def gene_bathy(xc, yc, xr, yr, n=500, amp=30.0):
noise = N.random.random(n)
a = N.random.random(n) * N.pi * 2
r = N.random.random(n)
x = xc + xr * r * N.cos(a)
y = yc + yr * r * N.sin(a)
return N.asarray([x, y, N.exp(-(x - xc) ** 2 / xr ** 2 - (y - yc) ** 2 / yr ** 2) * amp + noise])
def create_mv2_gridder_xyzt(nx=8,
ny=7,
nz=6,
nt=5,
xmin=-6.,
xmax=-3,
ymin=46,
ymax=48,
zmin=-200,
zmax=0,
tmin='2016',
tmax='2016-02',
tunits='days since 2016-01-01',
rotate=0):
"""Create a MV2 array on a grid
Return
------
MV2.array
"""
# Axes
shape = ()
axes = []
if nt != 0:
time = create_time(lindates(tmin, tmax, nt), tunits)
axes.append(time)
shape += nt,
if nz != 0:
dep = create_dep(N.linspace(zmin, zmax, nz))
axes.append(dep)
shape += nz,
if ny != 0:
lat = create_lat(N.linspace(ymin, ymax, ny))
axes.append(lat)
shape += ny,
if nx != 0:
lon = create_lon(N.linspace(xmin, xmax, nx))
axes.append(lon)
shape += nx,
# Array
data = MV2.array(N.arange(N.multiply.reduce(shape)).reshape(shape),
copy=False,
axes=axes,
id='temp',
dtype='d')
# Rotate grid
if rotate:
grid = data.getGrid()
if grid is not None:
grid = rotate_grid(grid, rotate)
set_grid(data, grid)
return data
def gene_bathy(xc, yc, xr, yr, n=500, amp=30.):
noise = N.random.random(n)
a = N.random.random(n) * N.pi * 2
r = N.random.random(n)
x = xc + xr * r * N.cos(a)
y = yc + yr * r * N.sin(a)
return N.asarray([
x, y,
N.exp(-(x - xc)**2 / xr**2 - (y - yc)**2 / yr**2) * amp + noise
])
def plot(xx,
yy,
target,
label,
figfiles,
figfile,
lon=None,
lat=None,
show=False):
xs, ys, mask = coord2slice(target, lon=lon, lat=lat)
P.figure(figsize=(6, 3.5))
P.title('Target=%(label)s / select: lon=%(lon)s, lat=%(lat)s' % locals())
add_grid((xx, yy))
xx = xx.asma()
yy = yy.asma()
if isinstance(lon, tuple):
P.axvline(lon[0], color='m', ls='--', lw=2)
P.axvline(lon[1], color='m', ls='--', lw=2)
elif isinstance(lon, slice):
i, j, k = lon.indices(xx.shape[1])
P.plot(xx[:, i], yy[:, i], 'c--', lw=2)
P.plot(xx[:, j - 1], yy[:, j - 1], 'c--', lw=2)
if isinstance(lat, tuple):
P.axhline(lat[0], color='m', ls='--', lw=2)
P.axhline(lat[1], color='m', ls='--', lw=2)
elif isinstance(lat, slice):
i, j, k = lat.indices(yy.shape[0])
P.plot(xx[i], yy[i], 'c--', lw=2)
P.plot(xx[j - 1], yy[j - 1], 'c--', lw=2)
P.xticks(N.arange(xx.min() - 1, xx.max() + 1))
P.yticks(N.arange(yy.min() - 1, yy.max() + 1))
xxi, yyi = xx, yy
xx = xx[ys, xs]
yy = yy[ys, xs]
# mask = mask[ys, xs]
xxb, yyb = meshbounds(xx, yy)
P.pcolormesh(xxb, yyb, mask, shading='faceted')
P.scatter(xx.ravel(), yy.ravel(), c=(0, 1, 0))
P.grid(True)
P.axis('image')
P.tight_layout()
i = len(figfiles)
savefig = figfile % i
if os.path.exists(savefig): os.remove(savefig)
P.savefig(savefig)
figfiles.append(savefig)
if show: P.show()
else: P.close()
def sync_norms(self, force=True):
"""Synchronise norms between variables whose id has the same prefix
Parameters
----------
force: bool
Force sync even if it seems at has already been synced.
"""
if not force and self._norm_synced:
return False
# Renorm for each prefix
for prefix in self.get_prefixes():
# Indices
indices = self.get_prefixed(prefix)
# Unified norm
psize = sum([self[i].psize for i in indices])
norm = N.sqrt(
sum([(self[i].psize * self[i].norm**2)
for i in indices]) / psize)
# Set it
for i in indices:
self[i].set_norm(norm)
self._norm_synced = True
return True
def create_mv2_scattered_xyzt(np=10,
nz=6,
nt=5,
xmin=-6.,
xmax=-3,
ymin=46,
ymax=48,
zmin=-200,
zmax=0,
tmin='2016',
tmax='2016-02',
tunits='days since 2016-01-01'):
"""Create a VM2 array of scattered data
Return
------
array: longitudes
array: latitude
MV2.array: data
"""
# Axes
shape = ()
axes = []
if nt != 0:
time = create_time(lindates(tmin, tmax, nt), tunits)
shape += nt,
axes.append(time)
if nz != 0:
dep = create_dep(N.linspace(zmin, zmax, nz))
axes.append(dep)
shape += nz,
shape += np,
axes.append(create_axis((np, )))
# Array
data = MV2.array(N.arange(N.multiply.reduce(shape)).reshape(shape),
copy=False,
axes=axes,
id='temp',
dtype='d')
# Positiions
lons = N.linspace(xmin, xmax, np)
lats = N.linspace(ymin, ymax, np)
return lons, lats, data
def plot(xx, yy, target, label, figfiles, figfile, lon=None, lat=None, show=False):
xs, ys, mask = coord2slice(target, lon=lon, lat=lat)
P.figure(figsize=(6, 3.5))
P.title('Target=%(label)s / select: lon=%(lon)s, lat=%(lat)s'%locals())
add_grid((xx, yy))
xx = xx.asma()
yy = yy.asma()
if isinstance(lon, tuple):
P.axvline(lon[0], color='m', ls='--', lw=2)
P.axvline(lon[1], color='m', ls='--', lw=2)
elif isinstance(lon, slice):
i, j, k = lon.indices(xx.shape[1])
P.plot(xx[:, i], yy[:, i], 'c--', lw=2)
P.plot(xx[:, j-1], yy[:, j-1], 'c--', lw=2)
if isinstance(lat, tuple):
P.axhline(lat[0], color='m', ls='--', lw=2)
P.axhline(lat[1], color='m', ls='--', lw=2)
elif isinstance(lat, slice):
i, j, k = lat.indices(yy.shape[0])
P.plot(xx[i], yy[i], 'c--', lw=2)
P.plot(xx[j-1], yy[j-1], 'c--', lw=2)
P.xticks(N.arange(xx.min()-1, xx.max()+1))
P.yticks(N.arange(yy.min()-1, yy.max()+1))
xxi, yyi = xx, yy
xx = xx[ys, xs]
yy = yy[ys, xs]
# mask = mask[ys, xs]
xxb, yyb = meshbounds(xx, yy)
P.pcolor(xxb, yyb, mask, shading='faceted')
P.scatter(xx.ravel(), yy.ravel(), c=(0, 1, 0))
P.grid('on')
P.axis('image')
P.tight_layout()
i = len(figfiles)
savefig = figfile%i
if os.path.exists(savefig): os.remove(savefig)
P.savefig(savefig)
figfiles.append(savefig)
if show: P.show()
else: P.close()
def assert_compatible_with_obs(self,
obsmanager,
sync_norms=True,
sync_missing_values=True):
"""Assert that the :class:`Ensemble` current instance is compatible
with a :class:`~sonat.obs.ObsManager` instance
It checks that observed variable are provided by the ensemble.
It optinonally synchonise norms between model and observations.
