Allow unexpected char for names in NSI

[Famlam]

Plugin Name_Script (item 5070, class 50701) reports when a name contains an character that's not common in the charset of the country. For example, in the case of n5782442218, I get

"name"="Søstrene Grene" onverwachte karakter "ø" (LATIN SMALL LETTER O WITH STROKE, 0x00f8)

(I marked it as false positive)

This shop however is also in the NameSuggestionIndex with the same spelling.
It would be good to whitelist names that are also in the NSI. We already do this for uppercase names:

osmose-backend/plugins/Name_UpperCase.py

Lines 59 to 84 in 07b1a01

	def _download_nsi(self):
	nsi_url = "https://raw.githubusercontent.com/osmlab/name-suggestion-index/main/dist/nsi.json"
	json_str = urlread(nsi_url, 30)
	results = json.loads(json_str)
	return results['nsi']
	def _whitelist_from_nsi(self, nsi, nsiprefix, country):
	whitelist = set()
	for tag, details in nsi.items():
	if tag.startswith(nsiprefix) and "items" in details:
	for preset in details["items"]:
	if "locationSet" in preset:
	if ("include" in preset["locationSet"] and
	country not in preset["locationSet"]["include"] and
	"001" not in preset["locationSet"]["include"]):
	continue
	if "exclude" in preset["locationSet"] and country in preset["locationSet"]["exclude"]:
	continue
	if "name" in preset["tags"]:
	for name in preset["tags"]["name"].split():
	if self.UpperTitleCase.match(name) and not self.RomanNumber.match(name):
	whitelist.add(name)
	for name in preset["displayName"].split():
	if self.UpperTitleCase.match(name) and not self.RomanNumber.match(name):
	whitelist.add(name)
	return whitelist

Since the code to do so would be (nearly) a duplicate, maybe it's an idea to move these NSI-parsing lines to a separate modules file, so that we can just call the same functions?

[frodrigo]

Deployed.

[Famlam]

I suspect that updating the plugin version doesn't clear the cached external resource (here: the NSI), correct?
Otherwise something is wrong: Søstrene Grene does still appear https://osmose.openstreetmap.fr/nl/issue/bc938e46-cf3f-c2aa-6c86-13cbc4d448da

[frodrigo]

Yes, cache and code update are not related.

[Famlam]

Ok, found the bug. The whitelist code that I took from Name_Uppercase splits by spaces, but I compare by full name, so it only worked for names without spaces

[Famlam]

Now the issue is fixed properly, the error is gone ;)