コード例 #1
0
def test_convert_compact_to_tiles(llc_mds_datadirs, mydir, fname, nk, nl,
                                  skip):
    """Read in compact data, convert to tiles.
    Verify this equals reading in face data
    """

    if mydir == 'xmitgcm':
        mydir, _ = llc_mds_datadirs

    data_compact = ecco.read_llc_to_compact(fdir=mydir,
                                            fname=fname,
                                            llc=90,
                                            nk=nk,
                                            nl=nl,
                                            filetype='>f4',
                                            skip=skip,
                                            less_output=False)
    data_tiles = ecco.read_llc_to_tiles(fdir=mydir,
                                        fname=fname,
                                        llc=90,
                                        nk=nk,
                                        nl=nl,
                                        filetype='>f4',
                                        skip=skip,
                                        less_output=False)

    data_converted = ecco.llc_compact_to_tiles(data_compact)
    assert np.all(np.equal(data_converted, data_tiles))
コード例 #2
0
ファイル: read_mds.py プロジェクト: timothyas/pych
def read_single_mds(fnamebase, xdalike, **kwargs):
    """Read meta/data that are not grid files but do not have time stamps.
    If coords, dims are provided, create xarray DataArray.
    Otherwise, return data as numpy ndarray

    Parameters
    ----------
    fnamebase : string
        the base filename <fname>.meta/data
    coords : dictionary, optional
        specify coordinates of the data in the file

    Returns
    -------
    data : numpy ndarray or xarray DataArray
    """

    # Read in as n_timesteps x n_vertical_levels x 1170 x 90 numpy.ndarray
    data = mut.rdmds(fnamebase, **kwargs)

    # Convert to the 13 tile format
    data = llc_compact_to_tiles(data, less_output=True)

    # Potentially convert to xarray.DataArray
    data = xr.DataArray(data, coords=xdalike.coords, dims=xdalike.dims)

    return data
コード例 #3
0
def test_convert_compact_to_tiles():
    """Read in compact data, convert to tiles. 
    Verify this equals reading in face data
    """

    # Loop through 2D, 3D, 4D cases
    for fname, nk, nl in zip(_TEST_FILES, _TEST_NK, _TEST_RECS):
        data_compact = ecco.read_llc_to_compact(fdir=_DATA_DIR,
                                                fname=fname,
                                                llc=90,
                                                nk=nk,
                                                nl=nl,
                                                filetype='>f4',
                                                less_output=False)
        data_tiles = ecco.read_llc_to_tiles(fdir=_DATA_DIR,
                                            fname=fname,
                                            llc=90,
                                            nk=nk,
                                            nl=nl,
                                            filetype='>f4',
                                            less_output=False)

        data_converted = ecco.llc_compact_to_tiles(data_compact)
        assert np.all(np.equal(data_converted, data_tiles))
コード例 #4
0
def interploate_one3d(infile, outfile, varnm, unitnm):
    # ## Load the Grid Data

    # In[2]:

    fdir = '/Volumes/data01/forum/user/owang/junk/nonblank/'
    fname = 'XC.data'
    # in compact format
    XC = ecco.load_llc_compact(fdir, fname)
    # in 13 tiles format
    XC_t = ecco.load_llc_compact_to_tiles(fdir, fname, less_output=True)

    fname = 'YC.data'
    YC = ecco.load_llc_compact(fdir, fname)
    YC_t = ecco.load_llc_compact_to_tiles(fdir, fname, less_output=True)

    fname = 'maskCtrlC.data'
    maskc = ecco.load_llc_compact(fdir, fname, nk=50)
    maskc_t = ecco.load_llc_compact_to_tiles(fdir,
                                             fname,
                                             less_output=True,
                                             nk=50)

    # This is a 1D array.
    RC = np.squeeze(-1 * ecco.load_binary_array(fdir, 'RC.data', 1, 50))

    # In[3]:

    fname = infile
    fdir = '/Volumes/data01/forum/user/owang/junk/h8i_i48/diags/STATE/'

    # as compact
    data = ecco.load_llc_compact(fdir, fname, nk=50)
    # as tiles
    data_t = ecco.llc_compact_to_tiles(data, less_output=True)

