This repo is a template for setting up a simple Question-Answering application using Django.
To get the application set up you will only need to be concerned with two files:
qa_gui/apps.py
: This is where the classes and the models are initated.qa_gui/code.py
: Here is where the actual algoritms are executed.
This application considers 3 methods:
- Extract the paragraphs from the documents/unstructured data. Here, this will consist of two functions that should be in
qa_gui/apps.py
. - Find the paragraphs best suited to the question given. This will require a model to ge a representation of the paragraphs and the question.
The initialization of the the vectorizer should be in
qa_gui/apps.py
and could also use the paragraphs. The function which will actually match the question to the paragraphs should be inqa_gui/code.py
. - Extract the answer given the paragraph and question. Here again the initialization of the QA model should be in
qa_gui/apps.py
and the function used to extract the answer should be inqa_gui/code.py
.