cedricAbonnel/folio
1
0
Fork 0
Code Issues 15 Pull Requests Actions Packages Projects Releases Wiki Activity

v1.6.32 : UA en entier + bouton « + bot » + filtrage bots des stats

Browse Source 
- Agents détectés : UA affiché sans troncature (drill-down et liste)
- Bouton « + bot » pour ajouter un agent aux patterns via AJAX (CSRF)
- Section Agents alimentée par all_uas (tous UAs publics, bots inclus)
- AccessLogParser : bots exclus des compteurs pages/livres/visiteurs
- Caches stats vidés après chaque modification des patterns

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Cédric Abonnel
2026-05-19 21:45:10 +02:00
parent d53b5da31a
commit b0f4814bb0
6 changed files with 199 additions and 84 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/AccessLogParser.php
+69 -33
View File
@@ -9,24 +9,31 @@ class AccessLogParser
private string $cacheFile;
private int $cacheTtl;
private int $days;
/** @var list<string> */
private array $botPatterns;
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}) \S+ "[^"]*" "([^"]*)"/u';
/**
* @param list<string> $botPatterns
*/
public function __construct(
string $logDir = '/var/log/apache2',
string $pattern = '*-access.log',
string $cacheFile = '',
int $cacheTtl = 600,
int $days = 14
string $logDir = '/var/log/apache2',
string $pattern = '*-access.log',
string $cacheFile = '',
int $cacheTtl = 600,
int $days = 14,
array $botPatterns = []
) {
$this->logDir = rtrim($logDir, '/');
$this->pattern = $pattern;
$this->cacheFile = $cacheFile !== '' ? $cacheFile : dirname(__DIR__) . '/_cache/access_stats.json';
$this->cacheTtl = $cacheTtl;
$this->days = $days;
$this->logDir = rtrim($logDir, '/');
$this->pattern = $pattern;
$this->cacheFile = $cacheFile !== '' ? $cacheFile : dirname(__DIR__) . '/_cache/access_stats.json';
$this->cacheTtl = $cacheTtl;
$this->days = $days;
$this->botPatterns = array_map('strtolower', $botPatterns);
}
/**
@@ -37,7 +44,8 @@ class AccessLogParser
* pages_by_day:array<string,list<int>>,
* ips_by_day:array<string,list<int>>,
* ip_top_paths:array<string,array<string,array{n:int,ts:int}>>,
* ip_agents:array<string,list<string>>
* ip_agents:array<string,list<string>>,
* all_uas:array<string,int>
* }
*/
public function stats(): array
@@ -52,24 +60,26 @@ class AccessLogParser
}
}
$cutoff = strtotime("-{$this->days} days midnight") ?: (time() - $this->days * 86400);
$pages = [];
$books = [];
$ips = []; // toutes requêtes publiques (tous chemins, tous statuts)
$dayPages = [];
$ipPaths = []; // chemins /post/ et /book/ avec statut 200 (pour les ts)
$ipPathTs = [];
$ipAllPaths = []; // tous chemins, tous statuts
$ipAllDays = []; // tous jours, tous statuts
$ipAgents = []; // tous user-agents par IP
$cutoff = strtotime("-{$this->days} days midnight") ?: (time() - $this->days * 86400);
$pages = [];
$books = [];
$ips = []; // requêtes publiques non-bot (tous chemins, tous statuts)
$dayPages = [];
$ipPaths = []; // chemins /post/ et /book/ avec statut 200 (pour les ts)
$ipPathTs = [];
$ipAllPaths = []; // tous chemins, tous statuts, non-bots
$ipAllDays = []; // tous jours, tous statuts, non-bots
$ipAgents = []; // user-agents non-bot par IP
$allUas = []; // tous UAs publics (bots inclus) pour "Agents détectés"
foreach ($this->logFiles() as $file) {
$this->parseFile($file, $cutoff, $pages, $books, $ips, $dayPages, $ipPaths, $ipPathTs, $ipAllPaths, $ipAllDays, $ipAgents);
$this->parseFile($file, $cutoff, $pages, $books, $ips, $dayPages, $ipPaths, $ipPathTs, $ipAllPaths, $ipAllDays, $ipAgents, $allUas);
}
arsort($pages);
arsort($books);
arsort($ips);
arsort($allUas);
$pagesByDay = [];
foreach ($dayPages as $path => $byOffset) {
@@ -82,7 +92,7 @@ class AccessLogParser
$pagesByDay[$path] = $arr;
}
