khaack / snlp-miniproject

SNLP-miniproject WS 2021/2022

Fact checker based on wikipedia corpus, StandfortNLP, TagMe and WordNet.

Steps for running the application

1. start docker container
2. run spring
3. 'access via UI' OR 'access the demos (training & test)'

pre requirements for the development environment

• java 11
• docker
  • docker-compose
  • maven 3.1 or higher

start docker container

Starts the docker containers necessary for the fact check.

docker-compose up

run unit tests

Maven command, for running the unit tests.

mvn clean test

run spring

Starts the spring boot application.

mvn spring-boot:run

access via UI

Access the UI via browser (with enabled js).

• upload => GET http://localhost:8080/upload
• factCheck => GET http://localhost:8080/factCheck

access the demos (training & test)

Run the the fact checking with the training & test data files of the execise.

• factCheck training file (.ttl-output) => GET http://localhost:8080/demo/runFileTraining
• factCheck test file (.ttl-output) => GET http://localhost:8080/demo/runFileTest
• factCheck training file (.html-output) => GET http://localhost:8080/demo/runFactCheck
• ner training file (.html-output) => GET http://localhost:8080/demo/runNer

api access

Internal API access.

• ping => GET http://localhost:8080/api/ping
• upload => POST http://localhost:8080/api/upload
• factCheck => GET http://localhost:8080/api/factCheck?sentence={text}

access wordnet

Internal WordNet API access.

• GET http://localhost:5679/synonyms/1/{word}
• GET http://localhost:5679/hypernyms/1/{word}
• GET http://localhost:5679/substance_meronyms/1/{word}
• GET http://localhost:5679/hyponyms/1/{word}
• GET http://localhost:5679/antonyms/1/{word}
• GET http://localhost:5679/substance_holonyms/1/{word}
• GET http://localhost:5679/causes/1/{word}

Controller explanation

• AlkApiController: for the API calls, to running a "real" fact check
• AlkController: for delivering the UI
• DemoController: for running the test & training datasets and provide a download of the result

licenses

• GNU v3 https://github.com/stanfordnlp/CoreNLP/blob/main/LICENSE.txt
• Apache License 2.0 https://github.com/gammaliu/tagme/blob/master/LICENSE
• WordNet 3.0 https://wordnet.princeton.edu/license-and-commercial-use