Tune up reproevents timing data in micropython/client

[vmdocua]

This issue addresses some problems in reproevents data described in

• Oddities with time sync collections #105 .

Proposal is to fix potential problems, review code again and extend device log with more detailed information related to time, to better understand where packets are lost or stuck:

• Use time_diff in device code callback in two places, use sleeps in waiting loop to pass control to CPU for other tasks.
• Review code in device, especially long sync calls like print or gc.collect() directly in device callback.
• Add sequencer for message send from device to client named as seq.
• Add device time (also known as server time) in to message send from device to client. Use utime.time/time_us APIs to calculate it, and convert it to ISO string.
• Add bounced flag in message from device to client. So we will know if bouncer code was executed for certain messages.
• Think how to add std python logger to event listener in client code/device and add also script version or git tag etc there. Because at this moment code produces only CSV file and details are unclear.
• Think how to name events.csv file in aspects of proposed migrations. Because it's not good idea to append events in new format to existing file. May be it can be named with datetime info as e.g. videocapture does or with version/tag info etc.
• Extend CSV file produced by client code with new columns like device_seq, device_isotime, device_bounced etc.