"Similarly named ways" Should also detect names with accent and case variations

[naoliv]

Have two highways, one named Rua São João and another Rua Sao Joao
Validate the data and see a proper warning about Similarly named ways

Now change the last highway name from Rua Sao Joao to Rua SAO JOAO and validate again.
We have no warnings anymore (which is wrong).

It seems that the Similarly named ways test should also catch such kind of variations (accents and case).

JOSM:

URL:http://josm.openstreetmap.de/svn/trunk
Repository:UUID: 0c6e7542-c601-0410-84e7-c038aed88b3b
Last:Changed Date: 2017-05-29 16:19:58 +0200 (Mon, 29 May 2017)
Build-Date:2017-05-29 14:25:03
Revision:12275
Relative:URL: ^/trunk

Identification: JOSM/1.5 (12275 en) Linux Debian GNU/Linux 9 (stretch)
Memory Usage: 247 MB / 10206 MB (84 MB allocated, but free)
Java version: 1.8.0_131-8u131-b11-2-b11, Oracle Corporation, OpenJDK 64-Bit Server VM
Screen: :0.0 1600x900, :0.1 1280x1024
Maximum Screen Size: 1600x1024
Java package: openjdk-8-jre:amd64-8u131-b11-2
Java ATK Wrapper package: libatk-wrapper-java:all-0.33.3-13
VM arguments: [-Dawt.useSystemAAFontSettings=on]
Program arguments: [--language=en]
Dataset consistency test: No problems found

[naoliv]

(I promise I will better review what I write before filling a ticket on the next time)

[Don-vip]

Comparing strings is a complex subject. Currently we raise a warning if the ​Levenshtein_distance is 1 or 2, with some normalization rules.

Basically it's the number of characters that differ. So it works well with your first example (two letters differ) but not with the second one because 5 letters differ.

Maybe we could add a normalization rule that handle the case where the difference concerns only the character case.

[Don-vip]

In 12283/josm:

fix #14858 - "Similarly named ways" test: detect accent and case variations for strings of same length