External Functions for Custom AI/ML

For advanced AI/ML scenarios, Databend supports external functions that connect your data with custom AI/ML infrastructure written in languages like Python.

Feature	Description	Benefits
Model Flexibility	Use open-source models or your internal AI/ML infrastructure	• Freedom to choose any model • Leverage existing ML investments • Stay up-to-date with latest AI advancements
GPU Acceleration	Deploy external function servers on GPU-equipped machines	• Faster inference for deep learning models • Handle larger batch sizes • Support compute-intensive workloads
Custom ML Models	Deploy and use your own machine learning models	• Proprietary algorithms • Domain-specific models • Fine-tuned for your data
Advanced AI Pipelines	Build complex AI workflows with specialized libraries	• Multi-step processing • Custom transformations • Integration with ML frameworks
Scalability	Handle resource-intensive AI operations outside Databend	• Independent scaling • Optimized resource allocation • High-throughput processing

Implementation Overview

1. Create an external server with your AI/ML code (Python with databend-udf)
2. Register the server with Databend using CREATE FUNCTION
3. Call your AI/ML functions directly in SQL queries

Example: Custom AI Model Integration

# Simple embedding UDF server demo
from databend_udf import udf, UDFServer
from sentence_transformers import SentenceTransformer

# Load pre-trained model
model = SentenceTransformer('all-mpnet-base-v2')  # 768-dimensional vectors

@udf(
    input_types=["STRING"],
    result_type="ARRAY(FLOAT)",
)
def ai_embed_768(inputs: list[str], headers) -> list[list[float]]:
    """Generate 768-dimensional embeddings for input texts"""
    try:
        # Process inputs in a single batch
        embeddings = model.encode(inputs)
        # Convert to list format
        return [embedding.tolist() for embedding in embeddings]
    except Exception as e:
        print(f"Error generating embeddings: {e}")
        # Return empty lists in case of error
        return [[] for _ in inputs]

if __name__ == '__main__':
    print("Starting embedding UDF server on port 8815...")
    server = UDFServer("0.0.0.0:8815")
    server.add_function(ai_embed_768)
    server.serve()

-- Register the external function in Databend
CREATE OR REPLACE FUNCTION ai_embed_768 (STRING)
    RETURNS ARRAY(FLOAT)
    LANGUAGE PYTHON
    HANDLER = 'ai_embed_768'
    ADDRESS = 'https://your-ml-server.example.com';

-- Use the custom embedding in queries
SELECT
    id,
    title,
    cosine_distance(
        ai_embed_768(content),
        ai_embed_768('machine learning techniques')
    ) AS similarity
FROM articles
ORDER BY similarity ASC
LIMIT 5;

Next Steps

1. External Functions Guide - Detailed setup instructions
2. Databend Cloud - Try AI functions with a free trial