How modlee-ai/uconn_modlee_ai_ml_club_hackathons works
Project type and purpose
• This is a starter template / sample code repository for the UCONN Modlee AI/ML Club Hackathon.
• Its purpose is to demonstrate how to build AI-agent-based applications by calling the Modlee Agents API.
• It provides example “client” scripts (in Python) that show how to invoke various agents:
• Core document agents (blog, social media)
• Data operator agents (REST connectors, SQL queries)
File structure
• Top-level:
• README.md – instructions for the hackathon and how to use the samples
• examples/ – all of the runnable client-side code samples
• modlee_agents_wiki.md – extended documentation link
• examples/core_docs_agent:
• blog_example.py – shows how to generate blog posts
• social_example.py – shows how to generate social posts
• examples/data_operator_agent:
• rest/ – subfolders for REST API connectors
– google_places_api, news_api, newsdata_api (each has docs, schema, Python examples, tests)
• sql/ – an SQLite mock database plus a Python example for SQL queries
• No “backend” code here – this repo is purely client-side integration with the Modlee Agents API.
README for additional insights
• Confirms that teams will use at least one Modlee Agent (SQL, blog generator, social media)
• Provides context on API keys, support sessions, submission guidelines
• No internal architecture details – so we infer based on examples that the key system under test is the Modlee Agents API itself, plus external data sources.
a. Main components to include:
1. Client Applications (Python scripts in examples/)
2. Modlee Agents API Platform
• API Gateway / Authentication Layer
• Agent Engine (orchestration of different agents)
• Agent Modules:
• Core Docs Agent
• Data Operator Agent – REST Connectors
• Data Operator Agent – SQL Connector
3. External Services
• Google Places REST API
• News API
• NewsData API
4. Local Data Store
• SQLite mock_database.db in examples/data_operator_agent/sql
b. Relationships and data flow:
• Client scripts → HTTP request → API Gateway
• API Gateway → routes to specific Agent Module
• Agent Module (e.g. REST connector) → calls External Service → returns data → Agent Engine processes response
• SQL connector → runs queries on SQLite DB → returns results
• Agent Engine → constructs final response → returns to Client
c. Architectural patterns and principles:
• Separation of Concerns: each agent is a self-contained module (docs vs data)
• Micro-style connectors: each external API or database is accessed via its own sub-agent
• API-driven design: all interactions are over HTTP via a centralized gateway
d. Key technologies / frameworks:
• Python client code (examples/*.py)
• Modlee Agents API (hosted service)
• External REST APIs (Google Places, News, NewsData)
• SQLite for local SQL examples