Files
AIclinicalresearch/backend/verify_system.ts
HaHafeng 40c2f8e148 feat(rag): Complete RAG engine implementation with pgvector
Major Features:
- Created ekb_schema (13th schema) with 3 tables: KB/Document/Chunk
- Implemented EmbeddingService (text-embedding-v4, 1024-dim vectors)
- Implemented ChunkService (smart Markdown chunking)
- Implemented VectorSearchService (multi-query + hybrid search)
- Implemented RerankService (qwen3-rerank)
- Integrated DeepSeek V3 QueryRewriter for cross-language search
- Python service: Added pymupdf4llm for PDF-to-Markdown conversion
- PKB: Dual-mode adapter (pgvector/dify/hybrid)

Architecture:
- Brain-Hand Model: Business layer (DeepSeek) + Engine layer (pgvector)
- Cross-language support: Chinese query matches English documents
- Small Embedding (1024) + Strong Reranker strategy

Performance:
- End-to-end latency: 2.5s
- Cost per query: 0.0025 RMB
- Accuracy improvement: +20.5% (cross-language)

Tests:
- test-embedding-service.ts: Vector embedding verified
- test-rag-e2e.ts: Full pipeline tested
- test-rerank.ts: Rerank quality validated
- test-query-rewrite.ts: Cross-language search verified
- test-pdf-ingest.ts: Real PDF document tested (Dongen 2003.pdf)

Documentation:
- Added 05-RAG-Engine-User-Guide.md
- Added 02-Document-Processing-User-Guide.md
- Updated system status documentation

