# Mountaineer - Topology-Driven Visual Analytics for Comparing Local Explanations
##### Accepted at IEEE Transactions on Visualization and Computer Graphics

<!---
[![Downloads](https://static.pepy.tech/personalized-badge/mountaineer?period=total&units=international_system&left_color=grey&right_color=blue&left_text=Downloads)](https://pepy.tech/project/mountaineer) [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](...) [![Build](https://github.com/pnxenopoulos/gale/actions/workflows/build.yml/badge.svg)](https://github.com/pnxenopoulos/gale/actions/workflows/build.yml) [![Documentation Status](https://readthedocs.org/projects/gale/badge/?version=latest)](https://gale.readthedocs.io/en/latest/?badge=latest) 
--->
[![MIT license](https://img.shields.io/badge/License-MIT-blue.svg)](https://github.com/PariSolunke/mountaineer/blob/main/LICENSE) [![Paper](https://img.shields.io/badge/read%20the-paper-blueviolet)](https://arxiv.org/pdf/2406.15613) 


Tested on Ubuntu 22.04.3 LTS

### Install
Prerequisites for the provided installation script:
* Linux based operating System
* Anaconda/Minoconda package manager
  
Clone this repository. Then navigate to the root of the cloned repository in a terminal and just run
```shell
bash install.sh
```
The required environment can now be activated by
```shell
conda activate mountaineer
```

### Recreating Case study 1

From the project root. First activate the installed environment and launch jupyter notebook.
```shell
conda activate mountaineer

jupyter notebook
```
The project root has notebooks with our experiments. You can use the loaded tool within 'Case_Study_1_Heloc_Explanations.ipynb' file directly to explore our experiments. The file includes everything needed to recreate Case Study 1. Simply run all the cells within the notebook. The final cell within the notebook will load the system with the generated explanation topologies.

The accompanying video 'CaseStudy1.mp4' will help guide the exploration to recreate figures from the paper for this case study.

### Usage
```python
conda activate mountaineer
import mountaineer as mnt

v = mnt.visualize(...)
```