Skip to content

Fabio-RibeiroB/macro-crew

BranchesTags

Repository files navigation

UK Macro Crew

UK Macro Crew is a multi-agent AI system built with crewAI that automates collection and reporting of UK macroeconomic data. The system uses specialized agents to research and produce a validated current snapshot JSON report of key UK economic indicators.

Installation

Ensure you have Python >=3.10 <3.14 installed on your system. This project uses UV for dependency management and package handling.

First, if you haven't already, install uv:

What It Does

The system is designed for regular execution to produce the latest available UK macroeconomic snapshot including:

  • Interest Rates: Bank of England base rates
  • CPIH: Consumer Price Index including Housing (month-over-month changes)
  • GDP: Gross Domestic Product monthly changes
  • Monetary Policy Report: Bank of England policy summaries
  • Financial Stability Report: BoE financial stability summaries

Final Outputs: research_report.json and history_report.json

The system produces two JSON files:

  • Location: research_report.json (in project root)
  • Growth Pattern: File is overwritten on each successful run with the latest validated snapshot
  • Structure: Snapshot format with current indicator values and report summaries
  • Usage: This file contains the current macro view used by the dashboard
  • Location: history_report.json (in project root)
  • Growth Pattern: Appends/upserts by publication date on each successful run
  • Structure: Chronological arrays per indicator/report for trend analysis
  • Usage: This file is the historical backbone for charts and longitudinal analysis

Agents

  • Researcher Agent: Searches for and retrieves key UK economic indicators from authoritative sources
  • Reporting Analyst Agent: Formats collected data into a current JSON snapshot

Output

The system generates:

  • research_report.json: a structured snapshot file with the latest available economic indicators and report summaries.
  • history_report.json: a validated historical dataset built from each successful snapshot run.
pip install uv

Next, navigate to your project directory and install the dependencies:

(Optional) Lock the dependencies and install them by using the CLI command:

crewai install

Environment Setup

Add your API keys to the .env file:

OPENAI_API_KEY=your_openai_api_key_here
EXA_API_KEY=your_exa_api_key_here

# Error Handling & Timeout Configuration
AGENT_MAX_ITERATIONS=8          # Maximum iterations per agent (default: 8)
AGENT_MAX_RETRIES=2            # Maximum retries on failure (default: 2)
AGENT_MAX_EXECUTION_TIME=900   # Timeout per agent in seconds (default: 900 = 15 min)
AGENT_MAX_RPM=150              # Rate limit per agent (default: 150)
CREW_FAIL_FAST=true           # Stop on first error (default: true)
CREW_MAX_RPM=100              # Rate limit for crew (default: 100)

The system requires:

  • OpenAI API Key: For the LLM agents (GPT-4o-mini by default)
  • Exa API Key: For web search and data retrieval

Error Handling & Reliability

The system includes robust error handling to prevent infinite loops and timeouts:

Built-in Safeguards

  • Iteration Limits: Each agent stops after 8 iterations by default
  • Retry Logic: Failed operations retry only twice before giving up
  • Execution Timeouts: Individual agents timeout after 15 minutes
  • Fail-Fast Mode: System stops immediately on unrecoverable errors
  • Rate Limiting: Prevents API quota exhaustion

Configuration Options

# Restrictive settings (for testing or quick runs)
AGENT_MAX_ITERATIONS=2
AGENT_MAX_RETRIES=1
AGENT_MAX_EXECUTION_TIME=300   # 5 minutes

# Production settings (more resilient)
AGENT_MAX_ITERATIONS=8
AGENT_MAX_RETRIES=2
AGENT_MAX_EXECUTION_TIME=900   # 15 minutes

# Disable fail-fast for partial results
CREW_FAIL_FAST=false

Error Scenarios Handled

  • API Rate Limits: Automatic rate limiting prevents quota exhaustion
  • Search Failures: Agents provide partial results rather than complete failure
  • Data Parsing Errors: Invalid data is skipped, valid data is preserved
  • Network Timeouts: Individual agent timeouts prevent system hangs
  • Infinite Loops: Iteration limits prevent runaway processes

Running the Project

To start the UK macro data collection process:

# Production run (searches for latest available data)
crewai run

# Alternative entry points
uv run uk_macro_crew
uv run run_crew

The system is now simplified to always search for the latest available data for each economic indicator. This ensures you always get the most recent UK macroeconomic data without needing to specify timeframes.

