Closed Short-bus closed 10 months ago
Next version to be released uses pandas dataframe, it reduces overhead time from 25% to 10% in testing.
Noticed during other testing that processing the Hipparcos catalog dataframe for visible stars hogs the RPi CPU, although the routine is overall faster to process, it's less friendly at sharing the CPU with other threads. This causes the other threads (Main observation loop and microcontroller communication) to pause noticeably. Pausing the microcontroller loop can cause the camera altitude/azimuth positions to become too old (stale). Investigating further.
Latest version using a small periodic delay in the Hipparcos handling seems to make the thread share resources better with the rest of the program activities.
Improvements in Summer2023 update. MarkupPreview remains slow if it's simulating images from scratch, but that's a different issue.
MarkupPreview function is taking quite a long time to process now. NGC, Messier and Hipparcos objects take quite a while to run through. Rework to speed up, switch from 'dictionary' lists to Pandas dataframes to speed up filtering and selection. Reduce excessive logging of the calculations.