These first weeks we focused on the HCI lectures.
We had our first meeting with our client and gathered some initial
requirements. Using these requirements, we completed our HCI report
which contains some initial sketches, prototypes and design
information.
We also done some reading on some of the algorithms we could
potentially use.
Once again we met with out client, this time to dicuss possible
solutions and to consolidate the requirements.
We received a dataset of scholoary articles from the client which we
will base our model on.
These weeks we were mainly focused on transforming and filtering the
data so that we could use it effectively in our model.
During our client meeting, we discussed the trade-offs of the models
that we would potentially be using, we then used a simple linear
regression model as a baseline and as a means to test the data.
We then considered alternative models to improve on the accuracy we
get using linear regression.
To connect the models we trained with the web app, we first needed to implement
the logic in the backend.
We decided on using Python for both the frontend and backend to make the web app
implementation easier and less prone to bugs.
We also started to draft out the overall architecture of the system. This was especially
since Python is not a statically-typed language, and poor planning would lead
to huge time spent on debugging.
The implementation of the backend was mostly smooth.
A key decision made during implementation was to make it easy to add or modify models. This is due to the
fact that better models can always be produced, and this makes the web app much more flexible.
Implementation took up more time than expected, and testing was pushed to the next week.
These weeks were mainly focused on researching about flaskr and implementing frontend. Along with
finishing off the backend with testing
We tried to finish off the implementation for backend. We added the search feature in the app where
the DOI is passed through an API which gives details corresponding to that particular DOI.
We finished off researching about flaskr and learning the basics, we started implementing the
frontend
We set up and the home page and the analysis page. The home page was finished first as the home page
is redirected to the analysis page.
These weeks were mainly focused on finishing off of the frontend of the web
app and testing.
We had code review this week to get feedback from our client and his edits. The UI
had an upload file options which was optional but the client suggested to add it as
a feature to upload file ( with the abstract).
We also tested our web app.
We added some unit tests to check whether the function we added are giving the
expected output.