replace unwrap() with error propagation across anise/src

[themouli]

Summary

Replaced unwrap() calls throughout anise/src/ with proper error propagation (?) or expect() carrying a meaningful invariant message where the call is provably infallible. Addresses #679 (follow-up to #677).

Improvements

• Replaces unwrap() with ? propagation across the almanac, ephemerides, DAF, KPL, orientations, structure, and math modules so failures surface with typed errors instead of panicking.
• Where an invariant genuinely guarantees success (e.g. a prior Some check, valid_up_to on UTF-8 decoding, a Keplerian-constructor contract), the unwrap() is replaced with expect("...") documenting why it cannot fail — making future debugging and audits easier.
• DAF::persist now returns Result<(), DAFError> instead of io::Result<()>, so I/O errors carry the file path / action context and decoding failures propagate cleanly. Breaking change for external callers of persist.
• KPL parser: file-open failure now reports as DataSetError::IO { action, source } preserving the underlying io::Error.

Bug Fixes

#679

Testing and validation

• cargo check --all-targets clean (pre-existing pyo3 deprecation warnings only).
• Existing test suite exercises the modified paths; no behavioral change on the happy path — only panics become recoverable errors.
• No new tests added; a reviewer suggestion for targeted negative-path tests against DAF::persist and the KPL parser would be welcome if desired.

[gemini-code-assist]

Code Review

This pull request replaces numerous instances of .unwrap() with .expect() or proper error handling throughout the codebase to improve robustness and error reporting. The review identified several areas for further improvement, including optimizing cloning operations in lookup table access, eliminating redundant parsing in file persistence, and correcting misleading variable naming in DER encoding logic.

[gemini-code-assist]

The destructuring of the tuple returned by der_encoding() has the names and id_entries variables swapped relative to the tuple's actual order (ids, id_entries, names, name_entries). While the subsequent addition in encoded_len is commutative and thus correct, the variable naming is misleading and could lead to bugs in future modifications.

Suggested change

	let (ids, names, id_entries, name_entries) = self.der_encoding()?;
	let (ids, id_entries, names, name_entries) = self.der_encoding()?;

[themouli]

Out of scope — the variable naming is pre-existing on master. My PR only added ? for error propagation, didn't touch the destructure or encode order.

[gemini-code-assist]

The variable names in this destructuring are swapped relative to the tuple returned by der_encoding(). This results in the names variable holding id_entries data and vice versa. Although they are encoded in the correct sequence (matching the Decode implementation), the naming is highly confusing for maintainers.

Suggested change

	let (ids, names, id_entries, name_entries) = self.der_encoding()?;
	let (ids, id_entries, names, name_entries) = self.der_encoding()?;

[themouli]

Out of scope — the variable naming is pre-existing on master. My PR only added ? for error propagation, didn't touch the destructure or encode order.