NearNode – Smart Travel Planning

NearNode is a Django-based travel planning application that addresses the "last mile" problem and supports multi-modal trip planning with natural-language search.

Features

1. Nearest Alternate Optimization

Finds cheaper flights to nearby airports and calculates total trip cost (flight + ground transport). Users enter their final address; the app evaluates airports within a configurable radius and prioritizes total journey time and cost.

2. Multi-Modal Connections

Goes beyond flight-to-flight connections by suggesting train links (e.g. fly into Brussels, train to Amsterdam, fly out) and scoring layovers by lounge access, sleeping pods, and city access.

3. AI Search

Natural-language trip search: users describe what they want (e.g. "warm beach under €300, max 5 hours from Milan"). The app parses the query and returns top matches using your flight database or Amadeus when configured.

4. Collaborative Planning

Couples can sync via a shared code, vote on options, and see matches that fit both preferences.

5. Delay Prediction & Self-Transfer Insurance

Route-level delay likelihood and self-transfer risk for unbundled connections.

Installation

Prerequisites: Python 3.8+, pip

1. Clone the repo and enter the project directory.
2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate   # Windows: venv\Scripts\activate

3. Install dependencies: pip install -r requirements.txt
4. Copy environment: cp .env.example .env and add your keys.
5. Run migrations: python manage.py migrate
6. (Optional) Create superuser: python manage.py createsuperuser
7. (Optional) Load airports: python manage.py load_world_airports
8. Start the server: python manage.py runserver

Access: Web UI at http://localhost:8000 · Admin at http://localhost:8000/admin · API at http://localhost:8000/api

API Overview

Feature	Method	Endpoint
Nearest alternate	POST	/api/nearest-alternate/
Multi-modal search	POST	/api/multi-modal/
AI Search	POST	/api/ai-search/
Collaborative	POST/GET	/api/collaborative/*
Delay prediction	GET	/api/delay-prediction/?flight_id=...

Request bodies use JSON (e.g. origin_airport_code, final_destination_address, date for nearest alternate; query for AI Search). Authentication required for protected endpoints.

Configuration

• AI Search (LLM): Set in .env: AI_SEARCH_LLM_BACKEND=ollama with optional OLLAMA_BASE_URL and AI_SEARCH_OLLAMA_MODEL. Without a backend, keyword-based parsing is used.
• Real flights: Set AMADEUS_API_KEY and AMADEUS_API_SECRET for live flight search; use load_world_airports so origins/destinations resolve.
• Weather: Optional WEATHER_API_KEY for destination weather in results.
• Database: SQLite by default; use PostgreSQL in production (see settings.py).

Project Structure

• nearnode/ – Django project settings
• core/ – Models, views, profile
• api/ – REST API (views, serializers, services)
• templates/ – HTML templates
• manage.py, requirements.txt

Tech Stack

Django, Django REST Framework, Geopy, Tailwind CSS. Optional: OpenAI-compatible LLM (Ollama/Groq/OpenAI), Amadeus API, Celery/Redis.

License

See LICENSE.