Skip to main content

Expandable and scalable OCR pipeline

Project description

Overview

https://travis-ci.org/OpenPhilology/nidaba.svg

Nidaba is the central controller for the entire OGL OCR pipeline. It oversees and automates the process of converting raw images into citable collections of digitized texts.

It offers the following functionality:

  • Grayscale Conversion

  • Binarization utilizing Sauvola adaptive thresholding, Otsu, or ocropus’s nlbin algorithm

  • Deskewing

  • Dewarping

  • Integration of tesseract, kraken, and ocropus OCR engines

  • Page segmentation from the aforementioned OCR packages

  • Various postprocessing utilities like spell-checking, merging of multiple results, and ground truth comparison.

As it is designed to use a common storage medium on network attached storage and the celery distributed task queue it scales nicely to multi-machine clusters.

Build

To easiest way to install the latest stable(-ish) nidaba is from PyPi:

$ pip install nidaba

or run:

$ pip install .

in the git repository for the bleeding edge development version.

Some useful tasks have external dependencies. A good start is:

# apt-get install libtesseract3 tesseract-ocr-eng libleptonica-dev liblept

Tests

Per default no dictionaries and OCR models necessary to runs the tests are installed. To download the necessary files run:

$ python setup.py download
$ python setup.py nosetests

Tests for modules that call external programs, at the time only tesseract, ocropus, and kraken, will be skipped if these aren’t installed.

Running

First edit (the installed) nidaba.yaml and celery.yaml to fit your needs. Have a look at the docs if you haven’t set up a celery-based application before.

Then start up the celery daemon with something like:

$ celery -A nidaba worker

Next jobs can be added to the pipeline using the nidaba executable:

$ nidaba batch -b otsu -l tesseract -o tesseract:eng -- ./input.tiff
Preparing filestore             [✓]
Building batch                  [✓]
951c57e5-f8a0-432d-8d77-8a2e27fff53c

Using the return code the current state of the job can be retrieved:

$ nidaba status 25d79a54-9d4a-4939-acb6-8e168d6dbc7c
PENDING

When the job has been processed the status command will return a list of paths containing the final output:

$ nidaba status 951c57e5-f8a0-432d-8d77-8a2e27fff53c
SUCCESS
14.tif → .../input_img.rgb_to_gray_binarize.otsu_ocr.tesseract_grc.tif.hocr

Documentation

Want to learn more? Read the Docs

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

nidaba-0.9.5.tar.gz (1.3 MB view hashes)

Uploaded Source

Built Distribution

nidaba-0.9.5-py2-none-any.whl (1.2 MB view hashes)

Uploaded Python 2

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page