laysense/proofdb
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

暂存

Browse Source
This commit is contained in:
Ziki Shay
2026-05-08 00:05:51 +08:00
parent 549b706fcc
commit ed70a140a2
40 changed files with 5590 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files
app/service/AdminConsole/AdminDocService.php
+76
View File
@@ -0,0 +1,76 @@
<?php
namespace app\service\AdminConsole;
use RuntimeException;
class AdminDocService
{
public function __construct(private readonly ?MarkdownRenderer $renderer = null)
{
}
public function list(): array
{
$items = [];
foreach (glob(base_path('apidoc/*.md')) ?: [] as $path) {
$name = basename($path);
$content = (string) file_get_contents($path);
$items[] = [
'name' => $name,
'title' => $this->title($content, $name),
];
}
usort($items, fn (array $a, array $b): int => strcmp($a['name'], $b['name']));
return $items;
}
public function read(string $name): array
{
$safeName = basename($name);
$path = base_path('apidoc/' . $safeName);
if (!is_file($path) || pathinfo($path, PATHINFO_EXTENSION) !== 'md') {
throw new RuntimeException('API doc not found.');
}
$content = (string) file_get_contents($path);
return [
'name' => $safeName,
'title' => $this->title($content, $safeName),
'content' => $content,
'html' => $this->renderer()->render($content),
];
}
public function readScriptDoc(string $name): array
{
$safeName = basename($name);
$path = base_path('scriptdoc/' . $safeName);
if (!is_file($path) || pathinfo($path, PATHINFO_EXTENSION) !== 'md') {
throw new RuntimeException('Script doc not found.');
}
$content = (string) file_get_contents($path);
return [
'name' => $safeName,
'title' => $this->title($content, $safeName),
'content' => $content,
'html' => $this->renderer()->render($content),
];
}
private function title(string $content, string $fallback): string
{
if (preg_match('/^#\s+(.+)$/m', $content, $matches)) {
return trim($matches[1]);
}
return $fallback;
}
private function renderer(): MarkdownRenderer
{
return $this->renderer ?? new MarkdownRenderer();
}
}
app/service/AdminConsole/ArchiveAdminService.php
+205
View File
@@ -0,0 +1,205 @@
<?php
namespace app\service\AdminConsole;
use InvalidArgumentException;
use support\Db;
class ArchiveAdminService
{
public function list(string $query = '', int $page = 1, int $pageSize = 20): array
{
$page = max(1, $page);
$pageSize = min(100, max(1, $pageSize));
$builder = Db::table('archives');
$query = trim($query);
if ($query !== '') {
$like = '%' . $query . '%';
$builder->where(function ($subQuery) use ($like): void {
$subQuery
->orWhere('archive_uid', 'like', $like)
->orWhere('title', 'like', $like)
->orWhere('summary', 'like', $like)
->orWhere('author', 'like', $like)
->orWhere('source', 'like', $like)
->orWhere('series', 'like', $like);
});
}
$total = (clone $builder)->count();
$rows = $builder
->orderByDesc('updated_time')
->offset(($page - 1) * $pageSize)
->limit($pageSize)
->get([
'archive_uid',
'title',
'summary',
'year',
'author',
'source',
'series',
'tags',
'created_time',
'updated_time',
Db::raw('jsonb_array_length(chunks) as chunk_count'),
])
->all();
return [
'items' => array_map(fn (object $row): array => $this->listItem($row), $rows),
'total' => (int) $total,
'page' => $page,
'page_size' => $pageSize,
];
}
public function detail(string $archiveUid): ?array
{
$row = Db::table('archives')->where('archive_uid', $archiveUid)->first();
if (!$row) {
return null;
}
return $this->detailItem($row);
}
public function update(string $archiveUid, array $payload): ?array
{
if (!$this->detail($archiveUid)) {
return null;
}
$updates = [];
foreach (['title', 'summary', 'author', 'source', 'series', 'content', 'raw'] as $field) {
if (array_key_exists($field, $payload)) {
$updates[$field] = $this->nullableText($payload[$field]);
}
}
if (array_key_exists('year', $payload)) {
