Adm error handling

[dihm]

This PR has adm respond with ready\r\n before listening for the binary read. This allows the user to check for issues in the command, without incurring the timeout penalty during normal operation.

It also prevents bugs where erroneous commands did not stop execution of the adm command.

Finally, I updated the examples in the readme to be python scripts, hopefully making the add and adm commands a bit easier to use.

[carterturn]

While we are handling the errors, I think it would also be good to take care of the invalid reps number case.
At the moment, I think this is handled in a very bad way: if the reps number is greater than zero and less than 5, it will subtract four anyway, which will loop around to ~2^32 reps.
As we do commit to reading all the expected commands, we should probably not stop the adm when this happens, and instead log the error and report it once it is over. I think I see three ways of handling this:

1. skip the insert, increment do_cmd_count, warn
2. skip the insert, don't increment do_cmd_count, warn
3. set reps to 5, warn
   At the moment, I think 1 is the best solution.
   It would probably be more user friendly to have the warning include the index of the problematic command(s), but this also increases the error handling complexity by quite a bit. As a compromise, we could just include the index of the most recent error and the number of problems.

Besides this, I think this looks good.

[carterturn]

Actually, having attempted to implement this, I think the third option above (with 0 reps or 5 reps) is actually the least problematic output if someone does not check for error messages.

[dihm]

Actually, having attempted to implement this, I think the third option above (with 0 reps or 5 reps) is actually the least problematic output if someone does not check for error messages.

@carterturn Thanks for tackling that. I had considered doing something about those error prints, but avoided it as I didn't have a very good idea. What you've implemented is simple and effective. Ultimately, there really shouldn't be reps errors in the binary write, and if there are it will likely be easier for the user to fix their table and re-write the whole thing than try to fix individual commands.

I'll fix the typo in the readme then merge.