feat: field rename and skip, plus four API-level fixes

[martinv13]

API fixes (breaking — minor bump)

Four bugs in the public API that warranted a minor version increment:

• db_identifier dead parameter removed — temp_prefix: bool = False on DatabaseDialect.db_identifier was never passed by any caller; the branch it
  guarded was also backwards (it added length instead of subtracting it). The parameter is removed.

• Table config validation now works — DataModel._build_model was validating user-provided table names against self.model_config.get("tables", {}),
  which is always empty because _validate_config strips the "tables" key. The fix uses self.tables_config instead. Before this fix, typos in table names
  silently passed without raising DataModelConfigError.

• DataModelColumn.model_config renamed to config — the attribute held a table-level config dict, not the top-level model config, making the name
  actively misleading. All internal references updated.

• XMLConverter.parse_xml default corrected — skip_validation defaulted to False (validate every document) despite every caller passing True. The
  default is now True, matching the documented and expected behaviour.

New features

rename — a new field-level config option that sets the physical database column name while keeping the original XSD name as the internal logical key:

model_config = {
    "tables": {
        "shiporder": {
            "fields": {
                "orderid": {"rename": "order_id"},
                "orderperson_city": {"rename": "contact_city"},  # elevated field
            }
        }
    }
}

All internal machinery (data dicts, FK lookups, bulk insert, merge statements) continues to use the original logical name; only the visible DB column name changes. Works on plain columns and on elevated fields (use the post-elevation prefixed name as the config key).

transform: "skip" — completely excludes a field from the data model:

model_config = {
    "tables": {
        "item": {
            "fields": {
                "note": {"transform": "skip"},     # column
                "delivery": {"transform": "skip"}, # relation
            }
        }
    }
}

A skipped column is absent from the DB schema, from fields_transforms (so the XML converter ignores it naturally via the existing not in fields_transforms guard), and from parsed records. A skipped relation is also removed from relations_1/relations_n so no FK columns or relation tables are created; if the child table has no other parents it is pruned from the model entirely. Note: skip is intentionally incompatible with round-trip XML reconstruction.

Documentation

docs/configuring.md gains two new subsections under Fields configuration: Renaming columns and Skipping fields, with examples including the elevated-field case and a warning about skip's round-trip incompatibility.