English → Portuguese
English to Brazilian Portuguese, built by pivoting Princeton WordNet 3.0 through OpenWN-PT and enriching with the Portuguese Wiktionary. Includes an inflection index, so looking up a word as it appears in a book works.
Download
en-pt.zip 14.9 MB · StarDict · CC-BY-SA-4.0
Install on your e-reader
Unzip and copy the folder onto your device so that you end up with:
koreader/data/dict/en-pt/en-pt.ifo
Restart KOReader. This dictionary includes an inflection index, so looking a word up exactly as it appears in a book works.
Coverage
How many of the most common words in the source language this dictionary can actually resolve, counting words reached through the inflection index.
| Frequency band | Found | As headword | Via inflection | Coverage |
|---|---|---|---|---|
| top 1,000 | 989 | 913 | 76 | 98.9% |
| top 5,000 | 4,857 | 4,133 | 724 | 97.14% |
| top 10,000 | 9,434 | 7,629 | 1,805 | 94.34% |
| top 50,000 | 38,383 | 27,510 | 10,873 | 76.77% |
Measured against OpenSubtitles 2018 English top 50k (hermitdave/FrequencyWords).
What is in an entry
| Entries with a real translation | 99,120 (57.35%) |
| Entries with only a definition, no translation | 73,722 |
| Entries with a usage example | 34,710 (20.08%) |
| Entries with inflected forms | 169,306 (97.95%) |
| Entries with pronunciation | 30,992 |
| Senses | 299,849 |
| Mean definition length | 61.9 characters |
| Mean translations per entry | 1.62 |
An entry counted as “definition only” is one where the source wordnet knows the concept but has no word for it in the target language. These are shipped because a definition still helps a reader, and counted separately because calling them translations would be dishonest.
Sources and licence
| Source | Licence | Credit |
|---|---|---|
omw-pt | CC-BY-SA-4.0 | OpenWN-PT (de Paiva & Rademaker), via Open Multilingual Wordnet 1.4 |
pwn30 | WordNet-3.0 | Princeton University WordNet 3.0 |
wikt-en-pt | CC-BY-SA-4.0 | English Wiktionary via kaikki.org (wiktextract) |
wikt-pt | CC-BY-SA-4.0 | Portuguese Wiktionary via kaikki.org (wiktextract) |
Bundled under CC-BY-SA-4.0, the least restrictive licence that covers every source above. Each source’s own terms are reproduced in the ATTRIBUTION file inside the download.