Add levenshtein distance. Fix #404

[vpinna80]

REOPENED DUE TO WRONG MERGE

• Add levenshtein to grammar

• Ignore build folder

• Add doc for levenshtein operator

• Add levenshtein tests

[vpinna80]

@NicoLaval Please read: #558 (comment)

You can use this branch (rebased on master) for further editing.

[vpinna80]

Comments copied from another thread:

@linardian said:

Good afternoon.
As decided in the last TF meeting, for this issue it will be defined a new operator with the possibility to indicate which string distance measure is going to be applied, something like:

string_distance(method, string1, string2)

with “method” having values in (levenshtein,….).
So please proceed along this direction.
Thank you

@NicoLaval said:

Good afternoon. As decided in the last TF meeting, for this issue it will be defined a new operator with the possibility to indicate which string distance measure is going to be applied, something like: string_distance(method, string1, string2) with “method” having values in (levenshtein,….). So please proceed along this direction. Thank you

Hi, I think it's harder because number of params is not fixed, so it seems to be easier to name directly levenshtein as an operator for the 2.2 version (and go deeper in this param issue for 3.0). Do you agree?

@linardian said:

No, sorry but I do not agree. As Franck remarked, we can start with two options (Levenshtein and another one), so the syntax won’t change and is more flexible to add more string distance measures in the next releases.

@antonio-olleros said:

venshtein and another one), so the syntax won’t change and is more flexible to add more string distance measures in the next releases.

Hi! Wouldn't it be easier to leave the method as third parameter, and optional with a default?

[linardian]

Ok for me. I just did not want to add just a Levenshtein operator, so the syntax:
string_distance(string1, string2, method)
with the "method" parameter optional with default is fine for me.

[NicoLaval]

string_distance(method, string1, string2)

Hello @linardian,

As number of parameter is not fix, isn't it better to have string_distance(method, string1, string2, ...) with the 3 first parameters mandatory?

I will update at the end of the week (quick holidays this week).

[linardian]

Anyway this fix will go in the 2.2 version... I will shortly copy alla the documentation in the new folder. Only fixes will be allowed in the 2.1 release.

[antonio-olleros]

string_distance(method, string1, string2)

Hello @linardian,

As number of parameter is not fix, isn't it better to have string_distance(method, string1, string2, ...) with the 3 first parameters mandatory?

I will update at the end of the week (quick holidays this week).

Hi! I think it is fixed, we can only compare two strings. right? So the proposal for me would be string1 and string2 as mandatory, method as optional, so a maximum of 3

[linardian]

@nico, since we have copied all files to 2.2, please re-apply your changes to the new folder. Thanks!

[linardian]

As said, the operator (string_distance or similar) will have 3 parameters: string1, string2 and (optional) a method (default is Levenshtein). Please take input from 2.2 folder.

[vpinna80]

Hi @NicoLaval,
I have rebased the levenshtein branch on the master's head.
Please use this to do the changes requested by @linardian .