Bash Search Function

search_shortcuts.sh

# Recursive File Contents Search function
# $1 = Search string (or -i to see list of ignored filetypes)
# $2 = (optional) File pattern (ex: *.js) (default is: *)
# $3 = (optional) Set to 1 for case-insensitive search (default is: 0 (case-sensitive))
# $4 = (optional) Comma-separated list of directories to ignore (format is: ".git,vendor,node_modules,bin")
# Bug: Backslashes are not displaying in results. (ex: "\n" shows up as "n")
function search(){
  # Define Vars
  local sep=$'\x01' # Obscure ascii character not likely to appear in files
  local col_spacing=3
  local bold='\e[1m'
  local end='\e[0m'
  local green='\e[32m'
  local purple='\e[35m'
  local start_red_hex='1b5b313b33316d' # printf '\e[1;31m' | xxd -p
  local stop_red_hex='1b5b306d'        # printf '\e[0m' | xxd -p
  local filter_swap_separators="s/^([^:]*):([^:]+):\s*(.*)$/\2$sep\1$sep\3/g"
  local col_line_w=0 # Column containing the line number's max width
  local col_path_w=0 # Column containing the file path's max width
  local fixed_strings='--fixed-strings '
  local search name case_sensitive find_array col_line col_path col_data error message usage filetypes_to_ignore escaped_search search_hex replace_hex ignore_paths count ignore_filetypes clip_cols

  # Ignore certain binary filetypes to speed up searching
  filetypes_to_ignore=(3gp 7z aac apng avi avif bmp class dll doc docx exe flac gif gpg gz gzip ico iso jar jfif jpeg jpg jrb mdb mkv mov mp3 mp4 mpeg odt ogg pdf pgp pgp_ png pp ppt pptx pyc rar 'so.*' sqlite svg tar tar tif tiff vob wav webm webp wma wmv xls xlsx zip)

  # Clip long results somewhere near the end of users' screen
  clip_cols=$(($COLUMNS-5))

  # Check for missing input
  if [[ -z "$1" ]]; then
    error="${bold}\e[37m\e[41m"
    message="${bold}\e[36m"
    usage="\n${bold}Recursive File Search${end}\n\n"
    usage+="${error}Error: %s${end}\n\n${message}Usage:\n${end}"
    usage+="search SEARCH_PATTERN [FILE_PATTERN] [CASE_INSENSITIVE] [IGNORE_DIRS]\n\n"

    usage+="${message}Parameters:${end}\n"
    usage+="SEARCH_PATTERN       The string to be matched\n"
    usage+="FILE_PATTERN         File matching pattern such as '*.php'\n"
    usage+="CASE_INSENSITIVE     0 for case-sensitive search (default), 1 for case-insensitive\n"
    usage+="IGNORE_DIRS          Comma-separated list of directories to ignore. Use '*/dirname' to exclude in nested directories.\n\n"

    usage+="${message}Examples:${end}\n"
    usage+="search '.ajax' '*.js' 1\n"
    usage+="search 'Fatal Error:' '*.log'\n"
    usage+="search '<div class=\"cart\">' '*' 0 '.git,vendor,node_modules,bin'\n\n"

    usage+="${message}Note:${end}\n"
    usage+="You can also pass in '-i' as the first parameter to see a list of ignored filetypes.\n\n"

    printf "${usage}" "No search string given"
    return 1
  fi

  # If user passes '-i' as the first parameter, show a list of the ignored filetypes
  if [[ "$1" == '-i' ]]; then
    printf "Filetypes that are ignored:\n"
    out="$(printf ".%s, " "${filetypes_to_ignore[@]}")"
    printf "%s\n\n" "$(echo $out | sed -E 's/(.*),/\1/')"
    return 1
  fi

  # Process user input
  search="$1"
  [[ -z "$2" ]] && name='*' || name="$2"
  [[ "$name" == "*." ]] && name='*'
  [[ "$3" == "1" ]] && case_sensitive='i'

  # Build out switches to ignore directories
  if [[ -n "$4" ]]; then
    IFS=',' read -r -a paths_to_ignore <<< "$4"

    # Expand out the array of paths to match this syntax:
    # \( -path "tmp/*" -o -path ".git/*" \) -prune -o
    ignore_paths='\('
    for path in "${paths_to_ignore[@]}"; do
      ignore_paths+=" -path \"${path}/*\" -o"
    done
    ignore_paths="$(echo $ignore_paths | sed -E 's/(.*)\-o/\1/')" # Trim off the last '-o'
    ignore_paths+='\) -prune -o'
  fi

