Perl script to list matplotlib examples ordered by last commit

findoldexamples.pl

#!/usr/bin/perl

# I don't claim credit for because I haven't written Perl before and just Frankensteined
# it together from these three sites:
# http://stackoverflow.com/a/17727887
# http://www.perlmonks.org/?node_id=97574
# http://perl.about.com/od/filesystem/qt/perl_file_exist.htm

# Note that the script should be placed inside the examples folder of the matplotlib repo.

# Ask git for a list of access times and files
my @logbook = `git whatchanged --pretty=%ai . `;

my %seen;
my $timestamp;
my $filename;

my $date;
my $relpath;

foreach (@logbook) {
    next if /^$/; # skip empty lines
    if (/^:/) {
        next unless /.py$/;          # skip files without .py extension
        chomp ($filename = (split /\t/)[1]);
        $relpath = substr($filename, 9); # because we're in examples/, drop that part from the path when testing existence
        next unless (-e $relpath);   # skip files no longer existing
        $date = substr($timestamp, 0, 10);
        next if ($date eq '2016-04-26');
        next if $seen{$filename};    # skip files already displayed with a newer commit
        print "$timestamp $filename\n";
        $seen{$filename} = 1;
    } else {
        chomp ($timestamp = $_);
    }
}

Updated to disregard commits from the date '2016-04-26', because the commit on that date wasn't thoroughly MEP12ing the files it affected, just removing shebang lines. Since the goal is to identify files which have gone a long time without receiving any love, let's not count the commit with that date. If we skip those lines when they appear, the files with that date will instead display elsewhere with the date of their next to last update.

See the results here (for the present state of matplotlib): https://gist.github.com/discardthree/2d1316c000d45946bdb4df12d8672070