perf: optimize sql queries and database connections

[m-truscott]

Summary

Refactor and optimize SQL queries in the SqlQuery class to prevent SQL injection vulnerabilities, ensure database connections are closed properly, and improve overall query execution performance.

Changes

• Implement contextlib.closing for all sqlite3.connect calls to ensure safe connection closures
• Update album_exists to use parameterized queries and SELECT 1 with fetchone() to avoid fetching full rows and prevent SQL injection
• Refactor join_albums to use LEFT JOINs instead of nested subqueries for improved execution time
• Update top_invoices to apply the LIMIT 10 clause in SQL rather than relying on Python list slicing

Metrics

Metric	Mean	Std	Improvement
benchmark runtime (s)	16.3826	-	-11.37%
benchmark memory (B)	87379069.52	-	-3.90%
benchmark cpu (%)	99.1296	-	+5.42%
unit runtime (s)	1.0040	-	-50.11%
unit memory (B)	3612672.00	-	-76.86%
unit cpu (%)	0.0000	-	-100.00%

Validation

All validation checks passed:

• Compilation: uv pip install poetry && python -m poetry install — passed (runtime 1.0060s, +0.03%; memory 26316800.00B, +1.39%)
• Unit tests: python -m poetry run pytest --benchmark-skip tests/ — passed
• Benchmark: python -m poetry run pytest --benchmark-only tests/ — passed