Recommendation API - context

User context

To personalize search and recommendation results or provide authorization feature, you can specify user information (anonymous and/or signed-in) in client context so they will be passed into API calls automatically.

User ID and user hash

You can set user_id and user_hash:

client.context.user_id = '...';
client.context.user_hash = '...';

See REST API for more details about user_hash.

User type

You can set user_type:

client.context.user_id = '...';
client.context.user_type = '...';

The user_type value is passed into API calls only when user_id is also set.

When including user_type, you must modify your user_hash generation process. Instead of hashing only the user_id, you should hash the combination of user_type and user_id in the format user_type:user_id. Here's a Python script demonstrating how to generate the user_hash:

import hashlib
import hmac

def generate_user_hash(secret_key, user_id, user_type=None):
    key_bytes = secret_key.encode('utf-8')
    
    if user_type:
        message = f"{user_type}:{user_id}".encode('utf-8')
    else:
        message = user_id.encode('utf-8')
    
    user_hash = hmac.new(key_bytes, message, hashlib.sha256).hexdigest()
    return user_hash

Anonymous ID

The SDK takes care of anonymous_id automatically using session storage, so you don't need to set it manually.

However, you can also override the anonymous_id value by yourself:

client.context.anonymous_id = 'my_anonymous_id';