$year = trim((string) ($payload['year'] ?? ''));
if ($year === '') {
$updates['year'] = null;
} elseif (!preg_match('/^\d{1,4}$/', $year)) {
throw new InvalidArgumentException('year must be empty or a 1-4 digit number.');
} else {
$updates['year'] = (int) $year;
}
}
if (array_key_exists('tags', $payload)) {
$updates['tags'] = json_encode($this->normalizeTags($payload['tags']), JSON_UNESCAPED_UNICODE | JSON_UNESCAPED_SLASHES);
}
if (array_key_exists('metadata', $payload)) {
$updates['metadata'] = json_encode($this->normalizeMetadata($payload['metadata']), JSON_UNESCAPED_UNICODE | JSON_UNESCAPED_SLASHES);
}
if ($updates !== []) {
Db::table('archives')->where('archive_uid', $archiveUid)->update($updates);
}
return $this->detail($archiveUid);
}
public function delete(string $archiveUid): bool
{
return (int) Db::table('archives')->where('archive_uid', $archiveUid)->delete() > 0;
}
private function listItem(object $row): array
{
$chunks = $this->decodeJson($row->chunks ?? null, []);
return [
'archive_uid' => (string) $row->archive_uid,
'title' => $row->title,
'summary' => $row->summary,
'year' => $row->year === null ? null : (int) $row->year,
'author' => $row->author,
'source' => $row->source,
'series' => $row->series,
'tags' => $this->decodeJson($row->tags ?? null, []),
'chunk_count' => property_exists($row, 'chunk_count')
? ($row->chunk_count === null ? 0 : (int) $row->chunk_count)
: count(is_array($chunks) ? $chunks : []),
'created_time' => $row->created_time,
'updated_time' => $row->updated_time,
];
}
private function detailItem(object $row): array
{
$data = $this->listItem($row);
$data['metadata'] = $this->decodeJson($row->metadata ?? null, []);
$data['content'] = $row->content;
$data['raw'] = $row->raw;
$data['chunks'] = $this->decodeJson($row->chunks ?? null, []);
return $data;
}
private function normalizeTags(mixed $value): array
{
if (is_array($value)) {
$items = $value;
} else {
$text = trim((string) $value);
if ($text === '') {
return [];
}
$items = preg_split('/[\r\n,]+/', $text) ?: [];
}
$tags = [];
foreach ($items as $item) {
$tag = trim((string) $item);
if ($tag !== '') {
$tags[] = $tag;
}
}
return array_values(array_unique($tags));
}
private function normalizeMetadata(mixed $value): array
{
if (is_array($value)) {
return $value;
}
$text = trim((string) $value);
if ($text === '') {
return [];
}
$decoded = json_decode($text, true);
if (!is_array($decoded)) {
throw new InvalidArgumentException('metadata must be a JSON object or array.');
}
return $decoded;
}
private function nullableText(mixed $value): ?string
{
$text = trim((string) $value);
return $text === '' ? null : $text;
}
private function decodeJson(mixed $value, mixed $fallback): mixed
{
if ($value === null) {
return $fallback;
}
if (is_array($value)) {
return $value;
}
$decoded = json_decode((string) $value, true);
return $decoded === null && json_last_error() !== JSON_ERROR_NONE ? $fallback : $decoded;
}
}
app/service/AdminConsole/MaintenanceScriptService.php
+148
View File
@@ -0,0 +1,148 @@
<?php
namespace app\service\AdminConsole;
use RuntimeException;
class MaintenanceScriptService
{
private const ARG_PATTERN = '/^--[A-Za-z0-9_]+(?:=[A-Za-z0-9._:@\/-]+)?$/';
public function list(): array
{
$docs = new AdminDocService();
$items = [];
foreach ($this->definitions() as $definition) {
$item = $definition;
try {
$doc = $docs->readScriptDoc($definition['doc_name']);
$item['doc_title'] = $doc['title'];
$item['doc_html'] = $doc['html'];
$item['doc_content'] = $doc['content'];
} catch (RuntimeException) {
$item['doc_title'] = null;
$item['doc_html'] = null;
$item['doc_content'] = null;
}
$items[] = $item;
}
return $items;
}
public function describe(string $name): array
{
$definitions = $this->definitions();
if (!isset($definitions[$name])) {
