ST_Read: add normalize_names option (like read_csv)

[virgesmith]

Feature request

read_csv supports a normalize_names option that lowercases column names and replaces special characters with underscores. It would be useful if ST_Read supported the same option.

Motivation

Many geospatial datasets (shapefiles, GeoPackages, etc.) use mixed-case column names — e.g. RC_Name, Region_NM, H3_count, pctCloneTown. When loading these into DuckDB alongside other tables that use snake_case, users must manually alias every column:

SELECT
    rc.RC_ID AS rc_id,
    rc.RC_Name AS rc_name,
    rc.Region_NM AS region_nm,
    ...
FROM ST_Read('data.gpkg') rc

With normalize_names := true this would become:

SELECT * FROM ST_Read('data.gpkg', normalize_names := true)
-- columns: rc_id, rc_name, region_nm, ...

Proposed behaviour

Consistent with read_csv:

• Lowercase all column names
• Replace spaces and special characters with _
• Default to false for backwards compatibility