mkscowl

#!/bin/bash
# mkscowl
# Concatenate, CAPITALIZE, and sort the scowl word lists of a specific size and smaller.
# After running this script, delete the old data directory and rerun generate_puzzles.py

# You may pick the spelling categories, sub-categories, and list size.
#
# For spelling categories, pick "english" and then add in "american" or "british" or both.
# For sub-categories, pick "words" and optionally add in "proper-names" or "abbreviations".
# For list size, pick 35 for a reasonable, but small dictionary. Or try 50, 70, 80, or 95.

spelling_categories={english,american,british}
#sub_categories={words,proper-names,abbreviations}
sub_categories=words		# No brace expansion for a single item.
size=${1:-35}			# Default to 35 if no command line arg given.

###

cd $(dirname $0)

if [[ ! -d scowl ]]; then
    echo "Error: Scowl files not found in $(dirname $0)" >&2
    exit 1
fi

outputfile=${spelling_categories}-${sub_categories}.${size}
outputfile=$(echo $outputfile | tr -d '{}')

possiblelists=$(eval echo scowl/$spelling_categories-$sub_categories.*)

for f in $possiblelists; do
    s=${f##*.}
    if [[ -e $f && $s -le $size ]]; then
	wordlists+=($f)
    fi
done

if [[ ${#wordlists} -eq 0 ]]; then
    echo "Error: No word lists matched scowl/$spelling_categories-$sub_categories.*" >&2
    echo "Please edit $0." >&2
    exit 1
fi

echo "Creating $outputfile from these wordlists:"
for f in ${wordlists[@]}; do
    echo "    $f"
done

cat ${wordlists[@]} |
    LANG=C grep -v "'" |			# Skip words with apostrophes
    iconv -f latin1 -t ASCII//TRANSLIT |	# Remove accents (café -> cafe)
    tr '[a-z]' '[A-Z]' |
    sort -u > $outputfile

echo "Done."

if ln -sf $outputfile scowl.txt; then
    echo "Linked $outputfile to scowl.txt"
fi

echo "You may now remove the 'data' directory and run generate_puzzles.py"

#####################################################################################
# Scowl file naming convention:							    #
#   <spelling category>-<sub category>.<size>					    #
# 										    #
# Where the spelling category is one of						    #
#   english, american, british, british_z, canadian, australian			    #
#   variant_1, variant_2, variant_3,						    #
#   british_variant_1, british_variant_2,					    #
#   canadian_variant_1, canadian_variant_2,					    #
#   australian_variant_1, australian_variant_2					    #
# 										    #
# Sub-category is one of							    #
#   abbreviations, contractions, proper-names, upper, words			    #
# 										    #
# And size is one of								    #
#   10, 20, 35 (small), 40, 50 (medium), 55, 60, 70 (large), 80 (huge), 95 (insane) #
#####################################################################################