rdkx5-app-resources_skill

Overview

This skill provides direct access to the /app folder resources on RDK X5 developer boards, exposing GPIO, character device demos, multimedia tools, ISP tuning, and Python AI samples. Use it to run embedded demos, control hardware pins, capture and process video, or run prebuilt AI inference examples. It emphasizes using system Python and appropriate permissions for hardware access.

How this skill works

The skill maps to ready-made scripts and C/C++ demos stored under /app (40pin_samples, cdev_demo, multimedia_samples, pydev_demo, tuning_tool). It launches system Python examples, builds and runs native demos, and calls ISP tuning tools. It does not modify system configuration; it provides entry points and guidance for executing the bundled resources on-device.

When to use it

• Run GPIO demos to test LEDs, buttons, PWM, I2C, SPI, or UART
• Execute Python AI samples (YOLO, SSD, segmentation) for on-device inference
• Build and run C/C++ multimedia or character device demos for camera/video pipelines
• Launch ISP tuning tools to tune camera parameters on the board
• Quickly validate hardware pins and peripheral communication during development

Best practices

• Always run Python samples with the system Python (/usr/bin/python3.10), not a conda interpreter
• Run hardware demos as root or with proper hardware permissions to access GPIO, cameras, and device nodes
• Call GPIO.cleanup() at the end of GPIO scripts to release pins safely
• Use BOARD or BCM pin encoding consistently and verify pin mapping before driving hardware
• Check /app/model/basic for pretrained models and verify model paths used by AI samples

Example use cases

• Toggle LEDs or control motors via /app/40pin_samples simple_out.py and simple_pwm.py
• Run YOLOv5 detection on a USB or MIPI camera using /app/pydev_demo/07_yolov5_sample/main.py with system Python
• Build and test video input/output pipelines by compiling multimedia_samples and running sample_vin/sample_vot demos
• Tune camera ISP parameters interactively with /app/tuning_tool/run_tuning.sh
• Stream an RTSP feed to display or decode it using cdev_demo examples (rtsp2display, decode2display)