throw new RuntimeException('Script is not allowed.');
}
foreach ($this->list() as $item) {
if ($item['name'] === $name) {
return $item;
}
}
throw new RuntimeException('Script metadata not found.');
}
public function run(string $name, array $args = []): array
{
$definitions = $this->definitions();
if (!isset($definitions[$name])) {
throw new RuntimeException('Script is not allowed.');
}
$script = $definitions[$name];
$scriptPath = base_path('scripts/' . $script['file']);
if (!is_file($scriptPath)) {
throw new RuntimeException('Script file not found.');
}
$safeArgs = [];
foreach ($args as $arg) {
$arg = trim((string) $arg);
if ($arg === '') {
continue;
}
if (!preg_match(self::ARG_PATTERN, $arg)) {
throw new RuntimeException('Only --key=value style arguments are allowed.');
}
$safeArgs[] = $arg;
}
$command = array_merge([PHP_BINARY, $scriptPath], $safeArgs);
$descriptors = [
0 => ['pipe', 'r'],
1 => ['pipe', 'w'],
2 => ['pipe', 'w'],
];
$process = proc_open($command, $descriptors, $pipes, base_path());
if (!is_resource($process)) {
throw new RuntimeException('Failed to start script process.');
}
fclose($pipes[0]);
$stdout = (string) stream_get_contents($pipes[1]);
fclose($pipes[1]);
$stderr = (string) stream_get_contents($pipes[2]);
fclose($pipes[2]);
$exitCode = proc_close($process);
return [
'script_name' => $name,
'command' => array_merge(['php', 'scripts/' . $script['file']], $safeArgs),
'exit_code' => $exitCode,
'stdout' => $stdout,
'stderr' => $stderr,
'ok' => $exitCode === 0,
];
}
private function definitions(): array
{
return [
'setup_database' => [
'name' => 'setup_database',
'file' => 'setup_database.php',
'label' => '初始化数据库',
'description' => '创建或补齐 archives、chunks 相关表结构与索引。',
'doc_name' => 'setup_database.md',
'args_hint' => '无参数',
],
'setup_opensearch' => [
'name' => 'setup_opensearch',
'file' => 'setup_opensearch.php',
'label' => '初始化 OpenSearch',
'description' => '创建或补齐 proofdb_chunks 索引与 mapping。',
'doc_name' => 'setup_opensearch.md',
'args_hint' => '无参数',
],
'reindex_opensearch' => [
'name' => 'reindex_opensearch',
'file' => 'reindex_opensearch.php',
'label' => '重建 OpenSearch 索引',
'description' => '把 PostgreSQL 中已向量化的数据重新写入 OpenSearch。',
'doc_name' => 'reindex_opensearch.md',
'args_hint' => '--archive_uid=01...',
],
'backfill_archive_content' => [
'name' => 'backfill_archive_content',
'file' => 'backfill_archive_content.php',
'label' => '回填 archive content',
'description' => '从 raw 或 chunks 回填 archives.content。',
'doc_name' => 'backfill_archive_content.md',
'args_hint' => '--archive_uid=01...',
],
'setup_admin_users' => [
'name' => 'setup_admin_users',
'file' => 'setup_admin_users.php',
'label' => '初始化管理员用户',
'description' => '创建 admin_users 表并写入或更新管理员账号。',
'doc_name' => 'setup_admin_users.md',
'args_hint' => '--username=admin --password=secret',
],
];
}
}
app/service/AdminConsole/MarkdownRenderer.php
+185
View File
@@ -0,0 +1,185 @@
<?php
namespace app\service\AdminConsole;
class MarkdownRenderer
{
public function render(string $markdown): string
{
$lines = preg_split('/\r\n|\n|\r/', $markdown) ?: [];
$html = [];
$paragraph = [];
$listType = null;
$table = null;
$inCodeBlock = false;
$codeLines = [];
$flushParagraph = function () use (&$paragraph, &$html): void {
if ($paragraph === []) {
return;
}
$text = implode(' ', array_map('trim', $paragraph));
$html[] = '<p>' . $this->renderInline($text) . '</p>';
$paragraph = [];
};
$flushList = function () use (&$listType, &$html): void {
if ($listType !== null) {
$html[] = '</' . $listType . '>';
$listType = null;
}
};
$flushTable = function () use (&$table, &$html): void {
if ($table === null) {
return;
}
$html[] = '<table class="markdown-table"><thead><tr>' .
implode('', array_map(fn (string $cell): string => '<th>' . $this->renderInline($cell) . '</th>', $table['headers'])) .
