Validator for punctuation characters

[daniel-j-h]

Can we write a validator for punctuation tokens in tags? I don't know how high the false positive rate would be but how often do punctuation characters such as

'!"#$%&\'()*+,-./:;<=>?@[\\]^_`{|}~'

occur.

If we can't do this on the name tag on all ways can we at least check it on a subset of all ways e.g. highways, ramps, etc. and then check specific tags such as name, ref, destination - basically all tags used for generating guidance instructions from?

What I'm mostly concerned about is routing engines announcing weird punctuation characters in guidance instructions to the user.

cc @srividyacb

[1ec5]

Note that some of these characters are expected in street names:

• ' – e.g., “O’Flaherty Street”
• ( ) ; – common in bilingual areas
• , – e.g., “Martin Luther King, Junior Drive”
• : – Finnish and Swedish
• - – e.g., “Cincinnati-Dayton Road”
• . – e.g., “8.Dezember-Straße”

A reasonably sophisticated text-to-speech engine can handle these punctuation characters well.

If we can't do this on the name tag on all ways can we at least check it on a subset of all ways e.g. highways, ramps, etc. and then check specific tags such as name, ref, destination - basically all tags used for generating guidance instructions from?

It makes sense to limit this check to street names. Business names can have a wider variety of punctuation (@ $ !).

[dannykath]

@Rub21 #256

[Rub21]

Thank you @dannykath, I merge the branch into master, now I am working to release the detection into to-fix.

[Rub21]

to-fix: task https://osmlab.github.io/to-fix/#/task/punctuationcharactersinhighways

[Rub21]

Working around those issues, I found some Telenav users are fixing those issues to in the highways: e.g: https://osmlab.github.io/osm-deep-history/#/way/206914585 , https://osmlab.github.io/osm-deep-history/#/way/8941557

@daniel-j-h, I found some issues which you could suggest to fix them.

• https://www.openstreetmap.org/way/348812150/history
• https://www.openstreetmap.org/way/236348364/history
• https://www.openstreetmap.org/way/236348361/history#map=15/37.8183/-122.3442

[1ec5]

As mappers attempt to map both destination:ref and destination:ref:lanes simultaneously, occasionally destination:ref has wound up with | instead of ; to delimit each destination (example). Can this validator catch those issues?

[dannykath]

@Rub21