As of September 15, 2015, the system version of python is 2.6.6.   To access a newer version (provided by Anaconda 4.3.0), use the module system:

module load python

This loads many packages, including Ipython, Jupyter and YT. If additional packages are needed in the Anaconda environment, contact Andrius.


Remote data analysis is possible by using Jupyter. To configure a Jupyter server do (after loading the python module):

 jupyter notebook --no-browser 

The output will show the port number and the token of the server:


In this case, the port is 8890 and the token is 509109... Note that both values will be different in your case. By default, the server only accepts connections from "localhost", so an ssh tunnel is needed. In a new terminal (on your computer) do:

ssh -L 9999:localhost:8890 -N

Replace "username" by your username in the cluster. Then, go to a web browser and put http://localhost:9999. Paste the token provided by server and click on login.


The preferred way of working with Python, at least in relation to RAMSES, is to use the YT project.

Examples showing how to use YT are collected in the (GIT) directory ramses/utils/examples/yt/, while reusable procedures should be placed in ramses/utils/yt/ (or in ramses/utils/py/ if they are general Python procedures).

Our local YT expert is Michael, who has passed on this useful code fragment in order to illustrate all parameters that YT determines from a single output file.:

import yt
ds = yt.load("YOURPATH/output_XXXXX/info_XXXXX.txt")
for i in sorted(ds.derived_field_list):
  print i