Nov 2019 (#10)

* First update for November 2019 training course

* Update links and re-do adding a new task to a suite section

* Further updates

* Update fcm tutorial to add an new module file. #4

* Fix typos and update cf-python exercises for python3

* Update cfa section

* Remove reference to cfdump

practicals/source/conf.py

@@ -44,7 +44,7 @@
 
 # General information about the project.
 project = u'NCAS Unified Model Introduction'
-copyright = u'2017, NCAS-CMS'
+copyright = u'2017-2019, NCAS-CMS'
 
 # The version info for the project you're documenting, acts as replacement for
 # |version| and |release|, also used in various other places throughout the
@@ -61,7 +61,7 @@
 
 # There are two options for replacing |today|: either, you set today to some
 # non-false value, then it is used:
-today = '03 April 2019'
+today = '06 November 2019'
 # Else, today_fmt is used as the format for a strftime call.
 #today_fmt = '%B %d, %Y'
 

practicals/source/fcm-tutorial.rst

@@ -151,18 +151,20 @@ Now make some code changes! Use the following scenario to take you through the b
 
 **Adding a new file**
 
-* Still in the ``src/control/top_level`` directory, add a new file with a subroutine in, called ``um_shell_sub.F90``.  (An example file is available on PUMA: ``~um/um-training/um_shell_sub.F90``.  The routine ``umPrint`` should be used for writing out messages rather than standard FORTRAN ``WRITE`` statements.)
+* Still in the ``src/control/top_level`` directory, add a new FORTRAN module file (``um_shell_mod.F90``) containing a subroutine called ``um_shell_sub()``.  (An example file is available on PUMA: ``~um/um-training/um_shell_mod.F90``.  The routine ``umPrint`` should be used for writing out messages rather than standard FORTRAN ``WRITE`` statements.)
 * Run **fcm add** on the command line, to let the repository know you're adding a new file at the next commit. Make sure you are still in ``src/control/top_level`` and then type: ::
 
-    fcm add um_shell_sub.F90
+    fcm add um_shell_mod.F90
 
  at the command prompt.
 
-* Amend ``um_shell.F90`` to call this new subroutine: ::
+* Modify ``um_shell.F90`` to use this new module.  You'll see lots of ``USE`` statements near the top of the file.  Add the following to use our new one. ::
 
-    CALL um_shell_sub()
+    USE um_shell_mod
+
+* Then add a line to call the ``um_shell_sub`` subroutine (suggest around line 375): ::  
 
-* Ensure you put a comment line **! DEPENDS ON: um_shell_sub** above the CALL statement to ensure the dependency on your new file is registered.
+    CALL um_shell_sub()
 
 **Deleting a file**
 
@@ -351,7 +353,7 @@ Version Control of Suites
 
 Just like the model code, your UM suites are also under version control in a subversion repository, usually *roses-u* which is on the MOSRS.  Once you have a working copy of your suite under ``~/roses`` you can use FCM commands in the same way as for your source code branches; i.e. commit changes, diff changes, etc.
 
-* Look in the roses-u repository via MOSRS Trac (https://code.metoffice.gov.uk/trac/roses-u) and find the suite you created in the previous section. (Hint: Go to *"Browse Source"* then drill down to find you suite. e.g. u-ag263 would be under a/g/2/6/3).  When was the suite last modified?
+* Look in the roses-u repository via MOSRS Trac (https://code.metoffice.gov.uk/trac/roses-u) and find the suite you created in the previous section. (Hint: Go to *"Browse Source"* then drill down to find your suite. e.g. u-ag263 would be under a/g/2/6/3).  When was the suite last modified?
 
 * Go to your suite working directory and type **fcm status** to see the changes you have made since you copied the suite.
 

practicals/source/further-exercises-2.rst

@@ -64,7 +64,7 @@ Navigate to *um -> namelist -> Model Input and Output -> NetCDF Output Options*
 
 Expand the *NetCDF Output Streams* section. A single stream - **nc0** - already exists; select it to display its content. As a useful comaprison, expand the *Model Output Streams* section and with the middle mouse button select **pp0**. Observe that the only significant differences between **pp0** and **nc0** are the values of ``file_id`` and ``filename_base``.  Data compression options for **nc0** are revealed if ``l_compress`` is set to true. NetCDF deflation is a computationally expensive process best handled asynchronously to computation and as yet not fully implemented through the UM IO Server scheme (but under active development.) For many low- to medium-resolution models and, depending precisely on output profiles, high-resolution models also, use of UM-NetCDF without IO servers still provides significant benefits over fields-file output since using it avoids the need for subsequent file format conversion.
 
-Right-click on **nc0** and select *Clone this section*. Edit the settings of the newly cloned section to make the new stream similar to **pp1**. It is sensible to change the name of the new stream from "1" to something more meaningful, nc1 for example (right click on "1", select *Rename a section*, and change ...nc(1) to ...nc (nc1)).
+Right-click on **nc0** and select *Clone this section*. Edit the settings of the newly cloned section appropriately to make the new stream similar to **pp1** (ie. edit ``filename_base`` and all the reinitialisation variables). It is sensible to change the name of the new stream from "1" to something more meaningful, nc1 for example (right click on "1", select *Rename a section*, and change ...nc(1) to ...nc (nc1)).
 
