1

I've got the following one-liner with Perl:

perl -MLingua::EN::Sentence=get_sentences -ne 'print "$_\n" for grep { /replace_me/i } @{get_sentences($_)}' "${filePath}"

How can I restructure this so that I'm able to replace the "replace_me" section with any string of text (aka, a variable) that I can pass into the script as an argument?

I understand that you cannot add variables within single quotes, but I'm not sure how to restructure this at the moment with double quotes. Any help would be appreciated.

Harrison Cramer
  • 2,860
  • 5
  • 25
  • 51

3 Answers3

6

A "one-liner" is a program that can take input like any other -- what is passed to the program at invocation is in @ARGV array.

Then there are two particular issues to take into account

  • With -n the code under '' is the body of a loop (over lines from the files that are processed) so you want to extract your input in a BEGIN block, before the runtime and so before the loop starts

  • As there are also filenames given on the command line the argument passed to the script must be removed from @ARGV, so that what remains are filenames, that can then be used as needed under -n

Then

perl -wnE'BEGIN { $input = shift };  ... '  "input string" filename(s)

where ... stand for your code, just as it was, but which now can use $input.

Another way is with -s switch which enables a rudimentary mechanism for arguments

perl -s -wnE'...code using $input...' -- -input="input string"  filename(s)

where the name given after - (I used input above) is the name of the variable in which input gets stored. The -- are there to mark the beginning of the arguments. All arguments passed to the script must be given before the filename(s).

Finally, you can set a shell variable which the script will then see

input="input string"  perl -wnE'...$ENV{input}...' filenames

Or, set it up beforehand and export it

export input="input string"
perl -wnE'... $ENV{input} ...'

See this post for details on all three ways.

Once this input is used in a regex escape it using quotemeta, /...\Q$input\E.../i, unless it is meant to be a ready regex pattern. Please see the linked documentation.

zdim
  • 56,772
  • 4
  • 49
  • 75
1

There are three primary ways of passing information to Perl without using STDIN or external storage.

  • Arguments

    perl -ne'BEGIN { $text=shift(@ARGV) } print if /\Q$text/i' -- "$text"

  • Command-line options

    In a full program, you'd use Getopt::Long, but perl -s will do fine here.

    perl -sne'print if /\Q$text/i' -- -text="$text" --

  • Environment variables

    TEXT="$text" perl -ne'print if /\Q$ENV{TEXT}/i' --

In your case, you could use the following:

perl -ne'
   use Lingua::EN::Sentence qw( get_sentences );
   CORE::say for grep { /\Q$text/i } @{ get_sentences($_) };
' -- -text="$text" -- "$filePath"

I used /Q$text/i instead of /$text/i because I assumed the input is text to find rather than a regex pattern. \Q..\E converts text into a pattern that will match that text. (The \E is optional at the end of the literal.)

ikegami
  • 343,984
  • 15
  • 249
  • 495
-1

Provided that you have REPLACEMENT environment variable defined:

export REPLACEMENT=replace_me

You should be able to do either:

perl -MLingua::EN::Sentence=get_sentences -ne 'print "$_\n" for grep { /'$REPLACEMENT'/i } @{get_sentences($_)}' "${filePath}"

or:

perl -MLingua::EN::Sentence=get_sentences -ne 'print "$_\n" for grep { /$ENV{REPLACEMENT}/i } @{get_sentences($_)}' "${filePath}"

It probably also would be wise to quote the string used in the regexp with \Q and \E, e.g. grep { /\Q$ENV{REPLACEMENT}\E/i }.

eyevan
  • 1,455
  • 8
  • 20