Report Generation Services

Purpose and Scope

The Report Generation Services provide AI-powered market analysis and trading insights for cryptocurrency trading signals. These services generate various types of reports including technical analysis, social sentiment analysis, and market data reports across different time ranges. The system leverages the AI agent swarm to create detailed, contextual reports for specific cryptocurrency symbols.

For information about signal processing and validation, see Signal Generation and Validation. For details about the AI agents that power these reports, see AI Agent Swarm System.

Architecture Overview

The report generation system follows a layered architecture with public service facades that orchestrate private implementation services through the AI agent swarm framework.

Report Types and Categories

The system provides six distinct types of reports, each specialized for different aspects of market analysis:

Report Type	Purpose	Data Sources	Time Frame
Short Range	Short-term technical analysis	Binance price data, technical indicators	Short-term
Swing Range	Swing trading analysis	Binance price data, technical indicators	Medium-term
Long Range	Long-term trend analysis	Binance price data, technical indicators	Long-term
Price Data	Price-focused technical analysis	Binance price data, technical indicators	Variable
Volume Data	Volume-based market analysis	Binance volume data, technical indicators	Variable
Mastodon Search	Social sentiment analysis	Mastodon social media posts	Real-time

Service Architecture and Patterns

All report generation services follow a consistent architectural pattern with public facades, dependency injection, and standardized caching mechanisms.

Common Structure Pattern

Dependency Injection Pattern

Each public service uses the IoC container to inject its dependencies:

// Common injection pattern across all report services
private privateService = inject<PrivateServiceType>(TYPES.privateServiceType);
private binanceService = inject<BinanceService>(TYPES.binanceService);
private coinMetaService = inject<CoinMetaService>(TYPES.coinMetaService);
Copy

Agent Swarm Integration

The report generation services integrate deeply with the AI agent swarm system to provide contextual, AI-generated analysis.

Agent Override Configuration

All services configure the ReportAgent with dynamic system prompts that include real-time market context:

overrideAgent({
  agentName: AgentName.ReportAgent,
  systemDynamic: async () => {
    const averagePrice = await this.binanceService.formatPrice(
      symbol,
      await this.binanceService.getClosePrice(symbol)
    );
    return str.newline([
      `Текущая дата/время: ${new Date().toISOString()}`,
      `Текущая цена ${this.coinMetaService.getCoinNameForSymbol(symbol)}: ${averagePrice}`,
      `Отчет создается для монеты ${this.coinMetaService.getCoinNameForSymbol(symbol)}`,
      "В отчете обязательно напиши дату индикаторов, на которые ты ссылаешься",
    ]);
  },
  mcp: [],
});
Copy

Caching and Performance Optimization

The system implements sophisticated caching mechanisms using TTL (time-to-live) strategies to optimize performance and reduce redundant API calls.

TTL Configuration

Service	TTL Constant	Cache Key Pattern
SwingRangeReportPublicService	CC_SWING_RANGE_REPORT_TTL	${symbol}
ShortRangeReportPublicService	CC_SHORT_RANGE_REPORT_TTL	${symbol}
LongRangeReportPublicService	CC_LONG_RANGE_REPORT_TTL	${symbol}
PriceDataReportPublicService	CC_PRICE_DATA_REPORT_TTL	${symbol}
VolumeDataReportPublicService	CC_VOLUME_DATA_REPORT_TTL	${symbol}
MastodonSearchReportPublicService	CC_MASTODON_REPORT_TTL	${symbol}

TTL Implementation Pattern

public getReport = ttl(
  async (symbol: string) => {
    // Report generation logic
  },
  {
    timeout: TTL_CONSTANT,
    key: ([symbol]) => `${symbol}`,
  }
);
Copy

Data Sources and Dependencies

The report generation services integrate with multiple data sources to provide comprehensive market analysis.

Primary Data Sources

Service Dependencies

Most technical analysis services depend on:

• BinanceService for real-time price and volume data
• CoinMetaService for cryptocurrency metadata and symbol resolution

The MastodonSearchReportPublicService is unique in that it:

• Does not depend on BinanceService
• Focuses solely on social sentiment analysis
• Uses simplified system prompts without price context

Client ID Generation and Swarm Context

Each report generation creates a unique execution context within the agent swarm using structured client ID patterns and consistent swarm naming.

Client ID Patterns

// Format: {report-type}_{symbol}_{randomString()}
clientId: `swing-range-report_${symbol}_${randomString()}`
clientId: `short-range-report_${symbol}_${randomString()}`
clientId: `long-range-report_${symbol}_${randomString()}`
clientId: `price-data-report_${symbol}_${randomString()}`
clientId: `volume-data-report_${symbol}_${randomString()}`
clientId: `mastadon-search-report_${symbol}_${randomString()}`
Copy

All services use SwarmName.ReportSwarm as their execution context, ensuring consistent agent behavior and resource management across different report types.