Re: [Pflexible-users] backward.py

[John B. <jfb...@gm...>]

Francesco,

I need to update the tutorials to try and create something more useful 
and relevant to the new API. However, please use the source code. I've 
done what I can in general to document the different functions.

In generaly, my workflow is something like this:

H = pf.Header('/path/to/output')
H.fill_backwards()

Then the H object will have a 'C' attribute. The 'C' attribute is a 
dictionary keyed by tuples of the release time and species.

E.g.

data_00 = H.C[(0, 0)]
data_00.keys()
['itime',
  'min',
  'timestamp',
  'gridfile',
  'rel_i',
  'shape',
  'spec_i',
  'grid',
  'max',
  'species',
  'slabs']

The key things are the 'grid' which is an array of all the data for that 
release, and the 'slabs', which is a dictionary again.

data_00.slabs.keys()
[0, 1, 2, 3]

data_00.grid.shape
(360, 180, 3)

Note that each slab is a reference to the grid array for each vertical 
level.

e.g.
data_00.slabs[1] is the same as data_00.grid[:,:,0]

data_00.slabs[0] is the total column (so np.sum(data_00.grid, axis=2))

Once you have that information, you can start to play around with the 
plotting routines. In general, they expect the data_00 object.

so:

FIG = pf.plot_totalcolumn(H, data_00)

Should work...

--john




On Fri 06 Jul 2012 11:12:26 AM CEST, Francesco Graziosi wrote:
> Hi John,
>
> I followed your suggestion, thanks, so finaly (in flex_wrk ) the
> program can read the example data and the the grid (with the command
> read_grid).
> From this point the tutarial (0.8.1 version) said to lanch the
> backward.py, but I don't have in my directory of pflexin. What shoud I
> do for continue to elaborate the data ?
>
> Best regards,
>
> Francesco Graziosi