refactor: Optimize instruction_accounts creation in invoke_context

[codebender828]

Problem

PS: I am a fairly new contributor, and a student of the code base, so these changes may be reviewed with a grain of salt.

The current implementation, while functional, has room for improvement in terms of performance and readability. These changes aim to make the code more idiomatic and potentially more efficient, especially when dealing with large datasets or frequent calls.

Summary of Changes

This PR introduces a tiny optimization for the creation of instruction_accounts in the invoke_context file of the Anza client implementation for Solana. It also hopes to improve code readability, and adherence to general Rust idioms.

Changes

• Replace explicit clone() with cloned() method
• Reorder operations for more efficient error handling
• Simplify the closure passed to map()

Benefits

1. Improved Readability: The new version has a flatter structure in the closure, making it easier to understand at a glance.
2. More Idiomatic Rust: Using cloned() instead of explicit clone() is more idiomatic and potentially more efficient.
3. Optimized Error Handling: The error is now only created when necessary, potentially saving unnecessary allocations.

Future Optimizations

While not included in this PR, here are some ideas for future optimizations:

1. If the size of duplicate_indicies is known in advance, pre-allocating the Vec could provide a small performance boost.
2. If this operation is performed in a loop, consider using iterators throughout instead of collecting into a Vec, if possible.

Performance Impact

While the performance difference is likely minimal for small datasets or infrequent calls, it could be noticeable for large-scale operations. Benchmarking is recommended to quantify the improvement.

Please review and let me know if you have any questions or suggestions for further improvements!

[CriesofCarrots]

Definitely cleaner and rust-ier. Thanks for the contribution!