GW_MC/DreamChat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

chore: add pnpm workspace configuration for apps and packages

Browse Source
This commit is contained in:
GW_MC
2026-02-23 21:04:19 +08:00
parent ab02758382
commit 932f384f0d
31 changed files with 9081 additions and 203 deletions
Download Patch File Download Diff File Expand all files Collapse all files

327
doc/database-schema.md
View File

@@ -23,35 +23,41 @@ CREATE EXTENSION IF NOT EXISTS "pgvector";
│ username │ │ name │ │ user_id (FK) │
│ password_hash │ │ avatar_url │ │ title │
│ keycloak_sub │ │ personality │ │ created_at │
│ role │ │ backstory │ │ updated_at │
│ created_at │ │ attributes │ └────────┬────────┘
│ updated_at │ │ created_at │ │
└─────────────────┘ │ updated_at │ │
└─────────────────┘ │
┌─────────────────┐ │
│import_documents │ │
├─────────────────┤ │
│ id (PK) │ │
│ user_id (FK) │ │
│ source_type │ ┌─────────────────┐ │
│ source_name │ │ messages │◄───────────────┘
│ content │ ├─────────────────┤
│ metadata │ │ id (PK) │
│ vector_id │ │ conversation_id │
│ created_at │ │ role │
└─────────────────┘ │ content │
│ tokens_used │
│ model │
┌─────────────────┐ │ metadata │
│vector_memories │ │ created_at │
├─────────────────┤ └─────────────────┘
│ id (PK) │
│ conversation_id │ ┌─────────────────┐
│ content │ │ story_branches │ (Phase 2)
│ embedding │ ├─────────────────┤
│ metadata │ │ id (PK) │
│ created_at │ │ conversation_id │
└─────────────────┘ │ parent_id (FK) │
│ role │ │ attributes │ │ updated_at │
│ created_at │ │ created_at │ └────────┬────────┘
│ updated_at │ │ updated_at │ │
└─────────────────┘ └────────┬────────┘ │
│ │
┌────────┴────────┐ │
│character_knowledge│ │
├─────────────────┤ │
│ id (PK) │◄───────────────┤
│ character_id │ │
│ name │ │
│ source_type │ │
│ raw_content │ │
│ status │ │
└────────┬────────┘ │
│ │
┌─────────────────┐ ┌────────┴────────┐ ┌────────┴────────┐
│import_documents │ │ vector_memories│ │ messages │
├─────────────────┤ ├─────────────────┤ ├─────────────────┤
│ id (PK) │ │ id (PK) │ │ id (PK) │
│ user_id (FK) │ │ content │ │ conversation_id │
│ source_type │ │ embedding │ │ role │
│ source_name │ │ memory_type │ │ content │
│ content │ │ conversation_id │ │ tokens_used │
│ status │ │ character_id │ │ model │
└─────────────────┘ │ knowledge_id │ │ metadata │
│ created_at │ │ created_at │
└─────────────────┘ └─────────────────┘
│
┌────────┴────────┐
│ story_branches │ (Phase 2)
├─────────────────┤
│ id (PK) │
│ conversation_id │
│ parent_id (FK) │
│ content │
│ direction │
│ metadata │
@@ -89,7 +95,7 @@ CREATE INDEX idx_users_keycloak_sub ON users(keycloak_sub);
### 2. characters
Character definitions with complex attribute system (JSONB for flexibility).
Character definitions with complex attribute system. Character knowledge is stored separately in `character_knowledge` with embeddings.
```sql
CREATE TABLE characters (
@@ -101,9 +107,6 @@ CREATE TABLE characters (
-- Core personality prompt sent to LLM
personality_prompt TEXT NOT NULL,
-- Backstory context for the character
backstory TEXT,
-- Complex attribute system (structured JSON)
-- Example: {"traits": ["brave", "witty"], "age": 25, "species": "human"}
attributes JSONB DEFAULT '{}',
@@ -181,19 +184,53 @@ CREATE INDEX idx_messages_created_at ON messages(created_at);
CREATE INDEX idx_messages_conversation_created ON messages(conversation_id, created_at);
```
### 5. vector_memories
### 5. character_knowledge
Vector embeddings for conversation memory using pgvector. Stores chunked content for semantic search.
