Extend support for Unicode variable names

[chrispcampbell]

The antlr4-vensim package historically only allowed for ASCII / Latin characters when reading (Vensim) variable names. A request was submitted in the antlr4-vensim repo (see climateinteractive/antlr4-vensim#12) to extend the grammar to allow for non-Latin (e.g., Chinese) characters in variable names (identifiers).

Once the fix is merged and published in antlr4-vensim, we should update the parse package to use the latest. I will also add an integration test to verify that non-Latin characters can be used in mdl files and that they are accepted in the parse/compile and runtime packages without issue.

For the record, prior to this fix, the parse package would throw an error if it encounters non-Latin characters in identifier position:

Generating model...
  Error: Failed to parse Vensim model definition at line 15:
  中文变量名 X = TIME ~~|

  Detail:
    token recognition error at: '中'
      at parseVensimModel (file://.../sdeverywhere/packages/parse/dist/index.js:1408:13)
      at parseModel (file://.../sdeverywhere/packages/compile/src/parse-and-generate.js:166:16)
      at parseAndGenerate (file://.../sdeverywhere/packages/compile/src/parse-and-generate.js:70:21)
      at generate (file://.../sdeverywhere/packages/cli/src/sde-generate.js:97:9)

After the fix, the characters will be accepted without errors.