@whiteslove/parsing-lexicon 0.2.0 → 0.2.2

package/LICENSE

@@ -1,21 +1,21 @@
-MIT License
-
-Copyright (c) 2026 MARHARYTA KUBAI
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+MIT License
+
+Copyright (c) 2026 MARHARYTA KUBAI
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -1,137 +1,137 @@
-# @whiteslove/parsing-lexicon
-
-Deterministic multilingual parsing primitives shared by Whiteslove housing and hiring services.
-
-The package is the canonical home for lexical data, normalization and reusable deterministic extraction. It is intentionally network-free and does **not** contain UI copy, persistence rules, source orchestration or consumer-specific business logic.
-
-## What it covers
-
-### Geography
-
-- Countries and multilingual aliases for the markets used by Whiteslove services.
-- Canonical city, region, district, microdistrict, locality and metro matching.
-- Uzbekistan and Kazakhstan location extensions.
-- Ukraine city/region catalog with current and historical names, major/regional city detail and metro/location extensions.
-- Tashkent districts, metro, colloquial areas, residential complexes and POIs.
-- Odesa metropolitan entities with separate administrative, suburb/locality, informal-area, development-area, POI and search-cluster semantics.
-- Normalization for apostrophe/dash variants and Cyrillic search folding across RU/UZ/KZ/UA input.
-
-### Housing
-
-- Deal/action and listing-kind taxonomy.
-- Property/building type, condition, layout, furniture and listing-status context.
-- Rooms, bedrooms, floors, total floors, building year and area extraction.
-- Deposit, prepayment, utilities and commission/payment context.
-- Seller/agency/owner confidence and no-commission semantics.
-- Tenant policies, documents, financing, availability and location relations.
-- Infrastructure/proximity extraction such as walking time to metro.
-
-### Hiring
-
-- Candidate vs vacancy intent and non-content/spam classification.
-- Section-aware vacancy/CV parsing.
-- Profession taxonomy and profession context without duplicating consumer technology-skill catalogs.
-- Salary range, currency and pay-period extraction.
-- Experience requirements and seniority.
-- Work mode, employment types, schedules and probation.
-- Languages, CEFR levels and requirement relation (`required`, `preferred`, `notRequired`, `candidateHas`).
-- Work authorization, visa sponsorship, relocation, contracts, benefits and hiring status context.
-
-## Architecture
-
-Consumers should resolve data in this order:
-
-```text
-structured source fields
-  -> shared canonical lexicon
-  -> deterministic parser
-  -> AI fallback / enrichment
-```
-
-Lexical aliases belong here. Consumers keep source adapters, persistence, ranking/filtering semantics and UI/API contracts.
-
-The package must remain dependency-light and must never require a runtime lexicon HTTP service, Redis or a message broker.
-
-## Usage
-
-```js
-import {
-  canonicalCity,
-  canonicalTashkentMetro,
-  normalizeForMatch,
-  parseHousingStructured,
-  parseHiringContext,
-  parseHiringSalary,
-} from '@whiteslove/parsing-lexicon';
-
-canonicalCity('Қарағанды', 'KZ');
-// Karaganda
-
-canonicalTashkentMetro('метро Максим Горький');
-// Buyuk Ipak Yoli
-
-normalizeForMatch('Oʻzbekiston');
-// o zbekiston
-
-parseHousingStructured('3 xona, 5/9, 78 m², depozit 500$, metroga 5 minut piyoda');
-
-parseHiringSalary('Salary: $3,000–4,500 per month');
-
-parseHiringContext(
-  'Senior Frontend Engineer. English B2 required. Remote. Visa sponsorship available.',
-  { mode: 'vacancy', title: 'Senior Frontend Engineer' },
-);
-```
-
-Subpath exports are available for consumers that need narrower modules, including `./geo`, `./locations`, `./housing-context`, `./housing-structured`, `./hiring-context`, `./hiring-professions` and `./money`.
-
-## Data-quality rules
-
-- Add aliases to the existing canonical entity instead of creating source-specific regex tables.
-- Keep administrative entities distinct from colloquial/search concepts.
