Recognize more analog-button Orca inputs

[LaurencePutterman]

Summary

• Add a second analog-shape signal for analog-button inputs with little off-axis noise.
• Require strict rim-coordinate variety plus many distinct axis-aligned analog magnitudes before bypassing the box classifier.
• Add four Orca replay fixtures covering the travel-time and SDI false positives.

Related to #15

Testing

• cd libenforcer-wasm && cargo test --target aarch64-apple-darwin --test utils_test -- --nocapture
• cd libenforcer-wasm && cargo test --target aarch64-apple-darwin --test input_fuzzing_test -- --nocapture
• cd libenforcer-wasm && cargo test --target aarch64-apple-darwin

[LaurencePutterman]

I realize that I'm adding more inexact heuristics. If you have better ideas, feel free to reject this PR. I was just given these SLPs by Rienne and another Orca user who had replays that were still getting flagged with the previous fix, and this is what I came up with to solve it.