rename-ext/rename-ext.sh

#!/usr/bin/env bash

## rename-ext
# > changes file-extension based on mime-type

## also
# - removes `*?|^:"<>` characters from the file-name
# - reducing multiple spaces, dots, and underscores
# - adds rename-date suffix to ensure uniquness of the filename
# - strips exec-flag from non-executive types (usual leftover from FAT/NTFS)
# - optimizes images while at it (shows reduction size in `kb` and `%`)

## example
# rename-ext /Data/Pictures/_unsorted/**/*

## requires
# - file - to determine file type
# - `cargo install sd` – better sed
# - jpegoptim – to compress JPEGs
# - imagemagick or graphicsmagic – to mogrify PNGs
# - pngcrush – to compress PNGs (losless)
# // - pngquant – to compress PNGs (lossy)

ALL_EXTS=`echo \
  'gif|jpe?g|jp2|jfif|a?png|web(p|m)|svg
  |eps|tga|tiff?|psd|ico|xcf|heic
  |eot|otf|ttf|epub|doc|docx|xls|swf|pdf|odt
  |flac|opus|ogg|m4a|wav
  |mpe?g|mp\d|mov|mkv|avif?|asf|3gpp?|av1
  |html?|sh|py|php
  ' \
  |  tr -d '[:space:]' \
  `

mkdir -p ~/tmp


OPTIONS=''
CHX_CHANGELOG=''
NR='1'


calc_reduction()
{
  SIZE_PRE="$1"
  SIZE_POST="$2"
  PERECENT=` echo "scale=2; 100 - (100 * $SIZE_POST / $SIZE_PRE)" | bc `
  RESULT="$(( $SIZE_PRE / 1000 ))-$(( ( $SIZE_PRE - $SIZE_POST ) / 1000 )) kb \
($PERECENT%)
"
  echo "$RESULT"
}


ch_x()
{
  FILENAME="$1"
  # echo "FILENAME = '$FILENAME'"

  CHX_CHANGELOG=''

  IS_X_INTENDED='0'
  if [ ! -z "$2" ]; then
    IS_X_INTENDED='1'
  fi
  # echo "IS_X_INTENDED = '$IS_X_INTENDED'"

  IS_X=`stat -c '%A' "$FILENAME" | grep 'x' | wc -l`
  # echo "IS_X = '$IS_X'"

  MODE='-x'
  if [[ '1' == "$IS_X_INTENDED" ]]; then
    MODE='+x'
  fi
  # echo "MODE = '$MODE'"

  if [[ ! "$OPTIONS" =~ 'test-run' ]]; then
    # echo '(not a test-run)'
    if [[ ! "$OPTIONS" =~ 'only-ext' ]]; then
      # echo '(not only-ext)'
      if [[ ! "$IS_X" == "$IS_X_INTENDED" ]]; then
        CHX_CHANGELOG="chmod $MODE"
      fi
      chmod "$MODE" "$FILENAME"
    fi
  fi
}


rename_ext()
{
  NAME="$1"
  EXT="$2"
  NOTE=''
  TARGET=''

  ch_x "$FILENAME"

  if [[ "$OPTIONS" =~ 'only-ext' ]]; then
    NAME_NO_EXT=` echo "$NAME" \
                | sd -f i "(\.+($ALL_EXTS))$" '' \
                `
    TARGET="$NAME_NO_EXT.$EXT"
  else
    if [[ ! "$OPTIONS" =~ 'test-run' ]]; then

      SIZE_PRE=` stat "$FILENAME" -c '%s' `

      [ 'jpg' = "$EXT" ] \
        && jpegoptim "$FILENAME" -qsftp
      if [ 'png' = "$EXT" ]; then
        mogrify "$FILENAME"
        pngcrush -nolimits -s "$FILENAME" "$FILENAME.tmp.png"
        # pngquant "$FILENAME" --output "$FILENAME" \
        #   --quality=100 --force --skip-if-larger --speed 1
        if [ -f "$FILENAME.tmp.png" ] \
          && [ -s "$FILENAME.tmp.png" ]
        then
          mv "$FILENAME.tmp.png" "$FILENAME"
        else
          rm "$FILENAME.tmp.png"
        fi
      fi

