📄 README.md – Real Estate Assistant

Overview

The Real Estate Assistant is an intelligent command-line tool designed to analyze and summarize Mongolian real estate listings from Unegui.mn. It uses a multi-agent workflow powered by a powerful LLM backend (Meta-LLaMA 3) to generate human-readable reports from raw listings.

---

💡 Features

• Extracts data from individual property URLs
• Parses details such as title, price, area, bedrooms, and description
• Summarizes listings into structured reports
• Context-aware analysis using market averages
• Modular agent architecture (RetrieverAgent, WriterAgent)

---

🏁 How to Run

python main.py

Then enter a query such as:

https://www.unegui.mn/adv/9341198_bgd-4-khoroolold-17-mkv-azhlyn-bair/

---

📦 Dependencies

Ensure you're in a virtual environment and install:

pip install -r requirements.txt

---

🧠 Architecture

main.py
├── classifies query type
├── invokes workflow
│   ├── RetrieverAgent
│   │   ├── fetches and parses URL
│   │   ├── extracts property details
│   └── WriterAgent
│       ├── formats and prompts the LLM
│       └── prints final report

---

✅ Sample Output

URL: https://www.unegui.mn/adv/9341198...
Title: Бгд 4 хороололд 1 өрөө 17 мк байр
Price: MNT 65,000,000
Location: шууд нүүж ороход бэлэн...  ← ⚠️ Not a true location!
Area: 17.0 м²
Bedrooms: 1
...

---

🛠️ TODO

Fix: Incorrect "Location" Retrieval

The current location value is incorrectly pulled from a nearby description tag instead of the actual address. To fix this:

• Inspect the true HTML structure where location data like Баянгол дүүрэг, Хан-Уул, etc. is stored.

• Update the RetrieverAgent logic to extract from:

  • A <ul> block containing key-value pairs (like <li> where key is Байршил: or similar).
  • Or schema.org metadata (e.g. itemprop="address" if available).

• Add fallback logic:

  • If explicit address not found, look for district or khoroo names in breadcrumbs or tags.

• Write a helper function extract_detail_by_label(soup, label) to reuse for similar fields like area, floor, etc.

• Add a test case for listings that only include a vague or no location.

🚀 Getting Started

Follow these steps to set up and run the project locally.

1. Clone the Repository

git clone https://github.com/yourusername/langchain-app.git
cd langchain-app

2. Set Up a Virtual Environment

Python 3.10 or newer is required.

python3 -m venv .venv
source .venv/bin/activate  # On Windows use: .venv\Scripts\activate

3. Install Dependencies

If you're using uv:

uv pip install -r requirements.txt

Otherwise, install directly from pyproject.toml:

pip install .

Or install using editable mode for development:

pip install -e .

✅ Make sure pip is the one from your virtual environment.

4. Set Up Environment Variables

Create a .env file in the root directory and add your API keys:

OPENAI_API_KEY=your_openai_key
TOGETHER_API_KEY=your_together_api_key

5. Run the App

python main.py

You’ll be prompted to enter a real estate listing URL or a general query.