RelationalAI in Cloud Notebooks#

You can use RelationalAI in cloud notebook environments like Google Colab or Hex. This guide shows you how to set up the relationalai package in these environments.

Google Colab#

Installation#

Begin by installing relationalai from PyPI by pasting the following code into the first cell and doing shift+enter to run it:

#!pip install relationalai

Note that the exclamation mark ! marks the command as a shell command in Colab.

Setup#

Click on the key icon in the left sidebar to open the user secrets editor. Add your password with the key snowflake_password. Then fill in the following code snippet with your Snowflake account details:

#import relationalai as rai
from relationalai.clients.config import Config

from google.colab import userdata
snowflake_password = userdata.get('snowflake_password')

snowflake_account = ""
snowflake_user = ""
snowflake_role = ""
snowflake_warehouse = ""
rai_engine_name = ""

config = Config({
    "platform": "snowflake",
    "authenticator": "snowflake",
    "account": snowflake_account,
    "user": snowflake_user,
    "password": snowflake_password,
    "role": snowflake_role,
    "warehouse": snowflake_warehouse,
    "rai_app_name": "relationalai",
    "engine": rai_engine_name,
})

model = rai.Model("Example", config=config)

Run this cell to define the model object and connect to your Snowflake account.

Usage#

Now you can use model to run queries as usual:

#with model.query() as select:
    z = select(1 + 1)

print(z.results)

Hex#

Installation#

Run the following command in a Python cell in a Hex notebook to install the relationalai package:

#!pip install relationalai

Setup and Usage#

Using Hex’s Snowflake Integration#

If you have connected to your Snowflake account from Hex and enabled Snowpark, you can use the hextoolkit module to obtain a session object and use it to create a model object:

#import relationalai as rai
import hextoolkit

hex_snowflake_conn = hextoolkit.get_data_connection("<Your Snowflake Connection Name>")
hex_snowpark_session = hex_snowflake_conn.get_snowpark_session()

model = rai.Model("Example", connection=hex_snowpark_session)

with model.query() as select:
    z = select(1 + 1)

print(z.results)

Connecting to Snowflake from Python#

Alternatively, you can connect to Snowflake from Python using the relationalai package:

Click on the Variables menu item in the left sidebar. Enter your Snowflake password as a secret. Paste the code below and fill in your account details:

#import relationalai as rai
from relationalai.clients.config import Config

config = Config({
    "platform": "snowflake",
    "authenticator": "snowflake",
    "account": "<SNOWFLAKE_ACCOUNT>",
    "user": "<SNOWFLAKE_USER>",
    "password": snowflake_password,
    "role": "<SNOWFLAKE_ROLE>",
    "warehouse": "<SNOWFLAKE_WAREHOUSE>",
    "rai_app_name": "relationalai",
    "engine": "<RAI_ENGINE_NAME>",
})

model = rai.Model("Example", config=config)

with model.query() as select:
    z = select(1 + 1)

print(z.results)

Next steps#

To begin exploring RelationalAI’s capabilities, check out the Getting Started guide or the Example Notebooks.