How can I generate keyboard inputs for Python's 'input' function?

You can use unittest.mock to patch the output of a function (including streams).

#!/usr/bin/env python3

import unittest
from unittest.mock import patch

# Unit under test.
def get_input():
    my_input = input("Enter some string: ")
    if my_input == "bad":
        raise Exception("You were a bad boy...")
    return my_input

class MyTestCase(unittest.TestCase):
    # Force input to return "hello" whenever it's called in the following test
    @patch("builtins.input", return_value="hello")
    def test_input_good(self, mock_input):
        self.assertEqual(get_input(), "hello")

    # Force input to return "bad" whenever it's called in the following test
    @patch("builtins.input", return_value="bad")
    def test_input_throws_exception(self, mock_input):
        with self.assertRaises(Exception) as e:
            get_input()
            self.assertEqual(e.message, "You were a bad boy...")

if __name__ == "__main__":
    unittest.main()

If that is just a test case and not a "real part of big system" - meaning that you just need to pass a certain input to your command line executable ones (e.g. run test) and you are using Unix, one convenient way of doing this is using pipes:

// read.py 
val = raw_input()
print 'nice', val

then via console:

$ echo "hat" | python read.py 
nice hat

on Windows syntax is little different - should be something like

dir> python.exe read.py < file.txt

One other hacky-simple way to achieve same thing is to replace sys.stdin with a custom stream object:

sys.stdin = StringIO.StringIO("line 1\nline 2\nline 3")

One should consider using @erip 's answer if that is something bigger than a automatisation routine or testing students' home assignments against fixed set of tests, though.