  # Expand out the array of filetypes to ignore to match this syntax:
  # \( -name '*.gif' -o -name '*.png' \) -prune -o
  ignore_filetypes='\('
  for filetype in "${filetypes_to_ignore[@]}"; do
    ignore_filetypes+=" -name '*.${filetype}' -o"
  done
  ignore_filetypes="$(echo $ignore_filetypes | sed -E 's/(.*)../\1/')" # Trim off the last '-o'
  ignore_filetypes+='\) -prune -o'

  # Escape any special characters in search input for safety with grep
  escaped_search="$(printf -- '%s' "$search" | sed -e 's/\"/\\"/g' -e 's/`/\\`/g')"

  # Handle special case if search is just a semicolon
  if [[ "$search" == ';' ]]; then
    fixed_strings=''
    escaped_search='\;'
  fi

  # To avoid difficulty allowing all search characters without sed confusing them for regex characters,
  # convert search input to hex characters where replace is simple.
  search_hex="$(printf -- "%s" "$search" | xxd -p -c 1000000)" # Convert search into hex
  replace_hex="$start_red_hex$search_hex$stop_red_hex"  # Build replacement string in hex

  # Perform search and capture the results into an array
  # Match all paths except for ones that start with '..' (captures '.abc' and '..abc' as files/dirs)
  # Searching this way gets rid of the leading './'
  mapfile find_array < <( \
    eval "find * .[^.]* ..?* $ignore_paths $ignore_filetypes -type f -name '$name' -exec \
      grep -${case_sensitive}nH --color=never $fixed_strings -- \"$escaped_search\" {} + \
      2>/dev/null | grep -v -- '^Binary' | cut -c 1-$clip_cols | uniq | sed -r -e '$filter_swap_separators'" \
  )

  # Loop through the first time to determine max column widths and total count
  count=0
  while read -r line; do
    while IFS="$sep" read -r col_line col_path col_data; do
      [[ -n "$col_data" ]] && ((count ++))
      [[ $col_line_w -lt ${#col_line} ]] && col_line_w=${#col_line}
      [[ $col_path_w -lt ${#col_path} ]] && col_path_w=${#col_path}
    done < <(echo "${line[@]}")
  done < <(echo "${find_array[@]}")

  # Begin display results
  if [[ $count -eq 0 ]]; then
    printf "\nNo matches found\n\n"
  else
    # Add some padding
    ((col_line_w += col_spacing))
    if [[ $col_line_w -lt $((col_spacing + 4)) ]]; then
      col_line_w=$((col_spacing + 4)) # Because the heading "Line" is 4 chars, make it at least that long
    fi
    ((col_path_w += col_spacing))

    # Print heading
    printf "\n${bold}%-${col_line_w}s%-${col_path_w}s%s${end}\n" "Line" "File Path" "Search Results"
    printf   "${bold}%-${col_line_w}s%-${col_path_w}s%s${end}\n" "----" "---------" "--------------"

    # Loop through again to display output in columns
    while read -r line; do
      while IFS="$sep" read -r col_line col_path col_data; do
        if [[ -n "$col_data" ]]; then
          # Add color to search string in results (Do search/replace in hex mode and then swap back)
          col_data="$(printf -- "%s" "$col_data" | xxd -p -c 1000000 | sed -- "s/$search_hex/$replace_hex/g" | xxd -p -r)"
          printf -- "${green}%-${col_line_w}s$end${purple}%-${col_path_w}s$end%s\n" "$col_line" "$col_path" "${col_data//^\w/}"
        fi
      done < <(echo "${line[@]}")
    done < <(echo "${find_array[@]}")
    printf "\nMatches found: %s\n\n" "$count"
  fi
}

function se(){    search "$1" '*.'"$2"  0    "$3"; } # Search shortcut which puts in the *. prefix to a filetype for you
function si(){    search "$1" '*.'"$2"  1    "$3"; } # Case-insensitive shortcut function
function sphp(){  search "$1" '*.php'   "$2" "$3"; } # Search PHP files
function scss(){  search "$1" '*.css'   "$2" "$3"; } # Search CSS files
function sjs(){   search "$1" '*.js'    "$2" "$3"; } # Search JavaScript files

# Search for a count of matches within each file
function searchcount(){
  local matches
  matches="$(command grep -RHn "$1" 2>/dev/null | grep -v '^Binary' | cut -d: -f1 | uniq -c)"
  printf "Matches\tFilename\n-----\t--------------------------------\n%s\n" "$matches" | column -t
}

# Search for a count of case-insensitive matches within each file
function searchcounti(){
  local matches
  matches="$(command grep -RHni "$1" 2>/dev/null | grep -v '^Binary' | cut -d: -f1 | uniq -c)"
  printf "Matches\tFilename\n-----\t--------------------------------\n%s\n" "$matches" | column -t
}

These Bash functions search for contents within files starting from the current directory recursively. You need to source it initially then just invoke the functions. It is recommended that you add this to your bash startup files such as .bashrc or .bash_profile.