      SIZE_POST=` stat "$FILENAME" -c '%s' `

      [ $SIZE_PRE -gt $SIZE_POST ] \
        && NOTE=` calc_reduction "$SIZE_PRE" "$SIZE_POST" `
    fi
    DATE=` date -r "$NAME" -u "+%Y%m%d%H%M%S" `
    TCLEAN=` echo "$NAME" \
           | sd '[\\\?\*\|^:"<>]+' '_' \
           | sd '_\d{12,}\.' '.' \
           | sd -f i "(\.+($ALL_EXTS))+$" '' \
           `
    TARGET=` echo "${TCLEAN}_$DATE.$EXT" \
          | sd '([\s\.,_]){3,}' '$1' \
          `
  fi

  IS_SAME_NAME='0'
  if [[ "$NAME" == "$TARGET" ]]; then
    IS_SAME_NAME='1'
  fi
  # echo "IS_SAME_NAME = '$IS_SAME_NAME'"

  # echo "CHX_CHANGELOG = '$CHX_CHANGELOG'"
  if [[ '1' == "$IS_SAME_NAME" ]] \
    && [ -z "$CHX_CHANGELOG" ]
  then
    return
  fi

  echo "$NR:"
  NR="$(($NR+1))"
  echo "  $FILENAME"

  if [ ! -z "$CHX_CHANGELOG" ]; then
    echo "  $CHX_CHANGELOG"
    CHX_CHANGELOG=''
  fi

  if [[ '1' == "$IS_SAME_NAME" ]]; then
    return
  fi


  if [[ "$OPTIONS" =~ 'test-run' ]]; then
    if [ -f "$TARGET" ]; then
      if [[ "$PARAMS" =~ 'force-replace' ]]; then
        echo "  >> replacing existing file >>"
      else
        echo "  !! file already exist !!"
      fi
    fi
    echo "$TARGET"
  else
    if [ ! -f "$TARGET" ]; then
      echo "  >> $NOTE >>"
      echo "  $TARGET"
      mv "$NAME" "$TARGET"
    else
      if [[ "$PARAMS" =~ 'force-replace' ]]; then
        echo "  >> $NOTE ## replacing existing file >>"
        echo "  $TARGET"
        mv "$NAME" "$TARGET"
      else
        echo '  !! File already exist, use `--force-replace` to overwrite !! '
        echo "  $TARGET"
      fi
    fi
  fi

}

for PARAM in "$@"; do
  # echo "PARAM = '$PARAM'"
  if [[ '--only-ext' == "$PARAM" ]]; then
    OPTIONS="$OPTIONS;only-ext"
    # echo "OPTIONS = '$OPTIONS'"
  elif [[ '--test-run' == "$PARAM" ]]; then
    OPTIONS="$OPTIONS;test-run"
    # echo "OPTIONS = '$OPTIONS'"
  elif [[ '--force-replace' == "$PARAM" ]]; then
    OPTIONS="$OPTIONS;force-replace"
    # echo "OPTIONS = '$OPTIONS'"
  fi
done

[ ! -z "$OPTIONS" ] \
  && echo "OPTIONS: '$OPTIONS'"


# echo "@ = '$@'"
for FILENAME in "$@"; do

  # echo "FILENAME = '$FILENAME'"

  # NOT_F_=`[ ! -f "$FILENAME" ]`
  # echo "NOT_F_ = '$NOT_F_'"

  # [ ! -f "$FILENAME" ] \
  #   && continue

  FILE_TYPE=` file -b --mime-type "$FILENAME" `
  # echo "FILE_TYPE = '$FILE_TYPE'"

  case "$FILE_TYPE" in

  # pixel-based
  'image/gif') rename_ext "$FILENAME" 'gif' ;;
  'image/jpeg') rename_ext "$FILENAME" 'jpg' ;;
  'image/jp2') rename_ext "$FILENAME" 'jp2' ;;
  'image/png') rename_ext "$FILENAME" 'png' ;;
  'image/webp') rename_ext "$FILENAME" 'webp' ;;
  'image/avif') rename_ext "$FILENAME" 'avif' ;;