"""
# Check varnames
for varname in obsmanager.suffixed_varnames:
if varname not in self.varnames:
raise SONATError(('Observed variable name "{}" not found '
' in ensemble').format(varname))
# Sync norms
if sync_norms:
self.sync_norms(force=False)
dnorms = self.get_named_norms()
obsmanager.set_named_norms(dnorms)
# Missing values
if sync_missing_values and self.missing_value != obsmanager.missing_value:
obsmanager.missing_value = self.missing_value
# Check projection
oens = self.project_on_obs(obsmanager)
poens = obsmanager.restack(oens)
mask = N.isclose(poens, obsmanager.missing_value)
if mask.ndim > 1:
mask = mask.any(axis=1)
valid = ~mask
if not valid.any():
raise SONATError("Ensemble does project at all onto observations")
if not valid.all():
self.warning("Ensemble does not fully projects onto observations")
return dict(ens_on_obs=oens,
packed_ens_on_obs=poens,
packed_ens_on_obs_valid=valid)
"""Test the fortran function :f:func:`nearest2d`"""
from vcmq import N, P, code_file_name, P, os, rotate_grid, add_grid, meshbounds
from vacumm.misc.grid._interp_ import nearest2d
# Input grid
gridi = rotate_grid((N.arange(5), N.arange(4)), 30)
xxi = gridi.getLongitude()[:].filled()
yyi = gridi.getLatitude()[:].filled()
vari = N.resize(yyi, (20, )+ yyi.shape)
nb = 10
xxbi, yybi = meshbounds(xxi, yyi)
# Output grid
grido = rotate_grid((N.linspace(0, 6, 50)-1, N.linspace(0, 4, 35)+1.), -20)
xxo = grido.getLongitude()[:].filled()
yyo = grido.getLatitude()[:].filled()
xxbo, yybo = meshbounds(xxo, yyo)
# Nearest
varo = nearest2d(vari, xxi, yyi, xxo, yyo, nb)
# Plot
vmin = varo.min()
vmax = varo.max()
P.figure(figsize=(8, 4))
P.subplot(121, aspect=1)
P.pcolor(xxbi, yybi, vari[0], vmin=vmin, vmax=vmax)
add_grid(grido)
P.title('original')
P.subplot(122, aspect=1)
var1.getAxis(0).id = 'time' var1.getAxis(1).id = 'lat' var1.getAxis(1).designateLatitude() var1.getAxis(2).id = 'lon' var1.getAxis(2).designateLongitude() var1.units = 'm' var1.id = 'ssh' var2 = var1.clone() var2[:] = N.random.random((nt, ny, nx)) var1[3:13, :1, :1] = MV2.masked var2[5:15, -1:, -1:] = MV2.masked var2.long_name = 'Sea level' var2.id = 'sla' mask = var1.mask|var2.mask # common mask vmax = var2.max() bins = N.linspace(-0.1*vmax, 0.9*vmax, 14) nbins = len(bins) restart_file5 = code_file_name(ext='5.nc') restart_file7 = code_file_name(ext='7.nc') print restart_file5 # Normal sab = StatAccum(tall=True, sall=True, bins=bins) sab += var1[:5], var2[:5] #print sab.get_tmean() # Dump sab.dump(restart_file5) sab += var1[5:7], var2[5:7] sab.dump(restart_file7)
"""Test function :func:`~vacumm.misc.grid.kriging.krig` for grid refinement"""
nxi = 15
nyi = 10
r = 3
from vcmq import P, savefigs, code_file_name, N, auto_scale, add_grid
from vacumm.misc.grid.kriging import gridded_gauss3, random_gauss3, random_points, krig
# Generate random gridded field
xi, yi, zzi = gridded_gauss3(nx=nxi, ny=nyi)
xxi, yyi = N.meshgrid(xi, yi)
# Refined grid
xo = N.linspace(xi[0], xi[-1], (nxi-1)*r+1)
yo = N.linspace(yi[0], yi[-1], (nyi-1)*r+1)
xxo, yyo = N.meshgrid(xo, yo)
# Interpolate
zzo = krig(xxi.ravel(), yyi.ravel(), zzi.ravel(), xxo.ravel(), yyo.ravel())
zzo.shape = xxo.shape
# Section
P.figure(figsize=(8, 4))
iyis = [3, 4]
for iyi in iyis:
label = iyi==iyis[0]
P.plot(xi, zzi[iyi], 'ob-', markersize=8, label='Original' if label else None)
P.plot(xo, zzo[iyi*r], 'or-', markersize=5, lw=.8, label='Interpolated' if label else None)
P.legend(loc='best', framealpha=0.5)
P.grid()
nx = ny = 50
np = 500
mtype = 'gauss'
distmax = 2.5
from vcmq import N, P, code_file_name, savefigs
from vacumm.misc.grid.kriging import gridded_gauss3, random_gauss3, variogram, variogram_fit
# Generate random field
xxg, yyg, zzg = gridded_gauss3(nx=nx, ny=ny)
x, y, z = random_gauss3(np=np)
# Variogram from data
d, v = variogram(x, y, z, distmax=distmax)
# Variogram fit with null nugget
vm = variogram_fit(x, y, z, mtype, n=0, distmax=distmax)
D = N.linspace(0, d.max())
V = vm(D)
# Compare
P.figure(figsize=(6, 4))
P.title('Variogram')
P.plot(d, v, 'ob', label='From data')
P.plot(D, V, '-r', label='Fitted model (%s)' % mtype)
P.legend(loc='best')
P.ylim(ymin=0)
savefigs(code_file_name(ext=False), pdf=True, verbose=False)
P.close()
"""Test the fortran function :f:func:`interp1dxx`""" from vcmq import N, P,meshcells, minmax, code_file_name, os from vacumm.misc.grid._interp_ import interp1dxx nx = nyi = 10 mv = 1.e20 u, v = N.mgrid[-3:3:nx*1j, -3:3:10j]-2 vari = N.ma.asarray(u**2+v**2) vari.set_fill_value(mv) yi = N.linspace(-1000.,0., nyi) yo = N.linspace(-1200, 100, 30.) vari[nx/3:2*nx/3, nyi/3:2*nyi/3] = N.ma.masked x = N.arange(nx) dyi = (yi[1]-yi[0])*0.49 dyo = (yo[1]-yo[0])*0.49 yyi = N.resize(yi, vari.shape)+N.random.uniform(-dyi, dyi, vari.shape) yyo = N.resize(yo, (nx, len(yo)))+N.random.uniform(-dyo, dyo, (nx, len(yo))) yyib, xxib = meshcells(yyi, x) yyob, xxob = meshcells(yyo, x) varon = N.ma.masked_values(interp1dxx(vari.filled(), yyi, yyo, mv, 0, extrap=0), mv) varol = N.ma.masked_values(interp1dxx(vari.filled(), yyi, yyo, mv, 1, extrap=0), mv) varoh = N.ma.masked_values(interp1dxx(vari.filled(), yyi, yyo, mv, 3, extrap=0), mv) kw = dict(vmin=vari.min(), vmax=vari.max()) axlims = [x[0], x[-1], yo[0], yo[-1]] P.figure(figsize=(8, 8)) P.subplot(221) P.pcolor(xxib, yyib, vari) P.axis(axlims)
"""Test :func:`~vacumm.misc.plot.hov2` with a TZ variable"""
# Imports
from vcmq import N, MV2, create_dep, create_time, hov2, os, rc, code_file_name
# Init data with z 1D
nt = 10
nz = 8
var = N.dot(N.hanning(nt).reshape(nt, 1), N.hanning(nz).reshape(1, nz))
var = MV2.array(var)
time = create_time((0., nt), units="days since 2000")
z1d = create_dep((-nz+1, 1.))
