Frequency filter - wordfreq (origin and licence)
================================================

Source:    wordfreq 3.1.1 by Robyn Speer, https://github.com/rspeer/wordfreq
           sdist https://pypi.org/project/wordfreq/3.1.1/ (wordfreq-3.1.1.tar.gz,
           SHA-256 7943098975f25c2a70e1151ee5a62083b14a5f86f6cc5703cc9526f716ceb408),
           data file wordfreq/data/large_pl.msgpack.gz
Retrieved: 2026-09-28
Licence:   The wordfreq data files "may be redistributed under a Creative Commons
           Attribution-ShareAlike 4.0 license" (https://creativecommons.org/licenses/by-sa/4.0/),
           code under the Apache License 2.0 (wordfreq README, section "License").
           wordfreq contains data derived from Wikipedia, OPUS OpenSubtitles 2018 (with attribution to
           OpenSubtitles, http://www.opensubtitles.org/), ParaCrawl, the Leeds Internet Corpus,
           Google Books Ngrams (http://books.google.com/ngrams) and other sources listed in its README.

Extract:   vendor/wordlists/pl/wordfreq-pl.tsv ("word<TAB>Zipf value"), made with
           scripts/wordlist-wordfreq-extrait.py: only entries written in lower-case letters of the
           language (no digits, punctuation or spaces), otherwise unchanged. Zipf = log10 of the
           frequency per billion words (1 = rare, 7 = very common).

Use:       scripts/build-wordlist.mjs keeps only the word forms of the spelling dictionary that
           occur in this frequency list with Zipf >= 1.3 (word finder, <length>.txt; size limit), and for the Wordle list
           (wordle.txt) only forms with Zipf >= 3 (about one occurrence per million words).
           wordfreq does not distinguish upper and lower case, so a rare word spelled like a common
           name or place (e.g. lower-case place or person names in the spelling dictionary) can pass the filter.

Because the lists combine the spelling dictionary with CC BY-SA 4.0 data, the word lists in this
folder (<length>.txt, wordle.txt) are distributed under the GNU General Public License version 3
(GPLv3 is a compatible licence for CC BY-SA 4.0 adaptations, see
https://creativecommons.org/share-your-work/licensing-considerations/compatible-licenses/).
Licence text: COPYING-GPL-3.0.txt.

Word-list pages (added 2026-09-28): scripts/build-wordlist.mjs (haeufigListe) also writes
haeufig-<length>.txt for the word-list pages: the words of the word finder list with Zipf >= 3.0,
2-15 letters, without country names (Unicode CLDR) and hand-checked name look-alikes,
one line "word<TAB>Zipf value", most frequent first. Licence of these files: GPL-3.0-or-later
(same as the word lists in this folder).