Python SDK

The Python SDK spicepy is the easiest way to use and query Spice.ai in Python.

The Python SDK uses Apache Apache Flight to efficiently stream data to the client and Apache Arrow Records as data frames which are then easily converted to Pandas data frames.

Requirements

• Python 3.11+

The following packages are required and will be automatically installed by pip:

• pyarrow

• pandas

• certify

• requests

Apple M1 Mac Requirements - How do I know if I have an M1?

Apple M1 Macs require an arm64 compatible version of pyarrow which can be installed using miniforge. We recommend the following procedure:

• Install Homebrew

• Install miniforge with:

brew install --cask miniforge

• Initialize conda in your terminal with:

conda init "$(basename "${SHELL}")"

• Install pyarrow and pandas with:

conda install pyarrow pandas

While Anaconda can be used to install pyarrow, the installed version is old (4.0.0) so we recommend using the miniforge distribution.

Installation

Install the spicepy package directly from the Spice Github Repository at https://github.com/spiceai/spicepy:

pip install git+https://github.com/spiceai/[email protected]

Usage

Import spicepy and create a Client by providing your API Key.

You can then submit queries using the query function.

from spicepy import Client

client = Client('API_KEY')
data = client.query('SELECT * FROM eth.recent_blocks LIMIT 10;', timeout=5*60)
pd = data.read_pandas()

Querying data is done through a Client object that initializes the connection with the Spice.ai endpoint. Client has the following arguments:

• api_key (string, optional): Spice.ai API key to authenticate with the endpoint.

• url (string, optional): URL of the endpoint to use (default: grpc+tls://flight.spiceai.io)

• tls_root_cert (Path or string, optional): Path to the tls certificate to use for the secure connection (ommit for automatic detection)

Once a Client is obtained queries can be made using the query() function. The query() function has the following arguments:

• query (string, required): The SQL query.

• timeout (int, optional): The timeout in seconds.

A custom timeout can be set by passing the timeout parameter in the query function call. If no timeout is specified, it will default to a 10 min timeout then cancel the query, and a TimeoutError exception will be raised.

Usage with local Spice runtime

Follow the quickstart guide to install and run spice locally.

from spicepy import Client

client = Client()
data = client.query('SELECT trip_distance, total_amount FROM taxi_trips ORDER BY trip_distance DESC LIMIT 10;', timeout=5*60)
pd = data.read_pandas()

Contributing

Contribute to or file an issue with the spicepy library at: https://github.com/spiceai/spicepy