Fix failing transport property cases

[markleader]

Summary

Addressing failing transport problems identified by @immamex in Issue #34.
Fix transport-property failures and legacy-behavior mismatches in CH4/O2 equilibrium/rocket cases, including an out-of-bounds crash path and interval-selection differences.

Changes

• Fixed the transport indexing crash in compute_transport_properties by replacing sparse pure_idx logic with explicit aligned maps and reverse lookup.
• Added legacy-style transport basis caching in equilibrium.
• Added safer binary assembly using transport_to_local to skip unmapped pairs (prevents OOB).
• Adjusted species seeding toward legacy intent (dominant carrier per active element, then monatomic, then threshold expansion).
• Aligned transport fit interval selection to legacy upper-bound behavior.
• Reworked interval-count inference in database read path.
• Improved main-interface test executable resolution.
• Updated transport DB parsing and interval handling in transport.f90.

Testing

• Added interval-boundary regression
• Added sparse/high-index CH4/O2 transport regression
• Added equilibrium CH4/O2 transport regression
• Updated test_simple_transport expected equilibrium transport outputs after legacy interval-alignment
• Ran ctest, pytest, and test_main.py with expected outputs.

Compatibility / Numerical behavior

• No expected changes to numerical results
• Expected changes: Expected numerical changes are limited to transport-property calculations where legacy interval/basis behavior was previously mismatched; these changes fix crash/OOB behavior and correct ionized transport divergence. Validation is provided by the new and updated transport regressions listed above.