Multiple knowledge sources for characters. Each source is chunked and stored with embeddings in `vector_memories`.
```sql
-- Create vector extension first
CREATE EXTENSION IF NOT EXISTS vector;
CREATE TYPE import_source_type AS ENUM ('file', 'url', 'manual');
CREATE TYPE import_status AS ENUM ('pending', 'processing', 'completed', 'failed');
CREATE TABLE character_knowledge (
id UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
character_id UUID NOT NULL REFERENCES characters(id) ON DELETE CASCADE,
-- Knowledge source info
name VARCHAR(255) NOT NULL, -- Display name
source_type import_source_type NOT NULL,
source_name VARCHAR(255) NOT NULL, -- Original filename or URL
mime_type VARCHAR(100),
file_size BIGINT,
-- Raw content (before chunking)
raw_content TEXT,
-- Processing status
status import_status DEFAULT 'pending',
processing_info JSONB, -- chunks count, errors, etc.
created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
updated_at TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);
CREATE INDEX idx_character_knowledge_character ON character_knowledge(character_id);
CREATE INDEX idx_character_knowledge_status ON character_knowledge(status);
```
### 6. vector_memories
Unified vector embeddings storage for:
- **Character knowledge** - Background info, imported documents (linked to `character_knowledge`)
- **Conversation history** - Chat context (linked to `conversations`)
```sql
CREATE TYPE memory_type AS ENUM ('conversation', 'character');
CREATE TABLE vector_memories (
id UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
conversation_id UUID NOT NULL REFERENCES conversations(id) ON DELETE CASCADE,
-- The text content
-- The text chunk
content TEXT NOT NULL,
-- Vector embedding (configurable dimension based on model)
@@ -201,14 +238,21 @@ CREATE TABLE vector_memories (
-- Must match the EMBEDDING_DIMENSION env var
embedding VECTOR({{EMBEDDING_DIMENSION}}),
-- Metadata for filtering
metadata JSONB DEFAULT '{
"chunk_index": 0,
"source": "conversation",
"timestamp": null
}',
-- Memory type determines the context
memory_type memory_type DEFAULT 'conversation',
created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW()
-- Metadata (chunk_index, source_info, etc.)
metadata JSONB,
created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
-- Polymorphic relations (at least one must be set)
-- For conversation context
conversation_id UUID REFERENCES conversations(id) ON DELETE CASCADE,
-- For character knowledge
character_id UUID REFERENCES characters(id) ON DELETE CASCADE,
knowledge_id UUID REFERENCES character_knowledge(id) ON DELETE CASCADE
);
-- HNSW index for efficient similarity search
@@ -216,22 +260,22 @@ CREATE TABLE vector_memories (
-- CREATE INDEX idx_vector_memories_embedding ON vector_memories
-- USING hnsw (embedding vector_cosine_ops);
CREATE INDEX idx_vector_memories_conversation ON vector_memories(conversation_id);
CREATE INDEX idx_vector_memories_conversation ON vector_memories(conversation_id) WHERE conversation_id IS NOT NULL;
CREATE INDEX idx_vector_memories_character ON vector_memories(character_id) WHERE character_id IS NOT NULL;
CREATE INDEX idx_vector_memories_knowledge ON vector_memories(knowledge_id) WHERE knowledge_id IS NOT NULL;
CREATE INDEX idx_vector_memories_type ON vector_memories(memory_type);
```
### 6. import_documents
### 7. import_documents
Raw imported documents from files or web scraping.
General-purpose imported documents (not linked to characters). For character knowledge, use `character_knowledge`.
