Skip to content

Latest commit

 

History

History
142 lines (89 loc) · 4.89 KB

README.md

File metadata and controls

142 lines (89 loc) · 4.89 KB
Raw

Document AI Modular Application

This Document AI application is an end-to-end application that shows how to integrate Document AI into full-stack application. This is the starting point and uses modular architecture so developers could easily demo Document AI and add components that fit their needs to quickly create a tailored demo.

This application uses a Flask backend server to handle Document AI API calls and a Angular Frontend.

Quickstart

Before clicking make sure Document AI API is enabled and your project has at least one processor created.

This application currently support OCR, Invoice and Form processors.

Run on Google Cloud

Prerequisites

Before you can run the application you first need to make sure you have the Google CLI installed.

once you have the CLI installed run

 gcloud init

This will create a empty project with billing not setup. To use the APIs this application needs you will need to setup your billing information

Once billing is enabled and setup you can then enable the Document AI API by running

gcloud services enable documentai.googleapis.com

You will need to create processors inorder to use this application, currently this application supports OCR, Form and Invoice processors.Follow these instructions to create the processors you need.

Now you can setup the application default authentication by running

gcloud auth application-default login

Running Application

Now that you have your project setup and you're logged into your Google Cloud project we can run the application. To run the application you will need to first start the Backend server and then the Frontend server. Follow the steps below.

Running Backend Server

Make sure you're in the Backend directory

cd Backend

When you're in the Backend server directory you'll need to install all the dependencies in a virtualenv using pip.

virtualenv is a tool to create isolated Python environments.

With virtualenv, it's possible to install this library without needing system install permissions, and without clashing with the installed system dependencies.

Mac/Linux

    pip install virtualenv
    virtualenv <your-env>
    pip install -r requirements.txt

Windows

    pip install virtualenv
    virtualenv <your-env>
    <your-env>\Scripts\pip.exe install -r requirements.txt

Once the dependencies are installed you can then run the Backend server.

Before running the server make sure the location variable is changed to the location of your project (us or eu) and you set the FLASK_APP environment variable

export FLASK_APP=main.py

Then you can run the flask application.

python -m flask run

Now that you have the Backend server running you can now start the Frontend server.

Running Frontend Server

Make sure you're in the Frontend directory and you've opened a new terminal since we want both servers to run at the same time.

cd Frontend

When you're in the Frontend server directory you'll need to install all the dependencies

npm install

Once the dependencies are installed you can then run the Frontend server.

To start a local development server run

ng serve

The app will automatically reload if you change any of the source files in the Frontend directory.

Navigate to http://localhost:4200/ and now you should now be able to see the application.

Document AI Modular E2E App

Building the project

Run ng build to build the project. The build artifacts will be stored in the dist/ directory.

Components

The following is a list of existing components and their functionality

  • Base Layer - Base Layer holds all imported components
  • Canvas - This component handles document display and annotation
  • Entity Tab - This component extracts text from the Document Proto and displays the text and highlights the bounding boxes depending on selection
  • Processor Selection - This component retrieves available processors and displays them, allows for processor selection, validates inputs of processors and the uploaded PDF, and sends the uploaded PDF for document processing
  • Upload File - This component handles document upload, updates the data service with the uploaded file, and validates that the uploaded file is a PDF and less them 20mb
  • Data Sharing Service - This service handles the sharing of data between different components
  • Document Annotation Class - This class handles the bounding box drawing and highlight

Contributing