Files
zyt/server/app/common/service/DifyRequiredAttachmentUpload.php
T
2026-10-23 10:40:43 +08:00

425 lines
23 KiB
PHP

<?php
declare(strict_types=1);
namespace app\common\service;
/**
* Turn an already-authorized mandatory attachment list into Dify local_file references.
*
* Trust options are SERVER configuration, never request parameters. Local origins are
* exact scheme/host/port aliases of this application's public/uploads directory;
* storage_origins are exact origins permitted for remote downloads. Request Host is
* deliberately not a source of trust. This class does not authorize patient access.
*/
final class DifyRequiredAttachmentUpload
{
private const DEFAULT_MAX_BYTES = 15 * 1024 * 1024;
private const MAX_RESPONSE_BYTES = 262144;
private const VIDEO_EXTENSIONS = ['mp4', 'mov', 'm4v', 'avi', 'mkv', 'webm', 'flv', 'wmv', 'mpeg', 'mpg', '3gp', 'ts'];
/**
* Options: local_origin/local_origins, storage_origins, public_root, temp_root,
* max_file_bytes, on_upload(array{index:int,total:int,status:string}). resolve_host
* is an optional server-owned DNS resolver (primarily for deterministic tests).
* No file ID cache survives this call, including between different Dify users.
*
* @return array{ok:bool,files?:array,error_code?:string,error?:string}
*/
public static function prepare(
string $chatEndpoint,
string $apiKey,
string $user,
array $files,
int $timeout,
?callable $shouldAbort = null,
array $options = []
): array {
$temporary = [];
$index = 0;
$total = 0;
$deadline = microtime(true) + max(0, $timeout);
try {
self::check($deadline, $shouldAbort);
$options = self::options($options);
$plan = self::plan($files);
$total = count($plan);
if ($plan === []) return ['ok' => true, 'files' => [], 'attachment_transport' => 'uploaded_files'];
if ($timeout < 1 || $timeout > 300 || $user === '' || strlen($user) > 1024
|| trim($apiKey) === '' || strpbrk($apiKey . $user, "\r\n\0") !== false) {
self::fail('CONFIG_INVALID');
}
if (!function_exists('curl_init') || !function_exists('curl_multi_init')) self::fail('CURL_UNAVAILABLE');
$endpoint = self::uploadEndpoint($chatEndpoint);
$publicRoot = realpath((string) ($options['public_root'] ?? dirname(__DIR__, 3) . '/public'));
if ($publicRoot === false || !is_dir($publicRoot)) self::fail('CONFIG_INVALID');
$localOrigins = self::origins($options['local_origins'] ?? []);
if (!empty($options['local_origin'])) $localOrigins += self::origins([(string) $options['local_origin']]);
$storageOrigins = self::origins($options['storage_origins'] ?? []);
$maxBytes = (int) ($options['max_file_bytes'] ?? self::DEFAULT_MAX_BYTES);
if ($maxBytes < 1 || $maxBytes > 100 * 1024 * 1024) self::fail('CONFIG_INVALID');
$tempRoot = self::tempRoot((string) ($options['temp_root'] ?? dirname(__DIR__, 3) . '/runtime/dify-required-attachments'), $publicRoot);
// Validate the whole manifest before the first upload. Never truncate it.
