Elasticsearch for vector search

from elasticsearch import Elasticsearch
import numpy as np

# Connect to Elasticsearch
es = Elasticsearch([{'host': 'localhost', 'port': 9200}])

# Index a document with its vector embedding
doc = {
    'text': 'The new EU AI regulations focus on transparency and accountability.',
    'vector': np.random.rand(384).tolist()  # Simulated 384-dimensional embedding
}
es.index(index='documents', body=doc)

# Perform a vector similarity search
query_vector = np.random.rand(384).tolist()  # Simulated query embedding
search_body = {
    "query": {
        "script_score": {
            "query": {"match_all": {}},
            "script": {
                "source": "cosineSimilarity(params.query_vector, 'vector') + 1.0",
                "params": {"query_vector": query_vector}
            }
        }
    }
}
results = es.search(index='documents', body=search_body)