def _get_draped(data, topodata, colormap):
"""Get array of data "draped" on topography.
Args:
data (ndarray): 2D Numpy array.
topodata (ndarray): 2D Numpy array.
colormap (ColorPalette): MMI color palette object.
Returns:
ndarray: Numpy array of data draped on topography.
"""
maxvalue = colormap.vmax
mmisc = data / maxvalue
rgba_img = colormap.cmap(mmisc)
del mmisc
rgb = np.squeeze(rgba_img[:, :, 0:3])
del rgba_img
if np.allclose(topodata, 0):
intensity = np.full_like(topodata, 0.5)
else:
# use lightsource class to make our shaded topography
ls1 = LightSource(azdeg=300, altdeg=45)
ls2 = LightSource(azdeg=45, altdeg=45)
intensity1 = ls1.hillshade(
topodata, fraction=0.25, vert_exag=VERT_EXAG)
intensity2 = ls2.hillshade(
topodata, fraction=0.25, vert_exag=VERT_EXAG)
intensity = intensity1 * 0.5 + intensity2 * 0.5
del intensity1, intensity2
ls = LightSource(azdeg=315, altdeg=45)
draped_hsv = ls.blend_hsv(rgb, np.expand_dims(intensity, 2))
return draped_hsv
def _get_shaded(ptopo, contour_colormap):
"""Get hill-shaded topo with topo colormap.
Args:
ptopo (ndarray): Projected topography numpy array.
contour_colormap (ColorPalette): Topography color palette object.
Returns:
ndarray: Hill-shaded topography RGB image.
"""
maxvalue = contour_colormap.vmax
ls1 = LightSource(azdeg=300, altdeg=45)
if np.allclose(ptopo, 0):
intensity = np.full_like(ptopo, 1.0)
else:
maxvalue = contour_colormap.vmax
ls2 = LightSource(azdeg=45, altdeg=45)
intensity1 = ls1.hillshade(ptopo, fraction=0.25, vert_exag=VERT_EXAG)
intensity2 = ls2.hillshade(ptopo, fraction=0.25, vert_exag=VERT_EXAG)
intensity = intensity1 * 0.5 + intensity2 * 0.5
del intensity1, intensity2
ptoposc = ptopo / maxvalue
rgba = contour_colormap.cmap(ptoposc)
del ptoposc
rgb = np.squeeze(rgba)
del rgba
draped_hsv = ls1.blend_hsv(rgb, np.expand_dims(intensity, 2))
return draped_hsv
def _getDraped(self, data, topodata):
"""Get array of data "draped" on topography.
Args:
data (ndarray): 2D Numpy array.
topodata (ndarray): 2D Numpy array.
Returns:
ndarray: Numpy array of data draped on topography.
"""
maxvalue = self.intensity_colormap.vmax
mmisc = data / maxvalue
rgba_img = self.intensity_colormap.cmap(mmisc)
rgb = np.squeeze(rgba_img[:, :, 0:3])
# use lightsource class to make our shaded topography
ls = LightSource(azdeg=135, altdeg=45)
# intensity = ls.hillshade(ptopo,fraction=0.25,vert_exag=1.0)
ls1 = LightSource(azdeg=120, altdeg=45)
ls2 = LightSource(azdeg=225, altdeg=45)
intensity1 = ls1.hillshade(
topodata, fraction=0.25, vert_exag=VERT_EXAG)
intensity2 = ls2.hillshade(
topodata, fraction=0.25, vert_exag=VERT_EXAG)
intensity = intensity1 * 0.5 + intensity2 * 0.5
draped_hsv = ls.blend_hsv(rgb, np.expand_dims(intensity, 2))
return draped_hsv
def _getShaded(self,ptopo):
maxvalue = self.contour_colormap.vmax
ls1 = LightSource(azdeg = 120, altdeg = 45)
ls2 = LightSource(azdeg = 225, altdeg = 45)
intensity1 = ls1.hillshade(ptopo, fraction = 0.25, vert_exag = VERT_EXAG)
intensity2 = ls2.hillshade(ptopo, fraction = 0.25, vert_exag = VERT_EXAG)
intensity = intensity1*0.5 + intensity2*0.5
ptoposc = ptopo/maxvalue
rgba = self.contour_colormap.cmap(ptoposc)
rgb = np.squeeze(rgba)
draped_hsv = ls1.blend_hsv(rgb,np.expand_dims(intensity,2))
return draped_hsv
def _getDraped(self,data,topodata):
maxvalue = self.intensity_colormap.vmax
mmisc = data/maxvalue
rgba_img = self.intensity_colormap.cmap(mmisc)
rgb = np.squeeze(rgba_img[:,:,0:3])
#use lightsource class to make our shaded topography
ls = LightSource(azdeg=135,altdeg=45)
# intensity = ls.hillshade(ptopo,fraction=0.25,vert_exag=1.0)
ls1 = LightSource(azdeg = 120, altdeg = 45)
ls2 = LightSource(azdeg = 225, altdeg = 45)
intensity1 = ls1.hillshade(topodata, fraction = 0.25, vert_exag = VERT_EXAG)
intensity2 = ls2.hillshade(topodata, fraction = 0.25, vert_exag = VERT_EXAG)
intensity = intensity1*0.5 + intensity2*0.5
draped_hsv = ls.blend_hsv(rgb,np.expand_dims(intensity,2))
return draped_hsv
def _getShaded(self, ptopo):
"""Get shaded topography.