foreach ($plan as &$item) {
$origin = self::origin($item['parts']);
if (isset($localOrigins[$origin])) {
$item['path'] = self::localPath($publicRoot, $item['parts']['path'] ?? '/');
} elseif (!isset($storageOrigins[$origin])) {
self::fail('REQUIRED_FILES_SOURCE_UNTRUSTED');
}
}
unset($item);
$prepared = [];
$cache = [];
foreach ($plan as $position => $item) {
$index = $position + 1;
self::check($deadline, $shouldAbort);
self::notify($options, $index, $total, 'uploading');
$cacheKey = hash('sha256', $item['type'] . "\0" . $item['url']);
if (!isset($cache[$cacheKey])) {
$path = @tempnam($tempRoot, 'required-');
if ($path === false) self::fail('REQUIRED_FILES_PREPARE_FAILED');
$temporary[] = $path;
if (!self::inside($tempRoot, (string) realpath($path))) self::fail('REQUIRED_FILES_PREPARE_FAILED');
@chmod($path, 0600);
if (isset($item['path'])) {
self::copyLocal($item['path'], $path, $maxBytes, $deadline, $shouldAbort);
} else {
self::download($item['url'], $item['parts'], $path, $maxBytes, $deadline, $shouldAbort, $options);
}
self::check($deadline, $shouldAbort);
$mime = @(new \finfo(FILEINFO_MIME_TYPE))->file($path) ?: 'application/octet-stream';
if (str_starts_with(strtolower($mime), 'video/')) self::fail('REQUIRED_FILES_INVALID');
$extension = strtolower(pathinfo(rawurldecode($item['parts']['path'] ?? ''), PATHINFO_EXTENSION));
if (!preg_match('/^[a-z0-9]{1,10}$/D', $extension)) {
$extension = ['image/png' => 'png', 'image/jpeg' => 'jpg', 'application/pdf' => 'pdf',
'audio/wav' => 'wav', 'audio/mpeg' => 'mp3', 'text/plain' => 'txt'][$mime] ?? 'bin';
}
$cache[$cacheKey] = self::upload($endpoint, $apiKey, $user, $path, $mime,
'attachment-' . $index . '.' . $extension, $deadline, $shouldAbort);
}
$prepared[] = ['type' => $item['type'], 'transfer_method' => 'local_file', 'upload_file_id' => $cache[$cacheKey]];
self::notify($options, $index, $total, 'uploaded');
}
self::check($deadline, $shouldAbort);
return ['ok' => true, 'files' => $prepared, 'attachment_transport' => 'uploaded_files'];
} catch (\Throwable $error) {
// Never expose cURL response bodies, URLs, local paths, credentials or exception text.
$code = $error instanceof \RuntimeException ? $error->getMessage() : 'REQUIRED_FILES_PREPARE_FAILED';
$messages = [
'CONFIG_INVALID' => '附件上传服务配置无效',
'CURL_UNAVAILABLE' => '服务器尚未启用附件上传所需的 cURL 扩展',
'AI_ASSISTANT_CANCELLED' => '本次请求已取消',
'REQUIRED_FILES_TIMEOUT' => '必需附件上传超时,请重试',
'REQUIRED_FILES_INVALID' => '必需附件格式或来源地址无效',
'REQUIRED_FILES_SOURCE_UNTRUSTED' => '必需附件来源未在服务器可信列表中',
'REQUIRED_FILES_SOURCE_UNAVAILABLE' => '必需附件不可读取,请检查资料后重试',
'REQUIRED_FILES_TOO_LARGE' => '必需附件超过服务器允许的文件大小',
'REQUIRED_FILES_DOWNLOAD_FAILED' => '必需附件下载失败,请检查资料后重试',
'REQUIRED_FILES_UPLOAD_FAILED' => '必需附件上传失败,本次未提交分析',
'REQUIRED_FILES_PREPARE_FAILED' => '必需附件准备失败,本次未提交分析',
];
if (!isset($messages[$code])) $code = 'REQUIRED_FILES_PREPARE_FAILED';
if ($index > 0 && $total > 0) {
try { self::notify($options, $index, $total, 'failed'); } catch (\Throwable $ignored) {}
}
return ['ok' => false, 'error_code' => $code, 'error' => $messages[$code],
'failed_index' => $index, 'attachment_count' => $total];
} finally {
foreach ($temporary as $path) if (is_file($path)) @unlink($path);
}
}
private static function options(array $options): array
{
$configured = [];
if (function_exists('config')) {
