DEEPREAD Extract - AI OCR - RapidAPI Samples

This repository provides an example Python script along with some sample images to test the DEEPREAD Extract - AI OCR RapidAPI endpoints.

Installation

The easiest way to run samples is within a virtual environment using Python3.

python3 -m venv .env
source .env/bin/activate
pip install -r requirements.txt

Once all the requirements are installed the samples can be processed.

Command line args

Required params:

-k|--key: X-RapidAPI-Key header required to access rapidapi.

One of these are required:

--all: when selected, all images in samples/ folder will be process.

-f|--file: to specify a specific file you want to process.

Required for English analysis when specifying individual file:

-p|--process-type: to specify a process type for analysis. in set (form, invoice, receipt).

Other args:

--vis: when selected, a visualisation output will be created.

-l|--language: ACCEPT-LANGUAGE value passed to rapidapi/deepread. (default en)

-h|--help: output details of command line inputs.

Usage

To process a specific image (notice how process-type argument is required): python run_extract_samples.py -k <X-RapidAPI-Key> -f samples/form/DummyApplicationForm.jpg -p form

To process a specific image with a visualisation output created: python run_extract_samples.py -k <X-RapidAPI-Key> -f samples/form/DummyApplicationForm.jpg -p form --vis

To process all images in samples/ folder. python run_extract_samples.py -k <X-RapidAPI-Key> --all

To process a specific image in japanese (notice how process-type argument is not allowed): python run_extract_samples.py -k <X-RapidAPI-Key> -f samples/invoice/invoice-sample-ja.png -l ja

Samples

The /samples directory contains sample files for processing. Running the script with the --all argument will process all files in this directory. The directory is organized into subdirectories using the process-type as their name - this is how the process-type is inferred.

Outputs

The script outputs are all sent to the outputs/ folder.

There are two outputs for every image processed:

i. outputs/<process-type>/<filename>.json: json output returned by DEEPREAD Extract via RapidAPI.

ii. (optional) outputs/<process-type>/<filename>.<image extension>: a visualisation of the processed images with side-by-side comparison. The visualisation places all of the extracted text in the approximate location of the bounding boxes found over a blank image. If the image requested for processing is a pdf the output visualisation will be created as a jpeg as it needs to be converted for visualisation.