CamelAI

🐫 CAMEL is an open-source community dedicated to finding the scaling laws of agents. We believe that studying these agents on a large scale offers valuable insights into their behaviors, capabilities, and potential risks. To facilitate research in this field, we implement and support various types of agents, tasks, prompts, models, and simulated environments.

Setup

You can run SurrealDB locally or start with a free SurrealDB Cloud account.

For local, two options:

Install SurrealDB and run SurrealDB. Run in-memory with:
surreal start -u root -p root

Run with Docker.

    docker run --rm --pull always -p 8000:8000 surrealdb/surrealdb:latest start

Example

import os

from camel.storages.vectordb_storages import (
    SurrealStorage,
    VectorDBQuery,
    VectorRecord,
)


def main():
    url = os.getenv("SURREAL_URL", "ws://localhost:8000/rpc")
    table = os.getenv("SURREAL_TABLE", "tb")
    vector_dim = int(os.getenv("SURREAL_VECTOR_DIM", 4))
    namespace = os.getenv("SURREAL_NAMESPACE", "ns")
    database = os.getenv("SURREAL_DATABASE", "db")
    user = os.getenv("SURREAL_USER", "user")
    password = os.getenv("SURREAL_PASSWORD")

    # Raise an error if password is not set
    if not password:
        raise ValueError(
            "Environment variable SURREAL_PASSWORD is not set. "
            "Please set it before running."
        )

    # Initialize the SurrealStorage instance with provided parameters
    storage = SurrealStorage(
        url=url,
        table=table,
        namespace=namespace,
        database=database,
        user=user,
        password=password,
        vector_dim=vector_dim,
    )

    # Clear existing data in storage
    storage.clear()

    # Print the current status after clearing
    print("[Step 1] After clear:", storage.status())

    vec1 = VectorRecord(vector=[1, 2, 3, 4], payload={"name": "test_1"})
    vec2 = VectorRecord(vector=[5, 6, 7, 8], payload={"name": "test_2"})
    vec3 = VectorRecord(vector=[9, 10, 11, 12], payload={"name": "test_3"})
    vec4 = VectorRecord(vector=[13, 14, 15, 16], payload={"name": "test_4"})
    storage.add([vec1, vec2, vec3, vec4])
    print("[Step 2] After add:", storage.status())

    res = storage.client.query_raw(
        "SELECT * FROM lyz_tb WHERE payload.name = 'test_3';"
    )["result"][0]["result"][0]["id"].id
    print("[Step 3] Query Result ID for 'test_3':", res)

    storage.delete(ids=[res])
    print("[Step 4] After delete 'test_3':", storage.status())

    res = storage.query(
        VectorDBQuery(query_vector=[1.1, 2.1, 3.1, 4.1], top_k=2)
    )
    print("[Step 5] Vector Query Result:", res)

The output should look like this:

[Step 1] After clear: vector_dim=4 vector_count=0oceanbasech
[Step 2] After add: vector_dim=4 vector_count=4
[Step 3] Query Result ID for 'test_3': lov5h16x6uog7l2xtsqp
[Step 4] After delete 'test_3': vector_dim=4 vector_count=3
[Step 5] Vector Query Result:
[VectorDBQueryResult(record=VectorRecord(vector=[],
id='9803ae3a-18da-4152-a522-48e1939a3604',
payload={'name': 'test_2'}), similarity=0.027665393972965968),
VectorDBQueryResult(record=VectorRecord(vector=[],
id='de3d085c-ed1b-4d23-9d12-d9fc64ae1e00',
payload={'name': 'test_1'}), similarity=0.00010404203326297434)]

Resources

Camel documentation