Ingestion
Prime tables, OEIS sequences, and reference conversions are mirrored daily into a read-only warehouse.
Data methodology
How the NumberPedia math engine works
This guide summarizes the deterministic functions and QA safeguards we apply before a number page ships to production.
Computation
Node workers run deterministic helpers (factorization, Roman numerals, divisibility checks) and snapshot results.
Validation
Results are diffed against golden datasets and fuzz-tested with random integers before publishing.
Publishing
Clean JSON feeds trigger ISR so new or updated numbers deploy without blocking the entire site.
Key computations
Prime factorization uses a hybrid trial division + wheel factor method up to √n, then falls back to deterministic Miller–Rabin for very large inputs.
Divisibility tests rely on digit sums, alternating sums, and modular arithmetic to produce explanations, not just booleans.
Sequence detection (triangular, Fibonacci, perfect numbers) uses integer-safe checks rather than floating approximations to avoid rounding errors.
Roman numerals, binary, hexadecimal, and base conversions always operate on strings to preserve leading zeros when relevant.
Quality gates
Every deploy runs automated unit tests plus random sampling of JSON payloads. We also log discrepancies flagged by users or partners and feed them back into the test suite.
Unit tests
150+
Random samples per release
1,000
Max tolerance for drift
<0.01%