Experiment: Communication between two Python scripts

In the “Experiment” series, I write about weird “I’ve always wanted to do this, but I am not sure if this will work” things

For a while now, I have wanted to conduct a very simple experiment. Can two Python scripts communicate with each other while both of them are running? Since both of the scripts are running in their respective shells, there is no immediate way I can think of that would allow a piece of code in one script to pass a variable to a piece of code in another script. Except via a shared space on the disk, such as a file.

Communication between two Python scripts via a shared file

Experiment 1

My first experimental setup involved the following two script files:

• publisher.py - a class that “publishes” some financial data by writing it to a file

• subscriber.py - a class that “subscribes” to financial data by reading it from the shared file

#publisher.py
import random
import time

publication_buffer=open("publication.buf", 'w')
trade_number=0

print "Starting exchange"

while True:
    time.sleep(5)
    probability=random.uniform(0,1)
    if probability>=0.5:
        print "Making trade"
        trade_number=trade_number+1
        publication_buffer.write("Trade number"+str(trade_number)+"\n")

#subscriber.py
number_of_lines=0
print "Opening trade channel"
while True:
    trade_buffer=open("publication.buf", 'r')
    lines=trade_buffer.readlines()
    if len(lines)>number_of_lines:
        print "Update received"
        print lines[len(lines)-1]
        number_of_lines=number_of_lines+1

The publisher is started first to give it a chance to “warm up” (ie. open the file and start publishing trades). Then after a few seconds, the subscriber starts listening to the published trades. In theory, this seems like something that might work. This is what happened:

#shell 1

$ python publisher.py
Starting exchange
Making trade

#shell 2
$ python subscriber.py
Opening trade channel

For some reason, nothing is received by the other file. If subscriber would be able to read the values published by publisher, some values would be printed in the console. But nothing happens! Also, immediately after launching this operation, my computer starts to whir and hangs. Something is going on!

Let’s look at the output of top to see if something is consuming a lot of computational power and indeed python is the top process (eating up 80% of the CPU). Now I have to admit, my hardware could use a bit of upgrading (it dates back to the stone age, aka 2010), but 80% CPU for something as simple as the above two scripts seems

Investigation will continue. :)