comments.php

## comments.php
<?php

$file = $argv[1];
$header = null;

$spec = array(
    '_id' => function ($data) {
        return intval($data['comment_ID']);
    },
    'name' => 'comment_author',
    'email' => function ($data) {
        return (empty($data['comment_author_email']) ? '' : md5($data['comment_author_email']));
    },
    'comment' => function ($data) {
        return trim($data['comment_content']);
    },
    'date' => function ($data) {
        return strtotime($data['comment_date_gmt'] . ' UTC');
    },
);

if ($fh = fopen($file, 'r')) {
    while ($data = fgetcsv($fh)) {
        if ($header) {
            // convert data to YAML
            $comment = array_combine($header, array_slice($data, 0, count($header)));
            $yaml = substr(yaml_emit(map($comment, $spec), YAML_UTF8_ENCODING), 4, -4);

            // create directory
            $dir = $argv[3] . '/' . slug($comment['comment_post_ID'], $argv[2]);
            echo "Creating directory $dir\n";
            if (!file_exists($dir)) { mkdir($dir); }

            echo "Creating comment $dir/${comment['comment_ID']}.yml\n\n";
            file_put_contents("$dir/" . $comment['comment_ID'] . '.yml', $yaml);
        }
        else {
            $header = $data;
        }
    }
}

function map($data, $spec) {
    $return = array();

    foreach ($spec as $field => $value) {
        $return[$field] = (is_callable($value) ? $value($data) : $data[$value]);
    }

    return array_filter($return);
}

function slug($id, $site, $pattern = '~^/blog/(?P<year>\d{4})/(?P<month>\d{2})/(?P<day>\d{2})/(?P<slug>[^/]+)/$~') {
    echo "Getting URL for $site?p=$id\n";

    $ch = curl_init($site . "?p=$id");
    curl_exec($ch);
    $url = curl_getinfo($ch, CURLINFO_REDIRECT_URL);
    curl_close($ch);

    echo "Got URL $url\n";
    if (preg_match($pattern, str_replace($site, '', $url), $matches)) {
        return $matches['year'] . '-' . $matches['month'] . '-' . $matches['day'] . '-' . $matches['slug'];
    }
}