Skip to content

Latest commit

 

History

History
252 lines (142 loc) · 12 KB

Install.md

File metadata and controls

252 lines (142 loc) · 12 KB

Python setup

You need to use Python for several of the practicals. You've verious options for this.

  1. Use the UCL JupyterHub
  2. Use Python on your own computer
  3. Use binder

We recommend you use the first of these, UCL JupyterHub as we can more easily work out any coding problems you have on that platform.

However, it is not very difficult to up your own Python installation and to use that, should you so wish.

If you want to just run through the practicals without saving material, then you can just make straight use of the binder facility (online read-only but runnable versions of the practicals).

Installation of these notes on UCL JupyterHub

This is the recommended option. You need internet access with reasonable reliability to use this, and you need to do a one-time setup of the course notes. After that, it should be easy to use.

There are three main ways you can access the UCL JupyterHub:

i. whilst on campus ii. through the UCL VPN iii. through UCL Desktop

The first is recommended and requires no other step. Follow the instructions below to log in to UCL JupyterHub and open a shell. After that, we will set up the notes.

Login using UCL VPN

First, set up and test the UCL VPN. This will take you some time to do if you haven't done it before, as it needs current virus definitions and has other requirements. Only proceecd if you have set up and tested the VPN. If you have problems with the VPN, contact ISD.

  1. Make sure you have the UCL VPN installed and running
  2. Log on the the UCL JupyterHub with your UCL credentials.
  3. Open a shell (New->Terminal) on JupyterHub and follow the installation instructions below.

Login using UCL Desktop

N.B. You do not need the VPN set up for this method.

  1. In a browser, log on to UCL DesktopAnywhere. Note, if you haven't done this before, it will take some time as it needs to download some software (or you can use the 'light' version).
  2. Open Desktop@UCL Anywhere and log in with your UCL credentialsi (this will take a few minutres to achieve).
  3. Open a browser in Desktop@UCL and log on the the UCL JupyterHub with your UCL credentials.
  4. You may need to click on 'Launch Server' to start
  5. You may need to 'choose your interface' before you can start (if so, select 'classic'), then click start
  6. Open a shell (New->Terminal) on JupyterHub and follow the installation instructions below.

Initial setup on UCL JupyterHub

The first time you want to use these coursenotes on UCL JupyterHub, you need to do some setup f the materials and codes. You should have a terminal (shell) open on UCL JupyterHub at this point. You should see a still cursor at a prompt, probably on a blqack background shell. This is a unix (linux) shell where you can type unix commands. To execute a command, you hit the key. Try that first, with the command:

whoami

It should show you username (e.g. ucfalew)

Next check to see where you are on the system:

pwd

It will show something like /nfs/.../home3/Uucfa6/ucfalew and should end with your username. This is you home directory.

If you have got to that point The first time you are using these notes, you should go through the items below. These are needed to set up the notes and various course settings.

  1. In the Terminal (shell), type:

     wget https://github.com/UCL-EO/geog0133/archive/refs/heads/main.zip

    This will download this repository as a zip file. Extract using:

     cd ~ && unzip main.zip

  2. Set up anaconda. In the Terminal (shell), type:

     conda init
 conda config --prepend envs_dirs /shared/groups/jrole001/geog0027/envs
 conda env list

    This should now show:

     # conda environments:
 #
 base                     /opt/miniconda-jhub/4.8.3
 jhubcode                 /opt/miniconda-jhub/4.8.3/envs/jhubcode
 geog0113              *  /shared/groups/jrole001/geog0027/envs/geog0113

    If that isn't the case, try opening a new shell again (bash), and/or stop and restart the notebook server (see 5. below)

  3. Now, set up notebook extensions by running the following in shell (Terminal):

     cd ~/geog0133-main/docs && ./postBuild

    That runs the scrip postBuild in the directory ~/geog0133-main/docs.

  4. Finally, make sure the kernel you need for the notebooks exists:

     python -m ipykernel install --name=conda-env-geog0113  --display-name 'conda env:geog0113' --user

  5. This should all be good to go now, but you should make sure that the new settings have taken place by stopping are restarting the notebook server. To do this:

    • click on the Control Panel button at the top right of the notebook page.
    • then click the big red button to stop the server
    • next, click start my server to restart (you may have to also then click launch my server)
    • if you are asked to choose an interface, choose classic, then click start

You should now see the link geog0113-main in the JupyterHub browser. Select that link to enter the course materials, then go to docs/notebooks to see the practical notebooks.