```sql
CREATE TYPE import_source_type AS ENUM ('file', 'url');
CREATE TYPE import_status AS ENUM ('pending', 'processing', 'completed', 'failed');
CREATE TABLE import_documents (
id UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
user_id UUID NOT NULL REFERENCES users(id) ON DELETE CASCADE,
source_type import_source_type NOT NULL,
source_type import_source_type NOT NULL, -- file, url, manual
source_name VARCHAR(255) NOT NULL, -- filename or URL
-- Mime type for files
@@ -258,7 +302,7 @@ CREATE INDEX idx_import_documents_user_id ON import_documents(user_id);
CREATE INDEX idx_import_documents_status ON import_documents(status);
```
### 7. story_branches (Phase 2)
### 8. story_branches (Phase 2)
Tree structure for branching narratives.
@@ -291,7 +335,7 @@ CREATE INDEX idx_story_branches_conversation ON story_branches(conversation_id);
CREATE INDEX idx_story_branches_parent ON story_branches(parent_id);
```
### 8. conversation_participants (Phase 3 - Multi-Character)
### 9. conversation_participants (Phase 3 - Multi-Character)
Supports multiple characters in a single conversation.
@@ -313,12 +357,73 @@ CREATE TABLE conversation_participants (
CREATE INDEX idx_participants_conversation ON conversation_participants(conversation_id);
```
### Enums
```sql
-- User roles
CREATE TYPE user_role AS ENUM ('USER', 'ADMIN');
-- Message roles
CREATE TYPE message_role AS ENUM ('user', 'assistant', 'system');
-- Import/knowledge source types
CREATE TYPE import_source_type AS ENUM ('file', 'url', 'manual');
-- Processing status
CREATE TYPE import_status AS ENUM ('pending', 'processing', 'completed', 'failed');
-- Vector memory types
CREATE TYPE memory_type AS ENUM ('conversation', 'character');
```
## Prisma Schema (Reference)
Prisma schema uses the [multi-file schema](https://www.prisma.io/docs/orm/prisma-schema/overview/location) feature. Models are organized in `prisma/models/` folder and imported into `schema.prisma`.
### Schema Structure
```
prisma/
├── schema.prisma # Main schema file with imports
├── seed.ts # Database seeding
└── models/
├── user.prisma # User model + UserRole enum
├── character.prisma # Character + CharacterKnowledge models
├── conversation.prisma # Conversation + ConversationParticipant
├── message.prisma # Message model + MessageRole enum
├── vectorMemory.prisma # VectorMemory + MemoryType enum
├── importDocument.prisma # ImportDocument model
└── storyBranch.prisma # StoryBranch model
```
### Main Schema (schema.prisma)
```prisma
// schema.prisma
generator client {
provider = "prisma-client-js"
}
datasource db {
provider = "postgresql"
url = env("DATABASE_URL")
}
// Import all models from the models folder
import { User } from "./models/user"
import { Character, CharacterKnowledge } from "./models/character"
import { Conversation, ConversationParticipant } from "./models/conversation"
import { Message } from "./models/message"
import { VectorMemory } from "./models/vectorMemory"
import { ImportDocument } from "./models/importDocument"
import { StoryBranch } from "./models/storyBranch"
```
### Full Schema Definition
```prisma
// schema.prisma
// models/user.prisma
generator client {
provider = "prisma-client-js"
@@ -344,6 +449,7 @@ enum MessageRole {
enum ImportSourceType {
file
url
manual
}
enum ImportStatus {
@@ -353,6 +459,11 @@ enum ImportStatus {
failed
}
enum MemoryType {
conversation
character
}
// Models
model User {
id String @id @default(uuid())
@@ -378,7 +489,6 @@ model Character {
name String
avatarUrl String?
personalityPrompt String
backstory String?
attributes Json @default("{}")
config Json @default("{}")
isPublic Boolean @default(false)
@@ -386,30 +496,53 @@ model Character {
updatedAt DateTime @updatedAt
userId String
user User @relation(fields: [userId], references: [id], onDelete: Cascade)
user User @relation(fields: [userId], references: [id], onDelete: Cascade)
conversations Conversation[]
knowledgeSources CharacterKnowledge[]
vectorMemories VectorMemory[]
@@index([userId])
@@index([name])
}
model Conversation {
id String @id @default(uuid())
title String?
