<?php

declare(strict_types=1);

class AccessLogParser
{
    private string $logDir;
    private string $pattern;
    private string $cacheFile;
    private int    $cacheTtl;
    private int    $days;

    private static ?array $memo = null;

    // Apache COMBINED : IP - - [timestamp] "METHOD /path HTTP/x" STATUS bytes "ref" "ua"
    private const RE = '/^(\S+) \S+ \S+ \[(\d{2}\/\w+\/\d{4}:\d{2}:\d{2}:\d{2} [+-]\d{4})\] "[A-Z-]+ ([^\s"?]+)[^"]*" (\d{3}) /';

    public function __construct(
        string $logDir    = '/var/log/apache2',
        string $pattern   = '*-access.log',
        string $cacheFile = '',
        int    $cacheTtl  = 600,
        int    $days      = 14
    ) {
        $this->logDir    = rtrim($logDir, '/');
        $this->pattern   = $pattern;
        $this->cacheFile = $cacheFile !== '' ? $cacheFile : dirname(__DIR__) . '/_cache/access_stats.json';
        $this->cacheTtl  = $cacheTtl;
        $this->days      = $days;
    }

    /**
     * @return array{pages:array<string,int>,books:array<string,int>,ips:array<string,int>}
     */
    public function stats(): array
    {
        if (self::$memo !== null) {
            return self::$memo;
        }
        if ($this->cacheValid()) {
            $d = json_decode((string) file_get_contents($this->cacheFile), true);
            if (is_array($d)) {
                return self::$memo = $d;
            }
        }

        $cutoff = strtotime("-{$this->days} days midnight") ?: (time() - $this->days * 86400);
        $pages  = [];
        $books  = [];
        $ips    = [];

        foreach ($this->logFiles() as $file) {
            $this->parseFile($file, $cutoff, $pages, $books, $ips);
        }

        arsort($pages);
        arsort($books);
        arsort($ips);

        $result = compact('pages', 'books', 'ips');
        @mkdir(dirname($this->cacheFile), 0755, true);
        @file_put_contents($this->cacheFile, json_encode($result, JSON_UNESCAPED_UNICODE | JSON_UNESCAPED_SLASHES));
        return self::$memo = $result;
    }

    public function isReadable(): bool
    {
        return count($this->logFiles()) > 0;
    }

    private function cacheValid(): bool
    {
        return file_exists($this->cacheFile)
            && (time() - filemtime($this->cacheFile)) < $this->cacheTtl;
    }

    /** @return list<array{path:string,type:string}> */
    private function logFiles(): array
    {
        $files  = [];
        $cutoff = time() - ($this->days + 1) * 86400;

        foreach (glob($this->logDir . '/' . $this->pattern) ?: [] as $base) {
            if (str_ends_with($base, '.gz') || preg_match('/\.\d+$/', $base)) {
                continue;
            }
            foreach (array_merge([$base], glob($base . '.*') ?: []) as $path) {
                if ($path !== $base && filemtime($path) < $cutoff) {
                    continue;
                }
                if (!is_readable($path)) {
                    continue;
                }
                if (str_ends_with($path, '.tar.gz')) {
                    $files[] = ['path' => $path, 'type' => 'tgz'];
                } elseif (str_ends_with($path, '.gz')) {
                    $files[] = ['path' => $path, 'type' => 'gz'];
                } else {
                    $files[] = ['path' => $path, 'type' => 'plain'];
                }
            }
        }

        return $files;
    }

    private static function parseTimestamp(string $raw): int
    {
        // "15/May/2026:00:41:01 +0200"
        if (!preg_match('/(\d{2})\/(\w{3})\/(\d{4}):(\d{2}:\d{2}:\d{2}) ([+-]\d{4})/', $raw, $m)) {
            return 0;
        }
        return (int) strtotime("{$m[1]} {$m[2]} {$m[3]} {$m[4]} {$m[5]}");
    }

    private function parseLine(string $line, int $cutoff, array &$pages, array &$books, array &$ips): void
    {
        if (!preg_match(self::RE, $line, $m)) {
            return;
        }
        [, $ip, $ts, $path, $status] = $m;

        if ($status !== '200') {
            return;
        }
        if (self::parseTimestamp($ts) < $cutoff) {
            return;
        }

        if (str_starts_with($path, '/post/') && strlen($path) > 6) {
            $pages[$path] = ($pages[$path] ?? 0) + 1;
            $ips[$ip]     = ($ips[$ip] ?? 0) + 1;
        } elseif (str_starts_with($path, '/book/') && strlen($path) > 6) {
            $books[$path] = ($books[$path] ?? 0) + 1;
            $ips[$ip]     = ($ips[$ip] ?? 0) + 1;
        }
    }

    private function parseFile(array $file, int $cutoff, array &$pages, array &$books, array &$ips): void
    {
        if ($file['type'] === 'tgz') {
            try {
                $phar = new PharData($file['path']);
                foreach ($phar as $entry) {
                    $content = @file_get_contents('phar://' . $file['path'] . '/' . $entry->getFilename());
                    if ($content === false) {
                        continue;
                    }
                    foreach (explode("\n", $content) as $line) {
                        $this->parseLine($line, $cutoff, $pages, $books, $ips);
                    }
                }
            } catch (\Exception $e) {
            }
        } elseif ($file['type'] === 'gz') {
            $h = @gzopen($file['path'], 'rb');
            if (!$h) {
                return;
            }
            while (!gzeof($h)) {
                $line = gzgets($h, 8192);
                if ($line !== false) {
                    $this->parseLine($line, $cutoff, $pages, $books, $ips);
                }
            }
            gzclose($h);
        } else {
            $h = @fopen($file['path'], 'rb');
            if (!$h) {
                return;
            }
            while (($line = fgets($h)) !== false) {
                $this->parseLine($line, $cutoff, $pages, $books, $ips);
            }
            fclose($h);
        }
    }
}