# Phở Learn dataset and license notices

This build does not claim to know every Vietnamese or English word. Results are lexical aids for entertainment and light learning, not a professional or authoritative dictionary.

## English spelling: SCOWL / en_US Hunspell

- Source package: `dictionary-en-us` 2.2.1
- Upstream: [SCOWL / English Speller Database](https://wordlist.aspell.net/)
- Use: spelling membership, inflected forms and suggestion candidates
- Coverage in this build after Hunspell expansion and WordNet merging: about 223,000 normalized forms
- License: SCOWL permission notice plus BSD-covered affix components; the package declares `(MIT AND BSD)`
- Commercial use: permitted, including use, modification, distribution and sale, provided notices are retained
- Required notice: `SCOWL-HUNSPELL-LICENSE.txt`

## English definitions: Princeton WordNet

- Source package: `wordnet-db` 3.1.14, containing WordNet 3.1 files
- Upstream: [Princeton WordNet](https://wordnet.princeton.edu/)
- Use: English definitions and parts of speech
- Coverage in this build: about 145,961 normalized entries have at least one definition
- License: Princeton WordNet license
- Commercial use: permitted for any purpose without fee or royalty, provided the copyright notice and disclaimer remain with copies
- Required notice: `WORDNET-LICENSE.txt`

## Vietnamese words and Vietnamese definitions: VNTK Dictionary

- Source package: `@vntk/dictionary` 1.0.0
- Upstream: [VNTK Dictionary](https://github.com/vntk/dictionary)
- Use: Vietnamese word membership, expressions, definitions and parts of speech
- Published package contents: a roughly 74,000-item word list and about 31,327 definition keys
- License: MIT, copyright 2017 Vietnamese Natural Language Processing (VNTK)
- Commercial use and redistribution: permitted with the copyright and permission notice
- Required notice: `VNTK-LICENSE.txt`
- Limitation: the package README does not document the individual provenance of every compiled entry. Phở Learn supplements this data with separately attributed Wiktionary-derived data and reports unknown words cautiously.

## Vietnamese English glosses and additional entries: Kaikki / Wiktionary

- Source: [Kaikki Vietnamese machine-readable dictionary](https://kaikki.org/dictionary/Vietnamese/index.html), extracted from English Wiktionary
- Use: English glosses, additional Vietnamese expressions and parts of speech
- Coverage: merged output contributes to roughly 52,502 Vietnamese entries with definitions or glosses; the combined Vietnamese spelling index contains about 96,240 forms
- License: the same licenses as Wiktionary — Creative Commons Attribution-ShareAlike and GNU Free Documentation License
- Commercial use: permitted subject to attribution and share-alike/license requirements
- Attribution: Wiktionary contributors; extraction by Tatu Ylonen / Wiktextract / kaikki.org
- Derived data: `site/data/vi/*.json` is redistributed under CC BY-SA 4.0
- Required notice: `KAIKKI-WIKTIONARY-NOTICE.txt`

## Partial English → Vietnamese translations: tratu-core

- Source package: `tratu-core` 2.2.0 by Hieu Dong
- Package page: [npm](https://www.npmjs.com/package/tratu-core)
- Use: short Vietnamese equivalents where available
- License declared by the published package: ISC
- Commercial use: permitted with the copyright and permission notice
- Required notice: `TRATU-ISC-LICENSE.txt`
- Limitation: this source is partial, older, and its package does not document per-entry provenance. Translations are shown only when present and are not claimed complete.

## Coverage and known limitations

- English spellings: broad US-English coverage, but specialist terms, new slang, names and other dialect spellings may be absent.
- English definitions: broad but not complete for all spelling forms; inflected forms may be recognized without their own definition.
- Vietnamese spellings: broad single-word and multi-syllable expression coverage, but regional usage, names, slang and recent terms may be absent.
- Vietnamese definitions/glosses: partial. Some entries have only a Vietnamese definition, only an English gloss, or no short definition.
- Translations: partial and context-dependent. The app never creates a translation for an unknown entry.
- Suggestions are ranked heuristically; an unknown word is never labeled definitely incorrect.

This report is an engineering license review, not legal advice. All required source notices are shipped with the test build.

## V81 English pronunciation recordings

English pronunciation may use individual audio recordings hosted on Wikimedia Commons. The file page, contributor and license are shown next to the pronunciation button and in listening-game rounds. License conditions vary by recording; follow the linked Commons file page for full credits and terms. MP3 transcodes are played from Wikimedia, not bundled into this package. If a recording is absent or unsupported, the device's speech synthesis is used.