try { $configured = \config('prescription_ai.required_attachment_upload', []); }
catch (\Throwable $ignored) { $configured = []; }
}
return array_replace(is_array($configured) ? $configured : [], $options);
}
private static function plan(array $files): array
{
$plan = [];
foreach ($files as $file) {
if (!is_array($file)) self::fail('REQUIRED_FILES_INVALID');
$type = strtolower((string) ($file['type'] ?? ''));
$mime = strtolower((string) ($file['mime_type'] ?? $file['mime'] ?? $file['content_type'] ?? ''));
$url = $file['url'] ?? '';
$path = is_string($url) ? (parse_url($url, PHP_URL_PATH) ?: '') : '';
$extension = strtolower(pathinfo(rawurldecode($path), PATHINFO_EXTENSION));
if ($type === 'video' || str_starts_with($mime, 'video/') || in_array($extension, self::VIDEO_EXTENSIONS, true)) continue;
if (!in_array($type, ['image', 'document', 'audio'], true)
|| ($file['transfer_method'] ?? 'remote_url') !== 'remote_url' || !is_string($url)) self::fail('REQUIRED_FILES_INVALID');
$plan[] = ['type' => $type, 'url' => $url, 'parts' => self::url($url)];
}
return $plan;
}
private static function url(string $url): array
{
if ($url === '' || strlen($url) > 8192 || preg_match('/[\x00-\x20\x7f\\\\]/', $url)) self::fail('REQUIRED_FILES_INVALID');
$parts = parse_url($url);
if (!is_array($parts) || !in_array(strtolower($parts['scheme'] ?? ''), ['http', 'https'], true)
|| empty($parts['host']) || isset($parts['user']) || isset($parts['pass']) || isset($parts['fragment'])) self::fail('REQUIRED_FILES_INVALID');
$host = strtolower($parts['host']);
$ip = trim($host, '[]');
if (filter_var($ip, FILTER_VALIDATE_IP) === false && !preg_match('/^(?=.{1,253}$)[a-z0-9](?:[a-z0-9.-]*[a-z0-9])?$/D', $host)) self::fail('REQUIRED_FILES_INVALID');
$parts['host'] = $host;
$parts['scheme'] = strtolower($parts['scheme']);
return $parts;
}
private static function origin(array $parts): string
{
$port = (int) ($parts['port'] ?? ($parts['scheme'] === 'https' ? 443 : 80));
if ($port < 1 || $port > 65535) self::fail('REQUIRED_FILES_INVALID');
return $parts['scheme'] . '://' . $parts['host'] . ':' . $port;
}
private static function origins($origins): array
{
if (!is_array($origins)) self::fail('CONFIG_INVALID');
$result = [];
foreach ($origins as $origin) {
if (!is_string($origin)) self::fail('CONFIG_INVALID');
$parts = self::url($origin);
if (($parts['path'] ?? '/') !== '/' || isset($parts['query'])) self::fail('CONFIG_INVALID');
$result[self::origin($parts)] = true;
}
return $result;
}
private static function uploadEndpoint(string $endpoint): string
{
$parts = self::url($endpoint);
if (isset($parts['query']) || preg_match('~/chat/completions/?$~', $parts['path'] ?? '')) self::fail('CONFIG_INVALID');
$path = rtrim($parts['path'] ?? '', '/');
$path = preg_replace('~/chat-messages$~', '', $path);
$authority = $parts['scheme'] . '://' . $parts['host'] . (isset($parts['port']) ? ':' . $parts['port'] : '');
return $authority . $path . '/files/upload';
}
private static function inside(string $root, string $path): bool
{
$root = rtrim(str_replace('\\', '/', $root), '/') . '/';
$path = str_replace('\\', '/', $path);
if (PHP_OS_FAMILY === 'Windows') { $root = strtolower($root); $path = strtolower($path); }
return str_starts_with($path, $root);
}
private static function localPath(string $publicRoot, string $urlPath): string
