DocAI helps developers quickly build document, image and text processing pipelines using open source and cloud-based machine learning models for a wide range of applications.

🚧 DocAI is still a work-in-progress and undergoing early development.

Our goal is to put modern machine-learning technology in the hands of the 20+ million developers in the world. If this excites you, we are looking for early adopters to come along for the ride!

Python >= 3.7

To install DocAI with pip:

pip install docai-py

• Install poppler

• Install poppler-utils via your package manager

Please follow the installation procedure and then run the following:

from docai import PredictionClient



# Get API Key from https://docs.butlerlabs.ai/reference/uploading-documents-to-the-rest-api#get-your-api-key

api_key = '<api-key>'

# Get Queue ID from https://docs.butlerlabs.ai/reference/uploading-documents-to-the-rest-api#go-to-the-model-details-page

queue_id = '<queue_id>'

# Path to a local JPEG, PNG, or PDF file

local_file_path = 'example.pdf'



extraction_results = PredictionClient(api_key).extract_document(queue_id, local_file_path)

print(extraction_results)

Install poetry on your host machine

poetry install

To regenerate code updates to REST API:

openapi-python-client update --url https://app.butlerlabs.ai/api/docs-json --config codegen.yaml

To run all unit tests:

poetry run pytest -v -m unit_tests

Alternatively, you can also use VSCode's "Testing" tab to run/debug individual tests

To add a new pip package dependency, see poetry add.

For versioning, it is best to use the minimum version that works, combined with ^, ~, or >= and < checks.

For example:

• poetry add my-package@^1.2.3 is a shorthand for >=1.2.3,<2.0.0

• poetry add my-package@~1.2.3 is a shorthand for >=1.2.3,<1.3.0

• poetry add "my-package>=1.2.3,<4.5.6"

For development only dependencies, make sure to include the --dev flag.

# setup for testpypi

poetry config repositories.testpypi https://test.pypi.org/legacy/

poetry config pypi-token.testpypi <testpypi token>



# setup for pypi

poetry config repositories.pypi https://upload.pypi.org/legacy/

poetry config pypi-token.pypi <pypi token>

Update pyproject.toml and docai/__init__.py to have a new version number

# build packages

poetry build



# upload to test pypi

poetry publish -r testpypi



# test install from test pypi

pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple docai-py



# upload to real pypi

poetry publish -r pypi