
 
@@ -90,9 +95,11 @@ Then to start the system server, and it provides HTTP backend services.
 
 - **Install the Python packages**
 
+> Please note the Milvus version should [match pymilvus](https://milvus.io/docs/release_notes.md#Release-Notes) version in [requirements.txt](./server/requirements.txt). And this tutorial uses [milvus 2.2.10](https://milvus.io/docs/v2.2.x/install_standalone-docker.md) and [pymilvus 2.2.11](https://milvus.io/docs/release_notes.md#2210).
+
 ```bash
 $ git clone https://github.com/milvus-io/bootcamp.git
-$ cd solutions/reverse_image_search/quick_deploy/server
+$ cd bootcamp/solutions/image/reverse_image_search/server
 $ pip install -r requirements.txt
 ```
 
@@ -165,7 +172,7 @@ First, build a container by pulling docker image.
 $ docker run -d \
 -p 8001:80 \
 -e "API_URL=http://${API_HOST}:${API_PORT}" \
- milvusbootcamp/img-search-client:1.0
+ milvusbootcamp/img-search-client:2.2.10
 ```
 
 ## How to use front-end
@@ -199,24 +206,22 @@ Select an image to search.
 If you are interested in our code or would like to contribute code, feel free to learn more about our code structure.
 
 ```bash
-server
-├── Dockerfile
-├── requirements.txt
-└── src
-    ├── __init__.py
-    ├── config.py # Configuration file
-    ├── encode.py # Convert an image to embedding using towhee pipeline (ResNet50)
-    ├── encode_tf_resnet50.py # Old encoder file using ResNet50 by tensorflow
-    ├── logs.py
-    ├── main.py # Source code to start webserver
-    ├── milvus_helpers.py # Connect to Milvus server and insert/drop/query vectors in Milvus.
-    ├── mysql_helpers.py # Connect to MySQL server, and add/delete/query IDs and object information.
-    ├── operations
-    │   ├── __init__.py
-    │   ├── count.py
-    │   ├── drop.py
-    │   ├── load.py
-    │   ├── search.py
-    │   └── upload.py
-    └── test_main.py # Pytest file for main.py
+Dockerfile
+requirements.txt
+src
+├── __init__.py
+├── config.py # Configuration file
+├── encode.py # Convert an image to embedding using towhee pipeline (ResNet50)
+├── logs.py
+├── main.py # Source code to start webserver
+├── milvus_helpers.py # Connect to Milvus server and insert/drop/query vectors in Milvus.
+├── mysql_helpers.py # Connect to MySQL server, and add/delete/query IDs and object information.
+├── operations
+│   ├── __init__.py
+│   ├── count.py
+│   ├── drop.py
+│   ├── load.py
+│   ├── search.py
+│   └── upload.py
+└── test_main.py # Pytest file for main.py
 ```

solutions/image/reverse_image_search/docker-compose.yaml

@@ -3,20 +3,21 @@ version: '3.5'
 services:
   etcd:
     container_name: milvus-etcd
-    image: quay.io/coreos/etcd:v3.5.0
+    image: quay.io/coreos/etcd:v3.5.5
     networks:
       app_net:
     environment:
       - ETCD_AUTO_COMPACTION_MODE=revision
       - ETCD_AUTO_COMPACTION_RETENTION=1000
       - ETCD_QUOTA_BACKEND_BYTES=4294967296
+      - ETCD_SNAPSHOT_COUNT=50000
     volumes:
       - ${DOCKER_VOLUME_DIRECTORY:-.}/volumes/etcd:/etcd
     command: etcd -advertise-client-urls=http://127.0.0.1:2379 -listen-client-urls http://0.0.0.0:2379 --data-dir /etcd
 
   minio:
     container_name: milvus-minio
-    image: minio/minio:RELEASE.2020-12-03T00-03-10Z
+    image: minio/minio:RELEASE.2023-03-20T20-16-18Z
     networks:
       app_net:
     environment:
@@ -33,7 +34,7 @@ services:
 
   standalone:
     container_name: milvus-standalone
-    image: milvusdb/milvus:v2.0.2
+    image: milvusdb/milvus:v2.2.10
     networks:
       app_net:
         ipv4_address: 172.16.238.10
@@ -45,6 +46,7 @@ services:
       - ${DOCKER_VOLUME_DIRECTORY:-.}/volumes/milvus:/var/lib/milvus
     ports:
       - "19530:19530"
+      - "9091:9091"
     depends_on:
       - "etcd"
       - "minio"
@@ -62,7 +64,7 @@ services:
 
   webserver:
     container_name: img-search-webserver
-    image: milvusbootcamp/img-search-server:towhee0.6
