Algorithmic Trading Strategy Builder & Transpiler

This application is designed to close the gap between visual strategy intuition and executable algorithmic code. It enables traders to design, validate, and compile algorithmic trading strategies using a modular, visual dependency graph. Unlike simple drag-and-drop tools, this system functions as a Domain-Specific Language (DSL) Transpiler. It treats a visual trading strategy as an Directed Acyclic Graph (DAG), performing deep graph analysis to generate optimized Pine Script code.

Techincal Implementation

• Topological Dependency: Utilizes Kahn’s Algorithm to recursively resolve indicator-on-indicator dependencies (e.g., if an EMA strategy is also dependant on RSI values), ensuring that the generated code respects strict execution order.
• Signal Mutual Exclusion Architecture: Implements state-snapshotting (can_buy / can_sell primitives) to eliminate "signal flicker" and ensure that entry/exit transitions are atomically consistent within a single bar.
• 1:1 Signal-to-Execution Synchronization: Engineered to ensure that visual chart annotations precisely match the strategy's internal execution state, solving the common "Phantom Signals" problem.
• Graph-Based Compilation: Represents strategies as directed acyclic graphs (DAGs) using adjacency lists (Python dictionaries) for dependency tracking and in-degree counters for prerequisite validation, enabling O(1) lookups during topological sorting.
• Cycle Detection: Validates strategies as executable DAGs, preventing impossible circular dependencies in complex indicator chains

Setting up Demo for EMA12 and EMA26 nodes.

---

Key Capabilities

• Recursive Indicator Composition: Create "Indicator Chains" where any technical indicator can serve as a source for another, enabling deep multi-layer analysis.
• Modular Action Logic: Decouples entry triggers from exit logic, allowing for "Flip" strategies (reversals) or traditional fixed-risk frameworks.
• Dynamic Variable Scoping: Automatically manages unique variable namespace isolation in the generated code, preventing naming collisions in complex strategies.
• State-Persistent Risk Management: Handles conditional Stop-Loss and Take-Profit logic as independent state machines, omitting unused logic to minimize cumulative script latency.

---

Tech Stack & Algorithms

• Frontend: React with React Flow for dynamic graph visualization and state management.
• Backend: FastAPI (Python 3.10+) utilizing Pydantic for strict diagram validation of strategy graphs, exposing RESTful API endpoints for strategy management and compilation.
• Core Algorithms:
  • Topological Sorting for dependency resolution.
  • Graph Traversal for recursive source routing.
  • Transpilation Logic for mapping Abstract Nodes to Pine Script v5 primitives.
• Styling: Tailwind CSS with custom internal design for a clean interface experience.

---

Getting Started

Backend Execution

1. cd backend
2. pip install -r requirements.txt
3. python main.py (Default listener: port 8010)

Frontend Environment

1. cd frontend
2. npm install
3. npm run dev (Default listener: port 5173)

---

Roadmap & Future Scope

• Integrate a user database to save and share strategies.
• Potential Integration of C# (NinjaScript) and MQL5 transpilation layers.
• Direct integration with Interactive Brokers for live asset-exchange execution.
• Pre-compilation backtesting engine within the visual editor.
• Machine Learning Models to use within the workspace for trading strategies