Args:
ptopo (ndarray): Numpy array of projected topography data.
Returns:
ndarray: Numpy array of light-shaded topography.
"""
maxvalue = self.contour_colormap.vmax
ls1 = LightSource(azdeg=120, altdeg=45)
ls2 = LightSource(azdeg=225, altdeg=45)
intensity1 = ls1.hillshade(ptopo, fraction=0.25, vert_exag=VERT_EXAG)
intensity2 = ls2.hillshade(ptopo, fraction=0.25, vert_exag=VERT_EXAG)
intensity = intensity1 * 0.5 + intensity2 * 0.5
ptoposc = ptopo / maxvalue
rgba = self.contour_colormap.cmap(ptoposc)
rgb = np.squeeze(rgba)
draped_hsv = ls1.blend_hsv(rgb, np.expand_dims(intensity, 2))
return draped_hsv
def modelMap(grids, shakefile=None, suptitle=None, inventory_shapefile=None,
plotorder=None, maskthreshes=None, colormaps=None, boundaries=None,
zthresh=0, scaletype='continuous', lims=None, logscale=False,
ALPHA=0.7, maproads=True, mapcities=True, isScenario=False,
roadfolder=None, topofile=None, cityfile=None, oceanfile=None,
roadcolor='#6E6E6E', watercolor='#B8EEFF', countrycolor='#177F10',
outputdir=None, savepdf=True, savepng=True, showplots=False,
roadref='unknown', cityref='unknown', oceanref='unknown',
printparam=False, ds=True, dstype='mean', upsample=False):
"""
This function creates maps of mapio grid layers (e.g. liquefaction or
landslide models with their input layers)
All grids must use the same bounds
TO DO change so that all input layers do not have to have the same bounds,
test plotting multiple probability layers, and add option so that if PDF and
PNG aren't output, opens plot on screen using plt.show()
:param grids: Dictionary of N layers and metadata formatted like:
maplayers['layer name']={
'grid': mapio grid2D object,
'label': 'label for colorbar and top line of subtitle',
'type': 'output or input to model',
'description': 'detailed description of layer for subtitle'}.
Layer names must be unique.
:type name: Dictionary or Ordered dictionary - import collections;
grids = collections.OrderedDict()
:param shakefile: optional ShakeMap file (url or full file path) to extract information for labels and folder names
:type shakefile: Shakemap Event Dictionary
:param suptitle: This will be displayed at the top of the plots and in the
figure names
:type suptitle: string
:param plotorder: List of keys describing the order to plot the grids, if
None and grids is an ordered dictionary, it will use the order of the
dictionary, otherwise it will choose order which may be somewhat random
but it will always put a probability grid first
:type plotorder: list
:param maskthreshes: N x 1 array or list of lower thresholds for masking
corresponding to order in plotorder or order of OrderedDict if plotorder
is None. If grids is not an ordered dict and plotorder is not specified,
this will not work right. If None (default), nothing will be masked
:param colormaps: List of strings of matplotlib colormaps (e.g. cm.autumn_r)
corresponding to plotorder or order of dictionary if plotorder is None.
The list can contain both strings and None e.g. colormaps = ['cm.autumn',
None, None, 'cm.jet'] and None's will default to default colormap
:param boundaries: None to show entire study area, 'zoom' to zoom in on the
area of action (only works if there is a probability layer) using zthresh
as a threshold, or a dictionary defining lats and lons in the form of
boundaries.xmin = minlon, boundaries.xmax = maxlon, boundaries.ymin =
min lat, boundaries.ymax = max lat
:param zthresh: threshold for computing zooming bounds, only used if
boundaries = 'zoom'
:type zthresh: float
:param scaletype: Type of scale for plotting, 'continuous' or 'binned' -
will be reflected in colorbar
:type scaletype: string
:param lims: None or Nx1 list of tuples or numpy arrays corresponding to
plotorder defining the limits for saturating the colorbar (vmin, vmax) if
scaletype is continuous or the bins to use (clev) if scaletype if binned.
