BASH - Shuffle characters in strings from several rows - ruby

Question

I have a file (filename.txt) with the following structure:
>line1
ABC
DEF
GHI
>line2
JKL
MNO
PQR
>line3
STU
VWX
YZ
I would like to shuffle the characters in the strings that do not start wit >. The output would (for example) look like the following:
>line1
DGC
FEI
HBA
>line2
JRP
OKN
QML
>line3
SZV
YXT
UW
This is what I tried to shuffle the characters for each >line[number]: ruby -lpe '$_ = $_.chars.shuffle * "" if !/^>/' filename.txt. The command works (see my post BASH - Shuffle characters in strings from file) but it shuffles line by line. I was wondering how I could modify the command to shuffle characters between all strings of each >line[number]). Using ruby is not a requirement.

Answer 1

First, we need to solve the problem: how to shuffle all characters in multiple lines:
echo -e 'ABC\nDEF\nGHI' |grep -o . |shuf |tr -d '\n'
GDAFHEIBC
and, we also need an array to record the length of each line in origin strings.
s=GDAFHEIBC
lens=(3 3 3)
start=0
for len in "${lens[#]}"; do
echo ${s:${start}:${len}}
((start+=len))
done
GDA
FHE
IBC
So, the origin multiple lines:
ABC
DEF
GHI
have been shuffled to:
GDA
FHE
IBC
Now, we can do our jobs:
lens=()
string=""
function shuffle_lines {
local start=0
local shuffled_string=$(grep -o . <<< ${string} |shuf |tr -d '\n')
for len in "${lens[#]}"; do
echo ${shuffled_string:${start}:${len}}
((start+=len))
done
lens=()
string=""
}
while read -r line; do
if [[ "${line}" =~ ^\> ]]; then
shuffle_lines
echo "${line}"
else
string+="${line}"
lens+=(${#line})
fi
done <filename.txt
shuffle_lines
Examples:
$ cat filename.txt
>line1
ABC
DEF
GHI
>line2
JKL
MNO
PQR
>line3
STU
VWX
YZ
>line4
0123
456
78
9
$ ./solution.sh
>line1
HFG
BED
AIC
>line2
JOP
KMQ
RLN
>line3
UVW
TYZ
XS
>line4
1963
245
08
7

Answer 2

#!/bin/bash
# echo > output.txt # uncomment to write in a file output.txt
mix(){
{
echo "$title"
line="$( fold -w1 <<< "$line" | shuf )"
echo "${line//$'\n'}" | fold -w3
} # >> output.txt # uncomment to write in a file output.txt
unset line
}
while read -r; do
if [[ $REPLY =~ ^\> ]]; then
mix
title="$REPLY"
else
line+="$REPLY"
fi
done < filename.txt
mix # final mix after loop's exit, otherwise line3 will be not mixed
exit
edited with comment of gniourf-gniourf

Answer 3

First create a test file.
str =<<FINI
>line1
ABC
DEF
GHI
>line2
JKL
MNO
PQR
>line3
STU
VWX
YZ
FINI
File.write('test', str)
#=> 56
Now read the file and perform the desired operations.
result = File.read('test').split(/(>line\d+)/).map do |s|
if s.match?(/\A(?:|>line\d+)\z/)
s
else
a = s.scan(/\p{Lu}/).shuffle
s.gsub(/\p{Lu}/) { a.shift }
end
end.join
# ">line1\nECF\nHIA\nGBD\n>line2\nJNP\nKLR\nOQM\n>line3\nTXY\nUZV\nSW\n"
puts result
>line1
ECF
HIA
GBD
>line2
JNP
KLR
OQM
>line3
TXY
UZV
SW
To do this from the command convert the code to a string with statements separated by a semicolon.
ruby -e "puts (File.read('test').split(/(>line\d+)/).map do |s|; if s.match?(/\A(?:|>line\d+)\z/); s; else; a = s.scan(/\p{Lu}/).shuffle; s.gsub(/\p{Lu}/) { a.shift }; end; end).join"
The steps are as follows.
a = File.read('test')
#=> ">line1\nABC\nDEF\nGHI\n>line2\nJKL\nMNO\nPQR\n>line3\nSTU\nVWX\nYZ\n"
b = a.split(/(>line\d+)/)
#=> ["", ">line1", "\nABC\nDEF\nGHI\n", ">line2", "\nJKL\nMNO\nPQR\n",
# ">line3", "\nSTU\nVWX\nYZ\n"]
Notice that the regular expression that is split's argument places >line\d+ within a capture group. Without that, ">line1", ">line2" and ">line3" would not be included in b.
c = b.map do |s|
if s.match?(/\A(?:|>line\d+)\z/)
s
else
a = s.scan(/\p{Lu}/).shuffle
s.gsub(/\p{Lu}/) { a.shift }
end
end
#=> ["", ">line1", "\nEAC\nIHB\nDGF\n", ">line2", "\nKQJ\nROL\nMPN\n",
# ">line3", "\nSUY\nXTV\nZW\n"]
c.join
#=> ">line1\nEAC\nIHB\nDGF\n>line2\nKQJ\nROL\nMPN\n>line3\nSUY\nXTV\nZW\n"
Now consider more closely the calculation of c. The first element of b is passed to the block and the block variable s is set to its value:
s = ""
We then compute
s.match?(/\A(?:|>line\d+)\z/)
#=> true
so "" is returned from the block. The regular expression can be expressed as follows.
/
\A # match the beginning of the string
(?: # begin a non-capture group
# match an empty space
| # or
>line\d+ # match '>line' followed by one or more digits
) # end non-capture group
\z # match the end of the string
/x # free-spacing regex definition mode.
Within the non-capture group an empty space was matched.
The next element of b is then passed to the block.
s = ">line1"
Again
s.match?(/\A(?:|>line\d+)\z/)
#=> true
so s is return from the block.
Now the third element of b is passed to the block. (Finally, something interesting.)
s = "\nABC\nDEF\nGHI\n"
d = s.scan(/\p{Lu}/)
#=> ["A", "B", "C", "D", "E", "F", "G", "H", "I"]
a = d.shuffle
#=> ["D", "C", "G", "H", "B", "F", "I", "E", "A"]
s.gsub(/\p{Lu}/) { a.shift }
#=> "\nDCG\nHBF\nIEA\n"
The remaining calculations are similar.