How does SequenceMatcher.ratio works in difflib

Question

I was trying out python's difflib module and I came across SequenceMatcher. So, I tried the following examples but couldn't understand what is happening.

>>> SequenceMatcher(None,"abc","a").ratio()
0.5

>>> SequenceMatcher(None,"aabc","a").ratio()
0.4

>>> SequenceMatcher(None,"aabc","aa").ratio()
0.6666666666666666

Now, according to the ratio:

Return a measure of the sequences' similarity as a float in the range [0, 1]. Where T is the total number of elements in both sequences, and M is the number of matches, this is 2.0*M / T.

so, for my cases:

1. T=4 and M=1 so ratio 2*1/4 = 0.5
2. T=5 and M=2 so ratio 2*2/5 = 0.8
3. T=6 and M=1 so ratio 2*1/6.0 = 0.33

According to my understanding T = len(aabc) + len(a) and M=2 because a comes twice in aabc.

So, where am I getting wrong what am I missing.?

Here is the source code of SequenceMatcher.ratio()

Answer 1

You've got the first case right. In the second case, only one a from aabc matches, so M = 1. In the third example, both as match so M = 2.

[P.S.: you're referring to the ancient Python 2.4 source code. The current source code is at hg.python.org.]