How to run a script before and after a Python expression is executed in Django shell?

Question

I want to make Django's management command shell to run a script before and after every Python expression is executed. For example:

(Think of I'm using command python3 manage.py shell --ipython to get into shell.)

In [1]: from api.models import Student
In [2]: new_student = Student.objects.create(name="John Doe", number=123)
In [3]: new_student.number
Out[3]: 123
In [4]: new_student.name
Out[4]: "John Doe"

I want to send the expressions and outputs of them to a Slack channel.To do that, I need to run a Python script before and after every expression which takes expressions or outputs to log to my Slack channel.

For the example above, the script would log expressions, for example "new_student.number", and their outputs -if exist, of course-, for exapmle "123", to the Slack channel.

I've tried using django-extensions module, used its shell_plus command with pre and post signals. But it was only calling the signals before and after running the plus_shell. Therefore, what I want to do was running those signal handlers before and after every command/expression.

Is there a way to achieve this using a configuration, a module or even writing a custom management command?

Answer 1

I am not sure if the shell would really support it the way you want but there's a quick hack you can do by writing a wrapper function and then calling all commands via that wrapper function.

def wrap_fun(expr):
    val=exec(expr)
    ## run/call your script with argument expr and val
    print(val)

and then calling all your functions as

In [1]: wrap_fun('from api.models import Student')
In [2]: wrap_fun('new_student = Student.objects.create(name="John Doe", number=123)')
In [3]: wrap_fun('new_student.number')
Out[3]: 123
In [4]: wrap_fun('new_student.name')
Out[4]: "John Doe"

You can use the subprocess module to load your python script(which sends slack messages) and call it. Alternatively, you can add this function and slack script to your Django project in a module, control the sending behaviour via local setting like DEBUG=true and then call it directly from the shell.