Use polyglot-sql in Rust query generation

[nicosuave]

Summary

• Adds a target dialect to the pure Rust SQL generator and uses it for symmetric aggregate SQL instead of hardcoding DuckDB.
• Routes Rust filter expression rewriting through polyglot-sql when parsing succeeds, with the existing string rewrite as fallback for legacy filter forms.
• Broadens Rust SQL metric parsing so known top-level aggregates preserve complex arguments, MEDIAN() maps correctly, COUNT(DISTINCT ...) normalizes, and unknown function calls stay expression metrics.
• Keeps the related SQL correctness fixes for relative-date intervals, DuckDB symmetric hash casting, and running-total window frames.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c46035eff3

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Parse COUNT(DISTINCT(...)) as count_distinct

The new top-level metric parser only treats COUNT as distinct when DISTINCT is followed by whitespace, so inputs like COUNT(DISTINCT(customer_id)) now fall through as plain count with sql = "DISTINCT(customer_id)". In practice this generates invalid/incorrect SQL (DISTINCT(customer_id) as a projected expression plus COUNT(...)) instead of COUNT(DISTINCT ...), which regresses valid metric definitions that use the parenthesized DISTINCT form.

Useful? React with 👍 / 👎.