var.setAxis(0, time)
var.setAxis(1, z1d)
z2d = N.resize(z1d, var.shape)
z2d *= N.resize((N.arange(1., nt+1)/nt).reshape(1, nt), (nz, nt)).T
# Plot with z 1D
rc('font', size=8)
kw = dict(show=False, bgcolor='0.5', date_fmt="%a")
hov2(var, subplot=211, **kw)
# Plot with z 2D
figfile = code_file_name(ext='png')
if os.path.exists(figfile): os.remove(figfile)
hov2(var, xaxis=z2d, subplot=212, twin='x', savefig=figfile, close=True, **kw)
# Unittest
result = dict(files=figfile)
from vcmq import create_axis, isaxis, coord2slice, cdms2, N
def subaxis2slice(cdaxis, values):
if cdms2.isVariable(cdaxis):
return [
subaxis2index(cdax, vv)
for cdax, vv in zip(cdaxis.getAxisList(), values)
]
cdaxis = create_axis(cdaxis)
ijk = cdaxis.mapIntervalExt((values[0], values[-1], 'cc'))
if ijk is None:
return
return slice(*ijk)
cdaxis = create_axis(N.linspace(0, 11., 17))
subaxis = cdaxis[2:7]
res = subaxis2slice(cdaxis, subaxis)
"""Test function :func:`~vacumm.misc.grid.kriging.krig` for grid refinement"""
nxi = 15
nyi = 10
r = 3
from vcmq import P, savefigs, code_file_name, N, auto_scale, add_grid
from vacumm.misc.grid.kriging import gridded_gauss3, random_gauss3, random_points, krig
# Generate random gridded field
xi, yi, zzi = gridded_gauss3(nx=nxi, ny=nyi)
xxi, yyi = N.meshgrid(xi, yi)
# Refined grid
xo = N.linspace(xi[0], xi[-1], (nxi - 1) * r + 1)
yo = N.linspace(yi[0], yi[-1], (nyi - 1) * r + 1)
xxo, yyo = N.meshgrid(xo, yo)
# Interpolate
zzo = krig(xxi.ravel(), yyi.ravel(), zzi.ravel(), xxo.ravel(), yyo.ravel())
zzo.shape = xxo.shape
# Section
P.figure(figsize=(8, 4))
iyis = [3, 4]
for iyi in iyis:
label = iyi == iyis[0]
P.plot(xi,
zzi[iyi],
'ob-',
markersize=8,
"""Test fortran function :f:func:`dstwgt2dto1dc`""" from vcmq import P, N, code_file_name, os, add_grid, rotate_grid from vacumm.misc.grid._interp_ import dstwgt2dto1dc # Input grid and data nxy = 15 xi = N.arange(nxy*1.) yi = N.arange(nxy*1.) gridi = rotate_grid((xi, yi), 30) xxi = gridi.getLongitude().getValue() yyi = gridi.getLatitude().getValue() zzi = N.ma.array(yyi) zzi[int(nxy*0.3):int(nxy*0.8), int(nxy*0.3):int(nxy*0.8)] = N.ma.masked zzi.shape = 1, nxy, nxy # Output positions no = 1000 xo = N.random.uniform(-nxy/4., nxy+nxy/4., no) yo = N.random.uniform(-nxy/4., nxy+nxy/4., no) # Interpolate mv = zzi.get_fill_value() zo = dstwgt2dto1dc(xxi,yyi,zzi.filled(mv),xo,yo,mv) zo = N.ma.masked_values(zo, mv) # Plot kw = dict(vmin=zzi.min(), vmax=zzi.max()) P.figure(figsize=(6, 6)) P.subplot(111, aspect=1)
"""Test the traditionnal CDAT regrid2 regridder""" from vcmq import MV2, create_grid, meshbounds, P, add_grid, N, bounds1d, plot2d, savefigs,code_file_name from regrid2 import Horizontal # Input nx, ny = 6, 4 vari = MV2.array(N.arange(nx*ny*1.).reshape(ny, nx), fill_value=1e20) xi = vari.getAxis(-1) xi[:] *= 2 yi = vari.getAxis(-2) yi[:] *= 3 xi.designateLongitude() yi.designateLatitude() xi.setBounds(bounds1d(xi)) yi.setBounds(bounds1d(yi)) vari[1:2, 2:4] = MV2.masked gridi = vari.getGrid() # Output grido = create_grid(xi[:]+2*2.5, yi[:]+3*1.5) xo = grido.getLongitude() yo = grido.getLatitude() xo.setBounds(bounds1d(xo)) yo.setBounds(bounds1d(yo)) xxob, yyob = meshbounds(xo, yo) # Regridding varo, wo = vari.regrid(grido, tool='regrid2', returnTuple=1)
"""Test the fortran function :f:func:`cellerr1d`, :f:func:`cellerr1dx` et :f:func:`cellerr1dxx`"""
from vcmq import cdms2, data_sample, N, P, meshcells, minmax, code_file_name, os
from vacumm.misc.grid._interp_ import cellerr1d, cellerr1dx, cellerr1dxx
# Academic example pure 1d
# - data
yi = N.array([-2., 2., 6. , 12., 15.])
vari = N.ma.array(yi)
vari[-1] = N.ma.masked
errm = N.ma.array([1, 2, 1., 2., 3.])
yo = N.array([5., 15])
mv = 1e20
vari = vari.reshape(1, -1).filled(mv)
errm = errm.reshape(1, -1).filled(mv)
errl = N.ones(1)
# - interp
varo, erro = cellerr1d(vari, yi, yo, mv, errm, errl)
# - truth
errot = N.array([
1/N.sqrt(
1/(errm[0, 1]**2+N.abs(yo[0]-yi[1])*errl[0]) +
1/(errm[0, 2]**2+N.abs(yo[0]-yi[2])*errl[0])),
N.sqrt(errm[0, 3]**2+N.abs(yo[1]-yi[3])*errl[0])
])
varot = N.array([
vari[0, 1]/(errm[0, 1]**2+N.abs(yo[0]-yi[1])*errl[0]) +
vari[0, 2]/(errm[0, 2]**2+N.abs(yo[0]-yi[2])*errl[0]),
vari[0, 3]/(errm[0, 3]**2+N.abs(yo[1]-yi[3])*errl[0])
])*errot**2
# - check
result = [(N.testing.assert_allclose, [erro.ravel(), errot]),
"""Test the fortran function :f:func:`curv2rect`"""
from vcmq import N, P, code_file_name, P, os
from vacumm.misc.grid._interp_ import curv2rect
# Input
x1, y1 = 0., 0.
x2, y2 = 3., 1.
x3, y3 = 2., 4.
x4, y4 = -1., 2.
# Format and convert
xx, yy = N.meshgrid(N.arange(-2, 4, 0.25), N.arange(-1, 5, 0.25))
nxy = xx.shape
xx.shape = -1
yy.shape = -1
pp, qq = [], []
for x, y in zip(xx, yy):
p, q = curv2rect(x1,x2,x3,x4,y1,y2,y3,y4,x,y)
pp.append(p)
qq.append(q)
pp = N.array(pp)
qq = N.array(qq)
# Plot
xp = [x1, x2, x3, x4, x1]
yp = [y1, y2, y3, y4, y1]
P.subplot(211)
levels = N.array([-10, 0, 1, 10.])
o = P.contourf(xx.reshape(nxy), yy.reshape(nxy), pp.reshape(nxy), levels=levels)
P.colorbar(o)
P.plot(xp, yp, 'k')
"""Test :meth:`vacumm.data.misc.arakawa.CGrid.interp`"""
from vcmq import MV2, N, create_grid, create_dep, set_grid, map2, \
code_file_name, CGrid, minmax, curve2, add_grid
# Initial variable
grid = create_grid(N.arange(-7, 0.), N.arange(43, 50.))
dep = create_dep([-5000, -3000, -2000, -1000, -500, -300, -200, -100.])
var = {}
var['t'] = MV2.reshape(N.arange(grid.size()*len(dep))*1., (len(dep), )+grid.shape)
set_grid(var['t'], grid)
var['t'].setAxis(0, dep)
# Arakawa manager
ag = CGrid()
# Interpolations
for p in 'u', 'v', 'f', 'w':
var[p] = ag.interp(var['t'], 't', p, mode='extrap')
# Surface plots
vmin, vmax = minmax(*[var[p][-1] for p in ['u', 'v', 'f']])
kw = dict(show=False, res=None, vmin=vmin, vmax=vmax, colorbar=False, grid=False, cmap='jet')
m = map2(var['t'][-1], fill='pcolor',
title='Interpolations on an Arakawa C grid: T->U/V/F', **kw)
add_grid(var['t'], linestyle='-')
kw.update(fill='scatter', contour=False, fill_s=60)
markers = dict(u='>', v='^', f='D', t='o')
for p in 't', 'u', 'v', 'f':
m = map2(var[p][-1], fill_marker=markers[p], shadow=True, zorder=100, **kw)
"""Test the fortran function :f:func:`curv2rel_single`"""
from vcmq import N, P, code_file_name, P, os, meshbounds
from vacumm.misc.grid._interp_ import curv2rel_single
# Input grid
x0, y0 = 0., 2.
nxi = 3
nyi = 3
dxi = (2., 2.)
dyi = (-2., 2.)
xxi = N.zeros((nyi, nxi))
xxi[0] = x0 + N.arange(nxi) * dxi[0]
for j in range(1, nyi):
xxi[j] = xxi[j-1] + dyi[0]
yyi = N.zeros((nyi, nxi))
yyi[:, 0] = y0 + N.arange(nyi) * dyi[1]
for j in range(1, nyi):
yyi[:, j] = yyi[:, j-1] + dxi[1]
xxbi, yybi = meshbounds(xxi, yyi)
relpos2index = lambda fi, fj, nyi: fj * nyi + fi
ii, jj = N.meshgrid(N.arange(nxi)+.5, N.arange(nyi)+.5)
iib, jjb = meshbounds(ii, jj)
zzi = relpos2index(ii, jj, nyi)
zzbi = relpos2index(iib, jjb, nyi)
# Input random points
N.random.seed(0)
np = 100
xxo = N.random.random(np)*(xxbi.max()-xxbi.min()) + xxbi.min()
yyo = N.random.random(np)*(yybi.max()-yybi.min()) + yybi.min()
"""Create fake observation netcdf files
"""
from vcmq import (cdms2, MV2, os, create_dep, create_lon, create_lat, N,
masked_polygon, map2)
cdms2.setAutoBounds(False)
# Profiles
ncfile = '../data/obs.profiles.nc'
lons = MV2.array([-5.8, -5.7, -4.6, -2.8], id='lon')
lats = MV2.array([48.1, 47.5, 47.4, 47.3], id='lat')
daxis = create_dep((-100., 1, 5.))
nz = len(daxis)
np = len(lons)
temp_error = N.resize([.2, .5, .3, .8], (nz, np))
sal_error = N.resize([.3, .1, .15, .4], (nz, np))
temp_error = MV2.array(temp_error, id='temp_error')
sal_error = MV2.array(sal_error, id='sal_error')
temp_error[:nz / 2, 2] = MV2.masked
sal_error[:nz / 2, 2] = MV2.masked
temp_error[:3 * nz / 4, 3] = MV2.masked
sal_error[:3 * nz / 4, 3] = MV2.masked
mob = MV2.array([0, 1, 1, 0], id='mobility', fill_value=-1)
paxis = lons.getAxis(0)
paxis.id = 'station'
axes = [daxis, paxis]
f = cdms2.open(ncfile, 'w')
for var in lons, lats, mob, temp_error, sal_error:
var.setAxis(-1, paxis)
if var.ndim == 2:
from vcmq import SimpleCloudKriger, N, P, variogram_model, code_file_name
# Kriging params
sill = 9.
range = 40
farvalue = 20
# Input
xi = [20., 26., 50., 70]
yi = [20, 26., 70., 20]
zi = [15., 6., 2., 4.]