-- Preserve ambiguity guards where a bare term can mean several things.
-- Structured source fields take precedence over inferred values.
-- Add a regression for every production parsing bug.
-- Canonical dictionaries are immutable at runtime.
-- Duplicate canonical/alias entries and unsupported language-key drift are covered by invariants.
-
-## Lexicon statistics
-
-Snapshot for `c509d92` (2026-08-25). The audit traverses exported lexicon data from `src/`, excludes functions, regular expressions and obvious service/path strings, then normalizes strings with Unicode NFKC, trim, whitespace collapse and lowercase.
-
-| Metric | Count |
-| --- | ---: |
-| Alias/translation-like string occurrences across exports | 153,617 |
-| Unique alias/translation strings, exact | 17,681 |
-| **Unique alias/translation strings, normalized** | **17,397** |
-| All exported lexicon strings, normalized | 17,406 |
-| Normalization merge groups | 282 |
-
-The occurrence count is intentionally much larger than the unique count because aggregate modules re-export the same canonical datasets. For package size and vocabulary breadth, the normalized unique count is the useful number.
-
-### Vocabulary by domain
-
-| Domain | Occurrences across exports | Unique exact | Unique normalized |
-| --- | ---: | ---: | ---: |
-| Geography / locations | 64,898 | 10,149 | **9,967** |
-| Hiring — professions & skills | 6,738 | 2,350 | **2,324** |
-| Hiring — other semantics | 3,592 | 2,589 | **2,558** |
-| Housing | 2,554 | 1,975 | **1,971** |
-| Hiring — languages | 552 | 495 | **478** |
-| Country / source aliases | 170 | 126 | **126** |
-
-Domain rows are not additive: the same canonical string can be reachable through multiple exported structures and categories.
-
-Language totals are not inferred from script alone because many valid aliases are intentionally shared or untagged (for example names that are identical in RU/UK or EN/UZ Latin). A language-by-language table should therefore be generated only from explicitly language-tagged data rather than guessed from Unicode characters.
-
-## Development
-
-Requires Node.js 20+.
-
-```bash
-npm install
-npm test
-npm pack
-```
-
-`npm pack` produces the distributable `@whiteslove/parsing-lexicon` tarball from the files declared in `package.json`.
+# @whiteslove/parsing-lexicon
+
+Deterministic multilingual parsing primitives shared by Whiteslove housing and hiring services.
+
+The package is the canonical home for lexical data, normalization and reusable deterministic extraction. It is intentionally network-free and does **not** contain UI copy, persistence rules, source orchestration or consumer-specific business logic.
+
+## What it covers
+
+### Geography
+
+- Countries and multilingual aliases for the markets used by Whiteslove services.
+- Canonical city, region, district, microdistrict, locality and metro matching.
+- Uzbekistan and Kazakhstan location extensions.
+- Ukraine city/region catalog with current and historical names, major/regional city detail and metro/location extensions.
+- Tashkent districts, metro, colloquial areas, residential complexes and POIs.
+- Odesa metropolitan entities with separate administrative, suburb/locality, informal-area, development-area, POI and search-cluster semantics.
+- Normalization for apostrophe/dash variants and Cyrillic search folding across RU/UZ/KZ/UA input.
+
+### Housing
+
+- Deal/action and listing-kind taxonomy.
+- Property/building type, condition, layout, furniture and listing-status context.
+- Rooms, bedrooms, floors, total floors, building year and area extraction.
+- Deposit, prepayment, utilities and commission/payment context.
+- Seller/agency/owner confidence and no-commission semantics.
+- Tenant policies, documents, financing, availability and location relations.
+- Infrastructure/proximity extraction such as walking time to metro.
+
+### Hiring
+
+- Candidate vs vacancy intent and non-content/spam classification.
+- Section-aware vacancy/CV parsing.
+- Profession taxonomy and profession context without duplicating consumer technology-skill catalogs.
+- Salary range, currency and pay-period extraction.
+- Experience requirements and seniority.
+- Work mode, employment types, schedules and probation.
+- Languages, CEFR levels and requirement relation (`required`, `preferred`, `notRequired`, `candidateHas`).
