Movawikt-idioms — multiword expressions from English Wiktionary
Extracts idioms, proverbs, phrases, phrasal and multiword verbs, and register- or region-labelled words from an English Wiktionary XML dump read from stdin. Its output feeds the expressions base described in Pragmatics: frames — who did what to whom, why, and what came next and the idiom level of Worlds and knowledge.
Run
cargo build --release -p wikt-idioms
lbzip2 -dc $MOVA_DATA/raw/wiktionary/enwiktionary/<date>/*-multistream*.xml*.bz2 | target/release/wikt-idioms > idioms.tsv
A page counts if its title has a space (or a hyphen between letters) and its ==English== section has a multiword part of speech (Phrase, Proverb, Idiom, Verb, Noun, Adjective, Adverb, Interjection, Prepositional phrase) with an idiomatic label, an idiom or phrasal-verb category, the POS Phrase/Proverb/Idiom, or a verb + particle/object title. A single word counts only with a register label (slang, informal, formal, humorous, euphemistic, archaic …) or a regional variety label.
Output (TSV, no header)
| Column | Content |
|---|---|
| 1 | phrase |
| 2 | kind: idiom, phrasal-verb, verb-phrase, phrase, proverb, multiword, word |
| 3 | part of speech |
| 4 | first gloss |
| 5 | register labels with the gloss of the labelled sense (slang: …; humorous: …) |
| 6 | regional varieties with the gloss of the labelled sense (UK: …; Australia: …) |
Limits
Line-based parsing of export-0.11; glosses are crude wikitext-to-text (links kept as text, most templates dropped). The output is CC BY-SA (Wiktionary): attribution and share-alike apply, so it is for learning and tests, not for shipping; see Data and model licences.
This page in Markdown → https://md.mova.elamur.ai/free/en/docs/tools/wikt-idioms/