brendan-w
Time interval on Async loop.
An obd software is sometimes used on devices that require low energy consumption, so the cpu cycles are somewhat important. However, it appears that Async retrieves the data on a thread which cycles at full cpu speed. If a developer needed to limit this speed, he/she needs to override this run method on his script, but if this supplied as a parameter at init of Async, then used with time.sleep() (Or whatever method you think best), I think it would prove useful.
Yes, I've been catering to the "make it as fast as humanly possible" crowd, since OBD is notoriously slow. Shouldn't be too hard. Probably make it a keyword arg like interval or something. Would happily look at PRs.
I haven't checked it myself, but it shouldn't be consuming 100% cpu, since it's spending non-trivial time blocking on IO (I think best case is a ~90 msec turnaround time). Has this been giving you grief? If so, I'm curious to know how egregiously python-OBD ate the CPU.
I haven't tested Async in the car. Currenty I am using the regular OBD class because the blocking did not causing problems for now. However I did check how many calls do the loop make. It made 5 per second for a single command. It should not be cpu consuming, but I don't know properties of the ELM device fully. If it can keep up with such speed, there is no problem. Btw, seeing you online, I will add some other issues I encountered.
Tested Async in the car. I made a gui for myself using pygame. With all the drawing functions(with anti-aliasing) and Async, the cpu usage percentage in the Raspberry Pi was 25-60%. I think those cpu spikes caused by the Pygame, not Async. ELM device also seems to keep up with the communication speed. So everything is perfect. However, I think an interval parameter should still be added for those who may want to use it.