Use mapper-provided romanized names (ja-Latn, ko-Latn) instead of machine transliteration

[MiMoHo]

Fixes #25267

Problem

For objects without name:en, OsmAnd falls back to machine transliteration (Junidecode), which produces wrong results especially for Japanese — kanji get Chinese readings. Example from the issue, way 1175336500 (大山阿夫利神社, name:ja-Latn = "Oyama Afuri Jinja"):

• current English fallback: "Da Shan A Fu Li Shen She"
• mapper-provided romanization: "Oyama Afuri Jinja"

The mapper-provided names can't even be used today because name:ja-Latn (165,044 uses) and name:ko-Latn (138,850 uses) are not in MapRenderingTypes.langs, so they are dropped during map creation (unlike sr-latn / zh-hans / zh-hant, which are already indexed).

Changes

1. MapRenderingTypes.langs: add ja-latn and ko-latn (placed next to ja / ko, matching the existing sr-latn style). Takes effect once maps are regenerated.
2. EntityParser.parseMapObject(): lowercase the language suffix before the langsSet lookup and store names under the lowercase code. Production parsing already lowercases tag keys (Entity.putTag), but code paths using putTagNoLC — e.g. the OSM editing download, which sets setConvertTagsToLC(false) — preserve OSM's BCP 47 capitalization (name:ja-Latn, name:sr-Latn) and previously failed the lookup.
3. MapObject.getEnName(true): prefer a stored Latin-script name (key ending in -latn) over machine transliteration when name:en is missing. Explicit name:en still wins; behavior with transliteration disabled is unchanged; objects without a romanized name fall back to Junidecode exactly as before.

int_name (907k uses) could be a further fallback candidate, but it is not indexed today and not guaranteed to be Latin script, so it is left out of this PR.

Testing

Compiled the patched classes against the current master snapshot jar and ran a standalone test (class-shadowing), 14 checks — all pass with the patch; the unpatched jar shows the current behavior:

Check	Unpatched master	Patched
name:ja-Latn indexed (lowercased keys, production path)	dropped	ja-latn ✓
name:ko-Latn indexed	dropped	ko-latn ✓
name:ja-Latn/name:sr-Latn via case-preserving path (putTagNoLC)	dropped	indexed ✓
getEnName(true) for 大山阿夫利神社 with ja-Latn	"Da Shan A Fu Li Shen She"	"Oyama Afuri Jinja"
name:sr-latn, name:zh-Hans, name:de (production path)	indexed	indexed (unchanged)
unknown language suffix	dropped	dropped (unchanged)
explicit name:en precedence	wins	wins (unchanged)
Junidecode fallback without romanized name (Москва)	"Moskva"	"Moskva" (unchanged)
getEnName(false) without name:en	empty	empty (unchanged)

---

Claude Code — Claude Fable 5, xhigh