DocuDevs Python Notebook Examples

Interactive Jupyter notebooks demonstrating how to use DocuDevs for document processing.

Prerequisites

1. Install the SDK:

   pip install docu-devs-api-client pydantic

2. Get an API key from app.docudevs.ai

3. Set your API key:

   export DOCUDEVS_API_KEY="your-api-key-here"

Notebooks

Notebook	Description
01-basic-extraction.ipynb	Extract structured data using prompts and Pydantic schemas
02-map-reduce.ipynb	Process long documents (50+ pages) with chunking
03-knowledge-search.ipynb	Enrich extractions with your reference data
04-operations.ipynb	Error analysis and follow-up questions

Quick Start

import os
from docudevs import DocuDevsClient
from pydantic import BaseModel

client = DocuDevsClient(token=os.getenv("DOCUDEVS_API_KEY"))

class Invoice(BaseModel):
    invoice_number: str
    date: str
    total: float

# Process a document
with open("invoice.pdf", "rb") as f:
    job_id = await client.submit_and_process_document(
        document=f.read(),
        document_mime_type="application/pdf",
        schema=Invoice.model_json_schema()
    )

# Get the result
result = await client.wait_until_ready(job_id, result_format="json")
invoice = Invoice(**result)
print(f"Invoice {invoice.invoice_number}: ${invoice.total}")

Running the Notebooks

# Install Jupyter if needed
pip install jupyter

# Start Jupyter
jupyter notebook

Then open any notebook and run the cells. Make sure to set your API key in the first setup cell.

Support

• Documentation
• API Reference