datagrok-api 0.1.0

Client library for integration with datagrok API

Datagrok python client library

This library can be used for integration with datagrok. It is a python wrapper for public API, that has OpenAPI specification available here.

Refer to Help for more information.

Installation

To install package, use pip.

pip install datagrok-api

Usage

Initialization

To use the API client, import DatagrokClient. It is recommended to use it as a context manager to ensure proper connection handling:

from datagrok_api import DatagrokClient

# Initialize with your server URL and API token
with DatagrokClient("https://public.datagrok.ai/api", "Bearer <your-token>") as grok:
    me = grok.users.current()
    print(f"Logged in as: {me.first_name} {me.last_name}")

Working with Data (Tables & Files)

The package uses Pandas for representation of tables and dataframes.

import pandas as pd
from sklearn.datasets import load_iris

# Upload a DataFrame as a table
iris_df = pd.DataFrame(load_iris()['data'])
iris_id = grok.tables.upload('iris', iris_df)['ID']

# Download a table by ID or Name
df = grok.tables.download(iris_id)
print(df.head())

# Upload a local file
uploaded_file = grok.files.upload("data.csv")

Functions & Connections

You can call server-side functions and manage data connections directly through their respective managers.

from datagrok_api.models import DataConnection, FileDataSourceType, Credentials

# Call a Datagrok function
result = grok.functions.call("JohnDoe:MyFunction", {"a": 1, "b": 2})

# Create and save a connection (e.g., S3)
creds = Credentials(accessKey="aws-key", secretKey="aws-secret")
conn = DataConnection(
    name="My S3 Bucket",
    data_source=FileDataSourceType.S3,
    bucket="my-bucket",
    region="us-east-1",
    credentials=creds
)
saved_conn = grok.connections.save(conn)

Group Management

Manage user groups and metadata.

# Search for a group and update its properties
groups = grok.groups.find("Chemists")
if groups:
    group = groups[0]
    group.description = "Updated description for the team"
    grok.groups.save(group)

License

See License.