The week prior to spring break, our group spent most of efforts in producing our design report since this is a significant part of our capstone. Since most of my work involves knowledge from the class Web Application Development, I spent a lot of time combing through old lectures for React and JavaScript help since it had been a while since I’ve taken the class. The mockups added into the report for a registration page, a login page, and an inventory/main page can be seen below and have been used as inspiration for the actual frontend designs:
The rendering between these pages have been complete in view.py along with the logout button functionality. Users have the option of logging in using Google OAuth as opposed to creating a separate account. Many benefits come with this service, including the elimination of users sharing their passwords to third-party applications by using a token-based authentication system and providing a more seamless user experience that is one click away. While our product is pretty self explanatory to use, we still want our users to have a most simplified user experience as possible. This was set up on the Google Developer Console to configure the OAuth consent screen for how it will be displayed to users. During further testing processes, we will test that the authentication flow works as smoothly as possible with the best security measures.
Since most of the frontend application has been completed, I was able to take some time and start on setting up the database retrieval process on our website using Django. Upon evaluation and comparison of different databases in our design report, we ended up settling on MySQL for many reasons. It all ended up boiling down to its reliability in handling large datasets and high compatibility with web application and Arduino programming. The structure of our database will look similar to the personal inventory page as pictured above, but may be incline to slight changes depending on how we want the item status to be particularly displayed.
I also assisted Steven in gathering data for his ML testing. Specifically, this included feature data that would help distinguish between the packaging between fruits and canned foods, label data for training the SVM classifier, and some other relevant information that was used during the labeling of each food item. For the feature data, various attributes such as visual appearance and material composition were collected to help capture the differences in packaging appearances related to fruits and canned foods. Additionally, the labeling process involved the careful examination of the packaging and verification against established criteria to ensure enough accuracy. This comprehensive dataset served as a strong foundation for Steven’s ML testing and will enable him to develop an effective classification system.