JohnSnowLabs/spark-nlp-workshop

Java example is not runnable - no pom.xml or build.gradle, missing imports and more

almogtavor opened this issue · 3 comments

almogtavor commented

Description

The Java example only contains one file. It starts with a DocumentAssembler instance creation, but without any import of the DocumentAssembler class. There is also an import on the EmbeddingHelper class which is not supported on the "com.johnsnowlabs.nlp:spark-nlp_2.12:3.1.2" dependency.
image

The folder doesn't contain any gradle.build file or pom.xml, which makes the process much harder (and makes me wonder if the class was even tested).
I'd like to get an explanation of how to run this because the documentation also doesn't contain any Java references.

maziyarpanahi commented

@xusliebana It would be better to have a full Maven project as we did for the SBT project for Java example: https://github.com/maziyarpanahi/spark-nlp-starter

I understand those examples are for users to be familiar with the Java syntax, but I think most Java developers appreciate an end-to-end Java project with Spark NLP.

  • 👍2
xusliebana commented

Ahh ok I had the Gradle project but I need to set up manually the spark-internal-library. I can create a readme to show the proccess

  • 👍1
almogtavor commented

@maziyarpanahi said a correct thing to my opinion. It would be much more comfortable if there would be an end to end Java project of Spark NLP. I didn't find one in Github at all.

  • 👍2