#!/bin/bash

[ $1 ] && TV2HTML_DIR=$1
if [ ! -d "$TV2HTML_DIR" ]; then
  echo 'Usage: updatedb <directory>' >&2
  exit 1
fi

echo This is updatedb, started on `date`

rm -f $TV2HTML_DIR/All_stations.db

for db in $TV2HTML_DIR/*.db; do
  station=`basename $db .db`
  echo processing $station

  gawk -v "station=$station" '
    {									# Find hidden date
      append_station = 0
      if (match($0 "x", /{AL_MAGENTA}.*{CONCEAL.*[^0-9][0-9][0-9][0-9][0-9][0-9][0-9][^0-9]/)) {
        pos = match($0, /[^0-9][0-9][0-9][0-9][0-9][0-9][0-9][^0-9]/)
        date = substr($0, pos + 5, 2) substr($0, pos + 3, 2) substr($0, pos + 1, 2)
      } else if (match($0, /{AL_MAGENTA}{CONCEAL}[0-9][0-9][0-9][0-9][0-9][^0-9]/)) {
        append_station = 1
        dump = 0
      }

      if ($0 ~ /Satelliten-Empfang über ASTRA{GR_RED}/) {
        skippage = 1
      } else if ($0 ~ /bis.*[0-9][0-9]\.[0-9][0-9].*Uhr/ || $0 ~ /^{GR_RED}{GCHR_12}{GCHR_12}.*>>/ ||
          $0 ~ /ca.*um.*[0-9].?\.[0-9][0-9].*Ende/ || $0 ~ /^{PAGE [0-9][0-9][0-9]\/[0-9].*}$/ ||
          $0 ~ /[0-9]\.[0-9][0-9].*Sendeschluß/) {
        dump = 0
        skippage = 0
      } else if (!skippage && $0 ~ /^({[^}]*}| )[0-9][0-9]\.[0-9][0-9]( |{[^}]*})( |{[^}]*})[0-9][0-9][0-9][0-9]/) {
        dump = 1						# Find first line of telecast
        stripline = $0
        gsub(/{[^}]*}/, " ", stripline)
        firstline = 0
        tmp_station = station
        prefix = date "~" substr(stripline, 2, 2) substr(stripline, 5, 2) "~" substr(stripline, 9, 4) "~" station
        dumpprefix = 1
      }
      if (dump || append_station) {
        if (!date) {
          print "updatedb: Error: No date found." > "/dev/stderr"
          exit 1
        }
        stripline = $0						# Find sign for PALplus

        gsub(/{GR_YELLOW}{GCHR_12}/, " «BB» ", stripline)	# Replace symbols
        gsub(/{GR_YELLOW}{GCHR_63}/, " «UT» ", stripline)
        gsub(/{AL_YELLOW}\(16:9\)( *$|{AL_[A-Z]*})/, " «BB» ", stripline)
        gsub(/{AL_RED}UT( *$|{AL_[A-Z]*})/, " «UT» ", stripline)
        gsub(/{AL_YELLOW}d\/f( *$|{AL_[A-Z]*})/, " «ZK» ", stripline)
        gsub(/{AL_YELLOW}d\/spanisch( *$|{AL_[A-Z]*})/, " «ZK» ", stripline)
        gsub(/{AL_YELLOW}s\/w( *$|{AL_[A-Z]*})/, " «SW» ", stripline)
        gsub(/{AL_YELLOW}oo( *$|{AL_[A-Z]*})/, " «ZK» ", stripline)
        gsub(/{AL_YELLOW}II( *$|{AL_[A-Z]*})/, " «ZK» ", stripline)
        gsub(/{AL_YELLOW}DS( *$|{AL_[A-Z]*})/, " «DS» ", stripline)
        gsub(/{AL_YELLOW}DS oo( *$|{AL_[A-Z]*})/, " «DS» «ZK» ", stripline)
        gsub(/{AL_YELLOW}sw( *$|{AL_[A-Z]*})/, " «SW» ", stripline)
        gsub(/{AL_BLUE}Format 16:9/, " «BB» ", stripline)
        gsub(/{AL_GREEN}PAL PLUS/, " «BB» ", stripline)

        gsub(/{CONCEAL}/, "[", stripline)			# Remove hidden parts of page
        gsub(/{AL_[A-Z]*}/, "]", stripline)
        gsub(/{GR_[A-Z]*}/, "]", stripline)
        gsub(/{[^}]*}/, " ", stripline)
        stripline = stripline "]"
        while (match(stripline, /\[[^]]*\]/)) {
          sub(/\[[^]]*\]/, substr("                                        ", 1, RLENGTH), stripline)
        }
        gsub(/\]/, " ", stripline)
        gsub(/ *$/, "", stripline)				# Remove trailing whitespace
        gsub(/\.\.\.* ?[0-9][0-9][0-9]$/, "", stripline)	# Remove links to other pages
        gsub(/>>>* ?[0-9][0-9][0-9]$/, "", stripline)

        tmp_str = substr(stripline, 1, 10)			# Skip titles in the middle
        if (gsub(/[a-zA-Z]/, " ", tmp_str) >= 3) {		# of the page (arte)
          stripline = ""
        }

        if (!append_station) {
          if (stripline != "") {
            if (dumpprefix) {
              printf "\n" prefix
            }
            tmpline = stripline						# Find sign for subtitles
            if (gsub(/\*/, "«UT»", tmpline) > 1) {
              tmpline = stripline
              gsub(/\*$/, "«UT»", tmpline)
            }
            stripline = tmpline
            if (gsub(/°°/, "«ZK»", tmpline) > 1) {		# Find sign for 2-channel sound
              tmpline = stripline
              gsub(/°°$/, "«ZK»", tmpline)
            }
            tmpline = substr(tmpline, 14)
            gsub(/^ /, "", tmpline)
            printf "~" tmpline					# Write line for telecast to database
            if (length(append_str)) {
              printf "   (" append_str ")"
              append_str = ""
            }
          } else {
            dump = 0		# Skip everything after empty lines until next entry is found
          }
          dumpprefix = 0
        } else {
          append_str = stripline
          gsub(/^ */, "", append_str)
        }
      }
    }
    
    END {
      print ""
    }
  ' $TV2HTML_DIR/$station*[0-9] >> $db
done

cat $TV2HTML_DIR/*.db > $TV2HTML_DIR/All_stations.db