Explicity | Devpost

Inspiration

Twitter toxicity and trolls on the Internet have gotten too normalized in our online experiences. We wanted to create a product to combat that.

What it does

Running the user's specific text input through a Chrome extension, it will count how many explicit words or phrases are present within and report the analytics back to the user.

How we built it

The Python script, stored on a Flask server, reads from the locally stored explicit words text file specifically curated using existing databases of banned or offensive words. On the front-end, we created the Chrome extension that gathers the user input for the Python to compare against database.

Challenges we ran into

Being first time hackers, we had trouble ideating and planning the design due to the limited time difficulty of the first iteration of the idea. Additionally, we had very limited experience developing in HTML, Javascript, and Python, and often ended up learning on the job. One of the largest challenges we faced was in connecting the frontend and backend, as Chrome's recent update to Manifest V3 meant that many solutions we found online was out of date due to being for V2. Additionally, many V3 solutions were very complex and had little documentation, meaning that we had to reverse-engineer the solutions based on the provided code snippets.

Accomplishments that we're proud of

I think one of the things we're most poud of is getting both the frontend and backend to work. Despite having trouble establishing communication between the frontend and backend, both parts individually serve their purpose as intended.

What we learned

Each of us worked on the code, so we walked away with improved knowledge of Python, HTML, CSS, and Javascript, while also learning a little bit about Flask development for web apps.

What's next for Explicity

Although we were able to get a basic filter via a regex, there were some factors that we considered but didn't have the time to guard against. For example, since our filter is space-sensitive, a user would bypass it by separating each character with a space or other special character. Additionally, actually considering which text to filter is often context-sensitive, a task that is too subjective to be modeled by a regex generator. As such, we plan to scale up the scope by incorporating NLP models so it can adapt to user's creativity in attempting to bypass it. Additionally, due to uncertainty around Manifest V3 we could refactor the project to instead serve as a plugin for messaging platforms and social media.

Built With

Submitted to

HackTX 2022

Created by

I worked on the front-end. I worked on the Chrome extension, specifically the HTML and CSS file. I also helped get the front-end and back-end get connected.

Reese Welty
I worked on the Python Flask backend. Specifically, I worked on the text parser and the explicit word counter. That being said, Flask is hard and so is RegEx but I managed to figure half of it out somehow.

David Lu
I worked on the Chrome Extension and other front-end development requirements. Specifically, I worked on the manifest file and getting the extension connected to the backend.

Lizzie
I helped out a bit on the Flask part of the backend. Mostly getting the connection between the client and the host for our backend and getting the right output directly from our Python script.

Remzy Alvarado