allowUnicodeIdentifiers option not supported for Java enum var names

Description

Having enum strings that contain only not-latin characters results in a broken java code where all enum names are made of a single underscore character. There is an allowUnicodeIdentifiers option, but it is not working in this case.

Part of swagger.json:

        "form": {
          "type": "string",
          "enum": [
            "Государственная",
            "Муниципальная",
            "Иная"
          ],
          "description": "Форма собственности"
        },

Generated Java code:

  public enum FormEnum {
    _ ("Государственная"),    
    _ ("Муниципальная"),    
    _ ("Иная");

Swagger-codegen version

2.2.1
master

Swagger declaration file content or url

https://gist.github.com/afrolovsky/667098bb49863fa5592e68969b923a2d

Command line used for generation

java -jar swagger-codegen-cli-2.2.1.jar generate ^
  --input-spec "swagger.json" ^
  --lang java ^
  --additional-properties dateLibrary=java8 ^
  --additional-properties allowUnicodeIdentifiers=true ^
  --output temp

Steps to reproduce

Related issues/PRs

Suggest a fix/enhancement

diff --git a/modules/swagger-codegen/src/main/java/io/swagger/codegen/languages/AbstractJavaCodegen.java b/modules/swagger-codegen/src/main/java/io/swagger/codegen/languages/AbstractJavaCodegen.java
index a83acff..809ea8f 100644
--- a/modules/swagger-codegen/src/main/java/io/swagger/codegen/languages/AbstractJavaCodegen.java
+++ b/modules/swagger-codegen/src/main/java/io/swagger/codegen/languages/AbstractJavaCodegen.java
@@ -990,7 +990,12 @@ public abstract class AbstractJavaCodegen extends DefaultCodegen implements Code
         }
 
         // string
-        String var = value.replaceAll("\\W+", "_").toUpperCase();
+        String var;
+        if (allowUnicodeIdentifiers) {
+            var = Pattern.compile("\\W+", Pattern.UNICODE_CHARACTER_CLASS).matcher(value).replaceAll("_").toUpperCase();
+        } else {
+            var = value.replaceAll("\\W+", "_").toUpperCase();
+        }
         if (var.matches("\\d.*")) {
             return "_" + var;
         } else {

Start in modules/swagger-codegen/src/main/java/io/swagger/codegen/languages/AbstractJavaCodegen.java, at the enum variable-name conversion shown in the issue. Run the provided generation command with the linked Swagger declaration and allowUnicodeIdentifiers enabled; done means Unicode-only enum values produce distinct valid Java enum identifiers instead of repeated underscores.