ceekz

use List::Util;
sub cosine_similarity {
	my ($vector_1, $vector_2, $is_pearson) = @_;

	my %union;
	foreach (keys %{$vector_1}, keys %{$vector_2}) {
		$union{$_}++;
	}

	my $avg_1 = List::Util::sum(values %{$vector_1}) / scalar(keys %union);

ceekz

sub snippet {
	my ($text, $terms, $max, $tag_s, $tag_e) = @_;

	utf8::decode($text);

	if (length($text) < $max) {
		return $text;
	}

	if (! ref($terms)) {

ceekz

use Text::MeCab;

our $mecab = Text::MeCab->new({ node_format => "%M" });
sub get_terms {
	my $str = shift;

	my @terms;
	my @surfaces;
	for (my $node = $mecab->parse($str); $node; $node = $node->next) {
		my @feature = split /,/, $node->feature;

ceekz

use Data::Pageset;
	
$self->helper(
	pagination => sub {
		my ($self, $total_entries, $current_page) = @_;

		my $dp = Data::Pageset->new({
			'total_entries'    => $total_entries,
			'entries_per_page' => 20,
			'current_page'     => (! $current_page || $current_page < 1) ? 1 : $current_page,

ceekz

# -*- coding: utf-8 -*-

# See documentation in:
# https://doc.scrapy.org/en/latest/topics/commands.html#custom-project-commands

from scrapy.commands import ScrapyCommand
from scrapy.crawler import CrawlerProcess
from scrapy.utils.project import get_project_settings

class Command(ScrapyCommand):

ceekz

set ttt=%1
set tt=%ttt:~0,-4%
gswin32c -dSAFER -dBATCH -dNOPAUSE -sDEVICE=pdfwrite  -sFONTPATH=%windir%/fonts;xfonts;. -sPDFPassword= -dPDFSETTINGS=/prepress -dPassThroughJPEGImages=true -sOutputFile="%tt%_noPW.pdf"  %1

ceekz

/* mail.google.com */

table img[title*=".zip" i], table img[title*=".rar" i], table img[title*=".gz" i], table img[title*=".tar" i], table img[title*=".bz2" i] {
  width:0px!important; height:0px!important; padding-left:16px!important; padding-top:16px!important; background-position: top left !important;background-image: url(https://ssl.gstatic.com/docs/doclist/images/mediatype/icon_1_archive_x16.png) !important;
}
table img[title*=".txt" i], table img[title*=".rtf" i], table img[title*=".csv" i], table img[title*=".tsv" i] {
  width:0px!important; height:0px!important; padding-left:16px!important; padding-top:16px!important; background-position: top left !important; background-image:url(https://ssl.gstatic.com/docs/doclist/images/mediatype/icon_1_text_x16.png) !important;
}
table img[title*=".c" i], table img[title*=".pl" i], table img[title*=".sql" i] {
  width:0px!important; height:0px!important; padding-left:16px!important; padding-top:16px!important; background-position: top left !important; background-