stanfordnlp_dg_server

This project provides an HTTP server returning an enriched dependency graph returned by Stanford NLP represented as a property graph. This code was extracted from this project coming from Giacomo Bergami's dissertation: https://bitbucket.org/unibogb/sherlock/src

Java

While FileReading.java provides a program for parsing an entire file of sentences, SNLP_Python.java provides a simple entrypoint for all the services (Dependency Parsing and Time Unit extractions). The HTTP service variant is available through the following class: Main.java. This acts as the main entrypoint for the jar generated at compile time. To generate a jar with the full set of dependencies, please run the following command:

mvn clean compile assembly:single

As this uses an older version of the library, please use also an older version for Maven (3.6.3).

Python

This project can be also installed in Python from the remote repository using git at the following command:

pip3 install git+https://github.com/datagram-db/stanfordnlp_dg_server.git@main

If you prefer to download the sourcecode, you can always install it as follows:

pip3 install .

This provides a singleton class that can access all of the aforementioned services without using Java as a service:

from StanfordNLPExtractor.OldWrapper import OldWrapper
nlp = OldWrapper.getInstance()
gsm = nlp.generateGSMDatabase(["Hello, beautiful world!"])
tunits = nlp.getTimeUnits(["Yesterday I was feeling ill"])