Status: Production ready
2026-01-21 20:24:29 +08:00

177 lines
5.8 KiB
TypeScript

import { PrismaClient } from '@prisma/client';
const prisma = new PrismaClient();
async function main() {
console.log('🔍 系统完整性验证\n');
console.log('=' .repeat(60));
// 1. 检查 schema.prisma 定义的表 vs 数据库实际的表
console.log('\n📋 1. 检查表结构完整性\n');
// 获取数据库中所有用户表
const dbTables: any[] = await prisma.$queryRaw`
SELECT table_schema, table_name
FROM information_schema.tables
WHERE table_schema NOT IN ('pg_catalog', 'information_schema', 'pg_toast')
AND table_type = 'BASE TABLE'
ORDER BY table_schema, table_name;
`;
// 按 schema 分组显示
const tablesBySchema: Record<string, string[]> = {};
for (const t of dbTables) {
if (!tablesBySchema[t.table_schema]) {
tablesBySchema[t.table_schema] = [];
}
tablesBySchema[t.table_schema].push(t.table_name);
}
for (const [schema, tables] of Object.entries(tablesBySchema)) {
console.log(`\n${schema}:`);
tables.forEach(t => console.log(` - ${t}`));
}
// 2. 检查关键表的列结构
console.log('\n\n📋 2. 检查关键表的列结构\n');
const platformUsersCols: any[] = await prisma.$queryRaw`
SELECT column_name, data_type, is_nullable, column_default
FROM information_schema.columns
WHERE table_schema = 'platform_schema' AND table_name = 'users'
ORDER BY ordinal_position;
`;
console.log('platform_schema.users 列结构:');
platformUsersCols.forEach(c => {
console.log(` ${c.column_name}: ${c.data_type} ${c.is_nullable === 'NO' ? 'NOT NULL' : 'NULLABLE'}`);
});
// 3. 检查枚举类型
console.log('\n\n📋 3. 检查枚举类型\n');
const enums: any[] = await prisma.$queryRaw`
SELECT n.nspname as schema, t.typname as enum_name,
string_agg(e.enumlabel, ', ' ORDER BY e.enumsortorder) as values
FROM pg_type t
JOIN pg_enum e ON t.oid = e.enumtypid
JOIN pg_namespace n ON t.typnamespace = n.oid
WHERE n.nspname NOT IN ('pg_catalog', 'information_schema')
GROUP BY n.nspname, t.typname
ORDER BY n.nspname, t.typname;
`;
enums.forEach(e => {
console.log(` ${e.schema}.${e.enum_name}: [${e.values}]`);
});
// 4. 验证 Prisma 连接和基本操作
console.log('\n\n📋 4. 验证 Prisma ORM 基本操作\n');
try {
// 测试读取用户
const userCount = await prisma.user.count();
console.log(` ✅ User.count(): ${userCount}`);
// 测试读取租户
const tenantCount = await prisma.tenant.count();
console.log(` ✅ Tenant.count(): ${tenantCount}`);
// 测试读取部门
const deptCount = await prisma.department.count();
console.log(` ✅ Department.count(): ${deptCount}`);
// 测试读取权限
const permCount = await prisma.permission.count();
console.log(` ✅ Permission.count(): ${permCount}`);
// 测试读取 Prompt 模板
const promptCount = await prisma.promptTemplate.count();
console.log(` ✅ PromptTemplate.count(): ${promptCount}`);
// 测试关联查询
const userWithTenant = await prisma.user.findFirst({
include: { tenant: true, department: true }
});
if (userWithTenant) {
console.log(` ✅ 关联查询成功: ${userWithTenant.name} @ ${userWithTenant.tenant?.name || '无租户'}`);
}
} catch (e: any) {
console.log(` ❌ Prisma 操作失败: ${e.message}`);
}
// 5. 验证其他模块的表是否可访问
console.log('\n\n📋 5. 验证其他模块表的可访问性\n');
const moduleTests = [
{ name: 'AIA Projects', query: 'SELECT COUNT(*) as count FROM aia_schema.projects' },
{ name: 'ASL Screening Projects', query: 'SELECT COUNT(*) as count FROM asl_schema.screening_projects' },
{ name: 'ASL Literatures', query: 'SELECT COUNT(*) as count FROM asl_schema.literatures' },
{ name: 'DC Templates', query: 'SELECT COUNT(*) as count FROM dc_schema.dc_templates' },
{ name: 'DC Extraction Tasks', query: 'SELECT COUNT(*) as count FROM dc_schema.dc_extraction_tasks' },
{ name: 'IIT Projects', query: 'SELECT COUNT(*) as count FROM iit_schema.projects' },
{ name: 'PKB Knowledge Bases', query: 'SELECT COUNT(*) as count FROM pkb_schema.knowledge_bases' },
{ name: 'PKB Documents', query: 'SELECT COUNT(*) as count FROM pkb_schema.documents' },
{ name: 'RVW Review Tasks', query: 'SELECT COUNT(*) as count FROM rvw_schema.review_tasks' },
];
for (const test of moduleTests) {
try {
const result: any = await prisma.$queryRawUnsafe(test.query);
console.log(`${test.name}: ${result[0].count} 条记录 (表存在)`);
} catch (e: any) {
console.log(`${test.name}: 表不存在或查询失败`);
}
}
// 6. 检查外键约束
console.log('\n\n📋 6. 检查外键约束\n');
const fks: any[] = await prisma.$queryRaw`
SELECT
tc.table_schema,
tc.table_name,
kcu.column_name,
ccu.table_schema AS foreign_table_schema,
ccu.table_name AS foreign_table_name,
ccu.column_name AS foreign_column_name
FROM information_schema.table_constraints AS tc
JOIN information_schema.key_column_usage AS kcu
ON tc.constraint_name = kcu.constraint_name
AND tc.table_schema = kcu.table_schema
JOIN information_schema.constraint_column_usage AS ccu
ON ccu.constraint_name = tc.constraint_name
WHERE tc.constraint_type = 'FOREIGN KEY'
AND tc.table_schema IN ('platform_schema', 'admin_schema', 'capability_schema')
ORDER BY tc.table_schema, tc.table_name;
`;
fks.forEach(fk => {
console.log(` ${fk.table_schema}.${fk.table_name}.${fk.column_name} -> ${fk.foreign_table_schema}.${fk.foreign_table_name}.${fk.foreign_column_name}`);
});
console.log('\n\n' + '=' .repeat(60));
console.log('✅ 系统验证完成!');
}
main()
.catch(console.error)
.finally(() => prisma.$disconnect());