FlowZen
- 4 Devlogs
- 13 Total hours
An AI focus tracker where it tracks how good those late night study sessions actually were, by using your webcam to see if your there or your phone is there, and seeing what apps are open.
I worked on 2 things this time. One, I used Ollama to use a model, like Qwen 0.7b, a lightweight LLM model, to process the app, title, and study goal and then it would return S for safe, A for ask before, and D for distracted. The decisions for A also get stored in a cache with SQLite. Then, I made a rough program that will get the app name and title using windows and return it.
I have completed the beginning steps. I have first made a UI for it, show below using something called PySide Designer, where you can use drag and drop to make a UI to your liking. I then connected them to my code, like the timer label to the timer, the webcam label to the webcam, etc. Then I fixed some timer issues, improved the pause functionality, and made a total and distraction time length. Lastly, I made a database using SQLite, and I made a save dialog that can save it to the database.
PS: The webcam is blank because I did not want to show my face.
I polished the face detection, so it has a timer on for how long you were studying. I also added ultralytics object detection which detects phones for me at this early stage. Now I am working on the UI with the drag and drop PySide QT Widgets.
I finished working on my face detector using media pipe. This is very early stages as it can only detect if there was a face found and then it will print it out.