'</tr></thead><tbody>';
foreach ($table['rows'] as $row) {
$html[] = '<tr>' .
implode('', array_map(fn (string $cell): string => '<td>' . $this->renderInline($cell) . '</td>', $row)) .
'</tr>';
}
$html[] = '</tbody></table>';
$table = null;
};
foreach ($lines as $line) {
if (preg_match('/^```/', $line)) {
$flushParagraph();
$flushList();
$flushTable();
if ($inCodeBlock) {
$html[] = '<pre class="markdown-pre"><code>' . htmlspecialchars(implode("\n", $codeLines), ENT_QUOTES, 'UTF-8') . '</code></pre>';
$codeLines = [];
$inCodeBlock = false;
} else {
$inCodeBlock = true;
}
continue;
}
if ($inCodeBlock) {
$codeLines[] = $line;
continue;
}
$trimmed = trim($line);
if ($trimmed === '') {
$flushParagraph();
$flushList();
$flushTable();
continue;
}
if (preg_match('/^(#{1,6})\s+(.+)$/', $trimmed, $matches)) {
$flushParagraph();
$flushList();
$flushTable();
$level = strlen($matches[1]);
$html[] = sprintf('<h%d>%s</h%d>', $level, $this->renderInline($matches[2]), $level);
continue;
}
if (preg_match('/^>\s?(.+)$/', $trimmed, $matches)) {
$flushParagraph();
$flushList();
$flushTable();
$html[] = '<blockquote>' . $this->renderInline($matches[1]) . '</blockquote>';
continue;
}
if (preg_match('/^---+$/', $trimmed)) {
$flushParagraph();
$flushList();
$flushTable();
$html[] = '<hr>';
continue;
}
if ($this->isTableDelimiter($trimmed) && $table !== null) {
continue;
}
if (str_contains($trimmed, '|')) {
$cells = $this->tableCells($trimmed);
if (count($cells) >= 2) {
$flushParagraph();
$flushList();
if ($table === null) {
$table = ['headers' => $cells, 'rows' => []];
} else {
$table['rows'][] = $cells;
}
continue;
}
}
if (preg_match('/^[-*]\s+(.+)$/', $trimmed, $matches)) {
$flushParagraph();
$flushTable();
if ($listType !== 'ul') {
$flushList();
$listType = 'ul';
$html[] = '<ul>';
}
$html[] = '<li>' . $this->renderInline($matches[1]) . '</li>';
continue;
}
if (preg_match('/^\d+\.\s+(.+)$/', $trimmed, $matches)) {
$flushParagraph();
$flushTable();
if ($listType !== 'ol') {
$flushList();
$listType = 'ol';
$html[] = '<ol>';
}
$html[] = '<li>' . $this->renderInline($matches[1]) . '</li>';
continue;
}
$flushList();
$flushTable();
$paragraph[] = $trimmed;
}
if ($inCodeBlock) {
$html[] = '<pre class="markdown-pre"><code>' . htmlspecialchars(implode("\n", $codeLines), ENT_QUOTES, 'UTF-8') . '</code></pre>';
}
$flushParagraph();
$flushList();
$flushTable();
return implode("\n", $html);
}
private function renderInline(string $text): string
{
$text = htmlspecialchars($text, ENT_QUOTES, 'UTF-8');
$text = preg_replace('/`([^`]+)`/', '<code>$1</code>', $text) ?? $text;
$text = preg_replace('/\*\*([^*]+)\*\*/', '<strong>$1</strong>', $text) ?? $text;
$text = preg_replace('/\*([^*]+)\*/', '<em>$1</em>', $text) ?? $text;
$text = preg_replace('/\[(.+?)\]\((.+?)\)/', '<a href="$2" target="_blank" rel="noreferrer">$1</a>', $text) ?? $text;
return $text;
}
private function isTableDelimiter(string $line): bool
{
return (bool) preg_match('/^\|?[\s:-]+\|[\s|:-]*$/', $line);
}
private function tableCells(string $line): array
{
$line = trim($line);
$line = trim($line, '|');
return array_map(static fn (string $cell): string => trim($cell), explode('|', $line));
}
}
app/service/AdminConsole/OpenSearchAdminService.php
+153
View File
@@ -0,0 +1,153 @@
<?php
namespace app\service\AdminConsole;
use app\service\Search\OpenSearchClientFactory;
use support\Db;
use Throwable;
class OpenSearchAdminService
{
public function status(): array
{
$config = config('opensearch.default', []);
$indexName = config('opensearch.indices.chunks', 'proofdb_chunks');
$status = [
'config' => [
'hosts' => $config['hosts'] ?? [],
