editing multiple html files at once

phantom_cyph · 02-27-2008, 07:49 AM

I know there is a way to do it, but I can't remember how.

I need to be able to change this line:

Quote:

../white.jpg

to this line:

Quote:

35468/jpg/white.jpg

I have to do it with around 20 files and don't want to do it one by one. How do I do it?

bitpicker · 02-27-2008, 08:26 AM

If you're using Quanta Plus with the KFileReplace plugin you can do that for all files in a given directory for example.

Robin

forrestt · 02-27-2008, 09:22 AM

I wrote this perl script a few years ago that does this recursively.

Code:

#!/usr/bin/perl

#####
# I haven't figured out why, but '@' symbol must be escaped in $find and NOT
# escaped in $replace
#####

# The string we are looking for
$find = '../white.jpg';

# The string we are replacing it with
$replace = '35468/jpg/white.jpg';

# add any particular file type to search for
# (e.g. @extensions = ( "html", "shtml", "pl");)
@extensions = ( "html" );

foreach $extension (@extensions) {
    # Search the filesystem for all specified file types
    open (FILELIST, "/usr/bin/find . -name \\\*.$extension|");
    while (<FILELIST>) {
        $file = $_;
        # skip over the "notthisdir" pages (we already know we're moving)
        if ($file =~ m/notthisdir/) {
            next;
        };

        # we need to escape certain characters in file names
        # (and probably find out who used them and give 'em a beating)
        $file =~ s/ /\\ /g;
        $file =~ s/\&/\\\&/g;
        $file =~ s/\$/\\\$/g;
        $file =~ s/~/\\~/g;
        $file =~ s/\(/\\\(/g;
        $file =~ s/\)/\\\)/g;

        # let's lose the newline at the end of the file name
        chomp ($file);

        # if we have an empty file name skip it
        if ($file eq '') {
            next
        };

        # Rather than read and write each file, make sure the file has
        # the search term first
        open (GREP, "/usr/bin/grep -l $find $file|");
        while (<GREP>) {
            $skipthis = $_;
        }
        close (GREP);

        if ($skipthis eq '') {
            next;
        } else {
            print "Updating file: $skipthis;
            $skipthis = '';
        }

        # Read in the page or die
        open(PAGE,"$file") || die "I can't open $file";
        while(<PAGE>){
            $thisPage= $thisPage . $_;
        }
        close(PAGE);

        # do our string substitutions
        $thisPage =~ s#$find#$replace#g;

        # write out the same page with chages
        open(PAGENEW,">$file") || die "I can't write the new html to $file";
        print PAGENEW $thisPage;
        close(PAGENEW);

        # reset some values
        $thisPage="";
        $file="";
    }
    close (FILELIST);
}

Use at your own risk.

HTH

Forrest

pwc101 · 02-27-2008, 09:31 AM

Sounds like a job for sed, Batman!

Code:

sed -i 's/..\/white.jpg/34568\/jpg\/white.jpg/g' *.html

Recursively, I suppose this could be done in conjunction with find:

Code:

find ./ -iname "*.html" -exec sed -i 's/..\/white.jpg/34568\/jpg\/white.jpg/g' {} \;

Unless I've misinterpreted the question?...

forrestt · 02-27-2008, 09:31 AM

Oh, you may need to escape the "/" and "." characters in the $find and $replace. I'm not sure, so test it.

Forrest

forrestt · 02-27-2008, 09:34 AM

Sure pwc101, that would work, but it is WAY to easy a solution

Forrest

pwc101 · 02-27-2008, 09:42 AM

Quote:

Originally Posted by forrestt

Sure pwc101, that would work, but it is WAY to easy a solution

Forrest

Where there's a will, there's a perl-way: http://www.linuxquestions.org/questi...78#post2385378

phantom_cyph · 02-27-2008, 03:02 PM

Quote:

Originally Posted by pwc101

Sounds like a job for sed, Batman!

Code:

sed -i 's/..\/white.jpg/34568\/jpg\/white.jpg/g' *.html

This would be ideal, but it doesn't seem to work. Essentially the base of this command is:

Code:

sed -i 's/foo/foo_bar/g' somefile.module

But, both of these don't actually change the files. I have tried the following:

Code:

sed -i 's/../white.jpg/34568/jpg/white.jpg/g' *.html

Code:

sed -i 's/..\/white.jpg/34568\/jpg\/white.jpg/g' *.html

Any ideas why it might not work? Is it because of the "/"s in the pathnames?

pwc101 · 02-27-2008, 03:24 PM

Quote:

Originally Posted by phantom_cyph

Any ideas why it might not work? Is it because of the "/"s in the pathnames?

