# 💽 Installation

VeraGrid is a software made in the Python programming language.
Therefore, it needs a Python interpreter installed in your operative system.
Once [python](https://www.python.org/) is installed in the system, 
you only need to install the VeraGrid package and the others will be installed as dependencies.

```bash
pip install veragrid
```

## Standalone setup

If you don't know what is this Python thing, we offer a windows' installation:

[💻 Windows setup](https://eroots.tech/veragrid-form)

This will install VeraGrid as a normal windows program, and you don't need to worry
about any of the previous instructions. Still, if you need some guidance, the
following video might be of assistance: 

[📺 Setup tutorial (video)](https://youtu.be/SY66WgLGo54).

## Package installation

We recommend to install the latest version of [Python](www.python.org) and then,
install VeraGrid with the following terminal command:

```
pip install VeraGrid
```

You may need to use `pip3` if you are under Linux or MacOS, both of which
come with Python pre-installed already.

## Install into an environment

```bash
python3 -m venv vg5venv
source vg5venv/bin/activate
pip install VeraGrid
veragrid
```

## Run the graphical user interface

Once you install VeraGrid in your local Python distribution, you can run the
graphical user interface with the following terminal command:

```bash
veragrid
```

If this doesn't work, try:

```bash
python -c "from VeraGrid.ExecuteVeraGrid import runVeraGrid; runVeraGrid()"
```

You may save this command in a shortcut for easy future access.

## Install only the engine

Some of you may only need VeraGrid as a library for some other purpose
like batch calculations, AI training or simple scripting. Whatever it may be,
you can get the VeraGrid engine with the following terminal command:

```bash
pip install VeraGridEngine
```

This will install the `VeraGridEngine` package that is a dependency of `VeraGrid`.

Again, you may need to use `pip3` if you are under Linux or MacOS.


## Additional packages for local AI

To install llama-cpp on ubuntu linux:

```bash
sudo apt-get install nvidia-cuda-toolkit
python -m pip uninstall -y llama-cpp-python
set CMAKE_ARGS="-DGGML_CUDA=on -DGGML_CUDA_FORCE_CUBLAS=on -DLLAVA_BUILD=off -DCMAKE_CUDA_ARCHITECTURES=native"
set FORCE_CMAKE=1
python -m pip install --require-hashes --no-cache-dir \
  "llama-cpp-python==0.3.23" \
  --hash=sha256:85493cd887b543588941e8704640fef6a54c057443292e527559c30728375ffd
```


## Troubleshooting

This section includes known issues and their solutions.

### Clean crashes on ARM-based MacOS

You may find VeraGrid crashing without any explanation on MacOS. 
A reason we've found is that the standard Numpy package is compiled 
against OpenBlas and not Apple's native Accelerate framework for linear algebra.
To fix this, you'll need to compile numpy from source:

- uninstall your numpy: `[python folder]/python -m pip uninstall numpy`
- install from source: `[python folder]/python -m pip install -U --no-binary :all: numpy`

You may need to install `xcode` for this to work

Here `[python folder]/python` is the complete path to the python binary executable 
file that you're using to run VeraGrid.