Build SQL dialect auto linter / add in github workflows

[keenzarate213]

Description & motivation

Adds a new CI check (sql_dialect_linter) that catches Databricks/Snowflake SQL incompatibilities in the PR.

Note

This is the edu_wh version of the same CI tool in edu_edfi_source here: edanalytics/edu_edfi_source#206

For each dialect, the workflow:

• Compiles every model and test in the edu_wh package against a dummy dbt profile and dummy macro overrides for anything that would require a database connection during compile (e.g. is_incremental, get_column_values, union_relations).
• Auto detects and excludes any model/test that still needs a warehouse connection to compile; currently
  bld_ef3__student_programs, bld_ef3__student_indicators, bld_ef3__student_assessments_long_results, and cfg_assessment_scores.
• Enables tpdm_warehouse and finance_warehouse, which are disabled by default, so they get checked linted as well.
• Lints the compiled SQL with sqlfluff and reports pass/fail counts split between models and tests.

Also added a check for known Databricks incompatible function/type usage that sqlfluff can't catch since it only validates syntax, doesnt detect a given function or type actually exists. Currently checks fortry_to_date() and casting as time, both valid syntax everywhere but with no Databricks equivalent. I added a list directly in sql_dialect_linter.sh
(databricks_incompatible_patterns) so new ones are a one-line addition (we could also maybe store in a yaml?)

Versions (sqlfluff, dbt-core, dbt-databricks, dbt-snowflake) are pinned in
.github/workflows/requirements/sql-linter.txt.
Also had to add noqa: PRS (see documntation here) workaround (scoped to a single known macro) for an unsupported sqlfluff parsing gap around Databricks' variant_explode syntax,
and also allow for caching of virtualenv and dbt_packages in workflow so dependencies don't get
reinstalled/re-downloaded on every push in the branch.

Breaking changes introduced by this PR:

None, this is a new CI/tooling, no changes to warehouse models or logic.

PR Merge Priority:

• Low
• Medium
• High

Changes to existing files:

New files created:

• .github/scripts/sql_dialect_linter.sh: compiles + lints edu_wh against a dummy
  Databricks/Snowflake profile for the given dialect argument.
• .github/workflows/sql_dialect_linter.yml: runs the above on PRs for both dialects.
• .github/workflows/requirements/sql-linter.txt: pinned tool versions

Tests and QC done:

• Did a pressure test on a couple models, linter successfuly detected Databricks incompatibility (except for the try_to_date() function ~ which I had to add as a specific flag to error out because sqlfluff doesn't really check whether or not a function exist in a dialect

see most recent run here: https://github.com/edanalytics/edu_wh/actions/runs/30828507203