Sciserver setup (recommended)

Create a HEASOFT image on Sciserver as explained on the HEASOFT Sciserver documentation page. We have tested the exercises with HEASOFT 6.31.1.

Creating a conda environment for SPEX on Sciserver

Once you have the Jupyterlab environment running on Sciserver, write the following commands in the terminal:

(heasoft) idies@aaaa:~$ conda activate base

(base) idies@aaaa:~$ conda create -n spex python=3.9 astropy matplotlib ipykernel

(base) idies@aaaa:~$ conda activate spex

(spex) idies@aaaa:~$ conda install -c spexxray spex pyspextools

(spex) idies@aaaa:~$ python -m ipykernel install --user --name spex --display-name "(spex)"

Next we have to add the following line "conda", "run", "--no-capture-output", "-n", "spex", to this file ~/.local/share/jupyter/kernels/spex/kernel.json to try to make sure that conda starts the spex conda environment: 

(spex) idies@aaaa:~$ vi ~/.local/share/jupyter/kernels/spex/kernel.json

In vi, get into edit mode by typing i. Now, you have a cursor to edit the text file. Add the line "conda", "run", "--no-capture-output", "-n", "spex", after the "argv": [ line.

In the end, the contents of ~/.local/share/jupyter/kernels/spex/kernel.json should look like this:

{

 "argv": [

  "conda", "run", "--no-capture-output", "-n", "spex",

  "/home/idies/miniconda3/envs/spex/bin/python",

  "-m",

  "ipykernel_launcher",

  "-f",

  "{connection_file}"

 ],

 "display_name": "(spex)",

 "language": "python",

 "metadata": {

  "debugger": true

 }

}

When the contents of the file look OK, press the <Esc> key on your keyboard, and then type :x followed by the <return> key. The file should be saved now.

It is necessary to restart the Jupyter session on Sciserver to load the new SPEX environment into Jupyter properly. To do that, execute the following steps:

• Close the Jupyter window and go back to the page on Sciserver with the list of your images.

• On this page, you can stop running the image by clicking the red square after the image name. Wait until the image is stopped.

• When the image is fully stopped and the page reloaded, you can click the green triangle to start the image again.

• Click on the name of the image to start Jupyter again. In this new session, SPEX should work in Jupyter.

Installation on your own Laptop: 

SPEX can also be installed on your own computer. However, many more things can go wrong and during the school there will be very little time to debug anything. However, if you are interested you can take a look at the SPEX software package website and try to install SPEX yourself. 

Troubleshooting

Importing SPEX in Jupyter fails

When you have Jupyter installed in multiple places, it sometimes happens that the wrong Jupyter executable is called. If you have trouble importing SPEX, then try to run Jupyterlab like this to force it to use Jupyter from the SPEX environment:

(spex) user@unix:~> $CONDA_PREFIX/bin/jupyter-lab &

For more information please check: https://summerschool-ahead2020.github.io/hands-on-sessions/index.html

Start SPEX 

You can start the Sciserver image and switch to the (spex) conda environment:

(heasoft) idies@aaaa:~$ conda activate spex

(spex) idies@aaaa:~$