# Output positions
nx = ny = 101
xg = N.linspace(1, 100, 101)
yg = N.linspace(1, 100, 101)
# Some inits
xi = N.array(xi)
yi = N.array(yi)
zi = N.array(zi)
xxg, yyg = N.meshgrid(xg, yg)
xo = xxg.ravel()
yo = yyg.ravel()
vgm = variogram_model('linear', n=0, s=sill, r=range)
# Setup the kriger
sck = SimpleCloudKriger(xi, yi, zi, vgf=vgm, farvalue=farvalue)
# Interpolate
"""Test function :func:`~vacumm.misc.plot.taylor` in different configurations"""
from vcmq import taylor, rc, code_file_name, MV2, N
# Make some data
nt = 50
nens = 50
# - references
ref1 = MV2.sin(MV2.arange(nt, dtype='f')) * 10.
ref2 = ref1 * 1.2
# - ensemble 1
dat1 = MV2.resize(ref1, (nens, nt)) + N.resize(N.random.rand(nt * nens),
(nens, nt)) * 20.
dat1.long_name = 'Set 1'
dat1.units = 'meters'
# - modele 2
dat2 = MV2.resize(ref2, (nens, nt)) + N.resize(N.random.rand(nt * nens),
(nens, nt)) * 40.
dat2.long_name = 'Set 2'
# Plot
rc('font', size=8)
kwplot = dict(figsize=(5, 5),
savefigs_pdf=False,
show=False,
close=True,
savefigs_verbose=False)
basename = code_file_name(ext=False) + '_%i'
# - single variable
taylor(dat1[0],
ref1,
savefigs=basename % 0,
var1.getAxis(0).id = 'time' var1.getAxis(1).id = 'lat' var1.getAxis(1).designateLatitude() var1.getAxis(2).id = 'lon' var1.getAxis(2).designateLongitude() var1.units = 'm' var1.id = 'ssh' var2 = var1.clone() var2[:] = N.random.random((nt, ny, nx)) var1[3:13, :1, :1] = MV2.masked var2[5:15, -1:, -1:] = MV2.masked var2.long_name = 'Sea level' var2.id = 'sla' mask = var1.mask | var2.mask # common mask vmax = var2.max() bins = N.linspace(-0.1 * vmax, 0.9 * vmax, 14) nbins = len(bins) restart_file5 = code_file_name(ext='5.nc') restart_file7 = code_file_name(ext='7.nc') print restart_file5 # Normal sab = StatAccum(tall=True, sall=True, bins=bins) sab += var1[:5], var2[:5] #print sab.get_tmean() # Dump sab.dump(restart_file5) sab += var1[5:7], var2[5:7] sab.dump(restart_file7)
"""Test the fortran function :f:func:`cellerr1d`, :f:func:`cellerr1dx` et :f:func:`cellerr1dxx`"""
from vcmq import cdms2, data_sample, N, P, meshcells, minmax, code_file_name, os
from vacumm.misc.grid._interp_ import cellerr1d, cellerr1dx, cellerr1dxx
# Academic example pure 1d
# - data
yi = N.array([-2., 2., 6., 12., 15.])
vari = N.ma.array(yi)
vari[-1] = N.ma.masked
errm = N.ma.array([1, 2, 1., 2., 3.])
yo = N.array([5., 15])
mv = 1e20
vari = vari.reshape(1, -1).filled(mv)
errm = errm.reshape(1, -1).filled(mv)
errl = N.ones(1)
# - interp
varo, erro = cellerr1d(vari, yi, yo, mv, errm, errl)
# - truth
errot = N.array([
1 / N.sqrt(1 / (errm[0, 1]**2 + N.abs(yo[0] - yi[1]) * errl[0]) + 1 /
(errm[0, 2]**2 + N.abs(yo[0] - yi[2]) * errl[0])),
N.sqrt(errm[0, 3]**2 + N.abs(yo[1] - yi[3]) * errl[0])
])
varot = N.array([
vari[0, 1] /
(errm[0, 1]**2 + N.abs(yo[0] - yi[1]) * errl[0]) + vari[0, 2] /
(errm[0, 2]**2 + N.abs(yo[0] - yi[2]) * errl[0]), vari[0, 3] /
(errm[0, 3]**2 + N.abs(yo[1] - yi[3]) * errl[0])
]) * errot**2
# - check
result = [(N.testing.assert_allclose, [erro.ravel(), errot]),
"""Compare CDAT regridding speed with rectangular and rectangular grids"""
config = {
'esmf':['linear', 'patch', 'conserv'],
'libcf':['linear'],
}
# Imports
from vcmq import MV2, create_grid2d, code_file_name, os, CDATRegridder, N, set_grid, psinfo
from vacumm.misc.grid import rotate_grid
from time import time
# Input
nx = ny = 300
vari = MV2.array(N.arange(nx*ny*1.).reshape(ny, nx))
gridi = create_grid2d(vari.getAxis(1)[:]*50/nx, vari.getAxis(0)[:]*50/nx)
set_grid(vari, gridi)
# Output grid
gridor = create_grid2d(vari.getAxis(1)[:]*0.09*50/nx,
vari.getAxis(0)[:]*0.09*50/nx)
gridoc = rotate_grid(gridi, 30)
# Log
logfile = code_file_name(ext='log')
if os.path.exists(logfile): os.remove(logfile)
f = open(logfile, 'w')
print >>f, 'NY=%(ny)i, NX=%(nx)i'%locals()
# Loop on methods
for tool, methods in config.items():
"""Test the fortran function :f:func:`bilin`"""
from vcmq import N, P, meshcells, minmax, code_file_name, os
from vacumm.misc.grid._interp_ import dstwgt
nxi = 15
nyi = 10
mv = 1.e20
u, v = N.mgrid[-3:3:nyi*1j, -3:3:nxi*1j]-2
vari = N.ma.asarray(u**2+v**2)
vari.set_fill_value(mv)
xi = N.arange(nxi)
yi = N.arange(nyi)
vari[int(nyi*0.4):int(nyi*0.4)+3, int(nxi*0.4):int(nxi*0.4)+2] = N.ma.masked
xxib, yyib = meshcells(xi, yi)
nxo = 40
nyo = 25
xo = N.linspace(int(nxi*0.2),int(nxi*1.2),nxo)
yo = N.linspace(int(-nyi*0.2),int(nyi*0.8),nyo)
xxob, yyob = meshcells(xo, yo)
vari.shape = (1, )+vari.shape
varo = N.ma.masked_values(dstwgt(vari.filled(), xi, yi, xo, yo, mv, 0), mv)
kw = dict(vmin=vari.min(), vmax=vari.max())
axlims = [min(xi.min(), xo.min()), max(xi.max(), xo.max()),
min(yi.min(), yo.min()), max(yi.max(), yo.max())]
P.figure(figsize=(8, 4))
P.subplot(211)
P.pcolor(xxib, yyib, vari[0], **kw)
"""Test :func:`~vacumm.misc.plot.section2` in quiver mode"""
# Imports
from vcmq import create_lon, N, MV2, create_dep, os, code_file_name, section2
from vacumm.misc.phys.units import deg2m
# Init data with z 1D
nz = 11
nx = 11
x = create_lon(N.arange(nx))
xm = deg2m(x[:],lat=45.) # meters
dx = xm[:].ptp()
z = create_dep((-nz+1, 1.), units='m', long_name='Depth')
dz = z[:].ptp()
scale = dz/dx
u = MV2.ones((nz,nx)) # 1 m/s
w = u*scale # 1 m/s * scale
for var in u,w:
var.setAxis(0, z)
var.setAxis(1, x)
var.units = 'm/s'