'ssl_verify' => (bool) ($config['ssl_verify'] ?? true),
'index_name' => $indexName,
],
'database' => [
'archives_total' => (int) Db::table('archives')->count(),
'chunks_total' => (int) Db::table('chunks')->count(),
'embedded_chunks' => (int) Db::table('chunks')->where('embedding_status', 3)->count(),
'indexed_chunks' => (int) Db::table('chunks')->where('search_index_status', 3)->count(),
],
'opensearch' => [
'reachable' => false,
'index_exists' => false,
'cluster_name' => null,
'health' => null,
'docs_count' => 0,
'mapping_fields' => [],
'error' => null,
],
];
try {
$client = (new OpenSearchClientFactory())->make();
$health = $client->cluster()->health();
$status['opensearch']['reachable'] = true;
$status['opensearch']['cluster_name'] = $health['cluster_name'] ?? null;
$status['opensearch']['health'] = $health['status'] ?? null;
$exists = (bool) $client->indices()->exists(['index' => $indexName]);
$status['opensearch']['index_exists'] = $exists;
if ($exists) {
$stats = $client->indices()->stats(['index' => $indexName]);
$mapping = $client->indices()->getMapping(['index' => $indexName]);
$status['opensearch']['docs_count'] = (int) (($stats['_all']['primaries']['docs']['count'] ?? 0));
$status['opensearch']['mapping_fields'] = array_keys($mapping[$indexName]['mappings']['properties'] ?? []);
}
} catch (Throwable $exception) {
$status['opensearch']['error'] = $exception->getMessage();
}
return $status;
}
public function documents(string $query = '', int $size = 20): array
{
$size = min(50, max(1, $size));
$indexName = config('opensearch.indices.chunks', 'proofdb_chunks');
$client = (new OpenSearchClientFactory())->make();
if (!(bool) $client->indices()->exists(['index' => $indexName])) {
return [
'index_name' => $indexName,
'items' => [],
'total' => 0,
];
}
$body = [
'_source' => [
'includes' => [
'chunk_uid',
'archive_uid',
'chunk_index',
'page_start',
'page_end',
'title',
'summary',
'source',
'author',
'year',
'series',
'tags',
'text',
'embedding_model',
'embedding_dimensions',
'created_time',
'updated_time',
],
],
'size' => $size,
'sort' => [
['updated_time' => ['order' => 'desc']],
],
];
$query = trim($query);
if ($query === '') {
$body['query'] = ['match_all' => (object) []];
} else {
$body['query'] = [
'multi_match' => [
'query' => $query,
'fields' => ['text^3', 'title^2', 'summary^2', 'source', 'author', 'tags'],
'type' => 'best_fields',
],
];
}
$response = $client->search([
'index' => $indexName,
'body' => $body,
]);
$hits = $response['hits']['hits'] ?? [];
return [
'index_name' => $indexName,
'total' => (int) (($response['hits']['total']['value'] ?? 0)),
'items' => array_map(function (array $hit): array {
$source = $hit['_source'] ?? [];
$text = trim((string) ($source['text'] ?? ''));
return [
'score' => $hit['_score'] ?? null,
'chunk_uid' => $source['chunk_uid'] ?? ($hit['_id'] ?? null),
'archive_uid' => $source['archive_uid'] ?? null,
'chunk_index' => $source['chunk_index'] ?? null,
'page_start' => $source['page_start'] ?? null,
'page_end' => $source['page_end'] ?? null,
'title' => $source['title'] ?? null,
'summary' => $source['summary'] ?? null,
'source' => $source['source'] ?? null,
'author' => $source['author'] ?? null,
'year' => $source['year'] ?? null,
'series' => $source['series'] ?? null,
'tags' => $source['tags'] ?? [],
'text_preview' => mb_substr($text, 0, 320),
'embedding_model' => $source['embedding_model'] ?? null,
'embedding_dimensions' => $source['embedding_dimensions'] ?? null,
'created_time' => $source['created_time'] ?? null,
'updated_time' => $source['updated_time'] ?? null,
];
}, $hits),
];
}
}