{
if (preg_match('/%(?:2f|5c)/i', $urlPath)) self::fail('REQUIRED_FILES_INVALID');
$decoded = rawurldecode($urlPath);
if (!str_starts_with($decoded, '/uploads/') || preg_match('/[\x00-\x1f\x7f\\\\%:?#]/', $decoded)) self::fail('REQUIRED_FILES_INVALID');
foreach (explode('/', substr($decoded, 1)) as $segment) {
if ($segment === '' || $segment === '.' || $segment === '..' || trim($segment) !== $segment
|| str_ends_with($segment, '.')) self::fail('REQUIRED_FILES_INVALID');
}
$uploads = realpath($publicRoot . '/uploads');
$real = realpath($publicRoot . $decoded);
if ($uploads === false || !self::inside($publicRoot, $uploads) || $real === false
|| !self::inside($uploads, $real) || !is_file($real) || !is_readable($real)) self::fail('REQUIRED_FILES_SOURCE_UNAVAILABLE');
return $real;
}
private static function tempRoot(string $root, string $publicRoot): string
{
if (!is_dir($root) && !@mkdir($root, 0700, true) && !is_dir($root)) self::fail('REQUIRED_FILES_PREPARE_FAILED');
$real = realpath($root);
if ($real === false || $real === $publicRoot || self::inside($publicRoot, $real) || !is_writable($real)) self::fail('CONFIG_INVALID');
@chmod($real, 0700);
return $real;
}
private static function copyLocal(string $source, string $destination, int $maxBytes, float $deadline, ?callable $shouldAbort): void
{
$input = @fopen($source, 'rb');
$output = @fopen($destination, 'wb');
try {
if ($input === false || $output === false) self::fail('REQUIRED_FILES_SOURCE_UNAVAILABLE');
$bytes = 0;
while (!feof($input)) {
self::check($deadline, $shouldAbort);
$chunk = fread($input, 65536);
if ($chunk === false) self::fail('REQUIRED_FILES_SOURCE_UNAVAILABLE');
$bytes += strlen($chunk);
if ($bytes > $maxBytes) self::fail('REQUIRED_FILES_TOO_LARGE');
if (fwrite($output, $chunk) !== strlen($chunk)) self::fail('REQUIRED_FILES_PREPARE_FAILED');
}
if ($bytes === 0) self::fail('REQUIRED_FILES_SOURCE_UNAVAILABLE');
} finally {
if (is_resource($input)) fclose($input);
if (is_resource($output)) fclose($output);
}
}
private static function download(string $url, array $parts, string $path, int $maxBytes, float $deadline, ?callable $shouldAbort, array $options): void
{
self::check($deadline, $shouldAbort);
$host = trim($parts['host'], '[]');
if (filter_var($host, FILTER_VALIDATE_IP) !== false) {
$addresses = [$host];
} elseif (is_callable($options['resolve_host'] ?? null)) {
$addresses = ($options['resolve_host'])($host);
} else {
// Validate every returned address, then pin the actual transfer to one.
// Native PHP DNS is synchronous; its elapsed time still consumes this deadline.
$records = @dns_get_record($host, DNS_A | DNS_AAAA);
$addresses = [];
foreach (is_array($records) ? $records : [] as $record) {
if (isset($record['ip'])) $addresses[] = $record['ip'];
if (isset($record['ipv6'])) $addresses[] = $record['ipv6'];
}
}
self::check($deadline, $shouldAbort);
if (!is_array($addresses) || $addresses === []) self::fail('REQUIRED_FILES_DOWNLOAD_FAILED');
foreach ($addresses as $address) if (!is_string($address) || !self::publicIp($address)) self::fail('REQUIRED_FILES_SOURCE_UNTRUSTED');
$ip = $addresses[0];
$port = (int) ($parts['port'] ?? ($parts['scheme'] === 'https' ? 443 : 80));
$output = @fopen($path, 'wb');
if ($output === false) self::fail('REQUIRED_FILES_PREPARE_FAILED');
$ch = curl_init();
