aboutsummaryrefslogblamecommitdiffstatshomepage
path: root/%HOME%/.bash_utils/text.sh
blob: 0ea007fd153c3f6f6bafca56f411b227851be4cd (plain) (tree)
1
2
3
4
5
6
7
8
9

                   
                                                        
                                                
                                                             

                                    

                                                                
 
                                                                           
 

                                                                                         
 

                                                                 
 
                                                                      

        
                                         


           
                                               
 
 
                       

                                         




                                            

                                                         










                        

 
                            

                                         










                                            

 
                







                                                        
                                         

                  
                                              




                                                   
                     


                                         
                                                        



                  
                                         
         
                  
                                              
         
 
                                                         
 
 
















                                                    




                                         
                     







                                         
                     


        
                        

                                         




                                            

                                                                         






                      
 
 



                                         
                                                



                  


                             
                                          


                                 




                                         
                                                
                
      
 
                  


                             
                                          
 
                                

 











                                                
                                          
 


                                


                                         
                    


                            


                      
                      
                                                                                             

                        
                                

                          
               

                     
               
                                                                       

                        
              
                              






                                    
                                                                                         



                          
                            
 



                                                                             
 
                  
 




                                                                          





                                         
                                                       





                    









                              
 

























                                                                                                         












































                                                                                                      
#!/usr/bin/env bash

# Copyright (c) 2016 Egor Tensin <Egor.Tensin@gmail.com>
# This file is part of the "linux-home" project.
# For details, see https://github.com/egor-tensin/linux-home.
# Distributed under the MIT License.

alias dos2eol='sed --binary --in-place -- '"'"'s/\(\r\?\)$//'"'"
alias eol2dos='sed --binary --in-place -- '"'"'s/\r\?$/\r/'"'"

alias trim='sed --binary --in-place -- '"'"'s/[[:blank:]]*\(\r\?\)$/\1/'"'"

alias trimeol='sed --binary --in-place -e :a -e '"'"'/^\n*$/{$d;N;ba}'"'"' --'
alias trimdoseol='sed --binary --in-place -e :a -e '"'"'/^\(\r\n\)*\r$/{$d;N;ba}'"'"' --'

alias eol='sed --binary --in-place -- '"'"'$a\'"'"
alias doseol='sed --binary --in-place -- '"'"'$s/\r\?$/\r/;a\'"'"

alias trimbom='sed --binary --in-place -- '"'"'1 s/^\xef\xbb\xbf//'"'"

lint() {
    trim "$@" && trimeol "$@" && eol "$@"
}

doslint() {
    trim "$@" && trimdoseol "$@" && doseol "$@"
}

_sed_escape_pattern() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -ne 1 ]; then
        echo "usage: ${FUNCNAME[0]} STR" >&2
        return 1
    fi

    # Only $^*./\[] need to be escaped according to this:
    # https://unix.stackexchange.com/a/33005/60124
    local s="$1"
    s="${s//'\'/'\\'}"
    s="${s//'/'/'\/'}"
    s="${s//'$'/'\$'}"
    s="${s//'^'/'\^'}"
    s="${s//'*'/'\*'}"
    s="${s//'.'/'\.'}"
    s="${s//'['/'\['}"
    s="${s//']'/'\]'}"
    s="${s//$'\n'/'\n'}"
    echo "$s"
)

_sed_escape_substitution() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -ne 1 ]; then
        echo "usage: ${FUNCNAME[0]} STR" >&2
        return 1
    fi

    local s="$1"
    s="${s//'\'/'\\'}"
    s="${s//'/'/'\/'}"
    s="${s//'&'/'\&'}"
    s="${s//$'\n'/'\n'}"
    echo "$s"
)

file_replace() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -lt 3 ]; then
        echo "usage: ${FUNCNAME[0]} OLD NEW PATH..." >&2
        return 1
    fi

    local old="$1"
    old="$( _sed_escape_pattern "$old" )"
    shift
    local new="$1"
    new="$( _sed_escape_substitution "$new" )"
    shift

    sed --binary --in-place -- "s/$old/$new/g" "$@"
)

file_replace_word() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -lt 3 ]; then
        echo "usage: ${FUNCNAME[0]} OLD NEW PATH..." >&2
        return 1
    fi

    local old="$1"
    old="$( _sed_escape_pattern "$old" )"
    shift
    local new="$1"
    new="$( _sed_escape_substitution "$new" )"
    shift

    sed --binary --in-place -- "s/\\b$old\\b/$new/g" "$@"
)

str_replace() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -ne 3 ]; then
        echo "usage: ${FUNCNAME[0]} STR SUB REP" >&2
        return 1
    fi

    local str="$1"
    local sub="$2"
    local rep="$3"

    sub="$( _bash_escape_pattern "$sub" )"

    echo "${str//$sub/$rep}"
)

str_tolower() (
    set -o errexit -o nounset -o pipefail

    local s
    for s; do
        echo "${s,,}"
    done
)

str_toupper() (
    set -o errexit -o nounset -o pipefail

    local s
    for s; do
        echo "${s^^}"
    done
)

