My map didn't turn off tonight.  I looked at the code, and as near as I could tell, the problem is that you set the time to turn the map off and it includes the date when the script is started, so there's some complicated footwork in the script to work with the datetime objects.  I think I was able to simplify it using the technique in this post.

The biggest problem I ran into is that the datetime and time functions get overloaded.  Fortunately, Python lets us pull in functions with different names.  So, I make one import change, to import a different time function.

@@ -72,6 +72,7 @@
 import time
 from datetime import datetime
 from datetime import timedelta
+from datetime import time as time_
 from rpi_ws281x import * #works with python 3.7. sudo pip3 install rpi_ws281x
 import sys
 import os

By importing the time functions as "time_" (with the underscore), I can use them without changing any of the other code in the script.  This is important in my next chunk, where I define the time of day as a time_ type.  (See https://topherpedersen.blog/2019/04/18/attributeerror-type-object-datetime-time-has-no-attribute-time/ for why I needed to do this.  This took a lot of time to find!)

@@ -302,6 +303,8 @@

 timeoff = now.replace(hour=offhour, minute=offminutes, second=0, microsecond=0) #When to turn map off 
 timeon = now.replace(hour=onhour, minute=onminutes, second=0, microsecond=0)    #When to turn map back on
 diff = timeon - timeoff
+lights_out = time_(offhour, offminutes, 0)
+lights_on = time_(onhour, onminutes, 0)
delay_time = 10                         #Number of seconds to delay before retrying to connect to the internet.
temp_time_flag = 0                      #Set flag for next round if sleep timer is interrupted by button push.

Next, I define a new function that compares the current clock time to the start and end times.  This naturally takes care of the problem where you have times that might span over the day boundary.  You're either in between the two numbers (because it's a range like 10:00-14:00), or you're outside the range (because if you're trying to tell if you're between 23:00 and 08:00, you look to be either past the start time or earlier than the stop time).

def time_in_range(start, end, x):
   if start <= end:
      return start <= x <= end
   else:
      return start <= x or x <= end

Then, I can replace all the time delta stuff in the main loop, and the difference, and checking whether it was a day earlier or not with a call to my new function:

@@ -1219,13 +1233,17 @@
 end_time = temp_end_time
 temp_time_flag = 0 #reset flag for next round

- if timeoff <= datetime.now() <= end_time:
+
+# if timeoff <= datetime.now() <= end_time:
+# Alternate clock check using defined function
+ if time_in_range(lights_out, lights_on, datetime.now().time()):
     sys.stdout.write ("\n\033[1;34;40m Sleeping- ") #Escape codes to render Blue text on screen
     sys.stdout.flush ()
     turnoff(strip)
     logger.info("Map Going to Sleep")

-     while timeoff <= datetime.now() <= end_time:
+#  while timeoff <= datetime.now() <= end_time:
+    while time_in_range(lights_out, lights_on, datetime.now().time()):
          temp_timeoff = timeoff #store original timeoff time and restore later.
          sys.stdout.write ("z")
          sys.stdout.flush ()

So, same logic, but because I'm able to do it all as clock time, the code is a bit more readable, and it's more robust.  I think I was seeing problems with scripts not turning the lights off at night, because if you start the script after midnight, what will happen is that the timeoff variable gets set to the time later.  The time_in_range function will work no matter what kind of range you have, and it doesn't need to pull in new timestamps to keep the day in sync when all you want is the clock time.