The list can contain tuples, arrays, and Nones, e.g. lims = [(0., 10.),
None, (0.1, 1.5), np.linspace(0., 1.5, 15)]. When None is specified, the
program will estimate the limits, when an array is specified but the scale
type is continuous, vmin will be set to min(array) and vmax will be set
to max(array)
:param lims: None or Nx1 list of Trues and Falses corresponding to
plotorder defining whether to use a linear or log scale (log10) for
plotting the layer. This will be reflected in the labels
:param ALPHA: Transparency for mapping, if there is a hillshade that will
plot below each layer, it is recommended to set this to at least 0.7
:type ALPHA: float
:param maproads: Whether to show roads or not, default True, but requires
that roadfile is specified and valid to work
:type maproads: boolean
:param mapcities: Whether to show cities or not, default True, but requires
that cityfile is specified and valid to work
:type mapcities: boolean
:param isScenario: Whether this is a scenario (True) or a real event (False)
(default False)
:type isScenario: boolean
:param roadfolder: Full file path to folder containing road shapefiles
:type roadfolder: string
:param topofile: Full file path to topography grid (GDAL compatible) - this
is only needed to make a hillshade if a premade hillshade is not specified
:type topofile: string
:param cityfile: Full file path to Pager file containing city & population
information
:type cityfile: string
:param roadcolor: Color to use for roads, if plotted, default #6E6E6E
:type roadcolor: Hex color or other matplotlib compatible way of defining
color
:param watercolor: Color to use for oceans, lakes, and rivers, default
#B8EEFF
:type watercolor: Hex color or other matplotlib compatible way of defining
color
:param countrycolor: Color for country borders, default #177F10
:type countrycolor: Hex color or other matplotlib compatible way of defining
color
:param outputdir: File path for outputting figures, if edict is defined, a
subfolder based on the event id will be created in this folder. If None,
will use current directory
:param savepdf: True to save pdf figure, False to not
:param savepng: True to save png figure, False to not
:param ds: True to allow downsampling for display (necessary when arrays
are quite large, False to not allow)
:param dstype: What function to use in downsampling, options are 'min',
'max', 'median', or 'mean'
:param upsample: True to upsample the layer to the DEM resolution for better
looking hillshades
:returns:
* PDF and/or PNG of map
* Downsampled and trimmed version of input grids. If no
modification was needed for plotting, this will be identical to grids but
without the metadata
"""
if suptitle is None:
suptitle = ' '
plt.ioff()
defaultcolormap = cm.jet
if shakefile is not None:
edict = ShakeGrid.load(shakefile, adjust='res').getEventDict()
temp = ShakeGrid.load(shakefile, adjust='res').getShakeDict()
edict['eventid'] = temp['shakemap_id']
edict['version'] = temp['shakemap_version']
else:
edict = None
# Get output file location
if outputdir is None:
print('No output location given, using current directory for outputs\n')
outputdir = os.getcwd()
if edict is not None:
outfolder = os.path.join(outputdir, edict['event_id'])
else:
outfolder = outputdir
if not os.path.isdir(outfolder):
os.makedirs(outfolder)
# Get plotting order, if not specified
if plotorder is None:
plotorder = list(grids.keys())
# Get boundaries to use for all plots
cut = True
if boundaries is None:
cut = False
keytemp = list(grids.keys())
boundaries = grids[keytemp[0]]['grid'].getGeoDict()
elif boundaries == 'zoom':
# Find probability layer (will just take the maximum bounds if there is
# more than one)
keytemp = list(grids.keys())
key1 = [key for key in keytemp if 'model' in key.lower()]
if len(key1) == 0:
print('Could not find model layer to use for zoom, using default boundaries')
keytemp = list(grids.keys())
boundaries = grids[keytemp[0]]['grid'].getGeoDict()
else:
lonmax = -1.e10
lonmin = 1.e10
latmax = -1.e10
latmin = 1.e10
for key in key1:
# get lat lons of areas affected and add, if no areas affected,
# switch to shakemap boundaries
temp = grids[key]['grid']
xmin, xmax, ymin, ymax = temp.getBounds()
lons = np.linspace(xmin, xmax, temp.getGeoDict().nx)
