Immobiliare .it Analyser

This repo is an example of how you can make data accessible with Python and LLMs for FREE.

Let me exaplain

Immobiliare is a seamless site for finding houses for sell and rent.
It has a lot filter and settings, making the research the most specific as possible.

Yet it doesn't have some filters I was personally searching for, number of bedrooms, rent contract types and students-only flag.

How it works

• You define your basic filter on Immobiliare site
• You copy the url and paste it in settings.json file
• Browse the properly filtered results after running the tool a few minutes

So what does this tool do?

This tool acquire all search results, collect and analyze them in order to open the links properly filtered in your browser.

• It makes a request to Immobiliare search query (previously built on Immobiliare website).
• Foreach result it makes a request to it in order to get the house's description.
• Now instead of using regex or other so fancy yet shitty things,
• it calls Llama 3.1 405B via API to analyze the house details and answer some question instead of me.
• After waiting for all info to be collected and processed, it filters by sending all answers to Llama.

Finally, it opens all filtered results.

Setup

You need to setup a few things, but just once.

LLM

As I said, Llama 3.1 is used, but you don't need to hosts nor pays it.
I integrated Akash Chat Api for FREE, as it is actually free and offered from Akash Cloud.

1. Go to https://chatapi.akash.network/
2. Click Get Started
3. Fill the form
4. Click Generate Api Key
5. Copy it
6. Paste it in "api" field in settings.json file

Browser

1. Locate your Firefox or Chrome browser executable
2. If you don't know, just ask Akash Chat :)
3. Copy the path
4. Paste it in "firefox_path" or "chrome_path" field in settings.json file

Immobiliare link

1. Go to https://www.immobiliare.it/
2. Select Affitta / Rent
3. Insert every filter you need (such as districts, price range, ecc)
4. Copy the url from the browser
5. Paste it in "link" field in settings.json file

Python Package

1. Open a terminal
2. Run pip install -r -U requirements.txt

Run

1. Open a terminal inside the script's folder
2. Run python main.py