AVL/script/gethtml

134 lines
5.2 KiB
Plaintext
Raw Permalink Normal View History

#!/bin/zsh
# This script is just the initial step in porting the avl data to anki
# It does not work 100% reliable, but it's good enough for me
# Manual adjustments on the exported data should be expected
# This script is rather ugly in its Implementation
2020-09-18 15:05:47 +02:00
[ -f cards.csv ] && rm cards.csv
touch cards.csv
handle_cleaned_html(){
## Handling of multiple meanings
meaning_clean_func="$1"
no_func="$2"
word_func="$3"
pof_func="$4"
if echo "$meaning_clean_func" | grep -q "SUP"; then
count_func="$(echo "$meaning_clean_func" | grep -c "SUP")"
echo "Found $count_func multiple meanings for the word $word_func ($pof_func)"
for j in {1..$count_func}; do
meaning="$( echo "$meaning_clean_func" | grep "<SUP> $j" | perl -pe "s|<P> <SUP> $j </SUP> <TT> (.*?) </TT> </P>|\1|" )"
echo MEANING:
echo "$meaning"
echo DONE
[ -f categoriestmp ] && rm categoriestmp
touch categoriestmp
[ -f backofcardtmp ] && rm backofcardtmp
touch backofcardtmp
printout=false
while IFS= read -r line; do
if [ $printout = true ]; then
if echo "$line" | grep -q "<P> <SUP> $(( j + 1 )) </SUP>" ; then
printout=false
else
echo "$line" >> backofcardtmp
if echo "$line" | grep -q '<P> <U>'; then
echo "$line" | perl -pe "s|<P> <U> (.*?) </U>(.*?)$|<P> <U> \1 </U> </P>|" >> categoriestmp
fi
fi
fi
if echo "$line" | grep -q "SUP"; then
if echo "$line" | grep -q "<P> <SUP> $j </SUP>"; then
printout=true
fi
fi
done <<< $(echo "$meaning_clean_func")
echo BACKSIDE:
cat backofcardtmp
echo CATEGORIES
cat categoriestmp
echo DONE
categories="$(cat categoriestmp | tr '\n' ' ')"
backside="$(cat backofcardtmp | tr '\n' ' ' )"
printf "%s.%s;\"%s\";\"%s\";\"%s\";\"%s\";\"%s\"\n" "$no_func" "$j" "$word_func" "$meaning" "$categories" "$backside" "$pof_func" >> cards.csv
done
else
echo "Found only one meaning for the word $word_func ($pof_func)"
meaning="" #ozdic only provides meaning if there are different ones and the collocations for those different meanings differ.
[ -f categoriestmp ] && rm categoriestmp
touch categoriestmp
while IFS= read -r line; do
#echo current line:
#echo "$line"
echo "$line" | grep "<P> <U> " | perl -pe "s|<P> <U> (.*?) </U>(.*?)$|<P> <U> \1 </U> </P>|" >> categoriestmp
done <<< $(echo "$meaning_clean")
backside="$( echo "$meaning_clean_func" | grep "<P> <U> " | grep -vF '<DIV class="item"><P class="word"><B>'| tr '\n' ' ')"
#cat categoriestmp
categories="$(cat categoriestmp | tr '\n' ' ' )"
printf "%s;\"%s\";\"%s\";\"%s\";\"%s\";\"%s\"\n" "$no_func" "$word_func" "$meaning" "$categories" "$backside" "$pof_func" >> cards.csv
fi
#printf "%s;\"%s\";\"%s\"\n" "$no" "$word" "$meaning_clean" >> cards.csv
}
2020-09-18 15:05:47 +02:00
cat avl.csv | while read line || [[ -n $line ]];
do
no="$(echo "$line" | awk 'BEGIN{FS = "\t" } {print $1}' )"
word="$(echo "$line" | awk 'BEGIN{FS = "\t" } {print $2}' )"
printf "card:%s\t(%s)\n" "$no" "$word"
meaning_clean="$(curl --no-progress-meter http://www.ozdic.com/collocation-dictionary/"$word" | sed '6,35d; 38,48d' | tac | sed '6,30d' | tac | tr '\t' ' ' | tr -d '\r')"
#echo MEANINGCLEAN1:
#echo "$meaning_clean"
#echo DONE
echo "Checking whether there are multiple parts of speech..."
pof_count="$(echo "$meaning_clean" | grep -c "<P class=\"word\">" )"
if [ $pof_count -gt 1 ]; then
echo "$pof_count parts of speech found!"
# seperate into multiple "meaning_clean"
pofs="$( echo "$meaning_clean" | grep "<DIV class=\"item\"><P class=\"word\"><B>" | perl -pe "s|^.*?<DIV class=\"item\"><P class=\"word\"><B>[a-zA-Z]*? </B><I>([a-zA-Z\.]*?) </I> </P>|\1|g")"
#echo "pofs:"
#echo "$pofs"
pofs="$(printf "%s\nscript" "$pofs" )"
for i in {1..$pof_count}; do
current_pof=$( echo "$pofs" | sed "${i}q;d") #| sed 's/\./\\./g')
next_pof=$(echo "$pofs" | sed "$((i+1))q;d") #| sed 's/\./\\./g')
echo current_pof: $current_pof
echo next_pof: $next_pof
#start="<DIV class=\"item\"><P class=\"word\"><B>$word <\/B><I>$current_pof <\/I> <\/P>"
start="$( printf '<DIV class="item"><P class="word"><B>%s </B><I>%s </I> </P>\n' "$word" "$current_pof" )"
if [ $i -eq $pof_count ]; then
stop="$next_pof"
#echo MEANINGCLEAN_LASTPOF:
#echo "$meaning_clean"
#echo DONE
else
stop="$( printf '<DIV class="item"><P class="word"><B>%s </B><I>%s </I> </P>\n' "$word" "$next_pof" )"
fi
meaning_clean_temp="$( echo "$meaning_clean" | sed 's,'"$start"',START,')"
meaning_clean_temp="$( echo "$meaning_clean_temp" | sed 's,'"$stop"',STOP,' )"
echo MEANINGCLEAN_TEMP:
echo "$meaning_clean_temp"
echo DONE
pof_meaning_clean="$(echo "$meaning_clean_temp" | sed -n '/START/,/STOP/ { /START/n; /STOP/!p }' )"
echo POFMEANINGCLEAN
echo "$pof_meaning_clean"
echo DONE
notemp="$no"'.'"$i"
handle_cleaned_html "$pof_meaning_clean" "$notemp" "$word" "$current_pof"
done
else
echo "only one part of speech found"
pof="$( echo "$meaning_clean" | grep "<DIV class=\"item\"><P class=\"word\"><B>" | perl -pe "s|^.*?<DIV class=\"item\"><P class=\"word\"><B>[a-zA-Z]*? </B><I>([a-zA-Z\.]*?) </I> </P>|\1|g")"
echo POF:"$pof"
handle_cleaned_html "$meaning_clean" "$no" "$word" "$pof"
fi
2020-09-18 15:05:47 +02:00
done
rm backofcardtmp categoriestmp