Dynamic Filtering

How does it work

When creating the project mapping (see the environment creation step), there is a parameter called categories. This allows you to define how your products are categorized. Below is an example of how a skincare or cosmetics company might set up their categories. Note that all category keys must exist in your catalog.

"categories": [
    {
        "name": "product_categories",
        "separator": ","
    },
    {
        "name": "skin_types",
        "separator": ","
    },
    {
        "name": "product_brands",
        "separator": ""
    },
    {
        "name": "good_for",
        "separator": ","
    },
    {
        "name": "product_tags",
        "separator": ","
    },
    {
        "name": "key_ingredients",
        "separator": ","
    }
]

Once you've defined categories in the mapper, you can use them to generate filters dynamically in your search results.

Continuing from the previous example, suppose you hit the POST /v1/items/search endpoint with the following request body. Notice the dynamic_filtering parameter, where two values defined in the mapper is passed:

{
  "user_id": "user12345",
  "message": "lotion",
  "type": "individual",
  "price_lower": 1,
  "price_upper": 10000,
  "categories": [ ],
  "flags": {},
  "sort": {},
  "dynamic_filtering": [
    "product_categories",
    "skin_types"
  ],
  "details": false,
  "start": 0,
  "size": 3
}

In the response, along with the product list, you will see values under keywords that can be used to generate dynamic filters:

{
  "detail": {
    "response": [
      {
        "title": "example product 1"
      },
      {
        "title": "example product 2"
      }
    ]
  },
  "summary": {
    "keywords": {
      "product_categories": [
        "Skin Care",
        "Toner/Lotion",
        "International Brands"
      ],
      "skin_types": [
        "sensitive",
        "dry",
        "normal"
      ]
    },
    "max_price": 1980,
    "min_price": 1350,
    "total_hits": 102
  }
}