This PR adds support for unescaped underscores in filter expresssion keys similar to how SQL allows tables and columns to have underscores.

I have submitted as 3 commits so you can see it was only a small code change in first commit, the 2nd one is autogenerated by antlr4 maven plugin and the 3rd is a new test case showing unescaped underscores are now accepted.

Feel free to re run the autogeneration to verify change

Fixes https://github.com/spring-projects/spring-ai/issues/696 - Default metadata field names in PagePdfDocumentReader are can't be parsed in a filter expression

Comment From: tzolov

Thanks @dafriz ! Rebased, squashed and merged at 23a3d13b3937a56b1a20c359ac116a8ee502a8c7