Source code online books for professionals by professionals

Listing 2-4.  A dict with Adjacency Sets
N = {
    'a': set('bcdef'),
    'b': set('ce'),
    'c': set('d'),
    'd': set('e'),
    'e': set('f'),
    'f': set('cgh'),
    'g': set('fh'),
    'h': set('fg')
} 
Note
 
■
  if you drop the 
set
 constructor in Listing 2-4, you end up with adjacency strings, which would work as well, as 
immutable adjacency lists of characters, with slightly lower overhead. it’s a seemingly silly representation, but as i’ve said 
before, it depends on the rest of your program. Where are you getting the graph data from? is it already in the form of 
text, for example? how are you going to use it?
Adjacency Matrices
The other common form of graph representation is the adjacency matrix. The main difference is the following: Instead 
of listing all neighbors for each node, we have a row (an array) with one position for each possible neighbor (that is, 
one for each node in the graph), and store a value, such as True or False, indicating whether that node is indeed a 
neighbor. Again, the simplest implementation is achieved using nested lists, as shown in Listing 2-5. Note that this, 
again, requires the nodes to be numbered from 0 to V–1. The truth values used are 1 and 0 (rather than True and 
False), simply to make the matrix more readable.
14
Sets were introduced in Python 2.3, in the form of the sets module. The built-in set type has been available since Python 2.4.
15
This, a dictionary with adjacency lists, is what Guido van Rossum uses in his article “Python Patterns—Implementing Graphs,” 
which is found online at 
https://www.python.org/doc/essays/graphs/
.

---

Chapter 2 
■
 the BasiCs
28