Select the first of these, 005_Solar_Practical.ipynb and run the cell In [1]:

import numpy as np
import matplotlib.pyplot as plt
from geog0133.solar import solar_model,radiation
import scipy.ndimage.filters
from geog0133.cru import getCRU,splurge
from datetime import datetime, timedelta
# import codes we need into the notebook

If that works fine (i.e. doesn't come up with an error) then you have set things up correctly.

If the problem you get is that the notebook kernel cannot be found, try the following:

bash

to open a new shell, then:

source activate geog0113
python -m ipykernel install --name=geog0113  --display-name 'conda env:geog0113' --user

and test the page again.

If not, try logging out (of all shells and notebooks) again, log in again and re-check.

If you get ModuleNotFoundError: No module named 'ephem', then you have not got the geog0133 running for some reason. Log out and in again. Then try the solution above for if the notebook kernel cannot be found. If that test doesn't run, the practical won't run, so make sure to fix that first. Once you have it properly configured, it should work all other times you log in.

If other issues, contact the [course convenor](mailto: [email protected]), sending them exactly what you typed and the response you get when typing these commands. Alternately wait for the practical classes for assistance.

Another way you can access this directory, should you need, is to use ssh to log in from a terminal to the UCL system (ssh [email protected]).

Summary for running on UCL JupyterHub

  1. Make sure you have the UCL VPN installed and running OR that you are running from Desktop@UCL.
  2. Log on the the UCL JupyterHub.
  3. Navigate to the directory geog0133/docs/notebooks
  4. Access the notebooks you want directly

Use Python on your own computer

You can run these practicals locally on your own computer. The only problemn is that it is more copmnplicated for us to support any comnputing iussues you might have, which is why we suggest you use the UCL system.

However, it it is not so hard to set things up and to run it for yourself.

Install Python

Whilst you can probably use any version of Python, we suggest you install Anaconda Python. If you are very short on computer space, you might prefer the cut-down version Miniconda.

make it faster if you want

The command that installs packages (libraries of software) for Anaconda Python is called conda. You will see various calls to this in the instructions below. There is a faster alternative to this, called mamba. If you want to use this, after you have set up Python, type:

conda install mamba -n base -c conda-forge

Then, everywhere below in the text where it says conda, you can use mamba and it will work a lot faster.

Clone the repository

The code repository is on github, so we need to pull a local copy.

You can do this in several ways:

  • Using PyCharm and Anaconda Navigator (if you installed Anaconda)
  • Using GitHub Desktop
  • Using git in a shell

Using PyCharm and Anaconda Navigator

We do not support this here, but you will find various online notes if you want to follow thbis path.

Getting the code:

Download from github. In a temrminal, in a sensible location run:

wget https://github.com/UCL-EO/geog0133/archive/refs/heads/main.zip

This will download this repository as a zip file. Extract using:

unzip main.zip

This will now have set up the directory geog0133-main. If you type:

ls geog0133-main

You should see:

Install.md		bin			requirements.txt
LICENSE			docs
README.md		readthedocs.yml
Setting up the environment

Assuming you have downloaded the repository though, now open a shell (Terminal) and type:

cd ~plewis/Documents/GitHub/geog0133/docs

replacing ~plewis/Documents/GitHub/geog0133 with the location of your repository.

Then, set up the environment with:

conda env create  --force -n geog0113 -f environment.yml

This will take a few minutes, but will create the environment geog0113 which contains all of the libaries you need for this course.

Now, activate it:

conda activate geog0113

Next, we need to set up the correct kernel for the notes:

python -m ipykernel install --name=conda-env-geog0113  --display-name 'conda env:geog0113' --user

Next run the post-build configuration script (sets up itens for Jupyter notebooks):

./postBuild

Now start Jupyter:

jupyter notebook

This may open a browser window for you, or might just instruct you to copy and paste a URL, e.g.:

http://127.0.0.1:8888/?token=4afdc076ec49592ca1059d957f0bccbce86e17ab838f61e0

Make sure you have the Jupyter window running in the browser.

Navigate to docs/notebooks and start the notebook 005_Solar_Practical.ipynb (click on it in the browser).

Now, run the cell In [1]: to test that the required codes load correctly.

If there is a problem, go back over the steps above. If you still can't solve the problems, try connecting by a different route, and/or and contact the [course convenor](mailto: [email protected]?subject=[geog0133 setup problem]), explaining exactly what you did and what the problem was.

Use Binder

If you only want to run the practicals, and not save the results, you can use the links to the Binder service to run the notebooks. Links are given to the exercises, and to answers.

Binder

Answers

Binder

Answers