From 6f7849c9c26c14f578b0f85a7012306ac78bd47a Mon Sep 17 00:00:00 2001 From: Kevin Wallace Date: Fri, 5 Aug 2022 16:34:29 -0700 Subject: add phraser words500K and phrases5M lists https://lahosken.san-francisco.ca.us/frivolity/prog/phraser/ --- Makefile | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/Makefile b/Makefile index 9d52f79..398a518 100644 --- a/Makefile +++ b/Makefile @@ -5,6 +5,8 @@ all: enwiktionary-latest.version all: enwiktionary-latest.index all: enwiki-latest.version all: enwiki-latest.index +all: phraser-words500K.index +all: phraser-phrases5M.index .PHONY: newest newest: @@ -56,3 +58,15 @@ enwiktionary-%-snap.merge: nutrimatic/bin wikiextractor/WikiExtractor.py %-latest.index: %-latest.version make $*-$(shell cat $*-latest.version)-snap.index ln -sf $*-$(shell cat $*-latest.version)-snap.index $@ + +phraser-words500K.index: nutrimatic/bin + # ~ 3m + wget -O - https://lahosken.san-francisco.ca.us/frivolity/prog/phraser/words_500K.txt |\ + awk -F'\t' '{ for (i=0; i < $$1; i++) print $$2 }' |\ + ./merge.sh $@ + +phraser-phrases5M.index: nutrimatic/bin + # ~ 10m + wget -O - https://lahosken.san-francisco.ca.us/frivolity/prog/phraser/phrases_5M.txt |\ + awk -F'\t' '{ for (i=0; i < $$1; i++) print $$2 }' |\ + ./merge.sh $@ -- cgit v1.2.3