  # vector-based
  'image/svg+xml') rename_ext "$FILENAME" 'svg' ;;
  'image/x-eps') rename_ext "$FILENAME" 'eps' ;;

  # some less common image formats
  'image/heic') rename_ext "$FILENAME" 'heic' ;;
  'image/tiff') rename_ext "$FILENAME" 'tiff' ;;
  'image/vnd.adobe.photoshop') rename_ext "$FILENAME" 'psd' ;;
  'image/vnd.microsoft.icon') rename_ext "$FILENAME" 'ico' ;;
  'image/x-tga') rename_ext "$FILENAME" 'tga' ;;
  'image/x-xcf') rename_ext "$FILENAME" 'xcf' ;;

  # fonts
  'application/vnd.ms-fontobject') rename_ext "$FILENAME" 'eot' ;;
  'application/vnd.ms-opentype') rename_ext "$FILENAME" 'otf' ;;
  'font/sfnt') rename_ext "$FILENAME" 'ttf' ;;

  # documents
  'application/epub+zip') rename_ext "$FILENAME" 'epub' ;;
  'application/msword') rename_ext "$FILENAME" 'doc' ;;
  'application/pdf') ch_x "$FILENAME" ;;  # can be .ai
  'application/vnd.ms-excel') rename_ext "$FILENAME" 'xls' ;;
  'application/vnd.oasis.opendocument.text') rename_ext "$FILENAME" 'odt' ;;
  'application/vnd.openxmlformats-officedocument.wordprocessingml.document') rename_ext "$FILENAME" 'docx' ;;
  'application/x-shockwave-flash') rename_ext "$FILENAME" 'swf' ;;

  # audio
  'audio/flac') rename_ext "$FILENAME" 'flac' ;;
  'audio/mpeg') rename_ext "$FILENAME" 'mp3' ;;
  'audio/ogg') rename_ext "$FILENAME" 'ogg' ;;
  'audio/x-m4a') rename_ext "$FILENAME" 'm4a' ;;
  'audio/x-wav') rename_ext "$FILENAME" 'wav' ;;

  # video
  'video/3gpp') rename_ext "$FILENAME" '3gp' ;;
  'video/av1') rename_ext "$FILENAME" 'avi' ;;
  'video/MP2T') rename_ext "$FILENAME" 'mp2' ;;
  'video/mp4') rename_ext "$FILENAME" 'mp4' ;;
  'video/quicktime') rename_ext "$FILENAME" 'mov' ;;
  'video/webm') rename_ext "$FILENAME" 'webm' ;;
  'video/x-m4v') rename_ext "$FILENAME" 'mp4' ;;
  'video/x-matroska') rename_ext "$FILENAME" 'mkv' ;;
  'video/x-ms-asf') rename_ext "$FILENAME" 'asf' ;;

  # text
  'application/json') ch_x "$FILENAME" ;;  # can be any other language
  'text/html') ch_x "$FILENAME" ;;  # can be .htm, .htc, .mht, ...
  'text/xml') ch_x "$FILENAME" ;;  # can be .xml, .opml, .rss, ...
  'text/plain') ;;  # can be any file with not enough text
  'text/x-python') ;;  # can be .py, .py3, have no .ext
  'text/x-shellscript') ;;  # can be .sh, .zsh, have no .ext

  # binady data
  'application/x-sqlite3') ch_x "$FILENAME" ;;  # can be .db, .sql, .sq3

  # special
  'application/octet-stream') ;;  # can be anything
  'inode/directory')
    # echo "$FILENAME is a folder"
    # echo "about to ch_x"
    ch_x "$FILENAME" '+'
    ;;  # a folder
  'inode/x-empty') ch_x "$FILENAME" ;;  # zero-size

  # default
  *) echo "??  $FILENAME : $FILE_TYPE" ;;

  esac

done

# if [[ "@#" -eq 1 ]]; then
#   if [[ `file -b --mime-type "$1"` == "inode/directory" ]]; then
#     find "$1"
#   fi
# fi