I have been tinkering with my WFH project for the last few weekends. For those that are not aware, my WFH project has two main goals. First, I'm automating small tasks to improve my work from home experience, and secondly, it's an opportunity to improve my Python coding skills. I did not start out as a developer, but I often joke with my team that we're all developers with a focus on infrastructure and deployments.
My recent efforts have really been focused on making incremental improvements in my environment. I'm using recycled hardware for my dev environment so I do have some limitations. Overall, I'm fairly happy with my code. For example, the WFH image publication on this blog is fairly compute intensive. My current process involves watermarking and publishing one frame from the first video in the set and archiving raw frames from the remaining videos. The entire cycle takes about 15 seconds to process 4 minutes of video. I know some people that are working to beat 7 seconds for a 30 second clip so I'm pretty happy. I have also built some time based automations to control the lights and fans in my WFH office.
As far as improvements go, I did sign up for a free NewRelic account so I can get some baseline numbers across the various transactions. NewRelic is one tool that I already use in my day to day professional activities. I'm also hoping to get some better object detections as well in the future. My WFH background is fairly static, but I occasionally have new detections. I also get the occasional false positive detection so I need to retrain the model in some areas. I'm also tinkering with the idea of publishing the reassembled raw frames as a stop motion clip. I've already done the PoC work for that. It is personal project after all so I'll continue to tinker as I have time.