Project:LayserCake: Difference between revisions

From London Hackspace Wiki
No edit summary
(a worked example - now needs actually doing in RL!)
Line 13: Line 13:


Please don't make me cry.
Please don't make me cry.
--[[User:AndyE|AndyE]] 13:00, 13 January 2011 (UTC)


<pre>
<pre>
Line 77: Line 79:
</pre>
</pre>


--[[User:AndyE|AndyE]] 13:00, 13 January 2011 (UTC)


'''How-to'''
 
'''How-to - a worked example'''
 
Let's say you have this charming picture of a cake to begin with:
 
[[File:220px-Strawberry_Cake.JPG]]
 
So you feed it to the script at the URL at the top of this page, and the script will give you back 8 files numbered 0 to 7.
 
So first, laze number 7 at full power (whatever you've decided that should be for your material).
 
[[File:220px-Strawberry_Cake_7.JPG]]
 
Then laze number 6 at half power:
 
[[File:220px-Strawberry_Cake_6.JPG]]
 
Then number 5 at quarter power:
 
[[File:220px-Strawberry_Cake_5.JPG]]
 
and so on and so forth, halving the power each time.
 
4
 
[[File:220px-Strawberry_Cake_4.JPG]]
 
3
 
[[File:220px-Strawberry_Cake_3.JPG]]
 
2
 
[[File:220px-Strawberry_Cake_2.JPG]]
 
1
 
[[File:220px-Strawberry_Cake_1.JPG]]
 
0
 
[[File:220px-Strawberry_Cake_0.JPG]]
 
Then report back here to let us know how it went!
 
An alternative would be to double the speed each time, rather than halving the power. Would be interesting to see tests for both methods on the same image.
 
'''How it went'''
 
?

Revision as of 14:19, 13 January 2011

http://hack.rs/cgi-bin/threshold_grayscale.pl

People

Source

People on IRC expressed a wish to use this as an example for security auditing, so here it is. Will transfer it to git when I 'git' time.

Audit away.

Please don't make me cry.

--AndyE 13:00, 13 January 2011 (UTC)


#!/usr/bin/perl -w
use strict;
use CGI;
use CGI::Carp qw(fatalsToBrowser warningsToBrowser);
#use List::Util qw(max min);

my $q = CGI->new();

print $q->header;
print $q->start_html;

print $q->p("this is a thing for doing threshholding");
print $q->p("<small>or possibly <i>thresholding</i>?</small>");

print $q->start_form( -enctype => "multipart/form-data" );

print $q->p("file");
print $q->filefield('uploaded_file');
print $q->submit();

print $q->end_form;

# do we have an upload?
my $filehandle = $q->upload('uploaded_file');
if (defined $filehandle) {

    # do shit
    print $q->p("I'm doing shit");

    # no, actually do shit
    my ($filename, $extension) = ($q->param('uploaded_file') =~ /^([0-9A-Za-z_-]+)\.([0-9A-Za-z_-]+)$/);
    die "no stupid filenames" unless ($filename and $extension);
    my $tempfile = $q->tmpFileName($q->param('uploaded_file'));

    foreach my $i (0 .. 7) {

        my $n = 2 ** $i;

        my $outfile = "$filename"."_$i.$extension";

        my @ar = ("gm",
                  "convert",
                    "-operator", "Gray", "And", $n ,
                    "-operator", "Gray", "Threshold",  $n - 1  ,
                  $tempfile,
                  "/var/www/threshold_output/$outfile" );

        print $q->p("$i : " . join " ", @ar);

        system(@ar) == 0 or die "system call failed: $? $!"; #safer than passing a string to system(),
                                                             # because doing it this way bypasses the shell

        print $q->img({src => "/threshold_output/$outfile"});
    }
}


print $q->end_html;


How-to - a worked example

Let's say you have this charming picture of a cake to begin with:

220px-Strawberry Cake.JPG

So you feed it to the script at the URL at the top of this page, and the script will give you back 8 files numbered 0 to 7.

So first, laze number 7 at full power (whatever you've decided that should be for your material).

220px-Strawberry Cake 7.JPG

Then laze number 6 at half power:

220px-Strawberry Cake 6.JPG

Then number 5 at quarter power:

220px-Strawberry Cake 5.JPG

and so on and so forth, halving the power each time.

4

220px-Strawberry Cake 4.JPG

3

220px-Strawberry Cake 3.JPG

2

220px-Strawberry Cake 2.JPG

1

220px-Strawberry Cake 1.JPG

0

220px-Strawberry Cake 0.JPG

Then report back here to let us know how it went!

An alternative would be to double the speed each time, rather than halving the power. Would be interesting to see tests for both methods on the same image.

How it went

?