mig-dx-api 0.2.8

Dataset Exchange API Client Library

DX API Client Library

Welcome to the DX API Client Library! This library provides a convenient Python interface to interact with the DX API, allowing you to manage datasets, installations, and perform various operations with ease.

Table of Contents

• Features
• Installation
• Prerequisites
• Getting Started
  • Authentication
  • Initialization
• Usage
  • Who Am I
  • Managing Installations
    • Listing Installations
    • Accessing an Installation Context
  • Managing Datasets
    • Listing Datasets
    • Accessing Dataset Operations
    • Creating a Dataset
    • Creating a Dataset with specific tags
    • Uploading Data to a Dataset
    • Getting Dataset Jobs
    • Retrieving Records from a Dataset
• Asynchronous Usage
• Examples
• Contributing
• License

Features

• Authenticate with the DX API using JWT tokens.
• Manage installations and datasets.
• Upload and download data to and from datasets.
• Synchronous and asynchronous support.
• Context managers for handling authentication scopes.

Installation

You can install the library using pip:

pip install mig-dx-api

Prerequisites

• Python 3.10 or higher.
• An application ID (app_id),a workspace key (workspace_key), and a corresponding private key in PEM format from the Console.
• DX API access credentials.

Getting Started

Authentication

The library uses JWT tokens for authentication. You need to provide your app_id, workspace_key, and the path to your private key file when initializing the client.

Initialization

from mig_dx_api import DX

# Initialize the client
dx = DX(app_id='your_app_id', workspace_key='your_workspace_key', private_key_path='path/to/private_key.pem')

# OR
dx = DX(app_id='your_app_id',workspace_key='your_workspace_key', private_key='your_private_key')

Alternatively, you can set the environment variables DX_CONFIG_APP_ID, DX_CONFIG_WORKSPACE_KEY, and DX_CONFIG_PRIVATE_KEY_PATH:

export DX_CONFIG_APP_ID='your_app_id'
export DX_CONFIG_WORKSPACE_KEY='your_workspace_key'

export DX_CONFIG_PRIVATE_KEY_PATH='path/to/private_key.pem'
# OR
export DX_CONFIG_PRIVATE_KEY='your_private_key'

And initialize the client without arguments:

dx = DX()

Usage

Who Am I

Retrieve information about the authenticated user:

user_info = dx.whoami()
print(user_info)

Managing Installations

Listing Installations

installations = dx.get_installations()
for installation in installations:
    print(installation.name)

Accessing an Installation Context

Use the installation context to perform operations related to a specific installation:

# Find an installation by name or ID
installation = dx.installations.find(install_id=1)

# Use the installation context
with dx.installation(installation) as ctx:
    # Perform operations within the context
    datasets = list(ctx.datasets)
    for dataset in datasets:
        print(dataset.name)

Or enter context with a lookup by name:

with dx.installation(install_id=1) as ctx:
    # Perform operations within the context
    datasets = list(ctx.datasets)
    for dataset in datasets:
        print(dataset.name)

Managing Datasets

Listing Datasets

with dx.installation(installation) as ctx:
    for dataset in ctx.datasets:
        print(dataset.name)

Accessing Dataset Operations

Dataset operations may be accessed by name if the dataset name is unique within the installation, or by dataset_id.

with dx.installation(installation) as ctx:
    dataset_ops = ctx.datasets.find(name='My Dataset')

with dx.installation(installation) as ctx:
    dataset_ops = ctx.datasets.find(dataset_id='123adatasetid')

with dx.installation(installation) as ctx:
    dataset_ops = ctx.datasets.find(dataset_id=UUID('123adatasetid'))

Creating a Dataset

from mig_dx_api import DatasetSchema, SchemaProperty

# Define the schema
schema = DatasetSchema(
    properties=[
        SchemaProperty(name='my_string', type='string', required=True),
        SchemaProperty(name='my_integer', type='integer', required=True),
        SchemaProperty(name='my_boolean', type='boolean', required=False),
    ],
    primary_key=['my_string']
)

# Create the dataset
with dx.installation(installation) as ctx:
    new_dataset = ctx.datasets.create(
        name='My Dataset',
        description='A test dataset',
        schema=schema.model_dump()  # this can also be defined as a dictionary
    )

Creating a Dataset with specific tags

# Get tag options
with dx.installation(installation) as ctx:
    tag_options = ctx.datasets.get_tags
    tags = [tags['data'][0]['movementAppId'], tags['data'][1]['movementAppId']]
    new_dataset = ctx.datasets.create(
        name='My Tagged Dataset',
        description='A test tagged dataset',
        schema={
            "primaryKey": ["my_van_id"],
            "properties": [
                {"name": "my_van_id", "type": "string"},
                {"name": "first_name", "type": "string"},
                {"name": "last_name", "type": "string"},
            ],
        },
        tagIds=[tags]
    )

Uploading Data to a Dataset

data = [
    {'my_string': 'string1', 'my_integer': 1, 'my_boolean': True},
    {'my_string': 'string2', 'my_integer': 2, 'my_boolean': False},
    {'my_string': 'string3', 'my_integer': 3, 'my_boolean': True},
]

with dx.installation(installation) as ctx:
    dataset_ops = ctx.datasets.find(name='My Dataset')
    dataset_ops.load(data, validate_records=True)  # validate_records=True will validate the records against the schema using Pydantic
    

Load defaults to csv for newline-delimited json (NDJSON), but can support optional content types of tsv or other json types if the type is passed in, e.g. dataset_ops.load(data, validate_records=True, content_type='json')

Dataset Jobs

with dx.installation(installation) as ctx:
    dataset_ops = ctx.datasets.find(name='My Dataset')
    # Job Id comes from load, load_from_file or load_from_url
    job_id = dataset_ops.load(data, validate_records=True)["jobId"]
    # Get dataset job
    job = ctx.datasets.get_dataset_job(job_id)
    # Get logs for dataset job
    job_logs = ctx.datasets.get_dataset_job_logs(job_id)
    # Get latest status for dataset job
     job_logs_current = ctx.datasets.get_current_dataset_logs(job_id)

Retrieving Records from a Dataset

with dx.installation(installation) as ctx:
    dataset_ops = ctx.datasets.find(name='My Dataset')
    records = dataset_ops.records()
    for record in records:
        print(record)

Asynchronous Usage

The library supports asynchronous operations using async/await.

import asyncio

async def main():
    dx = DX()
    async with dx.installation(installation) as ctx:
        async for dataset in ctx.datasets:
            print(dataset.name)

        dataset = await ctx.datasets.find(name='My Dataset')

        data = [
            {'my_string': 'string1', 'my_integer': 1, 'my_boolean': True},
            {'my_string': 'string2', 'my_integer': 2, 'my_boolean': False},
            {'my_string': 'string3', 'my_integer': 3, 'my_boolean': True},
        ]

        await dataset.load(data)

        async for record in dataset.records():
            print(record)

asyncio.run(main())

Examples

Example: Loading Data from a File

with dx.installation(installation) as ctx:
    dataset_ops = ctx.datasets.find(name='My Dataset')
    dataset_ops.load_from_file('data.csv')

TSV and JSON files are also supported. Uses file extension to determine file type

Example: Uploading Data from a URL

with dx.installation(installation) as ctx:
    dataset_ops = ctx.datasets.find(name='My Dataset')
    dataset_ops.load_from_url('https://example.com/data.csv')

TSV and JSON files are also supported. Uses file extension to determine file type

---

Note: This README assumes that the package name is mig-dx-api and that the code is properly packaged and available for installation via pip. Adjust the instructions accordingly based on the actual package name and installation method.