GcinScripts01: scim2tab

#!/bin/sh
#
# This script is public domain.
#
# scim2tab

  VERSION=0.0.8

  COMPAT_SCIM_TABLES=0.5.8

  if [ "$1" = "" ]; then
    echo ${0##*/}, version $VERSION, compatible with \
         scim-tables-$COMPAT_SCIM_TABLES.
    echo "Usage: ${0##*/} file"
    exit 1
  fi

  if [ -f $1 ]; then
    CINFILE=`basename $1 .txt.in | tr A-Z a-z`.cin
  else
    echo ${0##*/}: $1: No such file.
    exit 2
  fi

  echo Generating $CINFILE.

# Remove null content lines. `cat -b' does not count null content lines.
  grep -v '^$' $1 > scim2tab.tmp

# Split scim2tab.tmp into xx00 and xx01.
#                       Print line numbers.
#                       Find SCIM_Generic_Table_Phrase_Library_TEXT.
#                       Remove comments.
#                       Make sure there is only one line.
#                       Get the line number and pass to csplit.
  csplit scim2tab.tmp ` cat -b scim2tab.tmp | \
                        grep SCIM_Generic   | \
                        grep -v \##         | \
                        head -n 1           | \
                        cut -f1             | \
                        tr -d ' '               ` > /dev/null

# These are SCIM's original messages, copyrights, licenses, etc.
  mv xx00 $CINFILE

# Append table author information.
  grep AUTHOR xx01   | \
  grep -v \##        | \
  tail -n 1          | \
  sed -e 's/=//g'      \
      -e 's/^/### /'   \
      -e "s/AUTHOR /This table\'s author is/" >> $CINFILE

# %gen_inp
  echo %gen_inp >> $CINFILE

# %ename
  grep 'NAME' xx01      | \
  grep -v 'NAME\.'      | \
  head -n 1             | \
  sed -e 's/NAME//'       \
      -e 's/ //g'         \
      -e 's/=//g'         \
      -e 's/^/%ename /' | \
  tr A-Z a-z >> $CINFILE

# %cname
  CNAME=` grep 'NAME\.' xx01      | \
          grep zh_TW              | \
          grep -v \##             | \
          sed -e 's/#//g'           \
              -e 's/ //g'           \
              -e 's/=//g'           \
              -e 's/NAME.zh_TW//'   \
              -e 's/^/%cname /'       `

  if [ "$CNAME" = "" ]; then
    echo %cname $CINFILE >> $CINFILE
  else
    echo $CNAME >> $CINFILE
  fi

# %selkey
  SELECT_KEY_DEF=` grep SELECT_KEYS xx01    | \
                   grep -v VoidSymbol       | \
                   grep -v \##              | \
                   head -n 1                | \
                   sed -e 's/SELECT_KEYS//'   \
                       -e 's/ //g'            \
                       -e 's/=//g'            \
                       -e 's/,//g'            \
                       -e 's/^/%selkey /'       `

  if [ "$SELECT_KEY_DEF" = "" ]; then
    echo %selkey 123456789 >> $CINFILE
  else
    echo $SELECT_KEY_DEF >> $CINFILE
  fi

# %endkey
  grep KEY_END_CHARS xx01    | \
  grep -v \##                | \
  head -n 1                  | \
  sed -e 's/KEY_END_CHARS//'   \
      -e 's/ //g'              \
      -e 's/=//g'              \
      -e 's/^/%endkey /' >> $CINFILE

# %space_style
  grep AUTO_COMMIT xx01    | \
  grep -v \##              | \
  grep TRUE                | \
  sed -e 's/AUTO_COMMIT//'   \
      -e 's/ //g'            \
      -e 's/=//g'            \
      -e 's/TRUE//'          \
      -e 's/^/%space_style 1/' >> $CINFILE

# %keep_key_case
  echo %keep_key_case >> $CINFILE

# %keyname begin
# .
# .
# %keyname end
  echo %keyname begin >> $CINFILE

  KEYNAME_BEGIN=`grep BEGIN_CHAR_PROMPTS_DEFINITION xx01 | grep -v \##`

  if [ "$KEYNAME_BEGIN" = "" ]; then
    grep VALID_INPUT_CHARS xx01 | \
    grep -v \##                 | \
    head -n 1                   | \
    cut -d ' ' -f3              | \
    fold -w 1 -                 | \
    sed 's/\(^.$\)/\1 \1/'      | \
    perl -nle '(print,$hash{$_}=1) unless defined $hash{$_}' >> $CINFILE
  else
    mv xx01 scim2tab.tmp
    csplit scim2tab.tmp ` cat -b scim2tab.tmp                | \
                          grep BEGIN_CHAR_PROMPTS_DEFINITION | \
                          grep -v \##                        | \
                          head -n 1                          | \
                          cut -f1                            | \
                          tr -d ' '                              ` > /dev/null
    mv xx01 scim2tab.tmp

    csplit scim2tab.tmp ` cat -b scim2tab.tmp              | \
                          grep END_CHAR_PROMPTS_DEFINITION | \
                          grep -v \##                      | \
                          head -n 1                        | \
                          cut -f1                          | \
                          tr -d ' '                            ` > /dev/null
    sort xx00                             | \
    grep -v BEGIN_CHAR_PROMPTS_DEFINITION | \
    perl -nle '(print,$hash{$_}=1) unless defined $hash{$_}' >> $CINFILE
  fi

  echo %keyname end >> $CINFILE

# %chardef begin
  echo %chardef begin >> $CINFILE

  mv xx01 scim2tab.tmp

  csplit scim2tab.tmp ` cat -b scim2tab.tmp | \
                        grep BEGIN_TABLE    | \
                        grep -v \##         | \
                        head -n 1           | \
                        cut -f1             | \
                        tr -d ' '               ` > /dev/null
  echo \#! /bin/bash > scim2tab.tmp

  sed 's/$/\t0/' xx01 | \
  tr -s ' ' ' '       | \
  sed -e 's/ /  /'      \
      -e 's/ /  /' > xx00
  sort -g -r -k 3 xx00                                     | \
  cut -f3                                                  | \
  perl -nle '(print,$hash{$_}=1) unless defined $hash{$_}' | \
  grep -v '^$'                                             | \
  sed -e 's/^/grep /'                                        \
      -e 's/$/ xx00 \>\> xx01/' >> scim2tab.tmp
  echo grep 0 xx00 \>\> xx01 >> scim2tab.tmp

  rm xx01
  bash scim2tab.tmp
  grep -v _TABLE xx01 | \
  grep -v \###        | \
  sed 's/\t0$//'   | \
  perl -nle '(print,$hash{$_}=1) unless defined $hash{$_}' >> $CINFILE

# Compatibility issue.
# This is a strange work around for Bopomofo inputmethod.
# Anyway, it works! :-)
  grep KEY_END_CHARS $1      | \
  grep -v \##                | \
  head -n 1                  | \
  sed -e 's/KEY_END_CHARS//'   \
      -e 's/ //g'              \
      -e 's/=//g'              \
      -e 's/^/linux     linux/' >> $CINFILE

  rm scim2tab.tmp xx00 xx01

  echo

  gcin2tab $CINFILE | grep -v 'gcin2tab encoding UTF-8' | \
                      grep -v 'please use iconv'        | \
                      grep -v 'skip gen_inp'

  GTABFILE=`basename $1 .txt.in | tr A-Z a-z`.gtab

  if [ -f $GTABFILE ]; then
    echo
    echo Generating $GTABFILE. Done.
  else
    echo
    echo Failed to generate $GTABFILE.
    exit 3
  fi

exit 0