Hmmm... odd. I tested this on some dummy files I made that contained a number of paths (including your example path to replace) as well as some other text, and after running the command I checked the files and the content had indeed been changed. So I'm a bit confused :|

Perhaps if you ran the command in a for loop instead, it might be the *.html that's throwing things:

Code:

for FILE in *.html; do 
   sed 's/..\/white.jpg/34568\/jpg\/white.jpg/g' $FILE
done

This is a dump of me running the command on a number of test files:

Code:

pwc101@archie:~/tmp/examples> sed --version
GNU sed version 4.1.5
Copyright (C) 2003 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE,
to the extent permitted by law.
pwc101@archie:~/tmp/examples> ll
total 16K
-rw-r--r-- 1 pwc101 users 101 2008-02-27 21:14 test
-rw-r--r-- 1 pwc101 users 171 2008-02-27 21:14 test2
-rw-r--r-- 1 pwc101 users 182 2008-02-27 21:15 test3
-rw-r--r-- 1 pwc101 users 171 2008-02-27 21:15 test3y
pwc101@archie:~/tmp/examples> for i in test*; do echo $i:; cat $i; echo; done
test:
this is a test file
./some/path/file.jpg
../white.jpg
../white/file.jpg
some more text
/another/path

test2:
this is a test file
./some/path/file.jpg
../white.jpg
../white/file.jpg
some more text
/another/path

adding in
more text
      this is a test
\checking@odd-characters%$

test3:
more tests
this is a test file
./some/path/file.jpg
../white.jpg
../white/file.jpg
some more text
/another/path

adding in
more text
      this is a test
\checking@odd-characters%$

test3y:
this is a test file
./some/path/file.jpg
../white.jpg
../white/file.jpg
some more text
/another/path

adding in
more text
      this is a test
\checking@odd-characters%$
pwc@archibald:~/tmp/examples> sed -i 's/..\/white.jpg/34568\/jpg\/white.jpg/g' test*
pwc101@archie:~/tmp/examples> for i in test*; do echo $i:; cat $i; done
test:
this is a test file
./some/path/file.jpg
34568/jpg/white.jpg
../white/file.jpg
some more text
/another/path

test2:
this is a test file
./some/path/file.jpg
34568/jpg/white.jpg
../white/file.jpg
some more text
/another/path

adding in
more text
      this is a test
\checking@odd-characters%$

test3:
more tests
this is a test file
./some/path/file.jpg
34568/jpg/white.jpg
../white/file.jpg
some more text
/another/path

adding in
more text
      this is a test
\checking@odd-characters%$

test3y:
this is a test file
./some/path/file.jpg
34568/jpg/white.jpg
../white/file.jpg
some more text
/another/path

adding in
more text
      this is a test
\checking@odd-characters%$
pwc101@archie:~/tmp/examples>

I've put the changed parts in bold to show it seems to have worked here.

phantom_cyph · 02-27-2008, 08:11 PM

I'm sorry, but I'm not really following. How would I set it up do it, lets say, one file at a time, just changing the filename?

For example: enter the command to change ../white.jpg to 34568/jpg/white.jpg in the file "security.html".

Then, I simply hit the up key, and change the filename to "contact.html".

Sorry for my ignorance, I just can't seem to get it.

osor · 02-27-2008, 09:31 PM

Also, what sed are you using (phantom_cyph)? The -i flag for in-place editing is a non-standard GNU option, and may not be present on Solaris or BSD sed. In that case, you will need to redirect to a temporary file manually and then replace the original file with the temporary file.

Btw, the in-place flag is cross-platform, standard perl:

Code:

perl -pi -e 's|../white.jpg|34568/jpg/white.jpg|g' foo.html

phantom_cyph · 02-27-2008, 11:08 PM

Quote:

Originally Posted by osor

Also, what sed are you using (phantom_cyph)? The -i flag for in-place editing is a non-standard GNU option, and may not be present on Solaris or BSD sed. In that case, you will need to redirect to a temporary file manually and then replace the original file with the temporary file.

Btw, the in-place flag is cross-platform, standard perl:

Code:

perl -pi -e 's|../white.jpg|34568/jpg/white.jpg|g' foo.html

Thank God! It worked! I just put an * in front of the .html, and it changed everything in the folder.

Thank-you!