Replace seedrandom package with inline ESM implementation

[andreialecu]

Summary

• Replaces the CommonJS-only seedrandom dependency with an inline ESM module based on the same ARC4 (RC4) cipher algorithm
• Eliminates the "Module 'seedrandom' is not ESM" warning in Angular and other ESM-first bundlers
• Enables proper tree-shaking of this code

Part of #3649

Details

The seedrandom package hasn't been updated in 7+ years and only ships CommonJS. Rather than waiting for an upstream ESM build, this inlines a minimal implementation that only includes what mathjs needs (the double() generator).

The implementation produces identical output to the original seedrandom package — verified by exact sequence stability tests for multiple seeds.

Test plan

• New unit tests for the seedrandom module (determinism, output range, distribution, exact sequence stability, edge cases)
• Existing seededrandom.test.js integration tests pass
• All 6656 unit tests pass
• Lint passes with zero warnings

[andreialecu]

You can verify here that the code in this PR produces the exact same sequence as the seedrandom npm lib:

https://codesandbox.io/p/devbox/frosty-pond-jq6qqt?workspaceId=ws_HeHwxRY4bpopppVUoiziuV

[andreialecu]

Benchmark script

npm install --no-save seedrandom@3.0.5

node --input-type=module --eval "
import { createRequire } from 'module';
const require = createRequire(import.meta.url + '/../');
const originalSeedrandom = require('seedrandom');
const { seedrandom } = await import('./src/function/probability/util/seedrandom.js');

const seeds = ['hello', 'test123', '42', 'a longer seed with spaces'];
const iterations = 500;
const numPerIter = 10000;

console.log('Generate 10,000 random numbers per iteration (' + iterations + ' iterations):');
console.log('');
console.log('Seed                     | Original | Inline  | Speedup');
console.log('-------------------------|----------|---------|--------');

for (const seed of seeds) {
  for (let i = 0; i < 50; i++) { const r = originalSeedrandom(seed); for (let j = 0; j < numPerIter; j++) r(); }
  for (let i = 0; i < 50; i++) { const r = seedrandom(seed); for (let j = 0; j < numPerIter; j++) r(); }

  let start = performance.now();
  for (let i = 0; i < iterations; i++) { const r = originalSeedrandom(seed); for (let j = 0; j < numPerIter; j++) r(); }
  const orig = (performance.now() - start) / iterations;

  start = performance.now();
  for (let i = 0; i < iterations; i++) { const r = seedrandom(seed); for (let j = 0; j < numPerIter; j++) r(); }
  const inl = (performance.now() - start) / iterations;

  const ratio = (orig / inl).toFixed(1);
  console.log(seed.padEnd(25) + '| ' + orig.toFixed(3).padStart(6) + 'ms | ' + inl.toFixed(3).padStart(5) + 'ms | ' + ratio + 'x');
}

console.log('');
console.log('Realistic usage (create + generate N numbers, 100k iterations):');
console.log('');
console.log('Numbers/call | Original | Inline  | Speedup');
console.log('-------------|----------|---------|--------');

for (const count of [1, 5, 10, 50, 100]) {
  const iter = 100000;
  for (let i = 0; i < 1000; i++) { const r = originalSeedrandom('seed'); for (let j = 0; j < count; j++) r(); }
  for (let i = 0; i < 1000; i++) { const r = seedrandom('seed'); for (let j = 0; j < count; j++) r(); }

  let start = performance.now();
  for (let i = 0; i < iter; i++) { const r = originalSeedrandom('seed'); for (let j = 0; j < count; j++) r(); }
  const orig = (performance.now() - start) / iter;

  start = performance.now();
  for (let i = 0; i < iter; i++) { const r = seedrandom('seed'); for (let j = 0; j < count; j++) r(); }
  const inl = (performance.now() - start) / iter;

  const ratio = (orig / inl).toFixed(1);
  console.log(String(count).padEnd(13) + '| ' + (orig * 1000).toFixed(1).padStart(6) + 'µs | ' + (inl * 1000).toFixed(1).padStart(5) + 'µs | ' + ratio + 'x');
}

console.log('');
for (const seed of seeds) {
  const o = originalSeedrandom(seed); const n = seedrandom(seed);
  let match = true;
  for (let i = 0; i < 100; i++) { if (o() !== n()) { match = false; break; } }
  console.log(seed + ': ' + (match ? 'identical' : 'MISMATCH'));
}
"

Results (MBP M1 Max):

Generate 10,000 random numbers per iteration (500 iterations):

Seed                     | Original | Inline  | Speedup
-------------------------|----------|---------|--------
hello                    |  0.437ms | 0.335ms | 1.3x
test123                  |  0.449ms | 0.363ms | 1.2x
42                       |  0.472ms | 0.351ms | 1.3x
a longer seed with spaces|  0.440ms | 0.365ms | 1.2x

Realistic usage (create + generate N numbers, 100k iterations):

Numbers/call | Original | Inline  | Speedup
-------------|----------|---------|--------
1            |    3.3µs |   1.3µs | 2.6x
5            |    3.2µs |   1.3µs | 2.4x
10           |    3.3µs |   1.5µs | 2.2x
50           |    5.0µs |   2.9µs | 1.7x
100          |    7.1µs |   4.6µs | 1.6x

hello: identical
test123: identical
42: identical
a longer seed with spaces: identical