Fix stale Rust API references in documentation examples

[dchud]

The Rust code examples across the docs referenced methods and call shapes that don't match the real API, so most wouldn't compile. Root cause: rustdoc doctests run in CI, but mkdocs Markdown fences are never compiled, so the drift went unnoticed. Reported by @acdha.

Every Rust fence across 11 files (quickstart, tutorials, reference) was corrected against src/ and the compiled examples/, and verified by compiling a throwaway example exercising every shape (since mkdocs can't compile the fences itself):

• Field — field.get_subfield('a') (method, char) for field.subfield("a"); tag/indicator1/indicator2/subfields are public fields (no parens); field.value() not field.data().
• Record — get_field(tag) not field(tag); get_control_field(tag) not control_field(tag); record.leader is a public field.
• Query DSL — FieldQuery::new().tag(...) + record.fields_matching(&q) (no static FieldQuery::tag, no query.find_all); tag_range(...) + fields_matching_range; with_subfield(...) → SubfieldValueQuery + fields_matching_value.
• Builder — Record::builder(leader) (no .leader()); Leader::from_bytes(...) (no Leader::default()); control_field_str/subfield_str for literals; Field::builder("245".to_string(), …).
• Format conversion — mrrc::json::record_to_json / mrrc::marcxml::record_to_marcxml / mrrc::marcjson::record_to_marcjson (+ inverses), not the nonexistent mrrc::formats::{to_json, …}.
• Added use mrrc::RecordHelpers; wherever record.title() is used; fixed the testing.md Option<&str> assert.

This also resolves the "loop structure oddities" on the query page (mixed find_all/hand-rolled loops → coherent examples).

Scope note / follow-up: filed a separate bead for the rust-api.md prose "Key Methods" reference tables (method-call notation + a few wrong return types) and two Python-fence stale refs — distinct from the Rust examples this bead targets.

.cargo/check.sh green (566 tests + doctests + mkdocs). CHANGELOG [Unreleased] credits @acdha.

Closes #233

Bead: bd-nj22

[acdha]

Isn't this supposed to be showing how to use SubfieldPatternQuery rather than a regular Rust regex?

[dchud]

Fixed in #239. That section now uses SubfieldPatternQuery::new("020", 'a', r"^97[89]") with record.fields_matching_pattern(&query) instead of the raw regex::Regex loop.

[acdha]

This still isn't using SubfieldValueQuery or partial matching. I'd think this would be more:

let query = SubfieldValueQuery::partial("650", 'a', subject);
if (query.matches(&field)) …

(except that might be record.fields_matching for performance)

[dchud]

Fixed in #239. It now uses SubfieldValueQuery::new("650", 'a', "History") for exact matches and SubfieldValueQuery::partial("650", 'a', "History") for substring matches, both via record.fields_matching_value(&query); the section also notes query.matches(&field) for testing a single field. Your suggested shape was right.