unstructured-platform 0.1.0

Python SDK for the Unstructured Platform API

Unstructured Platform SDK

A Python SDK for the Unstructured Platform API.

Unstructured Platform is an enterprise-grade ETL (Extract, Transform, Load) platform designed specifically for Large Language Models (LLMs). It provides a no-code UI and production-ready infrastructure to help organizations transform raw, unstructured data into LLM-ready formats. With over 18 million downloads and used by more than 50,000 companies, Unstructured Platform enables data scientists and engineers to spend less time collecting and cleaning data, and more time on modeling and analysis.

This SDK was developed by One Minute AI, the AI platform that provides a one-click (or one line of code) way to create AI assistants and apps using data ingested by Unstructured.

• Unstructured Platform SDK
  • Installation
  • Usage
    • Initialize the client
    • Working with Source Connectors
    • Working with Destination Connectors
    • Working with Workflows
    • Working with Jobs
  • License

Installation

pip install unstructured-platform-sdk

Usage

Initialize the client

from unstructured_platform_sdk import UnstructuredPlatformClient

client = UnstructuredPlatformClient(api_key="your-api-key")

Working with Source Connectors

from unstructured_platform_sdk.models import PublicSourceConnectorType

# List source connectors
sources = client.sources.list()

# Create a source connector
source = client.sources.create(
    name="My S3 Source",
    type=PublicSourceConnectorType.S3,
    config={
        "bucket": "my-bucket",
        "prefix": "documents/",
        "aws_access_key_id": "your-access-key",
        "aws_secret_access_key": "your-secret-key",
    },
)

# Get a source connector
source = client.sources.get(source_id="source-id")

# Update a source connector
updated_source = client.sources.update(
    source_id="source-id",
    config={
        "prefix": "new-prefix/",
    },
)

# Delete a source connector
client.sources.delete(source_id="source-id")

Working with Destination Connectors

from unstructured_platform_sdk.models import PublicDestinationConnectorType

# List destination connectors
destinations = client.destinations.list()

# Create a destination connector
destination = client.destinations.create(
    name="My Elasticsearch Destination",
    type=PublicDestinationConnectorType.ELASTICSEARCH,
    config={
        "host": "https://my-es-cluster.com",
        "index": "documents",
        "api_key": "your-es-api-key",
    },
)

# Get a destination connector
destination = client.destinations.get(destination_id="destination-id")

# Update a destination connector
updated_destination = client.destinations.update(
    destination_id="destination-id",
    config={
        "index": "new-index",
    },
)

# Delete a destination connector
client.destinations.delete(destination_id="destination-id")

Working with Workflows

from unstructured_platform_sdk.models import (
    WorkflowAutoStrategy,
    WorkflowSchedule,
    CronTabEntry,
)

# List workflows
workflows = client.workflows.list()

# Create a workflow
workflow = client.workflows.create(
    name="My Daily Workflow",
    source_id="source-id",
    destination_id="destination-id",
    workflow_type=WorkflowAutoStrategy.BASIC,
    schedule=WorkflowSchedule(
        crontab_entries=[
            CronTabEntry(cron_expression="0 0 * * *"),  # Daily at midnight
        ],
    ),
)

# Get a workflow
workflow = client.workflows.get(workflow_id="workflow-id")

# Update a workflow
updated_workflow = client.workflows.update(
    workflow_id="workflow-id",
    name="New Workflow Name",
)

# Delete a workflow
client.workflows.delete(workflow_id="workflow-id")

# Run a workflow
workflow = client.workflows.run(workflow_id="workflow-id")

Working with Jobs

# List jobs
jobs = client.jobs.list()

# Get a job
job = client.jobs.get(job_id="job-id")

# Cancel a job
client.jobs.cancel(job_id="job-id")

License

MIT