Question Generator

Learning to generate questions from text.
Blog on this project :
Link1 : https://software.intel.com/en-us/articles/using-natural-language-processing-for-smart-question-generation
Link2 : http://dynamichub.in/aditya/sqg/

Strategy

• Sentence Selection: This module selects topically important sentences from text document.
• Gap Selection: This module uses Standford Parser extract NP(noun phrase) and ADJP(Adjective Phrase) from important sentences as candidate gaps.
• Question Formation: This module generate actual questions from the fill in the blank type of question. It uses the NLTK parser and grammar syntax logics for the same.
• Question Classification: Classify question quality based on pre-trained SVM classifier (Conditional trained only for Blank type questions)

Build

Build Project

Some details about the project has also been mentioned in procedure.txt file which lies in the home directory itself.

Install Python2.7`in your system

git clone https://github.com/adityasarvaiya/Automatic_Question_Generation.git

cd Automatic_Question_Generation 

pip install -r requirements.txt

if you have problem with dotenv package then uninstall dotenv and install python-dotenv

pip install nltk
python 
import nltk
nltk.download("punkt")
nltk.download("stopwords")
nltk.download("averaged_perceptron_taggepython r")

Build Stanford Parser & NER

• Create a folder to host all the stanford models, e.g. mkdir /your-path-to-stanford-models/stanford-models.
• Download Stanford Parser at here, unzip, and:
  • Move stanford-parser.jar to stanford models folder, e.g. /your-path-to-stanford-models/stanford-models/stanford-parser.jar
  • Move stanford-parser-x-x-x-models.jar to stanford models folder.
  • Unzip stanford-parser-x-x-x-models.jar, move /edu/stanford/nlp/models/lexparser/englishPCFG.ser.gz to stanford-models/
• Download Stanford NER at here, unzip, and:
  • Move stanford-ner.jar to stanford models folder.
  • Move stanford-ner-x-x-x.jar to stanford models folder (e.g. 3.7.0).
  • Move /classifiers/english.all.3class.distsim.crf.ser.gz to stanford models folder.

The stanford models folder should looks like this:

- stanford-models/
    | - stanford-parser.jar
    | - stanford-parser-x-x-x-models.jar
    | - englishPCFG.ser.gz
    | - stanford-ner.jar
    | - stanford-ner-x-x-x.jar
    | - english.all.3class.distsim.crf.ser.gz

Environment Variables

Create environment variable file with: touch .env for configuration (in project root).

SENTENCE_RATIO = 0.05 #The threshold of important sentences

STANFORD_JARS=/path-to-your-stanford-models/stanford-models/
STANFORD_PARSER_CLASSPATH=/path-to-your-stanford-models/stanford-models/stanford-parser-x.x.x-models.jar

STANFORD_NER_CLASSPATH=/path-to-your-stanford-models/stanford-models/stanford-ner.jar

Important Variables

[embed] https://github.com/adityasarvaiya/Automatic_Question_Generation/blob/master/project.pdf [/embed]