Python doesn’t have a preprocessor, nor does it have constants in the sense that they can’t be changed – you can always change (nearly, you can emulate constant object properties, but doing this for the sake of constant-ness is rarely done and not considered useful) everything. When defining a constant, we define a name that’s upper-case-with-underscores and call it a day – “We’re all consenting adults here”, no sane man would change a constant. Unless of course he has very good reasons and knows exactly what he’s doing, in which case you can’t (and propably shouldn’t) stop him either way.

But of course you can define a module-level name with a value and use it in another module. This isn’t specific to constants or anything, read up on the module system.

# a.py
MY_CONSTANT = ...

# b.py
import a
print a.MY_CONSTANT

Sure, you can put your constants into a separate module. For example:

const.py:

A = 12
B = 'abc'
C = 1.2

main.py:

import const

print const.A, const.B, const.C

Note that as declared above, A, B and C are variables, i.e. can be changed at run time.

And ofcourse you can do:

# a.py
MY_CONSTANT = ...

# b.py
from a import *
print MY_CONSTANT

As an alternative to using the import approach described in several answers, have a look a the configparser module.

The ConfigParser class implements a basic configuration file parser language which provides a structure similar to what you would find on Microsoft Windows INI files. You can use this to write Python programs which can be customized by end users easily.

If you really want constants, not just variables looking like constants, the standard way to do it is to use immutable dictionaries. Unfortunately it’s not built-in yet, so you have to use third party recipes (like this one or that one).

Try to look Create constants using a "settings" module? and Can I prevent modifying an object in Python?

Another one useful link: http://code.activestate.com/recipes/65207-constants-in-python/ tells us about the following option:

from copy import deepcopy

class const(object):

    def __setattr__(self, name, value):
        if self.__dict__.has_key(name):
            print 'NO WAY this is a const' # put here anything you want(throw exc and etc)
            return deepcopy(self.__dict__[name])
        self.__dict__[name] = value

    def __getattr__(self, name, value):
        if self.__dict__.has_key(name):
            return deepcopy(self.__dict__[name])

    def __delattr__(self, item):
        if self.__dict__.has_key(item):
            print 'NOOOOO' # throw exception if needed

CONST = const()
CONST.Constant1 = 111
CONST.Constant1 = 12
print a.Constant1 # 111
CONST.Constant2 = 'tst'
CONST.Constant2 = 'tst1'
print a.Constant2 # 'tst'

So you could create a class like this and then import it from you contants.py module. This will allow you to be sure that value would not be changed, deleted.

create constant file with any name like my_constants.py
declare constant like that

CONSTANT_NAME = "SOME VALUE"

For accessing constant in your code import file like that

import my_constants as constant

and access the constant value as –

constant.CONSTANT_NAME

In commercial software, the constant(s) will frequently be buried many folders deep. Use solution below in combination with above solutions for best results. Here’s the syntax that worked for me:

folder_a
    folder_b
        constants.py
# -----------
# constants.py
# ------------
MAXVAL = 1000

folder_c
    folder_d
        my_file.py

# ------------------------
# my_file.py
# ------------------------
import folder_a.folder_b.constants as consts

print(consts.MAXVAL)

In more recent editions, it seems a bit pickier.

the mode in the top answer

Main.py

import constants
print(token)

constants.py
enter code heretoken = "12345"

failed.
Instead, I had to import individual variable from constants.

Main.py

from constants import token
print(token)

A constants.py file is good, but I ran into a situation where I would rather had my constants on top of the file.

A similar Python dict would need an access to the variable with a string key dict['A'], but I needed something with the same syntax as a module import.

(For the curious, I was working on Google Collab .ipynb multiple files, with different configuration variables for each of them, and you can’t edit quickly an imported constants.py from Collab, nor import directly another .ipynb file. )

from dataclasses import dataclass

@dataclass
class Constants:
    A = "foo"

cst = Constants()
print(cst.A)
>>> "foo"

You could also create a classic Python Class, but this needs to define a __init__

This is my solution:

Asum Directory structure

project
|-- package1
|   |-- param1.py
|
|-- package2
|   |-- somefile.py
|
|-- main.py
|-- param2.py

param1.py

MY_CONST = 'abc'

param2.py

MY_CONST = 'xyz'

somefile.py

from package1 import param1

my_var = "inside some file " + param1.MY_CONST

main.py

from package1 import param1
import param2
from package2 import somefile

print(param1.MY_CONST)
print(param2.MY_CONST)
print(somefile.my_var)

Run main.py

Result:

abc
xyz
inside some file abc

As you can see, you can use const from another package.
Hope this helps others!!