NLTK (Natural Language Toolkit) is a well-known platform for Python application dealing with human
language data. It includes many downloadable lexical resources (named corpora). If your
application requires some corpora to work, add a
nltk.txt file at the root of the
application containing a corpora name per line. For instance:
The Python buildpack will automatically download them during the deployment of your application.
Note that there is a known issue on nltk which leads to an unharmful warning message when your app starts:
RuntimeWarning: 'nltk.downloader' found in sys.modules after import of package 'nltk', but prior to execution of 'nltk.downloader'; this may result in unpredictable behaviour
You can safely ignore it.