# Plot
figfile = code_file_name(ext='png')
if os.path.exists(figfile): os.remove(figfile)
s = section2((u,w), quiver_norm=1, fill=False, show=False,
savefig=figfile, close=False, axes_aspect=1)
# Result
result = dict(files=figfile)
nx = ny = 50
np = 500
mtype = 'gauss'
dmax = 2.5
from vcmq import N, P, code_file_name, savefigs
from vacumm.misc.grid.kriging import gridded_gauss3, random_gauss3, variogram, variogram_fit
# Generate random field
xxg, yyg, zzg = gridded_gauss3(nx=nx, ny=ny)
x, y, z = random_gauss3(np=np)
# Variogram from data
d, v = variogram(x, y, z, dmax=dmax)
# Variogram fit with null nugget
vm = variogram_fit(x, y, z, mtype, n=0, dmax=dmax)
D = N.linspace(0, d.max())
V = vm(D)
# Compare
P.figure(figsize=(6, 4))
P.title('Variogram')
P.plot(d, v, 'ob', label='From data')
P.plot(D, V, '-r', label='Fitted model (%s)'%mtype)
P.legend(loc='best')
P.ylim(ymin=0)
savefigs(code_file_name(ext=False), pdf=True, verbose=False)
P.close()
#!/usr/bin/env python # -*- coding: utf8 -*- """Axes et grilles avec VACUMM""" from vcmq import N, MV2, create_lon, create_lat, create_grid, isgrid, isrect, islon, set_grid, get_grid, get_axis, varsel, resol, curv2rect, get_xy, meshgrid, meshcells, create_dep, isregular, P, rotate_grid, shiftgrid, extendgrid, create_axes2d, isdepthup, coord2slice, monotonic, xshift, depth2dz, get_closest # Créer # - axes lon = create_lon((2., 11, 2.)) # -> SPECIFIEZ LE LONG_NAME lat = create_lat(N.arange(43, 50.)) dep = create_dep((0., 10)) # -> AFFICHEZ LES INFOS xx, yy = N.meshgrid(N.arange(5.), N.arange(4.)) lon2d, lat2d = create_axes2d(xx, yy) ii = lon2d.getAxis(1) # - grille grid = create_grid(lon, lat) # -> ESSAYEZ AVEC LON EXPLICITE gridc = create_grid(lon2d, lat2d) # Verifier print islon(lon) print isgrid(grid) # -> TEST PARAM CURV=... print isrect(gridc) # -> CREEZ GRILLE NON RECT ET RETESTER print isdepthup(dep) # -> TESTEZ EN CHANGEANT ATTRIBUT POSITIVE ET VALEURS print isregular(lon) # Affecter
def create_Dthnc(self, fileout, TimeSeries):
if '2D' in fileout:
self.i23d = 2
else:
self.i23d = 3
# create file
if self.i23d == 3:
Nlev = self.zz.shape[1]
else:
Nlev = 1
time_Series, nc = create_ncTH(
fileout, len(self.llon), Nlev, self.ivs,
np.round((TimeSeries - TimeSeries[0]) * 24 * 3600))
for n in range(0, len(TimeSeries)):
tin = create_time(np.ones(len(self.llon) * Nlev) *
(TimeSeries[n] + 1),
units='days since 1-1-1')
total = np.zeros(shape=(self.ivs, len(self.llon), Nlev))
# get tide
if self.tidal:
var = self.HC.keys()
for i, v in enumerate(sorted(var)):
# horizontal interpolation
tmp = get_tide(self.constidx, self.tfreq, self.HC[v],
np.array(TimeSeries[n]), self.lat0)
if self.i23d > 2: # vertical interpolation
tmp = vertical_extrapolation(tmp, self.zz, z0=self.z0)
total[i, :, :] = total[i, :, :] + tmp
if self.residual:
var = self.res_vars
for i, v in enumerate(sorted(var)):
arri = self.res_file[v][:]
if self.i23d > 2:
dep = create_depth(arri.getAxis(1)[:])
extra = create_axis(N.arange(1), id='member')
arri2 = np.tile(arri, [1, 1, 1, 1, 1])
arri3 = MV2.array(arri2,
axes=[
extra,
arri.getAxis(0), dep,
arri.getAxis(2),
arri.getAxis(3)
],
copy=False,
fill_value=1e20)
zi = arri.getAxis(1)[:]
if np.mean(zi) > 0:
zi = zi * -1
tb = grid2xy(arri3,
xo=np.tile(self.llon,
[Nlev, 1]).T.flatten(),
yo=np.tile(self.llat,
[Nlev, 1]).T.flatten(),
zo=self.zz.flatten(),
method='linear',
to=tin,
zi=zi)
else:
tb = grid2xy(arri,
xo=self.llon,
yo=self.llat,
method='linear',
to=tin)
if np.any(tb.mask == True):
bad = tb.mask == True
if len(bad.shape) > 1:
bad = bad[0, :]
tin_bad = create_time(np.ones(len(bad)) *
(TimeSeries[n] + 1),
units='days since 1-1-1')
if self.i23d > 2:
llon = np.tile(self.llon, [Nlev, 1]).T.flatten()
llat = np.tile(self.llat, [Nlev, 1]).T.flatten()
zz = self.zz.flatten()
zi = arri.getAxis(1)[:]
if np.mean(zi) > 0:
zi = zi * -1
tb[0, bad] = grid2xy(arri3,
xo=llon[bad],
yo=llat[bad],
zo=zz[bad],
method='nearest',
to=tin_bad,
zi=zi)
else:
tb[bad] = grid2xy(
arri,
xo=np.array(self.llon)[bad].tolist(),
yo=np.array(self.llat)[bad].tolist(),
method='nearest',
to=tin_bad)
if np.any(tb.mask == True):
print('probleme')
total[i, :, :] = total[i, :, :] + np.reshape(
tb, (len(self.llon), Nlev))
total = np.transpose(total, (1, 2, 0))
if np.isnan(total).any():
import pdb
pdb.set_trace()
if n % 100 == 0:
self.logger.info(
'For timestep=%.f, max=%.4f, min=%.4f , max abs diff=%.4f'
% (TimeSeries[n], total.max(), total.min(),
abs(np.diff(total, n=1, axis=0)).max()))
time_Series[n, :, :, :] = total
nc.close()
lon1d = create_lon((0, 10))
result.append(('AssertEqual', (coord2slice(lon1d, lon=(2.5, 4., 'cc')), slice(3, 5, 1))))
result.append(('AssertEqual', (coord2slice(lon1d, lon=(2.5, 4., 'ccb')), slice(2, 5, 1))))
result.append(('AssertEqual', (coord2slice(lon1d, lon=slice(3, 6)), slice(3, 6, None))))
result.append(('AssertEqual', (coord2slice(lon1d, lat=(6, 8)), slice(0, 10, 1))))
result.append(('AssertEqual', (coord2slice(lon1d, lon=(60, 70)), None)))
# Rect grid
grid = create_grid((0, 10.), (20, 30.))
result.append(('AssertEqual', (coord2slice(grid, lon=(0., 3.5), lat=slice(3, 5)),
(slice(0, 4, 1), slice(3, 5, None), None))))
result.append(('AssertEqual', (coord2slice(grid, lat=(21,21, 'ccb')),
(slice(0, 10, 1), slice(1, 2, 1), None))))
# 2D axis
lon2d = N.empty((10, 10.))
