tick-stock-panel

Product Introduction

1. Definition: TSP (tick-stock-panel) is a self-hosted, zero-ops A-share quantitative workstation. It is a full-stack, containerized application that integrates stock screening, real-time market monitoring, and strategy backtesting into a single platform.
2. Core Value Proposition: It solves the pain points of individual investors and quantitative enthusiasts who lack professional tools and face operational complexity. By leveraging LLM capabilities for strategy customization, stock analysis, and post-market review, and allowing free integration of third-party data sources, it provides a systematic, local-first investment analysis solution.

Main Features

1. Capability Routing Matrix: This is the core data abstraction layer. It routes six categories of datasets (e.g., daily bars, financials, minute bars) independently based on the declared capabilities of each data source plugin (like TickFlow, fuyao, or custom YAML sources). This allows users to switch data providers without changing their indicator calculations or backtesting logic, ensuring consistent data口径.
2. Screening & Backtesting Engine: Powered by Polars for millisecond-level full A-share scanning. It features a unified strategy pool for daily and minute-level strategies, with automatic routing based on strategy declaration cycles. The backtesting engine supports T+1 settlement, commission fees, slippage, stop-loss, and minute-by-minute trade replay. Results include comprehensive metrics like Sharpe ratio, max drawdown, win rate, and profit factor.
3. AI-Powered Dialogue Assistant: A fully decoupled extension module that provides conversational data Q&A. It utilizes 18 read-only tools covering个股,大盘,板块,自选,策略, and因子 data. It offers token-by-token streaming output, traceable tool call cards (showing parameters and latency), and built-in compliance disclaimers. It requires an OpenAI-compatible API key for function calling.
4. Real-time Anomaly Monitoring Center: Covers three trading phases: pre-market auction anomalies, intraday signals (limit-ups, breakdowns, new highs/lows), and exchange-defined price deviation alerts. It supports real-time desktop notifications, voice alerts, and Feishu push notifications. Rules can be scoped to watchlist groups.
5. Factor Platform & Mining: Includes a DSL for自定义因子 creation with 25 built-in operators. The因子挖掘 system performs out-of-sample搜索 for multi-factor combinations, with explicit publishing controls to prevent accidental deployment. The市场环境 module identifies 6 emotional cycle stages (冰点,启动,主升,高潮,退潮,修复) and ranks leading sectors/concepts.

Problems Solved

1. Pain Point: Fragmented quantitative workflow using disparate scripts and tools, leading to inconsistent data口径 between screening, backtesting, and monitoring.
2. Target Audience: Individual developers and quantitative trading enthusiasts interested in the A-share market, who seek a systematic, self-hosted analysis platform without the cost and complexity of cloud-based professional terminals.
3. Use Cases: 1) Rapid prototyping and testing of quantitative stock selection strategies. 2) Real-time monitoring of portfolio holdings and specific market signals. 3) Conducting rigorous factor research and validity testing. 4) Using AI for automated market review and individual stock analysis.

Unique Advantages

1. Differentiation: Unlike bloated commercial software (e.g., Tonghuashun) or complex cloud platforms, TSP is lightweight, self-hosted, and keeps all data本地. Unlike拼凑 toolkits, it offers a unified enriched data pipeline and consistent backtesting framework.
2. Key Innovation: The plugin-based Capability Routing Matrix decouples data logic from business logic. The completely decoupled extension system (exemplified by the AI Assistant) allows features to be added or removed by simply adding/deleting directories, with zero modifications to core files. The交易日探针 intelligently halts real-time tasks during market holidays.

Frequently Asked Questions (FAQ)

1. Is tick-stock-panel free to use? Yes, the TSP platform itself is open-source (MIT license). However, to use its full functionality, you may need API keys for specific data sources (like TickFlow or fuyao) and AI services, which may have their own costs or usage limits.
2. What data sources does TSP support? It natively supports TickFlow SDK and fuyao (同花顺 REST). Its plugin architecture allows easy integration of custom data sources via YAML configuration or Python plugins (e.g., stock-sdk), enabling users to route different data categories to their preferred providers.
3. How does TSP handle real-time data and backtesting? It uses a交易日探针 to manage scheduling. Real-time monitoring runs only on trading days. Backtesting is executed in an isolated spawn worker process with persistent run IDs, allowing users to refresh the page without losing progress. Results are streamed via Server-Sent Events (SSE).
4. Can I use TSP without any coding knowledge? Basic usage for screening and monitoring with built-in strategies requires minimal coding. However, creating custom strategies, factors, or integrating custom data sources requires Python programming knowledge.
5. How does the AI Dialogue Assistant work and is it safe? The assistant is a tool-calling agent that queries the platform's本地真实数据. It does not provide trading advice. Every answer includes risk and data source disclaimers. All tool calls are logged and traceable for verification. It requires an AI provider with function-calling capability (e.g., OpenAI-compatible APIs).