+- Work authorization, visa sponsorship, relocation, contracts, benefits and hiring status context.
+
+## Architecture
+
+Consumers should resolve data in this order:
+
+```text
+structured source fields
+  -> shared canonical lexicon
+  -> deterministic parser
+  -> AI fallback / enrichment
+```
+
+Lexical aliases belong here. Consumers keep source adapters, persistence, ranking/filtering semantics and UI/API contracts.
+
+The package must remain dependency-light and must never require a runtime lexicon HTTP service, Redis or a message broker.
+
+## Usage
+
+```js
+import {
+  canonicalCity,
+  canonicalTashkentMetro,
+  normalizeForMatch,
+  parseHousingStructured,
+  parseHiringContext,
+  parseHiringSalary,
+} from '@whiteslove/parsing-lexicon';
+
+canonicalCity('Қарағанды', 'KZ');
+// Karaganda
+
+canonicalTashkentMetro('метро Максим Горький');
+// Buyuk Ipak Yoli
+
+normalizeForMatch('Oʻzbekiston');
+// o zbekiston
+
+parseHousingStructured('3 xona, 5/9, 78 m², depozit 500$, metroga 5 minut piyoda');
+
+parseHiringSalary('Salary: $3,000–4,500 per month');
+
+parseHiringContext(
+  'Senior Frontend Engineer. English B2 required. Remote. Visa sponsorship available.',
+  { mode: 'vacancy', title: 'Senior Frontend Engineer' },
+);
+```
+
+Subpath exports are available for consumers that need narrower modules, including `./geo`, `./locations`, `./housing-context`, `./housing-structured`, `./hiring-context`, `./hiring-professions` and `./money`.
+
+## Data-quality rules
+
+- Add aliases to the existing canonical entity instead of creating source-specific regex tables.
+- Keep administrative entities distinct from colloquial/search concepts.
+- Preserve ambiguity guards where a bare term can mean several things.
+- Structured source fields take precedence over inferred values.
+- Add a regression for every production parsing bug.
+- Canonical dictionaries are immutable at runtime.
+- Duplicate canonical/alias entries and unsupported language-key drift are covered by invariants.
+
+## Lexicon statistics
+
+Snapshot for `c509d92` (2026-08-25). The audit traverses exported lexicon data from `src/`, excludes functions, regular expressions and obvious service/path strings, then normalizes strings with Unicode NFKC, trim, whitespace collapse and lowercase.
+
+| Metric | Count |
+| --- | ---: |
+| Alias/translation-like string occurrences across exports | 153,617 |
+| Unique alias/translation strings, exact | 17,681 |
+| **Unique alias/translation strings, normalized** | **17,397** |
+| All exported lexicon strings, normalized | 17,406 |
+| Normalization merge groups | 282 |
+
+The occurrence count is intentionally much larger than the unique count because aggregate modules re-export the same canonical datasets. For package size and vocabulary breadth, the normalized unique count is the useful number.
+
+### Vocabulary by domain
+
+| Domain | Occurrences across exports | Unique exact | Unique normalized |
+| --- | ---: | ---: | ---: |
+| Geography / locations | 64,898 | 10,149 | **9,967** |
+| Hiring — professions & skills | 6,738 | 2,350 | **2,324** |
+| Hiring — other semantics | 3,592 | 2,589 | **2,558** |
+| Housing | 2,554 | 1,975 | **1,971** |
+| Hiring — languages | 552 | 495 | **478** |
+| Country / source aliases | 170 | 126 | **126** |
+
+Domain rows are not additive: the same canonical string can be reachable through multiple exported structures and categories.
+
+Language totals are not inferred from script alone because many valid aliases are intentionally shared or untagged (for example names that are identical in RU/UK or EN/UZ Latin). A language-by-language table should therefore be generated only from explicitly language-tagged data rather than guessed from Unicode characters.
+
+## Development
+
+Requires Node.js 20+.
+
+```bash
+npm install
+npm test
+npm pack
+```
+
+`npm pack` produces the distributable `@whiteslove/parsing-lexicon` tarball from the files declared in `package.json`.