+    image: milvusbootcamp/img-search-server:2.2.10
     networks:
       app_net:
         ipv4_address: 172.16.238.12
@@ -80,7 +82,7 @@ services:
 
   webclient:
     container_name: img-search-webclient
-    image: milvusbootcamp/img-search-client:1.0
+    image: milvusbootcamp/img-search-client:2.2.10
     networks:
       app_net:
         ipv4_address: 172.16.238.13

solutions/image/reverse_image_search/server/Dockerfile

@@ -8,5 +8,6 @@ WORKDIR /app/src
 COPY . /app
 
 RUN pip3 install -r /app/requirements.txt
+RUN python3 encode.py
 
 CMD python3 main.py

solutions/image/reverse_image_search/server/requirements.txt

@@ -4,8 +4,9 @@ diskcache==5.2.1
 uvicorn==0.13.4
 numpy==1.21.3
 pydantic==1.8.2
-pymilvus==2.0.2
-towhee==1.0.0rc1
+pymilvus==2.2.11
+towhee==1.1.0
+towhee.models==1.1.0
 fastapi==0.65.2
 python-multipart==0.0.5
 pillow==8.4.0

solutions/image/reverse_image_search/server/src/encode.py

@@ -14,3 +14,9 @@ def __init__(self):
     def resnet50_extract_feat(self, img_path):
         feat = self.image_embedding_pipe(img_path)
         return feat.get()[0]
+
+
+if __name__ == '__main__':
+    MODEL = Resnet50()
+    # Warm up the model to build image
+    MODEL.resnet50_extract_feat('https://raw.githubusercontent.com/towhee-io/towhee/main/towhee_logo.png')

solutions/image/reverse_image_search/server/src/logs.py

@@ -127,9 +127,3 @@ def write_log():
 
 
 LOGGER = write_log()
-# if __name__ == "__main__":
-#     message = 'test writing logs'
-#     logger = write_log()
-#     logger.info(message)
-#     logger.debug(message)
-#     logger.error(message)

solutions/image/reverse_image_search/server/src/main.py

@@ -81,7 +81,6 @@ async def upload_images(image: UploadFile = File(None), url: str = None, table_n
         # Save the upload image to server.
         if image is not None:
             content = await image.read()
-            print('read pic succ')
             img_path = os.path.join(UPLOAD_PATH, image.filename)
             with open(img_path, "wb+") as f:
                 f.write(content)
@@ -103,7 +102,6 @@ async def search_images(image: UploadFile = File(...), topk: int = Form(TOP_K),
     try:
         # Save the upload image to server.
         content = await image.read()
-        print('read pic succ')
         img_path = os.path.join(UPLOAD_PATH, image.filename)
         with open(img_path, "wb+") as f:
             f.write(content)

solutions/image/reverse_image_search/server/src/milvus_helpers.py

@@ -6,11 +6,7 @@
 
 class MilvusHelper:
     """
-    Say something about the ExampleCalass...
-
-    Args:
-        args_0 (`type`):
-        ...
+    Milvus Helper
     """
     def __init__(self):
         try:
@@ -23,60 +19,52 @@ def __init__(self):
 
     def set_collection(self, collection_name):
         try:
-            if self.has_collection(collection_name):
-                self.collection = Collection(name=collection_name)
-            else:
-                raise Exception(f"There is no collection named:{collection_name}")
+            self.collection = Collection(name=collection_name)
         except Exception as e:
-            LOGGER.error(f"Failed to load data to Milvus: {e}")
+            LOGGER.error(f"Failed to set collection in Milvus: {e}")
             sys.exit(1)
 
     def has_collection(self, collection_name):
         # Return if Milvus has the collection
         try:
             return utility.has_collection(collection_name)
         except Exception as e:
-            LOGGER.error(f"Failed to load data to Milvus: {e}")
+            LOGGER.error(f"Failed to get collection info to Milvus: {e}")
             sys.exit(1)
 
     def create_collection(self, collection_name):
         # Create milvus collection if not exists
         try:
-            if not self.has_collection(collection_name):
-                field1 = FieldSchema(name="id", dtype=DataType.INT64, descrition="int64", is_primary=True, auto_id=True)