if ($ch === false) { fclose($output); self::fail('CURL_UNAVAILABLE'); }
$bytes = 0;
$writeError = '';
try {
$curlOptions = self::curlOptions($url, $deadline, $shouldAbort);
$curlOptions[CURLOPT_RESOLVE] = [$host . ':' . $port . ':' . (str_contains($ip, ':') ? '[' . $ip . ']' : $ip)];
$curlOptions[CURLOPT_HEADERFUNCTION] = static function ($handle, string $line) use ($maxBytes, &$writeError): int {
if (preg_match('/^Content-Length:\s*(\d+)/i', $line, $match) && (float) $match[1] > $maxBytes) {
$writeError = 'REQUIRED_FILES_TOO_LARGE'; return 0;
}
return strlen($line);
};
$curlOptions[CURLOPT_WRITEFUNCTION] = static function ($handle, string $chunk) use ($output, $maxBytes, &$bytes, &$writeError): int {
$length = strlen($chunk);
$bytes += $length;
if ($bytes > $maxBytes) { $writeError = 'REQUIRED_FILES_TOO_LARGE'; return 0; }
$written = fwrite($output, $chunk);
if ($written !== $length) { $writeError = 'REQUIRED_FILES_PREPARE_FAILED'; return 0; }
return $written;
};
if (!curl_setopt_array($ch, $curlOptions)) self::fail('REQUIRED_FILES_DOWNLOAD_FAILED');
$result = self::execute($ch, $deadline, $shouldAbort);
if ($writeError !== '') self::fail($writeError);
if ($result['errno'] === CURLE_OPERATION_TIMEDOUT) self::fail('REQUIRED_FILES_TIMEOUT');
if ($result['errno'] !== CURLE_OK || $result['status'] !== 200 || $bytes === 0) self::fail('REQUIRED_FILES_DOWNLOAD_FAILED');
} finally { curl_close($ch); fclose($output); }
}
private static function publicIp(string $ip): bool
{
if (filter_var($ip, FILTER_VALIDATE_IP, FILTER_FLAG_NO_PRIV_RANGE | FILTER_FLAG_NO_RES_RANGE) === false) return false;
if (str_contains($ip, ':')) {
// Global unicast only; exclude transition/documentation/special-use ranges.
return self::cidr($ip, '2000::', 3) && !self::cidr($ip, '2001::', 23)
&& !self::cidr($ip, '2002::', 16) && !self::cidr($ip, '3fff::', 20);
}
foreach ([['0.0.0.0', 8], ['10.0.0.0', 8], ['100.64.0.0', 10], ['127.0.0.0', 8], ['169.254.0.0', 16],
['172.16.0.0', 12], ['192.0.0.0', 24], ['192.0.2.0', 24], ['192.88.99.0', 24], ['192.168.0.0', 16],
['198.18.0.0', 15], ['198.51.100.0', 24], ['203.0.113.0', 24], ['224.0.0.0', 4], ['240.0.0.0', 4]] as [$network, $bits]) {
if (self::cidr($ip, $network, $bits)) return false;
}
return true;
}
private static function cidr(string $ip, string $network, int $bits): bool
{
$packed = @inet_pton($ip); $base = @inet_pton($network);
if ($packed === false || $base === false || strlen($packed) !== strlen($base)) return false;
$bytes = intdiv($bits, 8); $remaining = $bits % 8;
return substr($packed, 0, $bytes) === substr($base, 0, $bytes)
&& ($remaining === 0 || (ord($packed[$bytes]) & (255 << (8 - $remaining))) === (ord($base[$bytes]) & (255 << (8 - $remaining))));
}
private static function upload(string $endpoint, string $apiKey, string $user, string $path, string $mime, string $name, float $deadline, ?callable $shouldAbort): string
{
self::check($deadline, $shouldAbort);
$ch = curl_init();
if ($ch === false) self::fail('CURL_UNAVAILABLE');
$body = '';
$tooLarge = false;
try {
$options = self::curlOptions($endpoint, $deadline, $shouldAbort);
$options[CURLOPT_POST] = true;
$options[CURLOPT_HTTPHEADER] = ['Authorization: Bearer ' . trim($apiKey), 'Accept: application/json', 'Expect:'];
$options[CURLOPT_POSTFIELDS] = ['user' => $user, 'file' => new \CURLFile($path, $mime, $name)];
