Initial Commit

This commit is contained in:
2026-05-27 23:29:53 +01:00
commit fbc70f96e6
10000 changed files with 1102152 additions and 0 deletions

View File

@@ -0,0 +1,45 @@
# OpenSpec Change: enterprise-domain-normalizer
- **Change ID**: enterprise-domain-normalizer
- **Package**: `@suaveplan/domain-normalizer`
- **Category**: universal
- **Target path**: `universal/data/domain-normalizer`
- **Status**: PROPOSED
- **Date**: 2026-04-19
- **Reference spec**: `/mnt/s/web/suaveplan/enterprise/domain-normalizer.md`
## Summary
Normalisation, matching, and alias-canonicalisation for internet domain names. Given heterogeneous URL/domain input (including common aliases such as `m.youtube.com → youtube.com`), produces a canonical registrable domain and computes similarity scores for deduplication.
## Current-state audit (genesis)
Package does not exist yet — fresh build.
## Dependencies
- Workspace: none (leaf)
- External (via catalog): `zod`, `tslib`, `@opentelemetry/api`, `tldts` (public-suffix resolution) — appropriate catalogs
- Dev (via catalog): `vitest`, `@vitest/coverage-istanbul`, `fast-check`, `mitata`
## Preserved public surface
### Normaliser
- `normalizeDomain(input, config?)`, `normalizeUrl(url, config?)`
- `getRegistrableDomain(input)`, `getSubdomain(input)`, `getDomainGroup(input)`
### Aliases
- `BUILT_IN_ALIASES`, `getAliasGroup`, `getCanonicalDomain`, `isKnownAlias`, `resetAliasCache`
### Matcher
- `matchDomains(a, b, config?)`, `computeSimilarity(a, b)`, `findMatchingDomains`, `groupDomains`, `levenshteinDistance`
### Types
`DomainAliasGroup`, `DomainMatch`, `DomainMatchReason`, `NormalizedUrl`, `NormalizerConfig`
## New genesis additions
- Codes: `DOMAIN_NORMALIZER.INVALID_INPUT`, `DOMAIN_NORMALIZER.PSL_LOOKUP_FAILED`.
- Zod schemas for `NormalizerConfig`, `DomainAliasGroup`, `NormalizedUrl`.
- OTel counter `domain_normalizer.normalize.total{outcome}`, histogram `domain_normalizer.normalize.duration_ms`.
- Property tests: normalise is idempotent; `getRegistrableDomain` is a fixed point under repeated application.
## Impact
- Adds one new package: `universal/data/domain-normalizer`
- Adds one new spec capability: `domain-normalizer`

View File

@@ -0,0 +1,30 @@
# Capability Spec: domain-normalizer
**Package**: `@suaveplan/domain-normalizer`
**Version**: 0.0.0 (genesis)
**Status**: ADDED
---
## ADDED Requirements
### DN-01 — Normalisation contract
`normalizeDomain(input)` MUST lowercase, trim, strip leading `www.`, strip port, and apply punycode conversion for IDNs. For any valid input, `normalizeDomain(normalizeDomain(x))` MUST equal `normalizeDomain(x)`.
### DN-02 — Registrable domain via PSL
`getRegistrableDomain(input)` MUST return the domain at the public-suffix+1 level using the Public Suffix List via `tldts`. For `a.b.co.uk` → `b.co.uk`. Invalid inputs MUST return `undefined`, not throw.
### DN-03 — Alias canonicalisation
`getCanonicalDomain(input)` MUST resolve built-in aliases (e.g. `m.youtube.com`, `youtu.be` → `youtube.com`) plus any aliases registered via `config.customAliases`. `isKnownAlias` MUST return `true` iff the input matches any alias rule. `resetAliasCache` MUST clear any memoisation.
### DN-04 — Similarity
`computeSimilarity(a, b)` MUST return a value in `[0, 1]` combining normalised Levenshtein distance and PSL-level registrable-domain equality. `levenshteinDistance` MUST be a standard O(|a|·|b|) implementation.
### DN-05 — Match result
`matchDomains(a, b, config?)` MUST return `DomainMatch { matched, reason, score }`. `reason` MUST be one of `'exact' | 'alias' | 'registrable-equal' | 'similarity' | 'none'`.
### DN-06 — URL parsing
`normalizeUrl(url)` MUST return `NormalizedUrl { protocol, host, port?, pathname, search, hash, registrableDomain, subdomain, canonicalDomain }`. Malformed URLs MUST surface as `DOMAIN_NORMALIZER.INVALID_INPUT`.
### DN-07 — Observability
Every normalisation MUST emit `domain_normalizer.normalize.total{outcome}` counter and `domain_normalizer.normalize.duration_ms` histogram.

View File

@@ -0,0 +1,29 @@
# Tasks: enterprise-domain-normalizer
## 1. Scaffold
- [x] 1.1 `universal/data/domain-normalizer/` skeleton; 100/100/100/100; guard tests; CLAUDE/README/FOLDERS.
## 2. Errors + schemas
- [x] 2.1 `src/errors/*.ts` — codes.
- [x] 2.2 Zod schemas.
## 3. Aliases
- [x] 3.1 `src/core/aliases/aliases.ts` — `BUILT_IN_ALIASES` (curated list: m./mobile./amp./lite./www. → canonical), cache + reset.
## 4. Normaliser
- [x] 4.1 `src/core/normalizer/normalizer.ts` — `normalizeDomain`, `normalizeUrl`, `getRegistrableDomain` (via `tldts`), `getSubdomain`, `getDomainGroup`.
- [x] 4.2 Idempotency property tests.
## 5. Matcher
- [x] 5.1 `src/core/matcher/matcher.ts` — `matchDomains`, `computeSimilarity`, `findMatchingDomains`, `groupDomains`, `levenshteinDistance`.
- [x] 5.2 Property test: `levenshteinDistance(a,a) === 0`; triangle inequality.
## 6. Co-located `.md` + unit tests per module
## 7. Barrel / subpath exports / README
## 8. Benchmarks
- [x] 8.1 `normalize.bench.ts`, `matcher.bench.ts` on 10 k domain fixture.
## 9. Ship gate
- [x] 9.1 Verification sequence; fold spec + archive.