Address formatting incorrect for non-US addresses

[sam-nolan-tearfund]

This is similar to other issues, but there are problems with the way that addresses are formatted for non-US addresses. Basically the validation framework seems to assume a limited amount of formatting possibilities, locking us into non-standard formats.

Here in Australia, our postal format is:
28 Buckingham Ave
Bentleigh VIC 3204
Australia

The Google geocoding API formats this correctly. When returned from the backend, it gives: "28 Buckingham Ave, Bentleigh VIC 3204, Australia". But NPSP will format it in a very US style, expanding Ave to Avenue and shortening Australia to AU:
28 Buckingham Avenue
Bentleigh, VIC 3204
AU

This is due to the way the address format is hardcoded in the validator:

NPSP/force-app/main/default/classes/ADDR_GoogleGeoAPI_Validator.cls

Lines 373 to 406 in 53157c1

	if (type_set.contains('postal_code')) {
	String verified5digs = addrCmp.long_name;
	String originalFull = address.MailingPostalCode__c;
	String original5digs = '';
	if (originalFull != null && originalFull.length() > 5) {
	original5digs = originalFull.substring(0, 5);
	}
	//Google only returns 5-digit zipcodes. We don't want to override the original, if the
	//first 5 match, and the original has the additional 4.
	if (!(verified5digs == original5digs)) {
	address.MailingPostalCode__c = addrCmp.long_name;
	}
	} else if (type_set.contains('country')) {
	address.MailingCountry__c = addrCmp.short_name;
	} else if (type_set.contains('administrative_area_level_1')) {
	address.MailingState__c = addrCmp.short_name;
	} else if (type_set.contains('administrative_area_level_2')) {
	address.County_Name__c = addrCmp.long_name;
	} else if (type_set.contains('administrative_area_level_3')) {
	address.Administrative_Area__c = addrCmp.long_name;
	} else if (type_set.contains('street_number')) {
	streetNumber = addrCmp.long_name;
	} else if (type_set.contains('route')) {
	streetName = addrCmp.long_name;
	} else if (type_set.contains('intersection')) {
	intersection = addrCmp.long_name;
	} else if (type_set.contains('premise')) {
	premise += addrCmp.long_name;
	} else if (type_set.contains('subpremise')) {
	premise += addrCmp.long_name;
	} else if (type_set.contains('locality')) {
	address.MailingCity__c = addrCmp.long_name;
	}
	}

I would recommend if possible, following "formatted_address" for the final format, particularly when choosing between short and long versions of the address segments. or providing the formatted address as a field on the address object for others to reference if need be. Or maybe offer a format string to allow the user to specify the address format they would like to have.

I'm very happy to contribute to a fix here, it would be good to just get guidance on the desired path to go down.