Implement a binary search querying of ordered NAIF IDs in DAF files

[ChristopherRabotin]

Summary

DAF parsing now builds an index for NAIF ID summaries which are ordered. Enables O(log N) searching for ordered DAF files (which is most). Non-ordered summaries still use the previous O(N) searching.

Architectural Changes

While it does not affect the API, the DAF parsing now builds an index for the NAIF IDs whose summaries are chronologically ordered throughout the DAF summary blocks. This enables the summary_id_at_epoch function to perform a binary search, at an O(log N) cost, when seeking a specific summary, at the expense of a constant O(N) cost parsing a DAF file. The trade-off is well worth it.

New Features

No change

Improvements

No change

Bug Fixes

No change

Testing and validation

No change in the tests, thereby ensuring identical behavior to the CSPICE and previous ANISE version.

Documentation

This PR does not primarily deal with documentation changes.

[gemini-code-assist]

Code Review

This pull request implements an indexing system for DAF files to enable binary search for NAIF ID summaries, improving lookup performance for chronologically ordered records. The DAF struct now includes an IndexMap, and the parse function has been updated to populate this index while detecting unsorted segments. Review feedback highlights a missing 100ns tolerance in the binary search logic, a potential performance bottleneck in the indexing loop, and the presence of a dbg! macro that should be removed.

[ChristopherRabotin]

Benchmark changes:

	master	this pr
ANISE DAF/BPC single hop to parent	101.31 µs	66.012 µs
JPL Ephemeris	220.39 µs	138.15 µs
Bench Spacecraft (Hermite Type 13)	20.295 ms	15.370 ms

Note: previously, Bench Spacecraft (Hermite Type 13) was the only benchmark where ANISE was slightly slower than SPICE. This PR allows ANISE to catch up to SPICE, which executes the benchmark in 17.663 ms.

Edited to update the number after switching to the default hash map from an IndexMap. We should expect further speed improvements if we can switch to the Fnv hasher, cf. dante19031999/fnv-rs#1. I considered using FxHash because it's supposed to be faster, but the crate has not been maintained in 7 years.