similar-images

API

Create an item

POST /items

{
  "imageUrl": "https://picsum.photos/1000.jpg",
  "sku": "SKU"
}

Search similar skus

GET /similar-skus ?imageUrl=https://picsum.photos/1000.jpg Response:

{
   "similarSkus": [{
      "sku": "sku",
      "score": 0.5
   }]
}

Run

Web Client:

cd public
npm i
npm run build 

Server:

pip3 install -r requirments.txt
export FLASK_APP=server.py
flask run

Redis:

1. build RedisAI
   1. clone the repository
   2. run make
   3. copy ./src/redisai.so
2. build RedisGears
   1. clone the repository
   2. run make
   3. copy ./src/redisgears.so
3. build VecSim
   1. clone the repository
   2. open ./src/vector_similarity.c
   3. change #define VEC_SIZE (line 20) value to 1280
   4. change #define VEC_HOLDER_SIZE (line 22) value to 100 * 1024
   5. run make
   6. copy ./src/vector_similarity.so
4. run redis-server --loadmodule ./redisai.so --loadmodule ./redisgears.so Plugin ./vector_similarity.so