Currently, in strings like "a a a", the bi-gram "a a" is counted twice even though it technically only appears completely once.
This could be corrected by keeping track of:
- The last match
- Whether we are currently overlapping with the last match
And checking whether the current match is equal to the last match.
It is arguable whether this is a bug in the first place. For example, in the string "toki pona ala", "toki pona" and "pona ala" technically "double count" the occurrence of "pona." Will need to do some Reading The Literature:tm: to find out.
Currently, in strings like "a a a", the bi-gram "a a" is counted twice even though it technically only appears completely once.
This could be corrected by keeping track of:
And checking whether the current match is equal to the last match.
It is arguable whether this is a bug in the first place. For example, in the string "toki pona ala", "toki pona" and "pona ala" technically "double count" the occurrence of "pona." Will need to do some Reading The Literature:tm: to find out.