This is the repository for the interoperability kit of EBU-TT Live.
The kit is envisaged to contain a set of components for generating, testing and distributing subtitle documents in EBU-TT Part 3 format.
This is an open source project. Anyone is welcome to contribute to the development of the components. Please see the wiki for the list of required components, guidelines and release plan.
The project home page is at http://ebu.github.io/ebu-tt-live-toolkit/ and links to the pre-built documentation.
We have a Slack team called ebu-tt-lit for day to day communications, questions etc. Please join up!
If you would like to contribute or join the Slack team, please contact subtitling@ebu.ch or nigel.megitt@bbc.co.uk
Make sure you have python 2.7+. Make sure you have python virtual environment capability.
If not you can install virtualenv systemwide from your operating system's package repository or by pip:
sudo pip install virtualenv
After that creating a virtual environment should be as simple as:
virtualenv env
Let's activate it (source makes sure the current shell executes the script and assumes the environment variables that the activation script sets):
source ./env/bin/activate
To build the project you will also need node.js. Please read the instructions for your system here.
After having created the python virtual environment, having activated it and having installed node.js the package can be built by typing make if you have GNU build tooling on your system.
make
Alternatively:
pip install -r requirements.txt
python setup.py develop
pyxbgen --binding-root=./ebu_tt_live/bindings -m __init__ --schema-root=./ebu_tt_live/xsd/ -r -u ebutt_all.xsd
npm install nunjucks
node_modules/nunjucks/bin/precompile ebu_tt_live/ui/user_input_producer/template/user_input_producer_template.xml > ebu_tt_live/ui/user_input_producer/template/user_input_producer_template.js
After this you are supposed to be able to launch the command line tools this python package provides i.e.:
ebu-dummy-encoder
Windows is not the best friend of Makefiles. So there is a make.bat file for those who would like to develop using Windows. Assuming python 2.7 and virtualenv is installed and are on the PATH. To build the project you will also need node.js. Please read the instructions for your system here. Then run :
make
This will make sure a virtual environment is created and activated and installs all the tools into it.
After that the following command should work:
ebu-dummy-encoder
The schema definitions are to be found embedded in the Python library in the xsd1.1 subfolder. The root schemadocument is called ebutt_live.xsd.
The library uses XSD schemas from the xsd1.1 subdirectory. The bindings will keep the validation sane and PyXB makes sure that updates are working as expected. Should the schema be modified a regeneration can be run and the bindings will respect the changes.
There are several scripts that emulate different components (nodes) in the infrastructure. They can be executed individually or in combinations by running ebu-run. Assuming the Makefile worked, the package is installed in a virtual environment and the virtual environment is active, the components should be available by running the ebu-run script and passing a configuration file. There are several example configuration files in examples/config. For the complete list see of scripts see docs/build/html/scripts_and_their_functions.html.
Below is a list of some of the key components. .
The simple producer is the beginning of the data pipeline. It generates EBU-TT-Live documents in a timed manner. In the repository root there is a test.html file that can be used for manual testing of the producer in any websocket capable browser. To run it use ebu-run:
`ebu-run --admin.conf ebu_tt_live/examples/config/simple_producer.json`
The simple consumer connects to the producer or later on in the pipeline, assuming there are more components inserted.
ebu-run --admin.conf ebu_tt_live/examples/config/simple_consumer.json
The User Input producer is a web page with a user interface that allows you to send subtitle documents and view the output of a downstream node. For complete documentation see docs/build/html/user_input_producer.html.
To run a configuration of components, use a configuration file with multiple nodes defined. For example, this will create 3 nodes: a distributer that listens to the UIP and two consumers that subscribe to the distributer:
ebu-run --admin.conf ebu_tt_live/examples/config/user_input_producer_dist_consumers.json
Go straight to the pre-built documentation for the current master branch.
The documentation framework uses the popular Sphinx documentation generating engine and autodoc plugins to give developers the flexibility of writing Extra documentation interleaved with the autogenerated documentation created by autodoc.
To display the images in the documentation, you need to have Graphviz installed and make sure the dot executable is on the PATH. For example, for users of homebrew:
brew install graphviz
Documentation can be generated based on the sources in the docs/source directory. After having installed the packages in requirements.txt (which is done automatically by the make command) documentation can be generated by one of the following three ways:
1 Calling setuptools
python setup.py build_sphinx2 Running make in the docs directory where separate makefiles and a make.bat file is giving a variety of options.
cd docs
make html3 Calling the sphinx-build command line script that comes with sphinx. WARNING: Platform-dependent path-separators.
sphinx-build -b html docs/source/ docs/build/htmlAfter sphinx finished with a successful execution log the generated documentation should be accessible by opening the docs/build/html/index.html in any web browser.
The test framework is described in CONTRIBUTING.md
Please refer to CONTRIBUTING.md