Github user MLnick commented on the pull request:
https://github.com/apache/spark/pull/10272#issuecomment-199700457
IMO more specific or complex domain-specific stuff should live outside of
core, until such time as there is clear demand across a wider user base that
justifies the maintenance cost of including it. Already Spark ML has a large
maintenance & code review burden just with the algos and feature transformers
that are already in there.
The whole point of an API for pipelines is to enable external libraries for
more specific use cases. This is doubly the case when well-known and robust
libraries already provide the functionality. As you can see from your PR,
implementing one's own stemmer transformer using one of the external NLP libs
is a few lines of code.
Things like NLP (and image, video and audio processing, for example) should
start life as a Spark package. How about looking at contributing to
https://github.com/mengxr/spark-corenlp and wrapping the CoreNLP stemmer
functionality as a transformer?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
