Enhance Your Search Capabilities Using Algoboost as a Vector Store

In the ever-evolving landscape of artificial intelligence and machine learning, efficiency and scalability are paramount. Introducing Algoboost, a cutting-edge platform designed to revolutionize embedding model inference and vector storage. Whether you're a data scientist, machine learning engineer, or a business leveraging AI, Algoboost is here to elevate your capabilities and streamline your workflows.

To get started with algoboost have a look at this post here

In this blog post we will be showing you how to use algoboost as a vector store.

What is a vector store and how is it used?

A vector store is a specialized data structure or database optimized for storing and retrieving vectors. Vectors are mathematical representations of data points in a multi-dimensional space, often used in machine learning, natural language processing (NLP), and information retrieval. Each vector consists of a list of numerical values, representing various features of the data.

Key Uses of Vector Stores

Similarity Search

Purpose: To find the most similar data points to a given query vector.
Applications: Image search (finding similar images), document retrieval, recommendation systems, and more.

Clustering

Purpose: To group similar vectors together.
Applications: Customer segmentation, topic modeling, and anomaly detection.

Classification

Purpose: To assign labels to data points based on their vector representations.
Applications: Sentiment analysis, spam detection, and product categorization.

Embedding Storage

Purpose: To store embeddings generated by machine learning models.
Applications: Word embeddings (e.g., Word2Vec, GloVe), sentence embeddings (e.g., BERT, Sentence Transformers), and graph embeddings.

Real-World Applications

Recommendation Systems

By storing user and item embeddings, systems can recommend items based on the similarity of user preferences and item features.

Semantic Search

Enhances search engines by retrieving results based on semantic meaning rather than just keyword matching.

Fraud Detection

Detects anomalous behavior by finding vectors that deviate significantly from typical patterns.

How to use algoboost as a vector store

step 1: Create a collection for your vectors

First, you need to create a collection and give your collection a name along with specifying the similarity metric it will use for vector similarity search.

import requests

api_key = ""  # Replace {API_KEY} with your API key


def create_vector_store(collection_name, model_type, dimension,
                        similarity_metric):
    url = "https://app.algoboost.ai/api/model/create_vector_store"
    headers = {
        "Authorization": f"Bearer {api_key}",
        "Content-Type":  "multipart/form-data"
    }

    try:
        data = {
            "collection_name": collection_name,
            "model_type": model_type,
            "dimension": dimension,
            "similarity_metric": similarity_metric
        }

        response = requests.post(url, headers=headers, data=data)

        # Check the HTTP status code
        if response.status_code == 200:
            # Parse the JSON response
            results = response.json()
            return results
        else:
            print(f"API request failed with status code: {response.status_code}")
            return None
    except Exception as e:
        print(f"An error occurred: {str(e)}")
        return None


# Example usage:
collection_name = 'your_collection_name'  # your collection name
model_type = 'image'  # text or image
dimension = 512  # vector dimension
similarity_metric = 'COSINE'  # L2 or COSINE

result = create_vector_store(collection_name, model_type, dimension,
                             similarity_metric)
print(result)  # or do something with the result

Step 2: Store Vectors

Once a collection is created, you can store your vectors via the Algoboost API.

import requests

def store_vectors(api_url, collection_name, model_type, partition, vectors, api_key):
    """
    Store vectors in the Algoboost vector store.

    Parameters:
    - api_url (str): The API endpoint URL.
    - collection_name (str): The name of the collection to store vectors in.
    - model_type (str): The type of model (e.g. 'image' or 'text').
    - partition (str): The partition name.
    - vectors (list of dict): The list of vectors to store, each with 'input' and 'vector' keys.
    - api_key (str): api_key.
    - session_cookie (str): The session cookie for authentication.

    Returns:
    - response (requests.Response): The response object from the API call.
    """

    # Define the payload with the required data
    payload = {
        "collection_name": collection_name,
        "model_type": model_type,
        "model": "",
        "partition": partition,
        "data": vectors
    }

    # Define the headers including the Authorization token
    headers = {
        "Content-Type": "application/json",
        "Authorization": f"Bearer {api_key}"
    }

    # Send the POST request to the API
    response = requests.post(api_url, json=payload, headers=headers)

    return response

# Example usage of the function
api_url = "https://app.algoboost.ai/api/model/store_vector"
collection_name = "your_collection_name"
model_type = "image"  # text or image
partition = "your_partition"
vectors = [
    {"input": "", "vector": [0.1, 0.2, 0.3]},
    {"input": "", "vector": [0.4, 0.5, 0.6]}
]
api_key = "your_api_key"

response = store_vectors(api_url, collection_name, model_type, partition, vectors, api_key)

# Print the response from the API
print(response.json())

Step 3: Similarity Search in Your Collection

To perform a similarity search, use the Algoboost API to query your collection with a given vector and retrieve the most similar vectors.

import requests

def vector_store_similarity(api_key, collection_name, model_type, limit, partition, vector):
    """
    Makes a request to the Algoboost vector similarity API.

    Parameters:
    - api_key (str): Your Algoboost API key.
    - collection_name (str): The name of your collection.
    - model_type (str): The type of model ('text' or 'image').
    - limit (int): The number of results to return.
    - partition (str): Your partition.
    - vector (list): The vector for similarity comparison.

    Returns:
    - dict: The JSON response from the API.
    """

    url = "http://app.algoboost.ai/api/model/vector_similarity"

    payload = {
        "collection_name": collection_name,
        "model_type": model_type,
        "limit": limit,
        "partition": partition,
        "vector": vector
    }
    headers = {
        "Content-Type": "application/json",
        "Authorization": f"Bearer {api_key}"
    }

    response = requests.post(url, json=payload, headers=headers)

    return response.json()

# Example usage:
api_key = "your_algoboost_api_key"  # Replace with your actual Algoboost API key
collection_name = "your_collection_name"
model_type = "image"  # or "text"
limit = 3
partition = "your_partition"
vector = [0.4, 0.5, 0.6]

result = vector_store_similarity(api_key, collection_name, model_type, limit, partition, vector)
print(result)

Conclusion

In the world of AI and machine learning, staying ahead means leveraging the best tools available. Algoboost offers a powerful, efficient, and scalable solution for embedding model inference and vector storage, empowering you to innovate and achieve more. Experience the future of AI infrastructure with Algoboost.
Ready to boost your AI capabilities? Sign up for Algoboost today.

Blog

Enhance Your Search Capabilities Using Algoboost as a Vector Store

Tshidiso

Key Uses of Vector Stores

Real-World Applications

How to use algoboost as a vector store

Join Our Newsletter. No Spam, Only the good stuff.

Related