for i in xrange(10):
lon2d[i] = lon1d[:]+i
lat2d = N.resize((N.arange(10)+20), (10, 10)).T
lon2d, lat2d = create_axes2d(lon2d, lat2d)
kw = dict(show=False)
plot(lon2d, lat2d, lon2d, 'lon2d', figfiles, figfile, lon=(2, 4), **kw)
plot(lon2d, lat2d, lon2d, 'lon2d', figfiles, figfile, lon=(2, 4), lat=slice(0, 2), **kw)
plot(lon2d, lat2d, lat2d, 'lat2d', figfiles, figfile, lat=(22, 26.6,'ccb'), **kw)
# Curv grid
grid = create_grid(lon2d, lat2d)
plot(lon2d, lat2d, grid, 'grid', figfiles, figfile, lon=(8, 11, 'cc'), lat=(21.9, 26., 'cc'), **kw)
plot(lon2d, lat2d, grid, 'grid', figfiles, figfile, lon=slice(2, 5), lat=(23.4, 23.6, 'ccb'), **kw)
res = coord2slice(grid,lon=(8,8,'ccb'),lat=(24,24,'ccb'))
result.append(('AssertEqual', (res[:2], (slice(3, 6, 1), slice(4, 5, 1)))))
nexz = 2
nxi = 7
nyi = 6
nzi = 5
nti = 4
mv = 1.e20
no = 1
#no = 1
vfunc = lambda t, z, y, x: (1*x + 2.35*y + 3.24*z -0.65*t)
tti, zzi, yyi, xxi = N.mgrid[0:nti-1:nti*1j, 0:nzi-1:nzi*1j,
0:nyi-1:nyi*1j, 0:nxi-1:nxi*1j]
zzi = zzi[None]
zzi = N.repeat(zzi, nexz, axis=0)
# %% Pure 1D axes
xi = xxi[0, 0, 0:1, :] # (nyix=1,nxi)
yi = yyi[0, 0, :, 0:1] # (nyi,nxiy=1)
zi = zzi[0:1, 0:1, :, 0:1, 0:1] # (nexz=1,ntiz=1,nzi,nyiz=1,nxiz=1)
ti = tti[:, 0, 0, 0] # (nti)
vi = vfunc(tti, zzi, yyi, xxi)
vi = N.resize(vi, (nex, )+vi.shape[1:]) # (nex,nti,nzi,nyi,nxi)
N.random.seed(0)
xyztomin = -0.5
xo = N.random.uniform(xyztomin, nxi-1.5, no)
yo = N.random.uniform(xyztomin, nyi-1.5, no)
sp = f("speed")
spe = f("speed_error")
f.close()
# Create hourly time axis
taxi = sp.getTime()
taxi.toRelativeTime("hours since 2000")
ctimesi = taxi.asComponentTime()
ct0 = round_date(ctimesi[0], "hour")
ct1 = round_date(ctimesi[-1], "hour")
taxo = create_time(lindates(ct0, ct1, 1, "hour"), taxi.units)
# Lag error
# - estimation
els = []
lags = N.arange(1, 6)
for lag in lags:
els.append(N.sqrt(((sp[lag:] - sp[:-lag]) ** 2).mean()))
els = N.array(els)
a, b, _, _, _ = linregress(lags, els)
# - plot
P.figure(figsize=(6, 6))
P.subplot(211)
P.plot(lags, els, "o")
P.plot([0, lags[-1]], [b, a * lags[-1] + b], "g")
P.axhline(b, color="0.8", ls="--")
P.ylim(ymin=0)
P.xlabel("Lag [hour]")
P.ylabel("Error [m s-1]")
add_key(1)
P.title("Linear lag error model")
"""Test :func:`~vacumm.misc.plot.plot2d` with specified cell bounds"""
# Imports
from vcmq import MV2, N, code_file_name, plot2d, os
# Init
var = MV2.reshape(MV2.arange(4*3), (3, 4))
var.units = 'Count'
var.long_name = 'Navets'
x = var.getAxis(1)
x.units = 'km'
x.long_name = 'Distance'
y = var.getAxis(0)
y.units = 'm'
y.long_name = 'Height'
y.designateLevel()
# Bounds
y[:] = [1.5, 3.5, 4.5]
y2db = N.array([0., 3, 4, 5])
# Plot
figfile = code_file_name(ext='png')
if os.path.exists(figfile): os.remove(figfile)
plot2d(var, savefig=figfile, show=False, close=True, y2db=y2db, fill='pcolor',
xmin=-.5, xmax=3.5, ymin=0, ymax=5, cmap='jet')
# Unittest
result = {"files":figfile}
nyi,nxi = vari2d.shape #if rank==0: print 'expand in time and depth' vari = MV2.resize(vari2d, (nt, nz)+vari2d.shape) cp_props(vari2d, vari) #if rank==0: print 'grido' loni = gridi.getLongitude() lati = gridi.getLatitude() xib, yib = bounds2d(loni, lati) loni.setBounds(xib) lati.setBounds(yib) xi = loni.getValue() yi = lati.getValue() dx = N.diff(xi[0]).mean() dy = N.diff(yi[:, 0]).mean() xo = N.arange(xi.min()+10*dx, -30*dx+xi.max(), dx) yo = N.arange(yi.min()-20*dy, yi.max()-20*dy, dy) lono = cdms2.createAxis(xo) lono.designateLongitude() ; lono.units= 'degrees_east' lato = cdms2.createAxis(yo) lato.designateLatitude() ; lato.units = 'degrees_north' xob = bounds1d(lono) ; lono.setBounds(xob) yob = bounds1d(lato) ; lato.setBounds(yob) grido = cdms2.createRectGrid(lato, lono) xmin, xmax = minmax(loni.asma(),lono) ymin, ymax = minmax(lati.asma(), lato) nyo,nxo = grido.shape #print 'rank',rank basefile = code_file_name(ext=False)
"""Test :func:`~vacumm.misc.plot.section2` with a Z- variable"""
# Imports
from vcmq import N, MV2, cdms2, create_dep, rc, section2, code_file_name, os
# Init data with z 1D
nz = 8
nd = 10
var = N.dot(N.hanning(nz).reshape(nz, 1), N.hanning(nd).reshape(1, nd))
var = MV2.array(var)
d = cdms2.createAxis(N.arange(nd))
d.units='km'
d.long_name='Distance'
z1d = create_dep((-nz+1, 1.))
var.setAxis(0, z1d)
var.setAxis(1, d)
z2d = N.resize(z1d[:].reshape(1, nz), (nd, nz)).T
z2d *= N.arange(1., nd+1)/nd
# Plot with z 1D
rc('font', size=8)
kw = dict(show=False, bgcolor='0.5')
section2(var, subplot=211, **kw)
# Plot with z 2D
figfile = code_file_name(ext='png')
if os.path.exists(figfile): os.remove(figfile)
section2(var, yaxis=z2d, subplot=212, savefig=figfile, close=True, **kw)
# Result
result = dict(files=figfile)
"""Test fortran function :f:func:`dstwgt2dto1d`""" from vcmq import P, N, code_file_name, os, add_grid from vacumm.misc.grid._interp_ import dstwgt2dto1d # Input grid and data nxy = 15 xi = N.arange(nxy * 1.) yi = N.arange(nxy * 1.) xxi, yyi = N.meshgrid(xi, yi) zi = N.ma.array(yyi) zi[int(nxy * 0.3):int(nxy * 0.8), int(nxy * 0.3):int(nxy * 0.8)] = N.ma.masked zi.shape = 1, nxy, nxy # Output positions no = 1000 xo = N.random.uniform(-nxy / 4., nxy + nxy / 4., no) yo = N.random.uniform(-nxy / 4., nxy + nxy / 4., no) # Interpolate mv = zi.get_fill_value() zo = dstwgt2dto1d(xi, yi, zi.filled(mv), xo, yo, mv) zo = N.ma.masked_values(zo, mv) # Plot kw = dict(vmin=zi.min(), vmax=zi.max()) P.figure(figsize=(6, 6)) P.subplot(111, aspect=1) P.contourf(xxi, yyi, zi[0], **kw) add_grid((xi, yi), edges=False, centers=True, marker='o') P.scatter(xo, yo, c=zo[0], s=50, **kw)
zi[mask] = N.ma.masked
# - output
xo = N.random.uniform(xmin, xmax, 500)
yo = N.random.uniform(ymin, ymax, 500)
zot = N.ma.masked_array(get_values(xo, yo))
mask = (xo > (xmin + xmax) * .5) & (yo > (ymin + ymax) * .5)
zot[mask] = 1e20
zot[mask] = N.ma.masked
# Call and plot
kw = dict(vmin=zi.min(), vmax=zi.max())
P.figure(figsize=(8, 8))
P.subplot(221)
trii = get_tri((xi, yi), ttype='mpl', mask=zi)
P.tripcolor(trii, zi, **kw)
cc = N.zeros(len(xi), '|S3')
cc[:] = 'k'
cc[zi.mask] = '.5'
P.scatter(xi, yi, c=cc, s=7, marker='o')
P.title('Input')
for i, method in enumerate(['nearest', 'linear', 'cubic']):
# Interp
zo = xy2xy(xi, yi, zi, xo, yo, method=method) #[0]
# Triangles
trio = get_tri((xo, yo), ttype='mpl', mask=zo)
# Plot
P.subplot(2, 2, i + 2)
P.tripcolor(trio, zo, **kw)
sp = f('speed')
spe = f('speed_error')
f.close()
# Create hourly time axis
taxi = sp.getTime()
taxi.toRelativeTime('hours since 2000')
ctimesi = taxi.asComponentTime()
ct0 = round_date(ctimesi[0], 'hour')
ct1 = round_date(ctimesi[-1], 'hour')
taxo = create_time(lindates(ct0, ct1, 1, 'hour'), taxi.units)
# Lag error
# - estimation
els = []
lags = N.arange(1, 6)
for lag in lags:
els.append(N.sqrt(((sp[lag:] - sp[:-lag])**2).mean()))
els = N.array(els)
a, b, _, _, _ = linregress(lags, els)
# - plot
P.figure(figsize=(6, 6))
P.subplot(211)
P.plot(lags, els, 'o')
P.plot([0, lags[-1]], [b, a * lags[-1] + b], 'g')
P.axhline(b, color='0.8', ls='--')
P.ylim(ymin=0)
P.xlabel('Lag [hour]')
P.ylabel('Error [m s-1]')
add_key(1)
P.title('Linear lag error model')
def get_values(xx, yy):
return (N.sin(xx * N.pi / 6) * N.sin(yy * N.pi / 6)) * 100.
