Installation Guide

Installation Guide#

Using `dysh` at GBO#

You do not need to install dysh if you have access to a GBO data reduction host (e.g., fourier, thales, etc.). At GBO, you can launch the latest version of dysh from a terminal by typing

dysh

The first time you run this, it may take a minute. To verify you are getting the latest version, check the output of these commands:

which dysh

Should print: /opt/local/bin/dysh. And

dysh --version

Should print 0.11.0 or higher (e.g., 0.11.1).

If you want to install your own version of dysh instead, read the following section.

Installing `dysh`#

dysh requires a Python version between 3.10 and 3.14 (dysh has not been tested in Python versions newer than 3.14) and recent versions of astropy, numpy, scipy, pandas, specutils, matplotlib, and astroquery.

dysh has been tested on Ubuntu, MacOS and Windows. Using dysh requires at least twice the amount of memory than the size of the data to be reduced. For example, if you are going to reduce observations that generated 8 GB of data, you require 16 GB of RAM.

We strongly recommend the use of a virtual environment for installing dysh.

Linux and MacOS#

Outside of GBO

The details of how to install dysh on your own depend on what package manager you use (e.g., uv, conda, pip). We strongly recommend that you install dysh in a separate Python environment, so you do not change your system Python (if you use your own anaconda3, you can of course do anything you like, though we still recommend a virtual environment). It is important to note that the latest version of dysh will only work on Python versions between 3.10 and 3.14. Here we provide a few examples, which require the ability to install Python packages in your system.

Using conda

If you use conda, then you can use the following commands

Create a dedicated virtual environment for dysh, with Python 3.12 (you could use a Python version between 3.10 and 3.14)
```
conda create -n dysh-venv python==3.12
```
Activate the newly created virtual environment
```
conda activate dysh-venv
```

Install dysh into the environment

pip install "dysh[nb] @ git+https://github.com/GreenBankObservatory/dysh"

dysh should now be available through the dysh command. You can make sure you installed the latest version by running
```
dysh --version
```
It should print 0.11.0 or higher (e.g., 0.11.1).

Using venv

If you already have Python 3.10, up to 3.14, installed you can create a new virtual environment using venv.

Create a dedicated virtual environment for dysh, with Python 3.14:
```
python3.12 -m venv dysh-venv
```
Having python3.12 available as above may change depending on your own system. Make sure you are using a Python version between 3.10 and 3.14 with python --version.

Activate the newly created virtual environment

pip install "dysh[nb] @ git+https://github.com/GreenBankObservatory/dysh"

dysh should now be available through the dysh command. You can check what version you installed by running:
```
dysh --version
```
It should print 0.11.0 or higher (e.g., 0.11.1).

Using uv

If you use uv then the following command will install dysh:

uv tool install "dysh[nb] @ git+https://github.com/GreenBankObservatory/dysh"

dysh should now be available through the dysh command. You can check what version you installed by running:

dysh --version

It should print 0.11.0 or higher (e.g., 0.11.1).

Warning

Installing dysh using this method will install it as a tool, not a package, so it won’t be possible to import dysh from a Python session.

At GBO

At GBO, on a linux machine from the terminal (assumes you’re using bash).

First, create a new Python virtual environment. For this example we will use Python 3.11 (you can change the Python version, but the latest dysh will only work on Python between 3.10 and 3.14):
```
/users/gbosdd/python/bin/python3.11 -m venv /home/scratch/$USER/dysh-env
```

Activate the virtual environment:

source /home/scratch/$USER/dysh-env/bin/activate

Install dysh, and the additional dependencies to run Jupyter lab:
```
pip install "dysh[nb] @ git+https://github.com/GreenBankObservatory/dysh"
```
If you do not want the additional dependencies, remove the [nb] part.
Check what dysh version was installed:
```
dysh --version
```
It should show 0.11.0 or higher (e.g., 0.11.1), if using Python between 3.10 and 3.14 (to check the Python version use python --version).
Launch the command line interface
```
dysh
```
or the Jupyter lab interface (only available if you used the [nb] option, or if you installed Jupyter lab)
```
dysh-lab
```

Windows#

We provide instructions on how to install the latest version of dysh on Windows using anaconda and git. Anaconda provides a way of installing Python and managing virtual environments, and git is used to download, and install, the source code for the latest version of dysh. If you want to install the latest stable version of dysh, then it is not necessary to use git.

The first step is to download and install Anaconda and git. Anaconda comes in two flavors, as Anaconda distribution or as Miniconda. The Anaconda documentation provides details on the differences between both (link ), but the short story is that Anaconda has a graphical user interface (GUI), it comes with more packages pre-installed, but it uses more disk space. Miniconda does not have a GUI and it comes with less packages pre-installed. Both provide access to conda, which is the tool we will use to create a new virtual environment. You can find the downloads for Anaconda distribution and Miniconda here. git can be downloaded here. git is highly customizable, so it will ask a bunch of questions during the installation. For our purposes, the defaults should work. The most important part is to have git available through the command line (e.g., PowerShell).
After installing Anaconda or Miniconda, there should be an Anaconda menu in the start menu.

In the dropdown select either the Anaconda Prompt or the Anaconda PowerShell Prompt. This will launch a new terminal with the base virtual environment activated.
In the terminal enter the following command
```
conda create -n dysh-venv python==3.10
```
This will create a new virtual environment with Python3.10. You can use other versions of Python, but not lower than 3.10 nor above 3.14 (dysh has not been tested on Python versions newer than 3.14). You can check the Python version using the command python --version in the terminal.
After creating the environment, activate it. Enter the following command
```
conda activate dysh-venv
```
If everything went well you should see that the command line prompt now says (dysh-env) at the start of the current line.
Install dysh, with all the dependencies required to run a Jupyter lab server (not required). Enter the following command
```
pip install "dysh[nb] @ git+https://github.com/GreenBankObservatory/dysh"
```
This will download and install dysh with all its dependencies to the dysh-env virtual environment. Depending on your internet connection, this may take a while. If you do not want the extra dependencies, remove the [nb] part in the command.
If everything went well, the following commnad should print the dysh version
```
dysh --version
```
This should print 0.11.0 or above (e.g., 0.11.1).
Launch the command line interface
```
dysh
```
or the Jupyter lab interface (only available if you used the [nb] option, or if you installed Jupyter lab)
```
dysh-lab
```
The Jupyter lab interface will launch in your default browser, unless you have changed the Jupyter settings. You can avoid launching a browser with the --no-browser option, so the command would be
```
dysh-lab --no-browser
```
Then you can copy and paste the url in your preferred browser.

Quick Test#

After installing dysh you can run the following commands to make sure everyhting is working fine. This quick example will download a 31 MB file to the current directory.

Launch dysh
```
dysh
```

Copy and paste the following

filename = dysh_data(example="getfs") # This will download 31 MB, if outside of GBO.
sdfits = GBTFITSLoad(filename)
sdfits.summary()
scan_block = sdfits.getfs(scan=6, ifnum=0, plnum=0, fdnum=0)
spectrum = scan_block.timeaverage()
spectrum.plot()

This should look like

Terminal showing how to activate dysh, and launch the CLI interface. In the dysh CLI it download a file, loads it and prints a summary of its contents. There is a single scan of M33.

And the plot command should produce the following figure

A frequency switched spectrum calibrated and displayed using dysh.