Eclipse Lamp Devlog #3 - Testing Mechanical Iris Code post 3D-printing and trying to finish Person Detection AI software for Raspberry Pi _________________________________________________________________________

Today I made somewhat strong progress in two areas but as always i sold on the software side lol. First, on the mechanical side, I tested the mechanical iris with an SG90 servo by wiring it to an Elegoo Uno to validate the Arduino code, which successfully controlled the opening and closing after uploading an existing sketch. I then translated the code to Python, wired it to a Raspberry Pi, installed the required libraries, and confirmed the iris worked the same way, completing the mechanical system. Luckily this made me failing on my AI seem easier to deal with. Working on this I genuinely insane and started raging on ChatGPT after making me do the same thing 20 times. I continued the person detection AI setup on the Raspberry Pi camera by confirming libcamera and Picamera2 were working, building a Flask live video stream that displayed real-time footage in the browser, and testing YOLO (Ultralytics), but I kept running into environment and dependency issues like missing modules, virtual environment conflicts, and broken installs that stopped full AI detection from running, and ChatGPT was just so bad at helping, so the next step is fixing the Python environment and system dependencies to properly run YOLO on the live camera feed.