This initializes the UK Macro Crew, where the Researcher Agent searches for the latest UK economic data and the Reporting Analyst Agent writes a fresh validated research_report.json snapshot.

Additional Commands

# Run tests
uv run pytest

# Run tests with coverage
uv run pytest --cov=uk_macro_crew

# Train the crew (for model improvement)
uv run train <n_iterations> <filename>

# Replay specific task execution
uv run replay <task_id>

# Seed/update history_report.json from current snapshot
uv run migrate_history

How It Works

The UK Macro Crew consists of two specialized AI agents working together:

Researcher Agent

  • Role: UK Macroeconomic Data Researcher
  • Goal: Find and retrieve the most current UK economic indicators
  • Tools: Web search capabilities via Exa API
  • Sources: Bank of England, ONS, and other authoritative UK economic data sources

Reporting Analyst Agent

  • Role: Economic Data Analyst and Report Maintainer
  • Goal: Build a current and validated economic data snapshot
  • Tools: JSON manipulation tool for structured data updates
  • Output: Snapshot JSON report with publication dates, next publication dates, and sources

Data Structure

The system maintains data in this JSON snapshot format:

{
  "metadata": {
    "generated_at": "2026-03-06T17:02:19.884768Z",
    "last_updated": "2026-03-06T17:02:19.884768Z"
  },
  "current_economic_indicators": {
    "interest_rate": {
      "value": "3.75%",
      "publication_date": "2026-02-05",
      "next_publication_date": "2026-03-19",
      "source": "https://www.bankofengland.co.uk/monetary-policy/the-interest-rate-bank-rate"
    },
    "cpih": {
      "value": "-0.3%",
      "publication_date": "2026-02-18",
      "next_publication_date": "2026-03-25",
      "source": "https://www.ons.gov.uk/economy/inflationandpriceindices/bulletins/consumerpriceinflation/january2026"
    },
    "gdp": {
      "value": "+0.1%",
      "publication_date": "2026-02-12",
      "next_publication_date": "2026-03-31",
      "source": "https://www.ons.gov.uk/economy/grossdomesticproductgdp"
    }
  },
  "current_report_summaries": {
    "monetary_policy_report": {
      "summary": "Brief summary of policy decisions...",
      "report_date": "2026-02-05",
      "next_publication_date": "2026-03-19",
      "source": "https://www.bankofengland.co.uk/monetary-policy-report/2026/february-2026"
    },
    "financial_stability_report": {
      "summary": "Brief summary of financial system assessment...",
      "report_date": "2025-12-15",
      "next_publication_date": "2026-06-15",
      "source": "https://www.bankofengland.co.uk/financial-stability-report"
    }
  }
}

Configuration

The system is configured through YAML files:

  • src/uk_macro_crew/config/agents.yaml: Agent definitions and capabilities
  • src/uk_macro_crew/config/tasks.yaml: Task definitions and expected outputs
  • knowledge/user_preference.txt: User preferences for data presentation

Support

For support, questions, or feedback regarding the UK Macro Crew:

Project Structure

uk-macro-tracker/          # Project root
├── .env                   # API keys (OPENAI_API_KEY, EXA_API_KEY)
├── .env.dist             # Environment template
├── research_report.json  # Generated current snapshot
├── history_report.json   # Generated historical time series
├── knowledge/            # Templates and configuration
├── src/uk_macro_crew/    # Main source code
│   ├── config/          # Agent and task YAML configurations
│   ├── tools/           # Custom JSON manipulation tools
│   ├── crew.py          # Agent orchestration
│   └── main.py          # CLI entry points
└── tests/               # Test suite

About

Multi-agent AI system for automated UK economic data collection and reporting. Built with CrewAI, tracks BoE rates, CPIH, GDP monthly.

Resources

Readme
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages