feat(wasm): sequential rayon shim for the web build

wasm32 has no threads, so rayon's global pool would panic at runtime.
Add `crate::par::prelude`: native re-exports `rayon::prelude`, wasm
provides sequential `std`-iterator equivalents (par_iter / into_par_iter
/ flat_map_iter over slices, Vec, HashMap, HashSet). All call sites now
`use crate::par::prelude::*;` instead of rayon directly.

- wire_gpu's rayon `reduce` (2-arg, no std equivalent) is cfg-split to
  a sequential flat_map concat on wasm.
- rayon moves to a non-wasm target dependency.

Native still runs in parallel; all 75 tests pass. wasm bin compiles.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Cargo.toml

@@ -43,7 +43,6 @@ acadrust = "0.3.4"
 printpdf = "0.9.1"
 flate2 = "1"
 image = { version = "0.25", default-features = false, features = ["png", "jpeg", "bmp", "tiff"] }
-rayon = "1"
 inventory = "0.3"
 truck-shapeops = { version = "0.4", optional = true }
 # Fast non-cryptographic hasher. Handle/u64 keys dominate the hot block,
@@ -77,6 +76,7 @@ windows-sys = { version = "0.61", features = ["Win32_UI_Shell", "Win32_UI_Window
 acadrust = { git = "https://github.com/HakanSeven12/acadrust", branch = "main" }
 
 [target.'cfg(not(target_arch = "wasm32"))'.dependencies]
+rayon = "1"
 open = "5"
 ureq = { version = "3", default-features = false, features = ["rustls"] }