    # In[4]:

    #   print(data_t.shape)
    #   print(data.shape)

    # # Plot level 11 of the data

    # In[5]:

    ecco.plot_tiles(data_t[:, 10, :],
                    layout='latlon',
                    rotate_to_latlon=True,
                    cmin=-.2,
                    cmax=.2)

    # # Resample to nearest latlon point

    # In[6]:

    #help(ecco.resample_to_latlon)

    # In[7]:

    new_grid_delta_deg = 0.25

    new_grid_min_lat = -90 + new_grid_delta_deg
    new_grid_max_lat = 90 - new_grid_delta_deg
    new_grid_min_lon = -180
    new_grid_max_lon = 180 - new_grid_delta_deg

    radius_of_influence = 120000  #m

    # In[8]:

    RC.shape

    # In[9]:

    for num, name in np.ndenumerate(RC):
        print num, name
        print type(num)

    # In[10]:

    data_latlon_projection = []

    for num, name in np.ndenumerate(RC):
        print(num[0])

        # create mask
        masktmp = 1. - maskc[num, :]
        datamsked = ma.array(data[num, :], mask=masktmp)

        new_grid_lon, new_grid_lat, tmp = ecco.resample_to_latlon(
            XC,
            YC,
            datamsked,
            new_grid_min_lat,
            new_grid_max_lat,
            new_grid_delta_deg,
            new_grid_min_lon,
            new_grid_max_lon,
            new_grid_delta_deg,
            nprocs_user=1,
            mapping_method='nearest_neighbor',
            radius_of_influence=radius_of_influence)

        print(tmp.shape)
        tmp = np.expand_dims(tmp, 0)
        if num[0] == 0:
            data_latlon_projection = tmp
        else:
            data_latlon_projection = np.append(data_latlon_projection, tmp, 0)

        print data_latlon_projection.shape

    # In[11]:

    lons_1d = new_grid_lon[0, :]
    lats_1d = new_grid_lat[:, 0]

    # In[12]:

    nz, ny, nx = np.shape(data_latlon_projection)
    print(lons_1d[0])
    print(lons_1d[-1])

    print(nx, ny, nz)

    # In[13]:

    plt.imshow(data_latlon_projection[0, :],
               origin='lower',
               vmin=0,
               vmax=.2,
               cmap='jet')

    # In[14]:

    plt.imshow(data_latlon_projection[25, :],
               origin='lower',
               vmin=0,
               vmax=.2,
               cmap='jet')

    # In[16]:

    foo = xr.DataArray(data_latlon_projection,
                       coords=[RC, lats_1d, lons_1d],
                       dims=['depth', 'latitude', 'longitude'])
    foo.isel(depth=1).plot()

    # In[17]:

    foo.attrs['xmin'] = np.min(new_grid_lon)
    foo.attrs['ymax'] = np.max(new_grid_lat)
    foo.attrs['spacing'] = new_grid_delta_deg
    foo.attrs['no_data'] = -9999.0
    foo.attrs['proj4'] = '+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs'
    foo.attrs['Projection'] = 'Lat Lon EPSG:4326'
    foo.attrs['Insitution'] = 'JPL'
    foo.attrs['original_filename'] = fname
    foo.attrs['author'] = 'Ou Wang'
    foo.attrs['product_version'] = '0.1'
    foo.attrs['nx'] = nx
    foo.attrs['ny'] = ny
    foo.attrs['nz'] = nz
    foo.attrs['units'] = unitnm
    foo.attrs['Conventions'] = 'CF-1.6'
    foo.attrs[
        'Metadata_Conventions'] = 'CF-1.6, Unidata Dataset Discovery v1.0, GDS v2.0'
    foo.attrs['cdm_data_type'] = 'Grid'
    foo.attrs['grid_method'] = 'bin average neighbor'
    foo.attrs['radius_of_influence'] = radius_of_influence
    foo.attrs['depth_units'] = 'meters'
    foo.name = varnm

    # In[18]:

    #foo

    # In[19]:

    foo.to_netcdf('/Volumes/data01/forum/user/owang/junk/output/' + varnm +
                  '/' + outfile + '.nc')