Exception raised for database-related errors

Standardized response model for all API endpoints. This model provides a consistent structure for all API responses with: - A status fie

Exception raised for authentication failures

Service for chat functionality, including SQL generation, data querying, and visualization generation using the repository pattern.

Exception raised when a requested resource is not found

Service for managing datasources with optimized database operations. Uses repository pattern, connection pooling, and standardized error

Generic dependency provider for repository classes. Automatically handles injecting the appropriate session type based on current contex

Enum for supported database types

Agent for classifying user question intent and detecting ambiguity

Detailed response DTO with full columns and relationships

Detailed response DTO with full models and metrics

Exception raised for data validation errors

Asynchronous PostgreSQL database access with connection pooling. Provides high-performance non-blocking database operations.

Exception raised for invalid request data

Data Transfer Object for chat requests.

Individual message within a chat session.

Chat session entity representing a conversation thread.

Connection info for DuckDB datasource.

In-memory cache implementation with TTL support. This implementation stores cache entries and their expiration times in memory. It's eff

OpenAI-compatible LLM provider

Connection info for PostgreSQL datasource.

Exception raised for authentication errors

用户信息响应

认证服务

Connection info for BigQuery datasource.

Response model for cache operations.

History of chat interactions and operations for analytics and audit. This entity consolidates the functionality from the previous Conversati

Connection info for ClickHouse datasource.

Response model for conversation operations.

Database model for storing datasource connection information. Represents a data source configuration that users can connect to for queries.

An object representing an operation that failed for the reason given in `message`.

Summary of the insight generated.

Observer for tracking LLM calls with Langfuse. Features: - Automatic trace/generation creation for LLM calls - Token counting and cos

Represents a business entity in the semantic layer

Container for MDL models and metrics

Connection info for Microsoft SQL Server datasource.

Connection info for MySQL datasource.

Synchronous PostgreSQL database access with connection pooling. Provides standardized methods for database access with proper connection

Exception raised when a required service is unavailable

Connection info for Snowflake datasource.

Token 响应

Connection info for Trino datasource.

API configuration settings.

Context manager for tracking agent executions

Agent for generating natural language answers

Complete question answering pipeline

Qdrant connection manager (asynchronous)

用户仓储（异步）

用户会话仓储（异步）

Cache configuration settings.

Agent for generating chart configurations

Domain entity representing a chat message. This is a pure domain entity that's not tied to the ORM, but can be converted to and from the

Repository for chat operations using synchronous database access. Provides methods for managing chat sessions, messages, history and statist

Domain entity representing a chat session. This is a pure domain entity that's not tied to the ORM, but can be converted to and from the

Common response format for chat operations.

Application configuration class.

Enhanced connection manager for database connections. This class provides a centralized interface for database connections with advanced

Generic connection URL.

Represents the context maintained for a conversation. This includes things like the current schema, generated SQL, and other state that

ORM entity for Correction Log.

Repository for accessing correction logs.

Domain entity representing a data source configuration. This is a pure domain entity that's not tied to the ORM, but can be converted to

Response model for listing datasources with pagination information

Response model for datasource information

Response model for connection test results

Database connection configuration.

Repository for datasource operations.

Repository for accessing diagnosis results.

ORM entity for Diagnosis Result.

Enhanced rate limiter with per-user limits and Redis support. Rate Limit Tiers: - Anonymous users: 10 requests/minute, 100 requests/hour

Pipeline for retrieving similar historical questions

JWT authentication configuration.

Context manager for tracking LLM calls

LLM service configuration.

Langfuse observability configuration.

Logging configuration settings.

Represents a column in an MDL model

Represents a predefined metric (KPI)

Represents a relationship between two MDL models

Pipeline for retrieving MDL context from Qdrant

MDL domain service

Response DTO for MDL sync operation

Centralized metrics collection for GenBI pipeline. Usage: ```python collector = MetricsCollector() # Record

Context manager for tracking pipeline steps

Qdrant vector database configuration.

Qdrant connection manager (synchronous)

Database model for storing query execution history. Keeps track of queries executed against datasources.

Agent for generating query reasoning plans

Type definition for query execution results.

Redis-based cache implementation with TTL support. This implementation stores cache entries in Redis, making it suitable for distributed

Context manager for tracking requests

Result model for SQL execution, used internally.

SQLite database adapter implementation with enhanced connection pooling.