-                field2 = FieldSchema(name="embedding", dtype=DataType.FLOAT_VECTOR, descrition="float vector",
-                                     dim=VECTOR_DIMENSION, is_primary=False)
-                schema = CollectionSchema(fields=[field1, field2], description="collection description")
-                self.collection = Collection(name=collection_name, schema=schema)
-                LOGGER.debug(f"Create Milvus collection: {collection_name}")
-            else:
-                self.set_collection(collection_name)
+            field1 = FieldSchema(name="id", dtype=DataType.INT64, descrition="int64", is_primary=True, auto_id=True)
+            field2 = FieldSchema(name="embedding", dtype=DataType.FLOAT_VECTOR, descrition="float vector",
+                                    dim=VECTOR_DIMENSION, is_primary=False)
+            schema = CollectionSchema(fields=[field1, field2], description="collection description")
+            self.collection = Collection(name=collection_name, schema=schema)
+            LOGGER.debug(f"Create Milvus collection: {collection_name}")
             return "OK"
         except Exception as e:
-            LOGGER.error(f"Failed to load data to Milvus: {e}")
+            LOGGER.error(f"Failed create collection in Milvus: {e}")
             sys.exit(1)
 
     def insert(self, collection_name, vectors):
         # Batch insert vectors to milvus collection
         try:
-            self.create_collection(collection_name)
-            data = [vectors]
             self.set_collection(collection_name)
+            data = [vectors]
             mr = self.collection.insert(data)
             ids = mr.primary_keys
-            self.collection.load()
             LOGGER.debug(
                     f"Insert vectors to Milvus in collection: {collection_name} with {len(vectors)} rows")
             return ids
         except Exception as e:
-            LOGGER.error(f"Failed to load data to Milvus: {e}")
+            LOGGER.error(f"Failed to insert data to Milvus: {e}")
             sys.exit(1)
 
     def create_index(self, collection_name):
         # Create IVF_FLAT index on milvus collection
         try:
             self.set_collection(collection_name)
-            default_index = {"index_type": "IVF_SQ8", "metric_type": METRIC_TYPE, "params": {"nlist": 16384}}
+            default_index = {"metric_type": METRIC_TYPE, "index_type": "IVF_FLAT", "params": {"nlist": 2048}}
             status = self.collection.create_index(field_name="embedding", index_params=default_index)
             if not status.code:
                 LOGGER.debug(
@@ -103,10 +91,9 @@ def search_vectors(self, collection_name, vectors, top_k):
         # Search vector in milvus collection
         try:
             self.set_collection(collection_name)
+            self.collection.load()
             search_params = {"metric_type": METRIC_TYPE, "params": {"nprobe": 16}}
-            # data = [vectors]
             res = self.collection.search(vectors, anns_field="embedding", param=search_params, limit=top_k)
-            print(res[0])
             LOGGER.debug(f"Successfully search in collection: {res}")
             return res
         except Exception as e:
@@ -117,6 +104,7 @@ def count(self, collection_name):
         # Get the number of milvus collection
         try:
             self.set_collection(collection_name)
+            self.collection.flush()
             num = self.collection.num_entities
             LOGGER.debug(f"Successfully get the num:{num} of the collection:{collection_name}")
             return num

solutions/image/reverse_image_search/server/src/operations/__init__.py

@@ -1,2 +0,0 @@
-import logging
-logging.basicConfig(filename='app.log', filemode='w', format='%(asctime)s - %(name)s - %(levelname)s - %(message)s', level=logging.INFO)