Evandro Fonseca Junior
0cc7a81319
- Update .gitignore to exclude database files, extensions, and .cursor directory - Enhance chat streaming API with safe enqueue and close functions to prevent errors - Integrate ConversationSidebar component into ClientHome for better UI - Add dependencies like better-sqlite3 and esbuild packages for database functionality This commit introduces database persistence for conversations, improves error handling in streaming responses, and adds a sidebar for managing chats, enhancing overall app robustness and user experience.
101 lines
2.3 KiB
TypeScript
101 lines
2.3 KiB
TypeScript
import Database from "better-sqlite3";
|
|
|
|
export interface Embedding {
|
|
id: string;
|
|
message_id: string;
|
|
vector: Buffer;
|
|
}
|
|
|
|
export interface SearchResult {
|
|
id: string;
|
|
content: string;
|
|
role: string;
|
|
distance: number;
|
|
conversation_id: string;
|
|
}
|
|
|
|
export class EmbeddingRepository {
|
|
constructor(private db: Database.Database) {}
|
|
|
|
async create(messageId: string, vector: Float32Array): Promise<void> {
|
|
const id = crypto.randomUUID();
|
|
const blob = Buffer.from(vector.buffer);
|
|
|
|
this.db
|
|
.prepare(
|
|
`
|
|
INSERT INTO embeddings (id, message_id, vector)
|
|
VALUES (?, ?, ?)
|
|
`
|
|
)
|
|
.run(id, messageId, blob);
|
|
}
|
|
|
|
async searchSimilar(
|
|
queryVector: Float32Array,
|
|
limit: number = 5
|
|
): Promise<SearchResult[]> {
|
|
const queryBlob = Buffer.from(queryVector.buffer);
|
|
|
|
try {
|
|
// Tentar busca vetorial se extensão estiver disponível
|
|
return this.db
|
|
.prepare(
|
|
`
|
|
SELECT
|
|
m.id, m.content, m.role, m.conversation_id,
|
|
vector_distance(e.vector, ?) as distance
|
|
FROM embeddings e
|
|
JOIN messages m ON m.id = e.message_id
|
|
ORDER BY distance ASC
|
|
LIMIT ?
|
|
`
|
|
)
|
|
.all(queryBlob, limit) as SearchResult[];
|
|
} catch (error) {
|
|
console.warn(
|
|
"Busca vetorial não disponível, usando busca simples:",
|
|
error
|
|
);
|
|
|
|
// Fallback: busca simples por conteúdo
|
|
return this.db
|
|
.prepare(
|
|
`
|
|
SELECT
|
|
m.id, m.content, m.role, m.conversation_id,
|
|
0.0 as distance
|
|
FROM messages m
|
|
WHERE m.content LIKE ?
|
|
ORDER BY m.timestamp DESC
|
|
LIMIT ?
|
|
`
|
|
)
|
|
.all(`%${queryVector.toString()}%`, limit) as SearchResult[];
|
|
}
|
|
}
|
|
|
|
findByMessage(messageId: string): Embedding | null {
|
|
return this.db
|
|
.prepare(
|
|
`
|
|
SELECT * FROM embeddings WHERE message_id = ?
|
|
`
|
|
)
|
|
.get(messageId) as Embedding | null;
|
|
}
|
|
|
|
deleteByMessage(messageId: string): void {
|
|
this.db
|
|
.prepare("DELETE FROM embeddings WHERE message_id = ?")
|
|
.run(messageId);
|
|
}
|
|
|
|
count(): number {
|
|
const result = this.db
|
|
.prepare("SELECT COUNT(*) as count FROM embeddings")
|
|
.get() as { count: number };
|
|
return result.count;
|
|
}
|
|
}
|