Load Pangea data in Python using dltHub

Build a Pangea-to-database or-dataframe pipeline in Python using dlt with automatic Cursor support.

In this guide, we'll set up a complete Pangea data pipeline from API credentials to your first data load in just 10 minutes. You'll end up with a fully declarative Python pipeline based on dlt's REST API connector, like in the partial example code below:

Example code
@dlt.source
def pangea_prompt_guard_source(access_token=dlt.secrets.value):
    config: RESTAPIConfig = {
        "client": {
            "base_url": "https://redact.aws.us.pangea.cloud/v1/openapi.json#/paths/~1v1beta~1config~1update",
            "auth": {
                "type": "bearer",
                "token": access_token,
            }
        },
        "resources": [
            "responses", "requestBody"
        ],
    }
    [...]
    yield from rest_api_resources(config)


def get_data() -> None:
    # Connect to destination
    pipeline = dlt.pipeline(
        pipeline_name='pangea_prompt_guard_pipeline',
        destination='duckdb',
        dataset_name='pangea_prompt_guard_data', 
    )
    # Load the data
    load_info = pipeline.run(pangea_prompt_guard_source())
    print(load_info)

Why use dltHub Workspace with LLM Context to generate Python pipelines?

  • Accelerate pipeline development with AI-native context
  • Debug pipelines, validate schemas and data with the integrated Pipeline Dashboard
  • Build Python notebooks for end users of your data
  • Low maintenance thanks to Schema evolution with type inference, resilience and self documenting REST API connectors. A shallow learning curve makes the pipeline easy to extend by any team member
  • dlt is the tool of choice for Pythonic Iceberg Lakehouses, bringing mature data loading to pythonic Iceberg with or without catalogs

What you’ll do

We’ll show you how to generate a readable and easily maintainable Python script that fetches data from pangea_prompt_guard’s API and loads it into Iceberg, DataFrames, files, or a database of your choice. Here are some of the endpoints you can load:

  • Redaction Endpoint: Handles the redaction of sensitive data.
  • Configuration Endpoints:
    • Create: Endpoint to create a new configuration.
    • Delete: Endpoint to delete an existing configuration.
    • List: Endpoint to list all configurations.
    • Update: Endpoint to update an existing configuration.
  • Schemas: Various schemas related to redaction and configuration, including:
    • Redact Service Config: Schema defining the configuration for the redaction service.
    • Redact Service Config V2: Updated schema for the redaction service configuration.
    • Redaction Method Overrides: Schema specifying overrides for redaction methods.
    • Validation Errors: Schema for validation errors associated with redaction and configuration requests.

You will then debug the Pangea pipeline using our Pipeline Dashboard tool to ensure it is copying the data correctly, before building a Notebook to explore your data and build reports.

Setup & steps to follow

💡
Before getting started, let's make sure Cursor is set up correctly:

Now you're ready to get started!

  1. ⚙️ Set up dlt Workspace

    Install dlt with duckdb support:

    pip install "dlt[workspace]"

    Initialize a dlt pipeline with Pangea support.

    dlt init dlthub:pangea_prompt_guard duckdb

    The init command will setup the necessary files and folders for the next step.

  2. 🤠 Start LLM-assisted coding

    Here’s a prompt to get you started:

    Prompt
    Please generate a REST API Source for Pangea API, as specified in @pangea_prompt_guard-docs.yaml 
Start with endpoints responses and requestBody and skip incremental loading for now. 
Place the code in pangea_prompt_guard_pipeline.py and name the pipeline pangea_prompt_guard_pipeline. 
If the file exists, use it as a starting point. 
Do not add or modify any other files. 
Use @dlt rest api as a tutorial. 
After adding the endpoints, allow the user to run the pipeline with python pangea_prompt_guard_pipeline.py and await further instructions.

  3. 🔒 Set up credentials

    The authentication process requires obtaining an access token from the token URL /v1beta/oauth/token, which can be found in the API management documentation, and should be applied as specified in the relevant sections of the documentation.

    To get the appropriate API keys, please visit the original source at https://pangea.cloud/docs/api/management/prompt-guard. If you want to protect your environment secrets in a production environment, look into setting up credentials with dlt.

  4. 🏃‍♀️ Run the pipeline in the Python terminal in Cursor

    python pangea_prompt_guard_pipeline.py

    If your pipeline runs correctly, you’ll see something like the following:

    Pipeline pangea_prompt_guard load step completed in 0.26 seconds
1 load package(s) were loaded to destination duckdb and into dataset pangea_prompt_guard_data
The duckdb destination used duckdb:/pangea_prompt_guard.duckdb location to store data
Load package 1749667187.541553 is LOADED and contains no failed jobs

  5. 📈 Debug your pipeline and data with the Pipeline Dashboard

    Now that you have a running pipeline, you need to make sure it’s correct, so you do not introduce silent failures like misconfigured pagination or incremental loading errors. By launching the dlt Workspace Pipeline Dashboard, you can see various information about the pipeline to enable you to test it. Here you can see:

    • Pipeline overview: State, load metrics
    • Data’s schema: tables, columns, types, hints
    • You can query the data itself
    dlt pipeline pangea_prompt_guard_pipeline show dashboard

  6. 🐍 Build a Notebook with data explorations and reports

    With the pipeline and data partially validated, you can continue with custom data explorations and reports. To get started, paste the snippet below into a new marimo Notebook and ask your LLM to go from there. Jupyter Notebooks and regular Python scripts are supported as well.

    import dlt

data = dlt.pipeline("pangea_prompt_guard_pipeline").dataset()
# get "responses" table as Pandas frame
data."responses".df().head()

Extra resources:

Next steps