NMEA TAG blocks part 2

In my previous post of NMEA TAG block encoded data, I discussed my parsing of the USCG old style NMEA metadata, how to calculate NMEA checksums, and parsing the timestamp. Here are the complete list of codes for the TAG components:

Letter code	NMEA Config Msg	Description	type
c	CPC	UNIX time in sec on millisec	int
d	CPD	Destination	string
g	CPG	Grouping param	?
n	CPN	line-count	int
r	CPR	relative time (Units?)	int
s	CPS	Source / station	string
t	CPT	Text string	string

Before, we had:

\\c:(?P<timestamp&gt\d{10,15})\*(?P<checksum>[0-9A-F]{2})?\\.*

Now, we would like to add the rest. We start by adding a 2nd by creating a string with a destination ("d") entry. The spec says this is an alphanumeric string up to 15 charactions. What exactly can be, I'm not totally sure. I presume that it can't have either a star ('*'), comma (','), and backslash ('\'), but am unsure about characters in the ASCII printable set.

body_parts = ['c:%d' % time.time(), 'd:%s' % 'MyStation12345']
body = ','.join(body_parts); body
# 'c:1372002373,d:MyStation12345'

'\{body}*{checksum}\\'.format(body=body, checksum=checksum(body))
# \c:1372002373,d:MyStation12345*15\

We can get the destination with theis regex string: 'd:[^*,\\]{1,15}'. It needs to be named and if we turn on "Verbose," kodos allows us to have the regular expression span multiple lines.

\\
c:(?P<timestamp>\d{10,15}),
d:(?P<dest>[^*,\\]{1,15})
\*(?P<checksum>[0-9A-F]{2})?\\.*

What if the order were reversed? And what if either of the two entries was missing? Please note in this next code block that I have setup the regex to take either a common or star between any parameter. This will let through '*' between parameters, which is not correct. I am sure there is a way to make the regex such that commas between elements and the start is outside of the repeated block, but I haven't figured it out yet.

\\
(
 (
  c:(?P<timestamp>\d{10,15}) |
  d:(?P<dest>[^*,\\]{1,15})
 )[,*]
)
(?P<checksum>[0-9A-F]{2})?\\.*

Now we need to flush out for the rest of the options. Grouping has substructure to it:

SentenceNum-TotalSentences-IntegerIdentifyingTheGroup

g:(?P<group>(?P<sent_num>\d)-(?P<sent_tot>\d)-(?P<group_id>\d+))

Then n for line number and r for relative time are just an integer for each.

\\
(
 (
  c:(?P<time>\d{10,15}) |
  d:(?P<dest>[^*,\\]{1,15}) |
  g:(?P<group>(?P<sent_num>\d)-(?P<sent_tot>\d)-(?P<group_id>\d+)) |
  n:(?P<line_num>\d+) |
  r:(?P<rel_time>\d+) |
  s:(?P<src>[^$*,!\\]{1,15}) |
  t:(?P<text>[^$*,!\\]+)
 )[,*]
)+
(?P<checksum>[0-9A-F]{2})?\\(?P<nmea_msg>.*)

At this point, we can parse any tag block in the front.