// Top 200 IPs par volume total de requêtes
// Top 200 IPs non-bot par volume total de requêtes
$topIpKeys = array_keys(array_slice($ips, 0, 200, true));
$ipsByDay = [];
$ipTopPaths = [];
@@ -119,9 +129,11 @@ class AccessLogParser
'ips_by_day' => $ipsByDay,
'ip_top_paths' => $ipTopPaths,
'ip_agents' => $ipTopAgents,
'all_uas' => array_slice($allUas, 0, 300, true),
];
@mkdir(dirname($this->cacheFile), 0755, true);
@file_put_contents($this->cacheFile, json_encode($result, JSON_UNESCAPED_UNICODE | JSON_UNESCAPED_SLASHES));
return self::$memo = $result;
}
@@ -136,6 +148,21 @@ class AccessLogParser
&& (time() - filemtime($this->cacheFile)) < $this->cacheTtl;
}
private function matchesBot(string $ua): bool
{
if ($ua === '' || $this->botPatterns === []) {
return false;
}
$lo = strtolower($ua);
foreach ($this->botPatterns as $p) {
if ($p !== '' && str_contains($lo, $p)) {
return true;
}
}
return false;
}
/** @return list<array{path:string,type:string}> */
private function logFiles(): array
{
@@ -172,6 +199,7 @@ class AccessLogParser
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]}");
}
@@ -186,7 +214,8 @@ class AccessLogParser
array &$ipPathTs,
array &$ipAllPaths,
array &$ipAllDays,
array &$ipAgents
array &$ipAgents,
array &$allUas
): void {
if (!preg_match(self::RE, $line, $m)) {
return;
@@ -200,19 +229,25 @@ class AccessLogParser
$publicIp = filter_var($ip, FILTER_VALIDATE_IP, FILTER_FLAG_NO_PRIV_RANGE | FILTER_FLAG_NO_RES_RANGE) !== false;
$dayOffset = (int) floor(($tsVal - $cutoff) / 86400);
$isBot = $this->matchesBot($ua);
// Toutes les requêtes publiques : comptage global, chemins, jours, agents
if ($publicIp) {
// Tous les UAs publics pour la section "Agents détectés" (bots inclus)
if ($publicIp && $ua !== '') {
$allUas[$ua] = ($allUas[$ua] ?? 0) + 1;
}
// Requêtes publiques non-bot : comptage visiteurs, chemins, jours, agents
if ($publicIp && !$isBot) {
$ips[$ip] = ($ips[$ip] ?? 0) + 1;
$ipAllPaths[$ip][$path] = ($ipAllPaths[$ip][$path] ?? 0) + 1;
$ipAllDays[$ip][$dayOffset] = ($ipAllDays[$ip][$dayOffset] ?? 0) + 1;
if ($ua !== '') {
$ipAgents[$ip][$ua] = ($ipAgents[$ip][$ua] ?? 0) + 1;
$ipAgents[$ip][$ua] = ($ipAgents[$ip][$ua] ?? 0) + 1;
}
}
// Comptage spécifique aux pages de contenu (statut 200, /post/ ou /book/)
if ($status !== '200') {
// Comptage spécifique aux pages de contenu (statut 200, non-bot)
if ($status !== '200' || $isBot) {
return;
}
@@ -247,7 +282,8 @@ class AccessLogParser
array &$ipPathTs,
array &$ipAllPaths,
array &$ipAllDays,
array &$ipAgents
array &$ipAgents,
array &$allUas
): void {
if ($file['type'] === 'tgz') {
try {
@@ -258,7 +294,7 @@ class AccessLogParser
continue;
}
foreach (explode("\n", $content) as $line) {
$this->parseLine($line, $cutoff, $pages, $books, $ips, $dayPages, $ipPaths, $ipPathTs, $ipAllPaths, $ipAllDays, $ipAgents);
$this->parseLine($line, $cutoff, $pages, $books, $ips, $dayPages, $ipPaths, $ipPathTs, $ipAllPaths, $ipAllDays, $ipAgents, $allUas);
}
}
} catch (\Exception $e) {
@@ -271,7 +307,7 @@ class AccessLogParser
while (!gzeof($h)) {
$line = gzgets($h, 8192);
if ($line !== false) {
$this->parseLine($line, $cutoff, $pages, $books, $ips, $dayPages, $ipPaths, $ipPathTs, $ipAllPaths, $ipAllDays, $ipAgents);
$this->parseLine($line, $cutoff, $pages, $books, $ips, $dayPages, $ipPaths, $ipPathTs, $ipAllPaths, $ipAllDays, $ipAgents, $allUas);
}
}
gzclose($h);
@@ -281,7 +317,7 @@ class AccessLogParser
return;
}
while (($line = fgets($h)) !== false) {
$this->parseLine($line, $cutoff, $pages, $books, $ips, $dayPages, $ipPaths, $ipPathTs, $ipAllPaths, $ipAllDays, $ipAgents);
$this->parseLine($line, $cutoff, $pages, $books, $ips, $dayPages, $ipPaths, $ipPathTs, $ipAllPaths, $ipAllDays, $ipAgents, $allUas);
}
fclose($h);
}