feat(parser): update to Substrait v0.87.0 with enum arg support

[benbellick]

Upgrades the substrait dependency from v0.85.0 to v0.87.0 and adds parser support for the enumArg syntax introduced in that version (IDENTIFIER::enum, e.g. SAMPLE::enum in std_dev).

The interesting change is widening CaseLiteral.Value from expr.Literal to types.FuncArg. This is necessary because enum args don't map to an Expression.Literal in the substrait proto. FunctionArgument is a oneof of {string enum, Type type, Expression value}, making enum a sibling of a literal expression rather than a subtype. So types.Enum cannot implement expr.Literal, and the field needs to hold the broader types.FuncArg interface instead.

std_dev and variance test cases remain skipped because the v0.87.0 test files use an invalid single-column compact table format fixed in substrait v0.88.0 (substrait-io/substrait#1043).

[codecov]

Codecov Report

❌ Patch coverage is 91.66667% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 68.71%. Comparing base (2242c28) to head (858423d).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
testcases/parser/nodes.go	75.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #222      +/-   ##
==========================================
+ Coverage   68.56%   68.71%   +0.15%     
==========================================
  Files          47       47              
  Lines       10847    10856       +9     
==========================================
+ Hits         7437     7460      +23     
+ Misses       3055     3042      -13     
+ Partials      355      354       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[bvolpato]

LGTM -- nice writeup on the CaseLiteral.Value widening from expr.Literal to types.FuncArg, the reasoning around the proto oneof is spot on. Good that extract.test gets unblocked too.

One thing worth fixing before merge: in testcases/parser/parse_test.go:436-438, the new if lit, ok := tc.AggregateArgs[0].Argument.Value.(expr.Literal); ok { assert.Equal(t, listType, lit.GetType()) } silently no-ops if the cast fails. Previously that was an unconditional assert.Equal(t, listType, tc.AggregateArgs[0].Argument.Value.GetType()), so a regression would now pass the test instead of failing it. The value here is always a float32 list literal, so use require.True(t, ok, ...) like you already do at parse_test.go:218-219.