What is Python?

This section will give a very brief introduction to the Python language.

See also

Executing Python code

You can execute Python code interactively by starting the interpreter like this:

$ python3
>>> print('hello')


ASE works with both Python 2 and the latest Python 3 — the differences are minor. We use Python 3 for this web-page.

You can also put the print('hello') line in a file (hello.py) and execute it as a Python script:

$ python3 hello.py

Or like this:

$ python3 -i hello.py
>>> print('hi!')

Finally, you can put #!/usr/bin/env python3 in the first line of the hello.py file, make it executable (chmod +x hello.py) and execute it like any other executable.


For interactive Python sessions, it is very convenient to have a personal .pythonrc file:

import rlcompleter
import readline
readline.parse_and_bind("tab: complete")
import numpy as np
from ase import *

and point the PYTHONSTARTUP environment variable at it (see rlcompleter for details).


For an even better interactive experience, use ipython.


Python has the following predefined types:

type description example
bool boolean False
int integer 117
float floating point number 1.78
complex complex number 0.5 + 2.0j
str string 'abc'
tuple tuple (1, 'hmm', 2.0)
list list [1, 'hmm', 2.0]
dict dictionary {'a': 7.0, 23: True}

A dict object is mapping from keys to values:

>>> d = {'s': 0, 'p': 1}
>>> d['d'] = 2
>>> d
{'p': 1, 'd': 2, 's': 0}
>>> d['p']

In this example all keys are strings and all values are integers. Types can be freely mixed in the same dictionary; any type can be used as a value and most types can be used as keys (mutable objects cannot be keys).

A list object is an ordered collection of arbitrary objects:

>>> l = [1, ('gg', 7), 'hmm', 1.2]
>>> l[1]
('gg', 7)
>>> l
[1, ('gg', 7), 'hmm', 1.2]
>>> l[-2]

Indexing a list with negative numbers counts from the end of the list, so element -2 is the second last.

A tuple behaves like a list - except that it can’t be modified in place. Objects of types list and dict are mutable - all the other types listed in the table are immutable, which means that once an object has been created, it can not change. Tuples can therefore be used as dictionary keys, lists cannot.


List and dictionary objects can change. Variables in Python are references to objects - think of the = operator as a “naming operator”, not as an assignment operator. This is demonstrated here:

>>> a = ['q', 'w']
>>> b = a
>>> a.append('e')
>>> a
['q', 'w', 'e']
>>> b
['q', 'w', 'e']

The line b = a gives a new name to the array, and both names now refer to the same list.

However, often a new object is created and named at the same time, in this example the number 42 is not modified, a new number 47 is created and given the name d. And later, e is a name for the number 47, but then a new number 48 is created, and e now refers to that number:

>>> c = 42
>>> d = c + 5
>>> c
>>> d
>>> e = d
>>> e += 1
>>> (d, e)
(47, 48)


Another very important type is the ndarray type described here: Numeric arrays in Python. It is an array type for efficient numerics, and is heavily used in ASE.


A loop in Python can be done like this:

>>> things = ['a', 7]
>>> for x in things:
...     print(x)

The things object could be any sequence. Strings, tuples, lists, dictionaries, ndarrays and files are sequences. Try looping over some of these types.

Often you need to loop over a range of numbers:

>>> for i in range(5):
...     print(i, i*i)
0 0
1 1
2 4
3 9
4 16

Functions and classes

A function is defined like this:

>>> def f(x, m=2, n=1):
...     y =  x + n
...     return y**m
>>> f(5)
>>> f(5, n=8)

Here f is a function, x is an argument, m and n are keywords with default values 2 and 1 and y is a variable.

A class is defined like this:

>>> class A:
...     def __init__(self, b):
...         self.c = b
...     def m(self, x):
...         return self.c * x
...     def get_c(self):
...         return self.c

You can think of a class as a template for creating user defined objects. The __init__() function is called a constructor, it is being called when objects of this type are being created.

In the class A __init__ is a constructor, c is an attribute and m and get_c are methods.

>>> a = A(7)
>>> a.c
>>> a.get_c()
>>> a.m(3)

Here we make an instance (or object) a of type A.

Importing modules

If you put the definitions of the function f and the class C in a file stuff.py, then you can use that code from another piece of code:

from stuff import f, C
print(f(1, 2))


import stuff
print(stuff.f(1, 2)(


import stuff as st
print(st.f(1, 2))

Python will look for stuff.py in these directories:

  1. current working directory
  2. directories listed in your PYTHONPATH
  3. Python’s own system directory (typically /usr/lib/pythonX.Y)

and import the first one found.