Signal Processing Pipeline

This document covers the automated signal generation, validation, and execution system that forms the core trading logic of the platform. The pipeline processes market data from multiple sources, applies AI-driven analysis, validates trading opportunities, and distributes actionable signals through various channels.

For information about the AI agents that provide market analysis, see AI Agent Swarm System. For details about external service integrations used by the pipeline, see External Service Integrations.

Overview

The signal processing pipeline is an automated system that continuously monitors cryptocurrency markets, generates trading signals using AI analysis, validates these signals against predefined rules, and distributes them to users through multiple channels. The pipeline operates on a scheduled basis and can process signals for multiple cryptocurrency symbols simultaneously.

Core Data Structures

The pipeline operates on several key data structures that define signal information and processing states:

Interface	Purpose	Key Fields
ISignalDto	Signal data transfer object	symbol, buyPrice, sellPrice, quantity, stopLossPrice, possibilityPercent
ISignalRow	Persisted signal with ID	Extends ISignalDto with id field
SignalContract	AI agent communication	action, stopLoss, possibility, comment, info
ISignalReport	Aggregated market data	longRangeReport, swingRangeReport, shortRangeReport, mastodonNewsReport

Signal Generation Flow

The signal generation process follows a structured workflow that combines market data analysis with AI-driven decision making:

Signal Validation Logic

The SignalValidationService implements sophisticated validation rules to determine when signals should be executed and when positions should be closed:

Execution Validation

The validation process checks multiple criteria before allowing signal execution:

• EMA Signal Strength: Analyzes recent EMA signals using validateShouldExecute() function
• Price Range Conflicts: Prevents overlapping orders within configured percentage ranges
• Timing Constraints: Enforces cooldown periods based on CC_LADDER_FAIL_HOURS and CC_LADDER_GROWTH_HOURS

Position Resolution Validation

The system automatically identifies profitable positions ready for closure:

• Revenue Calculation: Compares current market price against target sell prices
• Profit Threshold: Uses CC_LADDER_TRADE_REVENUE configuration for minimum profit requirements
• Multiple Position Handling: Can close multiple positions simultaneously if they meet criteria

Signal Types and Actions

The pipeline supports three primary signal actions defined in the SignalAction enum:

Action	Enum Value	Purpose	Generated By
Buy Signal	SignalAction.SignalBuy	Recommends purchasing cryptocurrency	AI analysis of bullish indicators
Wait Signal	SignalAction.SignalWait	Advises holding current positions	Mixed or bearish market conditions
Close Notification	N/A	Triggered when positions reach profit targets	Automated validation checks

Buy Signal Generation

Buy signals are generated when the AI agent calls the SignalRecommendBuy tool with required parameters:

• reason: Detailed analysis justifying the purchase recommendation
• stop_loss: Price level for loss mitigation
• possibility: Probability percentage of price increase (1-100)

Distribution Mechanisms

Once signals are validated and processed, they are distributed through multiple channels to ensure users receive timely notifications:

Telegram Publishing

The TelegramWebService formats and publishes signals to configured Telegram channels:

• Buy Signals: Include current price, stop-loss, sell target, and investment summary
• Wait Signals: Provide market analysis and reasoning for holding positions
• Sell Notifications: Display profit/loss calculations and performance metrics

Webhook Notifications

The WebhookService sends HTTP POST requests to external systems:

• Authentication: Uses JWT tokens generated by JwtService
• Endpoints: /api/v1/buy, /api/v1/wait, /api/v1/sell
• Error Handling: Comprehensive logging for failed webhook deliveries

Database Persistence

All signals are persisted using the SignalDbService which extends the base CRUD functionality:

• Active Signals: Tracks unresolved positions using getActiveSignals()
• Chart Data: Provides historical signal data via getChartSignals()
• Latest Signal: Retrieves most recent signal for timing validations

Automated Job Processing

The SignalJobService orchestrates the entire pipeline execution on a scheduled basis:

The service runs continuously with configurable intervals, processing multiple cryptocurrency symbols and handling both signal generation and position resolution in each cycle.