_bash_escape_pattern() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -ne 1 ]; then
        echo "usage: ${FUNCNAME[0]} STR" >&2
        return 1
    fi

    # Only *?\[] need to be escaped according to:
    # http://wiki.bash-hackers.org/syntax/pattern#normal_pattern_language
    local s="$1"
    s="${s//'\'/'\\'}"
    s="${s//'*'/'\*'}"
    s="${s//'?'/'\?'}"
    s="${s//'['/'\['}"
    s="${s//']'/'\]'}"
    echo "$s"
)

str_contains() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -ne 2 ]; then
        echo "usage: ${FUNCNAME[0]} STR SUB" >&2
        return 1
    fi

    local str="$1"
    local sub="$2"

    [ -z "$sub" ] && return 0
    sub="$( _bash_escape_pattern "$sub" )"

    test "$str" != "${str#*$sub}"
)

str_starts_with() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -ne 2 ]; then
        echo "usage: ${FUNCNAME[0]} STR SUB" >&2
        return 1
    fi

    local str="$1"
    local sub="$2"

    [ -z "$sub" ] && return 0
    sub="$( _bash_escape_pattern "$sub" )"

    test "$str" != "${str#$sub}"
)

str_ends_with() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -ne 2 ]; then
        echo "usage: ${FUNCNAME[0]} STR SUB" >&2
        return 1
    fi

    local str="$1"
    local sub="$2"

    [ -z "$sub" ] && return 0
    sub="$( _bash_escape_pattern "$sub" )"

    test "$str" != "${str%$sub}"
)

str_split() (
    set -o errexit -o nounset -o pipefail

    local fmt='%s\n'
    local -a args

    while [ "$#" -ne 0 ]; do
        local key="$1"
        shift
        case "$key" in
            -h|--help)
                echo "usage: ${FUNCNAME[0]} [-h|--help] [-0|--null|-z|--zero] [--] STR DELIM"
                return 0
                ;;
            -0|--null|-z|--zero)
                fmt='%s\0'
                ;;
            --)
                break
                ;;
            -*)
                echo "${FUNCNAME[0]}: unrecognized parameter: $key" >&2
                return 1
                ;;
            *)
                args+=("$key")
                ;;
        esac
    done

    args+=("$@")

    if [ "${#args[@]}" -ne 2 ]; then
        echo "usage: ${FUNCNAME[0]} [-h|--help] [-0|--null|-z|--zero] [--] STR DELIM" >&2
        return 1
    fi

    local str="${args[0]}"
    local delim="${args[1]}"

    if [ "${#delim}" -ne 1 ]; then
        echo "${FUNCNAME[0]}: delimiter must be exactly 1 character long" >&2
        return 1
    fi

    local -a xs=()

    # Thanks to this guy for this trick:
    # http://stackoverflow.com/a/24426608/514684
    IFS="$delim" read -a xs -d '' -r < <( printf -- "%s$delim\\0" "$str" )

    [ "${#xs[@]}" -gt 0 ] && printf -- "$fmt" ${xs[@]+"${xs[@]}"}
)

str_join() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -lt 1 ]; then
        echo "usage: ${FUNCNAME[0]} DELIM [STR]..." >&2
        return 1
    fi

    local delim="$1"
    shift

    [ "$#" -eq 0 ] && return 0

    local str="$1"
    shift

    echo -n "$str"
    for str; do
        echo -n "$delim$str"
    done
    echo
)

str_grep() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -ne 1 ]; then
        echo "usage: ${FUNCNAME[0]} PATTERN" >&2
        return 1
    fi

    local pattern="$1"

    find . -type f -exec grep --fixed-strings --binary-files=without-match --regexp="$pattern" -- {} +
)

str_grep_word() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -ne 1 ]; then
        echo "usage: ${FUNCNAME[0]} PATTERN" >&2
        return 1
    fi

    local pattern="$1"

    find . -type f -exec grep --basic-regexp --binary-files=without-match --regexp="\b$pattern\b" -- {} +
)

str_iconv_fromto() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -lt 3 ]; then
        echo "usage: ${FUNCNAME[0]} SRC_ENCODING DEST_ENCODING PATH..." >&2
        return 1
    fi

    local src_encoding="$1"
    shift
    local dest_encoding="$1"
    shift

    local path path_dir tmp_path rm_tmp_path

    for path; do
        path_dir="$( dirname -- "$path" )"
        tmp_path="$( mktemp -- "$path_dir/${FUNCNAME[0]}_XXX" )"
        if iconv --from-code="$src_encoding" --to-code="$dest_encoding" -- "$path" > "$tmp_path"; then
            mv -f -- "$tmp_path" "$path"
        else
            rm -f -- "$tmp_path"
            return 1
        fi
    done
)

str_iconv() (
    set -o errexit -o nounset -o pipefail

    if [ "$#" -lt 2 ]; then
        echo "usage: ${FUNCNAME[0]} DEST_ENCODING PATH..." >&2
        return 1
    fi

    local dest_encoding="$1"
    shift

    local src_encoding path
    for path; do
        src_encoding="$( file --brief --mime-encoding -- "$path" )"
        str_iconv_fromto "$src_encoding" "$dest_encoding" "$path"
    done
)