cspythoncoder

Work Completed, I improved the UI, made it smoother and minimalistic(our school manager likes minimalist things), also added a admin dashboard. To access the admin dashboard, go to https://alumni-network-sms.vercel.app/admin. For now the admin DASHBOARD is admin@1243

will change the password later and for testing the website the details are attached in the screenshot below

Big Work Done. This project would be one of the quickest ones I’ve ever done as it needs to be submitted to my school tomorrow…

I’ve made the basic website, initialized next.js etc etc. Also connected the Convex Database instead of using .csv which is insecure…
The basics are done. Now I think I’ll just add a admin dashboard, create dev and prod data bases for testing and production. And then finally host it on vercel…

Made a small change what I did is I tinkered with some Text To Speech things and gonna use edge-tts. Also now I added an integration with Discord. Now I can access my chatbot from anywhere using my discord server(on phone, another laptop etc…)

If many ppl are chatting at the same time, the first message is treated first and others are queued. That’s it as of now. Next up, adding memory

Okay just a sample devlog bcz I Had to resubmit.

The use of Python3.11 is stricty required as its the most stable version mediapipe runs on please use python3.11

Alright after many errors, finally the web based speech recognition is ready. This basically uses selenium to scrape the recognized text continuously. Also Status.data and Mic.data. The UniversalTranslator function uses mtranslate to translate any language to english.

This is the fastest speech recognition method I could think of. This is free, very fast and efficient. This is web based speech recognition and the accuracy is unmatched. Vosk was a good offline alternative but this is more accurate than Vosk + runs offline too…

I packaged this project as a pip now…

pip install hand3d

You can install the package and run the code block below to run my project and all set!
import hands
hands.hands()

Alrightyy, Big work done in the past few hours. Added the drag system and refined it. Added a grid floor so that motion of the cube(drag) can be perceived(motion is relative folks) + its better with the grid. Changed the cube color to blue as white was kinda looking odd with the white grids. New drag gestures are: DRAG XY: Just close your fist and move your hand DRAG Z/Depth: Pinch(thumb and index finger) and move hand up to pull the cube towards you. Down to push it away from you… That’s pretty much it as of now…

There is more to come but for testing purposes, I added a 3D cube and then added 3D model interaction and zoom logic for the 3D model. Now I can zoom perfectly and also I added rotate…

Added Zoom and made it very much smooth, responsive and accurate. Now the Zoom works faster as it just uses the distance between two index fingers.

As of now I’ve added a textbox kinda thing where you can view what action is being done. As of now I only have two actions… Drag and Rotate. Rotation can be done by only showing your index finger. Drag can be done by keeping your hand in the dragging position…

Okay the stepping stone of my project, installed mediapipe, set up my coding environment(I had some issues with my hackatime on MacOS so I have literally change my OS now. I’m using Zorin OS). But anyways. Now its all set, everything smooth. I tinkered a lot with mediapipe to get to know about all its features(basically learnt it a bit by documentation and a bit using Claude) and now im done learning mp. I’ve implemented tracking for both the hands with a green tip to the index finger because the index finger would be used a lot! Like a Lot in the future features and that’s pretty much it!

I’ve made a LOGO for my Chatbot. I wanted it to be minimalistic yet great to look at…