Run Python with Tkinter (sometimes) headless OR replacement for root.after() - python-2.7

Question

I have working code below.
I have a set of machines operated with Python. I have a gui in Tkinter but very often these machines are run headless with the python code auto-starting at boot.
I really like the design pattern of using root.after() to start multiple tasks and keep them going. My problem is that this comes from the Tkinter library and when running headless the line "root=Tk()" will throw an error.
I have two questions
Can I perform some trick to have the code ignore the fact there is no display?
OR
Is there a library that will match the design pattern of Tkinter "root.after(time_in_ms,function_to_call)".
I did try to poke around in the underlying code of Tkinter to see if there was simply another library wrapped by Tkinter but I don't have the skill to decode what is going on in that library.
This code works with a display connected: (it prints hello 11 times then ends)
from Tkinter import *
# def __init__(self, screenName=None, baseName=None, className='Tk', useTk=1, sync=0, use=None):
root = Tk() # error is thrown here if starting this command in headless hardware setup
h = None
count = 0
c = None
def stop_saying_hello():
global count
global h
global c
if count > 10:
root.after_cancel(h)
print "counting cancelled"
else:
c = root.after(200, stop_saying_hello)
def hello():
global h
global count
print "hello " + str(count)
count += 1
h = root.after(1000, hello)
h = root.after(1000, hello) # time in ms, function
c = root.after(200, stop_saying_hello)
root.mainloop()
If this is run headless - in an ssh session from a remote computer then this error message is returned
Traceback (most recent call last): File "tkinter_headless.py", line
5, in
root = Tk() File "/usr/lib/python2.7/lib-tk/Tkinter.py", line 1813, in init
self.tk = _tkinter.create(screenName, baseName, className, interactive, wantobjects, useTk, sync, use)
_tkinter.TclError: no display name and no $DISPLAY environment variable

Answer 1

You can use
threading and threating.timer()
shed
APSheduler
or create own taks manager with own after() and mainloop()
Simple example
import time
class TaskManager():
def __init__(self):
self.tasks = dict()
self.index = 0
self.running = True
def after(self, delay, callback):
# calcuate time using delay
current_time = time.time()*1000
run_time = current_time + delay
# add to tasks
self.index += 1
self.tasks[self.index] = (run_time, callback)
# return index
return self.index
def after_cancel(self, index):
if index in self.tasks:
del self.tasks[index]
def mainloop(self):
self.running = True
while self.running:
current_time = time.time()*1000
# check all tasks
# Python 3 needs `list(self.tasks.keys())`
# because `del` changes `self.tasks.keys()`
for key in self.tasks.keys():
if key in self.tasks:
run_time, callback = self.tasks[key]
if current_time >= run_time:
# execute task
callback()
# remove from list
del self.tasks[key]
# to not use all CPU
time.sleep(0.1)
def quit(self):
self.running = False
def destroy(self):
self.running = False
# --- function ---
def stop_saying_hello():
global count
global h
global c
if count > 10:
root.after_cancel(h)
print "counting cancelled"
else:
c = root.after(200, stop_saying_hello)
def hello():
global count
global h
print "hello", count
count += 1
h = root.after(1000, hello)
# --- main ---
count = 0
h = None
c = None
root = TaskManager()
h = root.after(1000, hello) # time in ms, function
c = root.after(200, stop_saying_hello)
d = root.after(12000, root.destroy)
root.mainloop()