# From Source

The installation consist two part: python-dependencies and external libraries.

## Core Software/Libraries

Before the installation, following software/libraries should be installed on the system.

- Python 3.8+
- GCC 7.8+, cmake 3.18+
- HDF5 (parallel build)
- MPICH
- OpenMPI (for cluster)

## Python Dependencies

The python-dependencies are managed through [`poetry`][link-poetry-website]. We provide important alias in `makefile`.

```sh
make poetry-downloads  # Install poetry on the system
make install           # Install MiV-Simulator
```

> **Note**: If any issues are encoutered, please make a report to [GitHub-Issue](https://github.com/GazzolaLab/MiV-Simulator/issues).

> The installation using `pip` is not yet supported due to the dependency compilation.

## External Libraries

1. Building and installing NEURON

```sh
pip install neuron
```

<details>
  <summary>How to install from the source</summary>
```sh
git clone https://github.com/neuronsimulator/nrn.git
cd nrn
mkdir build
cd build
cmake .. -DNRN_ENABLE_INTERVIEWS=OFF -DNRN_ENABLE_MPI=ON -DNRN_ENABLE_RX3D=ON -DNRN_ENABLE_CORENEURON=ON -DNRN_ENABLE_PYTHON=ON -DPYTHON_EXECUTABLE=$(which python3) -DCMAKE_C_COMPILER=mpicc -DCMAKE_CXX_COMPILER=mpicxx
make install
```
</details>

2. Building and installing NeuroH5

The NeuroH5 build system requires cmake.

```sh
git clone https://github.com/soltesz-lab/neuroh5.git
cd neuroh5
CMAKE_BUILD_PARALLEL_LEVEL=8 pip install .
```

## Cluster Usage

For architecture-level optimization, some of the core libraries should be built within the worker node.

:::{note}
Make sure to run on a worker node, not the login-node!
:::

### Necessary Modules

Following modules (or equivalent modules) are necessary before starting the installation.

```sh
module load python/3
module load cmake
module load gcc
module load openmpi
```

### Building HDF5 (Parallel)

Download the source from [here][source-hdf5].

```sh
wget https://support.hdfgroup.org/ftp/HDF5/releases/hdf5-1.12/hdf5-1.12.1/src/hdf5-1.12.1.tar.gz
tar -xzf hdf5-1.12.1.tar.gz
cd hdf5-1.12.1
CC=mpicc ./configure --prefix=$PWD/build --enable-parallel --enable-shared
make && make check
make install && make check-install
```

### H5PY

```sh
CC="mpicc" HDF5_MPI="ON" HDF5_DIR=/Users/skim0119/github/hdf5-1.12.1/build pip install --no-binary=h5py h5py
```

### Install/Build Python Dependencies

```sh
make poetry-download
make install
```

### Building NEURON simulator

```sh
git clone https://github.com/neuronsimulator/nrn.git
cd nrn
git submodule update --init --recursive
mkdir build
cd build
cmake .. -DNRN_ENABLE_INTERVIEWS=OFF -DNRN_ENABLE_MPI=ON -DNRN_ENABLE_RX3D=ON -DNRN_ENABLE_CORENEURON=ON -DPYTHON_EXECUTABLE=$(which python3) -DNRN_ENABLE_PYTHON=ON -DCMAKE_C_COMPILER=mpicc -DCMAKE_CXX_COMPILER=mpicxx -DCMAKE_INSTALL_PREFIX=../install
cmake --build . --parallel 8 --target install

export PATH=<Installation path>/bin:$PATH
export PYTHONPATH=<Installation path>/lib/python:$PYTHONPATH
```

<details>
  <summary>Possible Issues</summary>

- `Readline` cannot be found:
    - Try to install `Readline` using `apt` or `yum`. It can also be installed using `conda`.
    - `Readline` might already exist on the system. Search in `/usr/lib` or `/usr/lib64`.
    - Pass environment variable directly: `cmake -DReadline_INCLUDE_DIR=/usr/lib64 -DReadline_LIBRARY=/usr/lib64/libreadline.so.7 ....`

</details>
<br/>

### Install NeuroH5

```sh
# get the source
git clone https://github.com/soltesz-lab/neuroh5.git
cd neuroh5

# add HDF5 build to PATH
export HDF5_SOURCE=<HDF5 installation directory>
export PATH=$PATH:$HDF5_SOURCE/build

# (make sure the node has enough RAM and cores, otherwise the compilation will fail)
CMAKE_BUILD_PARALLEL_LEVEL=8 pip install .
```

Other _CLI tools_ in NeuroH5 can be installed using `cmake`.

```sh
cmake .
make -j4
export PATH=<NeuroH5 installation path>/bin:$PATH
```

## Tools and scripts to aid cluster usage

- [How to install and configure as module](https://github.com/GazzolaLab/MiV-Simulator/tree/latest/support/cluster)
- [Machinable: A modular system to manage research code](https://machinable.org)