Add a cross-runtime parity corpus for Swift and Rust

[DennisWayo]

Problem Statement

Studio exposes Swift and Rust runtime choices and attempts a reference-runtime parity comparison after a run. However, the public automated test surface is not yet a broad parity corpus. Rust currently has a very small unit-test surface, while Swift has much broader simulator tests.

Proposed Solution

Add a checked-in parity corpus and script:

• fixture circuits for Gaussian, loss, thermal, measurement, feedback, QEC, and supported Fock/hybrid paths,
• expected tolerance windows per metric,
• Swift and Rust execution for every supported fixture,
• JSON diff output for mean photon number, final state summary, measurement count, QEC metrics, backend used, and provenance,
• CI job that runs the fast subset on every PR.

Acceptance Criteria

• scripts/runtime-parity-suite.sh --runtime both exists and produces machine-readable output.
• Parity failures identify fixture, metric, expected tolerance, and actual delta.
• Unsupported runtime/backend combinations are explicit skips, not silent passes.
• Rust and Swift version metadata is included in the parity report.
• Studio can reuse the same parity comparison model.

Validation

cargo test -p schrosim-core
bash scripts/swift-test.sh
bash scripts/runtime-parity-suite.sh --runtime both

Evidence

• core-swift/Sources/schrosim-studio/CLIService.swift
• core-swift/Sources/schrosim-cli/main.swift
• core-rust/schrosim-core/src/main.rs
• docs/examples/