messageCount Int @default(0)
totalTokens Int @default(0)
settings Json @default("{}")
createdAt DateTime @default(now())
updatedAt DateTime @updatedAt
model CharacterKnowledge {
id String @id @default(uuid())
name String
sourceType ImportSourceType
sourceName String
mimeType String?
fileSize BigInt?
rawContent String?
status ImportStatus @default(pending)
processingInfo Json?
createdAt DateTime @default(now())
updatedAt DateTime @updatedAt
userId String
user User @relation(fields: [userId], references: [id], onDelete: Cascade)
characterId String
character Character @relation(fields: [characterId], references: [id], onDelete: Cascade)
messages Message[]
characterId String
character Character @relation(fields: [characterId], references: [id], onDelete: Cascade)
vectorMemories VectorMemory[]
@@index([characterId])
@@index([status])
}
model Conversation {
id String @id @default(uuid())
title String?
messageCount Int @default(0)
totalTokens Int @default(0)
settings Json @default("{}")
createdAt DateTime @default(now())
updatedAt DateTime @updatedAt
userId String
user User @relation(fields: [userId], references: [id], onDelete: Cascade)
characterId String
character Character @relation(fields: [characterId], references: [id], onDelete: Cascade)
messages Message[]
vectorMemories VectorMemory[]
storyBranches StoryBranch[]
participants ConversationParticipant[]
storyBranches StoryBranch[]
participants ConversationParticipant[]
@@index([userId])
@@index([characterId])
@@ -436,14 +569,24 @@ model Message {
model VectorMemory {
id String @id @default(uuid())
content String
embedding Unsupported("vector")? // pgvector extension
embedding Unsupported("vector")?
memoryType MemoryType @default(conversation)
metadata Json?
createdAt DateTime @default(now())
conversationId String
conversation Conversation @relation(fields: [conversationId], references: [id], onDelete: Cascade)
conversationId String?
conversation Conversation? @relation(fields: [conversationId], references: [id], onDelete: Cascade)
characterId String?
character Character? @relation(fields: [characterId], references: [id], onDelete: Cascade)
knowledgeId String?
knowledge CharacterKnowledge? @relation(fields: [knowledgeId], references: [id], onDelete: Cascade)
@@index([conversationId])
@@index([characterId])
@@index([knowledgeId])
@@index([memoryType])
}
model ImportDocument {
@@ -546,11 +689,16 @@ export class CharacterRepository {
```typescript
// Similarity search using pgvector with Prisma
async similaritySearch(
conversationId: string,
targetId: string,
queryEmbedding: number[],
memoryType: MemoryType,
k: number = 5
) {
// Using raw query for pgvector-specific operations
// Build the where clause based on memory type
const whereClause = memoryType === 'conversation'
? { conversationId: targetId, memoryType }
: { characterId: targetId, memoryType };
const results = await this.prisma.$queryRaw`
SELECT
id,
@@ -558,7 +706,7 @@ async similaritySearch(
metadata,
embedding <=> ${queryEmbedding}::vector as distance
FROM "VectorMemory"
WHERE "conversationId" = ${conversationId}
WHERE ${whereClause}
ORDER BY embedding <=> ${queryEmbedding}::vector
LIMIT ${k}
`;
@@ -566,20 +714,23 @@ async similaritySearch(
return results;
}
// Alternative: using cosine similarity
async similaritySearchCosine(
conversationId: string,
// Search character knowledge
async searchCharacterKnowledge(
characterId: string,
queryEmbedding: number[],
k: number = 5
) {
const results = await this.prisma.$queryRaw`
SELECT
id,
content,
metadata,
1 - (embedding <=> ${queryEmbedding}::vector) as similarity
FROM "VectorMemory"
WHERE "conversationId" = ${conversationId}
vm.id,
vm.content,
vm.metadata,
ck.name as source_name,
1 - (vm.embedding <=> ${queryEmbedding}::vector) as similarity
FROM "VectorMemory" vm
JOIN "CharacterKnowledge" ck ON vm."knowledgeId" = ck.id
WHERE vm."characterId" = ${characterId}
AND vm."memoryType" = 'character'
ORDER BY similarity DESC
LIMIT ${k}
`;