Collection of OCR-related python tools and wrappers from the OCR-D team

To bootstrap the tool, you'll need installed (Ubuntu packages):

• Python (python)
• pip (python-pip)
• exiftool (libimage-exiftool-perl)
• libxml2-utils for xmllint (libxml2-utils)

To install system-wide:

make deps-ubuntu deps-pip install

To develop, install to a virtualenv

pip install virtualenv
virtualenv --no-site-packages venv
source venv/bin/activate
make deps-pip install

pyocrd installs a binary ocrd that can be used to invoke the processors directly (ocrd process) or start (development) webservices (ocrd server)

TODO: Update docs here.

Examples:

# List available processors
ocrd process

# Region-segment with tesserocr all files in METS INPUT fileGrp
ocrd process -m /path/to/mets.xml segment-region/tesserocr

# Chain multiple processors
ocrd process -m /path/to/mets.xml characterize/exif segment-line/tesserocr recognize/tesserocr

# Start a processor web service at port 6543
ocrd server process -p 6543
http PUT localhost:6543/characterize url==http://server/path/to/mets.xml

Download ocrd-assets (make assets)

Test with local files: make test

Test with local asset server:

• Start asset-server: make asset-server
• make test OCRD_BASEURL='http://localhost:5001/'

Test with remote assets:

• make test OCRD_BASEURL='https://github.com/OCR-D/ocrd-assets/raw/master/data/'

• OCR-D Specifications
• pyocrd wiki