[minor]: Binomial guarantees

[tayloraswift]

document and test Binomial guarantees

[gemini-code-assist]

Code Review

This pull request adds validation to the Binomial subscript to ensure p is finite, adds documentation for the sample method, and introduces a new test suite BinomialGuarantees to verify the sampling range. The review feedback suggests adding a guard to prevent negative values of n in the subscript to avoid potential runtime crashes, and expanding the test suite to cover boundary edge cases like n = 0, p = 0, and p = 1.

[gemini-code-assist]

If n is negative, calling pdf(_:) will attempt to construct an invalid range 0 ... self.n, which triggers a runtime fatal error in Swift (Can't form Range with upperBound < lowerBound). Guarding that n >= 0 in the subscript prevents this crash and ensures the distribution is always initialized in a valid state.

Suggested change

	@inlinable public static subscript(n: Int64, p: Double) -> Self {
	guard p.isFinite else {
	fatalError("Binomial.subscript cannot be called with `p` = \(p)")
	}
	return .init(n: n, p: p)
	}
	@inlinable public static subscript(n: Int64, p: Double) -> Self {
	guard n >= 0 else {
	fatalError("Binomial.subscript cannot be called with negative `n` = \(n)")
	}
	guard p.isFinite else {
	fatalError("Binomial.subscript cannot be called with `p` = \(p)")
	}
	return .init(n: n, p: p)
	}

[gemini-code-assist]

To thoroughly test the binomial guarantees and boundary conditions, consider adding test cases for edge cases such as n = 0, p = 0, and p = 1.

    @Test mutating func OutputRange() {
        let n: Int64 = Int64.max - 1
        let distribution = Binomial[n, 1.nextDown]
        for _: Int in 0 ..< 1000 {
            let result: Int64 = distribution.sample(using: &self.random.generator)
            #expect(result >= 0)
            #expect(result <= n)
        }

        // Test boundary edge cases
        #expect(Binomial[0, 0.5].sample(using: &self.random.generator) == 0)
        #expect(Binomial[10, 0.0].sample(using: &self.random.generator) == 0)
        #expect(Binomial[10, 1.0].sample(using: &self.random.generator) == 10)
    }