"""Test the fortran function :f:func:`extrap1d`"""
from vcmq import N, P,meshcells, minmax
from vacumm.misc.grid._interp_ import extrap1d
mv = 999.
vari = N.zeros((5,5))+mv
vari[1,2:4] = [2,3]
vari[2,1] = 1
vari[2,3] = 3
vari[3,3:] = [3,4]
vari[4,:2] = [0,1]
vari = N.asfortranarray(vari)
varo0 = extrap1d(vari, mv, 0)
varop1 = extrap1d(vari, mv, 1)
varom1 = extrap1d(vari, mv, -1)
varop2 = extrap1d(vari, mv, 2)
result = [
('AssertTrue', N.allclose(vari, varo0)),
('AssertTrue', N.allclose(varop1[:, -1], [999., 3., 3., 4., 1.])),
('AssertTrue', N.allclose(varom1[:, 0], [999., 2., 1., 3., 0.])),
('AssertTrue', N.allclose(varop1[:, -1], varop2[:, -1])),
('AssertTrue', N.allclose(varom1[:, 0], varop2[:, 0])),
]
"""Test the fortran function :f:func:`interp1dxx`"""
from vcmq import N, P, meshcells, minmax, code_file_name, os
from vacumm.misc.grid._interp_ import interp1dxx
nx = nyi = 10
mv = 1.e20
u, v = N.mgrid[-3:3:nx * 1j, -3:3:10j] - 2
vari = N.ma.asarray(u**2 + v**2)
vari.set_fill_value(mv)
yi = N.linspace(-1000., 0., nyi)
yo = N.linspace(-1200, 100, 30.)
vari[nx / 3:2 * nx / 3, nyi / 3:2 * nyi / 3] = N.ma.masked
x = N.arange(nx)
dyi = (yi[1] - yi[0]) * 0.49
dyo = (yo[1] - yo[0]) * 0.49
yyi = N.resize(yi, vari.shape) + N.random.uniform(-dyi, dyi, vari.shape)
yyo = N.resize(yo, (nx, len(yo))) + N.random.uniform(-dyo, dyo, (nx, len(yo)))
yyib, xxib = meshcells(yyi, x)
yyob, xxob = meshcells(yyo, x)
varon = N.ma.masked_values(
interp1dxx(vari.filled(), yyi, yyo, mv, 0, extrap=0), mv)
varol = N.ma.masked_values(
interp1dxx(vari.filled(), yyi, yyo, mv, 1, extrap=0), mv)
varoh = N.ma.masked_values(
interp1dxx(vari.filled(), yyi, yyo, mv, 3, extrap=0), mv)
kw = dict(vmin=vari.min(), vmax=vari.max())
axlims = [x[0], x[-1], yo[0], yo[-1]]
P.figure(figsize=(8, 8))
P.subplot(221)
"""Test the :func:`~vacumm.misc.grid.masking.polygons` function"""
from vcmq import N, P, polygons, create_polygon, plot_polygon, create_grid
from _geoslib import Polygon
# Data
xx = N.array([0., 5., 4., 1.])
yy = N.array([0., 0., 2., 2.])
clip = [-3, -3, 0, 0]
xaxis = yaxis = N.linspace(-2., 2., 20.)
def plot_polygons(polys, **kwargs):
for p in polys:
plot_polygon(p, **kwargs)
## Single known argument
#pp0 = polygons(Polygon(N.array([xx, yy]).T))
#
## Classic data
#pp0 = polygons([N.array([xx, yy]), N.array([xx, yy]).T+6.])
#
## Classic with projection
#proj = lambda x, y: (x*1.5, y*1.5)
#pp1 = polygons([N.array([xx, yy])])
#
## Classic with clipping
#pp2 = polygons([-N.array([xx, yy])], clip=clip)
# From grid
pp3 = polygons([create_grid(xaxis, yaxis)])
"""Test the :func:`~vacumm.misc.grid.masking.create_polygon` function""" from vcmq import N, P, create_polygon, plot_polygon, code_file_name from _geoslib import Polygon # Data xx = N.array([0., 5., 4., 1.]) yy = N.array([0., 0., 2., 2.]) # From Polygon P.figure(figsize=(5, 5)) p0 = create_polygon(Polygon(N.array([xx, yy]).T)) plot_polygon(p0, color='b', label='From poly') # From data p1 = create_polygon(N.array([xx, yy]).T-.25) plot_polygon(p1, color='r', label='From data') # From transposed data p2 = create_polygon(N.array([xx, yy])-0.5) plot_polygon(p2, color='g', label='From transposed data') # From min/max p3 = create_polygon([xx.min(), yy.min(), xx.max(), yy.max()]) plot_polygon(p3, color='m', label='From min/max') # With projection proj = lambda x, y: (x*1.5, y*1.5) p4 = create_polygon(p0, proj=proj) plot_polygon(p4, color='cyan', label='With projection') # Save
"""Test the :func:`~vacumm.misc.grid.regridding.regrid2d` function"""
from vcmq import P, N, MV2, code_file_name, os, add_grid, rotate_grid, set_grid, \
create_grid, rc, rcdefaults, plot2d
from vacumm.misc.grid.regridding import regrid2d
# Input grid and data
nxi = 20
nyi = 15
# - rect
xi = N.arange(nxi*1.)
yi = N.arange(nyi*1.)
gridri = create_grid(xi, yi)
xxri, yyri = N.meshgrid(xi, yi)
zzri = N.ma.array(yyri)
zzri[int(nyi*0.3):int(nyi*0.6), int(nxi*0.3):int(nxi*0.6)] = N.ma.masked
varri = MV2.asarray(zzri)
set_grid(varri, gridri)
# - curv
gridci = rotate_grid(gridri, 30)
xxci = gridci.getLongitude().getValue()
yyci = gridci.getLatitude().getValue()
zzci = N.ma.array(yyci)
zzci[int(nyi*0.3):int(nyi*0.6), int(nxi*0.3):int(nxi*0.6)] = N.ma.masked
varci = MV2.asarray(zzci)
set_grid(varci, gridci)
# Output positions
nxo = 25
nyo = 18
# - rect
"""Test the fortran function :f:func:`nearest2d`"""
from vcmq import N, P, code_file_name, P, os, rotate_grid, add_grid, meshbounds
from vacumm.misc.grid._interp_ import nearest2d
# Input grid
gridi = rotate_grid((N.arange(5), N.arange(4)), 30)
xxi = gridi.getLongitude()[:].filled()
yyi = gridi.getLatitude()[:].filled()
vari = N.resize(yyi, (20, ) + yyi.shape)
nb = 10
xxbi, yybi = meshbounds(xxi, yyi)
# Output grid
grido = rotate_grid((N.linspace(0, 6, 50) - 1, N.linspace(0, 4, 35) + 1.), -20)
xxo = grido.getLongitude()[:].filled()
yyo = grido.getLatitude()[:].filled()
xxbo, yybo = meshbounds(xxo, yyo)
# Nearest
varo = nearest2d(vari, xxi, yyi, xxo, yyo, nb)
# Plot
vmin = varo.min()
vmax = varo.max()
P.figure(figsize=(8, 4))
P.subplot(121, aspect=1)
P.pcolor(xxbi, yybi, vari[0], vmin=vmin, vmax=vmax)
add_grid(grido)
P.title('original')
P.subplot(122, aspect=1)
P.pcolor(xxbo, yybo, varo[0], vmin=vmin, vmax=vmax)
"""Test :func:`~vacumm.misc.grid.regridding.extend1d` and :func:`~vacumm.misc.grid.regridding.extend2d`"""
from vcmq import N, create_grid2d, P, rc, plot2d, MV2, set_grid, savefigs, code_file_name
from vacumm.misc.grid.regridding import extend1d, extend2d
# Input data
nxi = 4
nyi = 3
xxi, yyi = N.meshgrid(N.arange(nxi) + .25, N.arange(nyi) - .25)
for j in xrange(nyi):
xxi[j, :] -= j * 0.5
for i in xrange(nxi):
yyi[:, i] += i * 0.5
gridi = create_grid2d(xxi, yyi) # input cdms grid
vari = MV2.array(N.arange(nyi * nxi).reshape(nyi, nxi)) + 10.
set_grid(vari, gridi) # set grid and axes
# Extend and plot
rc('font', size=9)
P.figure(figsize=(6, 6))
kw = dict(xmin=xxi.min() - 3,
xmax=xxi.max() + 3,
ymin=yyi.min() - 3,
ymax=yyi.max() + 3,
show=False,
xhide='auto',
yhide='auto')
# - original
plot2d(vari, title='Original', subplot=(2, 2, 1), **kw)
# - extend1d
for i, (axis, ext, mode) in enumerate([(-1, (2, 2), 'same'),
"""Test CDAT conservative regridding"""
from vcmq import N, meshbounds, bounds1d, cdms2, MV2, rc, P, add_grid, rcdefaults, \
create_lon, create_lat, savefigs, code_file_name
# Input grid
x0, y0, nx, ny, dx, dy = 0., 0., 20, 15, 5., 5.