$options[CURLOPT_WRITEFUNCTION] = static function ($handle, string $chunk) use (&$body, &$tooLarge): int {
if (strlen($body) + strlen($chunk) > self::MAX_RESPONSE_BYTES) { $tooLarge = true; return 0; }
$body .= $chunk;
return strlen($chunk);
};
if (!curl_setopt_array($ch, $options)) self::fail('REQUIRED_FILES_UPLOAD_FAILED');
$result = self::execute($ch, $deadline, $shouldAbort);
if ($result['errno'] === CURLE_OPERATION_TIMEDOUT) self::fail('REQUIRED_FILES_TIMEOUT');
if ($result['errno'] !== CURLE_OK || $tooLarge || !in_array($result['status'], [200, 201], true)) self::fail('REQUIRED_FILES_UPLOAD_FAILED');
$decoded = json_decode($body, true);
$id = is_array($decoded) ? ($decoded['id'] ?? '') : '';
if (!is_string($id) || !preg_match('/^[a-zA-Z0-9_-]{1,128}$/D', $id)) self::fail('REQUIRED_FILES_UPLOAD_FAILED');
return $id;
} finally { curl_close($ch); }
}
private static function curlOptions(string $url, float $deadline, ?callable $shouldAbort): array
{
$remaining = max(1, (int) ceil(($deadline - microtime(true)) * 1000));
$options = [CURLOPT_URL => $url, CURLOPT_RETURNTRANSFER => false,
CURLOPT_FOLLOWLOCATION => false, CURLOPT_MAXREDIRS => 0, CURLOPT_PROXY => '',
CURLOPT_PROTOCOLS => CURLPROTO_HTTP | CURLPROTO_HTTPS,
CURLOPT_SSL_VERIFYPEER => true, CURLOPT_SSL_VERIFYHOST => 2,
CURLOPT_TIMEOUT_MS => $remaining, CURLOPT_CONNECTTIMEOUT_MS => min(10000, $remaining),
CURLOPT_NOSIGNAL => true, CURLOPT_NOPROGRESS => false];
foreach (['CURLOPT_XFERINFOFUNCTION', 'CURLOPT_PROGRESSFUNCTION'] as $name) {
if (!defined($name)) continue;
$options[constant($name)] = static function (...$unused) use ($deadline, $shouldAbort): int {
try { self::check($deadline, $shouldAbort); return 0; }
catch (\Throwable $ignored) { return 1; }
};
break;
}
return $options;
}
/** curl_multi polling observes cancellation even while no headers/body arrive. */
private static function execute($ch, float $deadline, ?callable $shouldAbort): array
{
$multi = curl_multi_init();
$added = false;
try {
if (curl_multi_add_handle($multi, $ch) !== CURLM_OK) self::fail('REQUIRED_FILES_PREPARE_FAILED');
$added = true;
$errno = CURLE_OK;
do {
self::check($deadline, $shouldAbort);
do { $status = curl_multi_exec($multi, $running); } while ($status === CURLM_CALL_MULTI_PERFORM);
if ($status !== CURLM_OK) self::fail('REQUIRED_FILES_PREPARE_FAILED');
while (($info = curl_multi_info_read($multi)) !== false) $errno = (int) $info['result'];
self::check($deadline, $shouldAbort);
if ($running && curl_multi_select($multi, 0.05) === -1) usleep(1000);
} while ($running);
return ['errno' => $errno, 'status' => (int) curl_getinfo($ch, CURLINFO_HTTP_CODE)];
} finally {
if ($added) curl_multi_remove_handle($multi, $ch);
curl_multi_close($multi);
}
}
private static function check(float $deadline, ?callable $shouldAbort): void
{
if ($shouldAbort !== null && $shouldAbort()) self::fail('AI_ASSISTANT_CANCELLED');
if (microtime(true) >= $deadline) self::fail('REQUIRED_FILES_TIMEOUT');
}
private static function notify(array $options, int $index, int $total, string $status): void
{
if (is_callable($options['on_upload'] ?? null)) ($options['on_upload'])(['index' => $index, 'total' => $total, 'status' => $status]);
}
private static function fail(string $code): void { throw new \RuntimeException($code); }
}