From 393eebbf1f0571aa6255a30da714eabc60123e4f Mon Sep 17 00:00:00 2001 From: Laurent Mazet Date: Wed, 17 Jun 2020 12:02:38 +0200 Subject: [PATCH] add Czech dictionary --- dict-dists-extractor/Makefile | 26 ++++++++++++++++++-------- 1 file changed, 18 insertions(+), 8 deletions(-) diff --git a/dict-dists-extractor/Makefile b/dict-dists-extractor/Makefile index 1c06e3e..5a2832d 100644 --- a/dict-dists-extractor/Makefile +++ b/dict-dists-extractor/Makefile @@ -2,9 +2,8 @@ ROOT_DICT = ../public/dict -LANGS = bg br ca cy da de el en eo es et fi fr ga hr hy is it lv nl no pl pt ro sv uk -# to add: ar cs he hu ru sk sl -# cs: bug +LANGS = bg br ca cs cy da de el en eo es et fi fr ga hr hy is it lv nl no pl pt ro sv uk +# to add: ar he hu ru sk sl # hu: too big #LANGS = af ar bg br ca cs cy da de el en eo es et eu fi fr ga he hr hu hy is it la lv ms nl no pl pt ro ru sk sl sv tr uk xk @@ -116,11 +115,22 @@ $(ROOT_DICT)/cs.dict: cs.low $(DEPEND) cat $< | \ grep -v '^.$$' | \ grep -v "['-]" | \ - tr '[a-záčďéěíňóřšťúůýž]' '[A-ZÁČĎÉĚÍŇÓŘŠŤÚŮÝŽ]' | \ - sort | \ - uniq > $@ - -#perl -p -e 'tr/aábcčdďeéěfghiíjklmnňoópqrřsštťuúůvwxyýzž/AÁBCČDĎEÉĚFGHIÍJKLMNŇOÓPQRŘSŠTŤUÚŮVWXYÝZŽ/' | \ + sed -e 's/[äâ]/a/g' \ + -e 's/[çćĆ]/c/g' \ + -e 's/[ęë]/e/g' \ + -e 's/[î]/i/g' \ + -e 's/[ľĽłĺ]/l/g' \ + -e 's/[ń]/n/g' \ + -e 's/[őöŐÖô]/o/g' \ + -e 's/[śŚ]/s/g' \ + -e 's/[űüŰÜ]/u/g' \ + -e 's/[ź]/z/g' | \ + sed -e 's/á/Á/g' -e 's/č/Č/g' -e 's/ď/Ď/g' -e 's/é/É/g' -e 's/ě/Ě/g' \ + -e 's/í/Í/g' -e 's/ň/Ň/g' -e 's/ó/Ó/g' -e 's/ř/Ř/g' -e 's/š/Š/g' \ + -e 's/ť/Ť/g' -e 's/ú/Ú/g' -e 's/ů/Ů/g' -e 's/ý/Ý/g' -e 's/ž/Ž/g' | \ + tr '[a-z]' '[A-Z]' | \ + sort | \ + uniq > $@ $(ROOT_DICT)/cy.dict: cy.low $(DEPEND) echo Building $@ from $< -- 2.30.2