 **iii. Direct output to the nc streams**
 

practicals/source/getting-setup.rst

@@ -10,12 +10,12 @@ Unfortunately we can no longer pre-install MobaXterm for you due to licencing re
 * | From Chrome, go to page: https://mobaxterm.mobatek.net/download.html 
   | Google “mobaxterm download” and this should be the first entry (check the URL). 
 * Under “Home Edition” select “Download now”
-* | On next page select **“MobaXterm Home Edition v11.1 (Portable edition)”**. 
+* | On next page select **“MobaXterm Home Edition v12.2 (Portable edition)”**. 
   | This should download the package. 
 * Click the download icon in the bottom left hand corner. 
-* | Double-click on the **MobaXterm_Personal_11.1** application file, and select “Extract all”. 
+* | Double-click on the **MobaXterm_Personal_12.2** application file, and select “Extract all”. 
   | A new directory window will open up. 
-* Double-click **MobaXterm_Personal_11.1** to launch the application.
+* Double-click **MobaXterm_Personal_12.2** to launch the application.
 
 Next time, navigate to “Downloads” to open the application.
 

practicals/source/index.rst

@@ -6,7 +6,7 @@
 NCAS Unified Model Introduction: Practical sessions
 ===================================================
 
-Practical exercises for the UM training course in Reading, 3-5 April 2019. 
+Practical exercises for the UM training course in Reading, 6-8 November 2019. 
 
 NCAS Computational Modelling Services: http://cms.ncas.ac.uk/
 

practicals/source/post-processing.rst

@@ -94,8 +94,8 @@ Run ``um-convpp`` on a fieldsfile (E.g ba799a.pc19880901_00) ::
 
 Note the reduction in file size. Now use xconv to examine the contents of the PP file.
 
-cfa and cfdump
---------------
+cfa
+---
 
 There is an increasing use of python in the community and we have, and
 continue to develop, python tools to do much of the data processing
@@ -105,19 +105,18 @@ features - we'll use it to convert UM fields file or PP data to
 CF-compliant data in NetCDF format. You first need to set the
 environment to run ``cfa``: ::
 
- esPP001$ module load anaconda/2.2.0-python2 cf udunits
- esPP001$ module swap PrgEnv-cray PrgEnv-intel
+ esPP001$ export PATH=/home/n02/n02/ajh/anaconda3/bin:$PATH
  esPP001$ cfa -i -o ba799a.pc19880901_00.nc ba799a.pc19880901_00.pp
  
 Try viewing the NetCDF file with xconv.
 
 
-``cfdump`` is a tool to view CF fields. It can be run on PP or NetCDF
+``cfa`` can also view CF fields. It can be run on PP or NetCDF
 files, to provide a text representation of the CF fields contained in
 the input files. Try it on a PP file and its NetCDF equivalent,
 e.g. ::
 
-  archer$ cfdump ba799a.pc19880901_00.pp | less
+  archer$ cfa -vm ba799a.pc19880901_00.pp | less
   Field: long_name:HEAVYSIDE FN ON P LEV/UV GRID (ncvar%UM_m01s30i301_vn1100)
   ---------------------------------------------------------------------------
   Data           : long_name:HEAVYSIDE FN ON P LEV/UV GRID(time(5), air_pressure(17), latitude(145), longitude(192)) 
@@ -143,21 +142,21 @@ Many tools exist for analysing data from NWP and climate models and there are ma
 
 * Set up the environment and start python. ::
 
-   archer$ module load anaconda/2.2.0-python2 cf udunits
-   archer$ module swap PrgEnv-cray PrgEnv-intel
+   
+   archer$ export PATH=/home/n02/n02/ajh/anaconda3/bin:$PATH
    archer$ python
    >>> import cf
 
 We'll be looking at CRU observed precipitation data
 
 * Read in data files ::
 
-  >>> f = cf.read_field('~charles/UM_Training/cru/*.nc')
+  >>> f = cf.read('~charles/UM_Training/cru/*.nc')[0]
 
 * Inspect the file contents with different amounts of detail ::
 
   >>> f
-  >>> print f
+  >>> print(f)
   >>> f.dump()
   
 Note that the three files in the cru directory are aggregated into one
@@ -166,24 +165,24 @@ field.
 * Average the field with respect to time ::
 
   >>> f = f.collapse('T: mean')
-  >>> print f
+  >>> print(f)
 
 Note that the time coordinate is now of length 1.
 
 * Read in another field produced by a GCM, this has a different latitude/longitude grid to regrid the CRU data to ::
 
-  >>> g = cf.read_field('~charles/UM_Training/N96_DJF_precip_means.nc')
-  >>> print g
+  >>> g = cf.read('~charles/UM_Training/N96_DJF_precip_means.nc')[0]
+  >>> print(g)
 
 * Regrid the field of observed data (f) to the grid of the model field (g) ::
 
   >>> f = f.regrids(g, method='bilinear')
-  >>> print f
+  >>> print(f)
 
 * Subspace the regridded field, f, to a European region ::
 
   >>> f = f.subspace(X=cf.wi(-10, 40), Y=cf.wi(35, 70))
-  >>> print f
+  >>> print(f)
 
 Note that the latitude and longitude coordinates are now shorter in length.