PiedWeb\SeoPocketCrawler\ExtractExternalLinks

Complexity

Total Complexity

12

Size/Duplication

Total Lines	62
Duplicated Lines	0 %

Test Coverage

Coverage

96.43%

Importance

Changes

0

6 Methods

Rating	Name	Duplication	Size	Complexity
A	__construct()	0	9	2
A	filter()	0	3	1
A	scanDir()	0	9	4
A	getExternals()	0	3	1
A	harvestExternalLinks()	0	8	3
A	scan()	0	6	1

<?php

namespace PiedWeb\SeoPocketCrawler;

class ExtractExternalLinks
{

    protected $id;
    protected $dir;
    protected $base;
    protected $filter;
    protected $filterType;
    protected $external = [];

    protected function __construct(string $id)
    {
        $this->id = $id;
        $this->dir = __DIR__.'/../data/'.$id.'/links';

        if (!file_exists($this->dir.'/../config.json')) {
            throw new \Exception('no crawl results found for id `'.$id.'`');
        }
        $this->base = json_decode(file_get_contents($this->dir.'/../config.json'), true)['base'];
    }

    protected function filter($filename)
    {
        return strpos($filename, 'From_') === 0;
    }

    public static function scan(string $id)
    {
        $self = new self($id);
        $self->scanDir();

        return $self->getExternals();
    }

    protected function scanDir()
    {
        if ($resource = opendir($this->dir)){
            while (($filename = readdir($resource)) !== false){
                if ($this->filter($filename)) {
                    $this->harvestExternalLinks(file_get_contents($this->dir.'/'.$filename));
                }
            }
            closedir($resource);
        }
    }


    function harvestExternalLinks(string $content)

It is generally recommended to explicitly declare the visibility for methods.

    {
        $lines = explode(chr(10), $content);

        foreach ($lines as $line)
        {
            if (strpos($line, $this->base) !== 0) {
                $this->external[$line] = ($this->external[$line] ?? 0) + 1;
            }
        }
    }

    function getExternals()

It is generally recommended to explicitly declare the visibility for methods.

    {
        return $this->external;
    }
}