x = x0 + N.arange(0, nx * dx, dx)
y = y0 + N.arange(0, ny * dy, dy)
xxbi, yybi = meshbounds(x, y)
xb = bounds1d(x)
yb = bounds1d(y)
lon = create_lon(x)
lat = create_lat(y)
lon.setBounds(xb)
lat.setBounds(yb)
gridi = cdms2.createRectGrid(lat, lon)
# Input data
vari = MV2.ones(gridi.shape) * 50
vari[5:10, 7:13] = -10
#vari[:, -1] = MV2.masked # <<< THIS MAKES IT WORK!
vari.setAxisList([gridi.getLatitude(), gridi.getLongitude()])
vari.setGrid(gridi)
# Output grid
grido, xxbo, yybo = gridi, xxbi, yybi
# Regrid
diag = {'dstAreaFractions': None}
varo = vari.regrid(grido,
# Lecture de la température
f =cdms2.open(data_sample('mars3d.tsigyx.nc'))
data_in = f('TEMP', time=slice(0,2)) # T-YX (- = level)
# Détermination des profondeurs d'après les sigma
sigma_converter = NcSigma.factory(f) # détection auto et initialisation du convertisseur
# -> VERIFIER QUE sigma_class EST BIEN SigmaGeneralized
depths_in = sigma_converter.sigma_to_depths(selector=dict(time=slice(0,2))).filled() # lecture eta, etc + conversion
# (Equivalent à depths_in = sigma_converter(selector=dict(time=slice(0,2))).filled())
f.close()
# Creation de l'axe des profondeurs cibles
depths = N.array([0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,
22,24,26,28,30,32,34,36,38,40,45,50,55,60,65,70,75,80,85,90,95,100,120,140,160])
depths = -depths[::-1] # croissantes et négatives
depth_out = create_depth(depths)
# Interpolation
data_out = regrid1d(data_in, depth_out, axi=depths_in, axis=1, method='linear', extrap=1)
# Plot
kw = dict(vmin=10, vmax=14, xhide='auto', add_grid=True, ymax=0, fill='contourf') # FILL=PCOLOR ?
section2(data_in[0, :, 10], yaxis=depths_in[0, :, 10], subplot=211, title='Sigma', show=False, **kw)
s = section2(data_out[0, :, 10], subplot=212, title='Z', show=False, savefig=__file__, **kw)
for i, np in enumerate(nps):
xc = P.rand(1)*dxy*2.5
yc = P.rand(1)*dxy*2.5
x = P.rand(np)*dxy+xc
y = P.rand(np)*dxy+yc
xx.extend(x)
yy.extend(y)
P.scatter(x, y, c=colors[i:i+1], s=40)
P.title('Original (mixed clouds)')
# Spliting
indices, centroids = cloud_split(xx, yy, getdist=False, getcent=True, npmax=npmax)
# Plot
P.subplot(122, aspect=1)
xx = N.array(xx)
yy = N.array(yy)
msize = 0
for i, select in enumerate(indices):
x = xx[select]
y = yy[select]
P.scatter(x, y, c=colors[i:i+1], s=40, label='size=%i'%len(x))
P.scatter(centroids[i:i+1, 0], centroids[i:i+1, 1], c=colors[i:i+1], s=100, lw=2)
msize = max(len(x), msize)
P.legend(loc='upper left').get_frame().set_alpha(.5)
P.title('Split clouds: max size=%i'%npmax)
savefigs(code_file_name(ext='png'), pdf=True, verbose=False)
#P.tight_layout()
P.close()
# For unitest
nexz = 2
nxi = 7
nyi = 6
nzi = 5
nti = 4
mv = 1.e20
no = 1
#no = 1
vfunc = lambda t, z, y, x: (1 * x + 2.35 * y + 3.24 * z - 0.65 * t)
tti, zzi, yyi, xxi = N.mgrid[0:nti - 1:nti * 1j, 0:nzi - 1:nzi * 1j,
0:nyi - 1:nyi * 1j, 0:nxi - 1:nxi * 1j]
zzi = zzi[None]
zzi = N.repeat(zzi, nexz, axis=0)
# Pure 1D axes
xi = xxi[0, 0, 0:1, :] # (nyix=1,nxi)
yi = yyi[0, 0, :, 0:1] # (nyi,nxiy=1)
zi = zzi[0:1, 0:1, :, 0:1, 0:1] # (nexz=1,ntiz=1,nzi,nyiz=1,nxiz=1)
ti = tti[:, 0, 0, 0] # (nti)
vi = vfunc(tti, zzi, yyi, xxi)
vi = N.resize(vi, (nex, ) + vi.shape[1:]) # (nex,nti,nzi,nyi,nxi)
N.random.seed(0)
xyztomin = -0.5
xo = N.random.uniform(xyztomin, nxi - 1.5, no)
yo = N.random.uniform(xyztomin, nyi - 1.5, no)
zo = N.random.uniform(xyztomin, nzi - 1.5, no)
"""Test fortran function :f:func:`nearest2dto1d`""" from vcmq import P, N, code_file_name, os, add_grid from vacumm.misc.grid._interp_ import nearest2dto1d # Input grid and data nxy = 15 xi = N.arange(nxy*1.) yi = N.arange(nxy*1.) xxi, yyi = N.meshgrid(xi, yi) zi = N.ma.array(yyi) zi[int(nxy*0.3):int(nxy*0.8), int(nxy*0.3):int(nxy*0.8)] = N.ma.masked zi.shape = 1, nxy, nxy # Output positions no = 1000 xo = N.random.uniform(-nxy/4., nxy+nxy/4., no) yo = N.random.uniform(-nxy/4., nxy+nxy/4., no) # Interpolate mv = zi.get_fill_value() zo = nearest2dto1d(xi,yi,zi.filled(mv),xo,yo,mv) zo = N.ma.masked_values(zo, mv) # Plot kw = dict(vmin=zi.min(), vmax=zi.max()) P.figure(figsize=(6, 6)) P.subplot(111, aspect=1) P.contourf(xxi, yyi, zi[0], **kw) add_grid((xi, yi), edges=False, centers=True, marker='o')
"""Test the :func:`~vacumm.misc.grid.masking.create_polygon` function""" from vcmq import N, P, create_polygon, plot_polygon, code_file_name from _geoslib import Polygon # Data xx = N.array([0., 5., 4., 1.]) yy = N.array([0., 0., 2., 2.]) # From Polygon P.figure(figsize=(5, 5)) p0 = create_polygon(Polygon(N.array([xx, yy]).T)) plot_polygon(p0, color='b', label='From poly') # From data p1 = create_polygon(N.array([xx, yy]).T - .25) plot_polygon(p1, color='r', label='From data') # From transposed data p2 = create_polygon(N.array([xx, yy]) - 0.5) plot_polygon(p2, color='g', label='From transposed data') # From min/max p3 = create_polygon([xx.min(), yy.min(), xx.max(), yy.max()]) plot_polygon(p3, color='m', label='From min/max') # With projection proj = lambda x, y: (x * 1.5, y * 1.5) p4 = create_polygon(p0, proj=proj) plot_polygon(p4, color='cyan', label='With projection') # Save
"""Test the :func:`~vacumm.misc.grid.regridding.regrid2d` function"""
from vcmq import (P, N, MV2, code_file_name, os, add_grid, rotate_grid, set_grid,
create_grid, rc, rcdefaults, plot2d, regrid2d, create_time)
# Input grid and data
nxi = 20
nyi = 15
nt = 5
# - rect
xi = N.arange(nxi*1.)
yi = N.arange(nyi*1.)
ti = create_time((nt, ), 'years since 2000')
gridri = create_grid(xi, yi)
xxri, yyri = N.meshgrid(xi, yi)
zzri = N.ma.resize(yyri, (nt, nyi, nxi))
zzri[:, int(nyi*0.3):int(nyi*0.6), int(nxi*0.3):int(nxi*0.6)] = N.ma.masked
zzri[1] = N.ma.masked
varri = MV2.asarray(zzri)
varri.setAxis(0, ti)
set_grid(varri, gridri)
# - curv
gridci = rotate_grid(gridri, 30)
xxci = gridci.getLongitude().getValue()
yyci = gridci.getLatitude().getValue()
zzci = N.ma.resize(yyci, (nt, nyi, nxi))
zzci[:, int(nyi*0.3):int(nyi*0.6), int(nxi*0.3):int(nxi*0.6)] = N.ma.masked
zzci[1] = N.ma.masked
varci = MV2.asarray(zzci)
varci.setAxis(0, ti)
set_grid(varci, gridci)