lats = np.linspace(ymax, ymin, temp.getGeoDict().ny) # backwards so it plots right
row, col = np.where(temp.getData() > float(zthresh))
lonmin = lons[col].min()
lonmax = lons[col].max()
latmin = lats[row].min()
latmax = lats[row].max()
# llons, llats = np.meshgrid(lons, lats) # make meshgrid
# llons1 = llons[temp.getData() > float(zthresh)]
# llats1 = llats[temp.getData() > float(zthresh)]
# if llons1.min() < lonmin:
# lonmin = llons1.min()
# if llons1.max() > lonmax:
# lonmax = llons1.max()
# if llats1.min() < latmin:
# latmin = llats1.min()
# if llats1.max() > latmax:
# latmax = llats1.max()
boundaries1 = {'dx': 100, 'dy': 100., 'nx': 100., 'ny': 100} # dummy fillers, only really care about bounds
if xmin < lonmin-0.15*(lonmax-lonmin):
boundaries1['xmin'] = lonmin-0.1*(lonmax-lonmin)
else:
boundaries1['xmin'] = xmin
if xmax > lonmax+0.15*(lonmax-lonmin):
boundaries1['xmax'] = lonmax+0.1*(lonmax-lonmin)
else:
boundaries1['xmax'] = xmax
if ymin < latmin-0.15*(latmax-latmin):
boundaries1['ymin'] = latmin-0.1*(latmax-latmin)
else:
boundaries1['ymin'] = ymin
if ymax > latmax+0.15*(latmax-latmin):
boundaries1['ymax'] = latmax+0.1*(latmax-latmin)
else:
boundaries1['ymax'] = ymax
boundaries = GeoDict(boundaries1, adjust='res')
else:
# SEE IF BOUNDARIES ARE SAME AS BOUNDARIES OF LAYERS
keytemp = list(grids.keys())
tempgdict = grids[keytemp[0]]['grid'].getGeoDict()
if np.abs(tempgdict.xmin-boundaries['xmin']) < 0.05 and \
np.abs(tempgdict.ymin-boundaries['ymin']) < 0.05 and \
np.abs(tempgdict.xmax-boundaries['xmax']) < 0.05 and \
np.abs(tempgdict.ymax - boundaries['ymax']) < 0.05:
print('Input boundaries are almost the same as specified boundaries, no cutting needed')
boundaries = tempgdict
cut = False
else:
try:
if boundaries['xmin'] > boundaries['xmax'] or \
boundaries['ymin'] > boundaries['ymax']:
print('Input boundaries are not usable, using default boundaries')
keytemp = list(grids.keys())
boundaries = grids[keytemp[0]]['grid'].getGeoDict()
cut = False
else:
# Build dummy GeoDict
boundaries = GeoDict({'xmin': boundaries['xmin'],
'xmax': boundaries['xmax'],
'ymin': boundaries['ymin'],
'ymax': boundaries['ymax'],
'dx': 100.,
'dy': 100.,
'ny': 100.,
'nx': 100.},
adjust='res')
except:
print('Input boundaries are not usable, using default boundaries')
keytemp = list(grids.keys())
boundaries = grids[keytemp[0]]['grid'].getGeoDict()
cut = False
# Pull out bounds for various uses
bxmin, bxmax, bymin, bymax = boundaries.xmin, boundaries.xmax, boundaries.ymin, boundaries.ymax
# Determine if need a single panel or multi-panel plot and if multi-panel,
# how many and how it will be arranged
fig = plt.figure()
numpanels = len(grids)
if numpanels == 1:
rowpan = 1
colpan = 1
# create the figure and axes instances.
fig.set_figwidth(5)
elif numpanels == 2 or numpanels == 4:
rowpan = np.ceil(numpanels/2.)
colpan = 2
fig.set_figwidth(13)
else:
rowpan = np.ceil(numpanels/3.)
colpan = 3
fig.set_figwidth(15)
if rowpan == 1:
fig.set_figheight(rowpan*6.0)
else:
fig.set_figheight(rowpan*5.3)
# Need to update naming to reflect the shakemap version once can get
# getHeaderData to work, add edict['version'] back into title, maybe
# shakemap id also?
fontsizemain = 14.
fontsizesub = 12.
fontsizesmallest = 10.
if rowpan == 1.:
fontsizemain = 12.
fontsizesub = 10.
fontsizesmallest = 8.
if edict is not None:
if isScenario:
title = edict['event_description']
else:
timestr = edict['event_timestamp'].strftime('%b %d %Y')
title = 'M%.1f %s v%i - %s' % (edict['magnitude'], timestr, edict['version'], edict['event_description'])
plt.suptitle(title+'\n'+suptitle, fontsize=fontsizemain)
else:
plt.suptitle(suptitle, fontsize=fontsizemain)
clear_color = [0, 0, 0, 0.0]
# Cut all of them and release extra memory
xbuff = (bxmax-bxmin)/10.
ybuff = (bymax-bymin)/10.
cutxmin = bxmin-xbuff
cutymin = bymin-ybuff
cutxmax = bxmax+xbuff
cutymax = bymax+ybuff
if cut is True:
newgrids = collections.OrderedDict()
for k, layer in enumerate(plotorder):
templayer = grids[layer]['grid']
try:
newgrids[layer] = {'grid': templayer.cut(cutxmin, cutxmax, cutymin, cutymax, align=True)}
except Exception as e:
print(('Cutting failed, %s, continuing with full layers' % e))
newgrids = grids
continue
del templayer
gc.collect()
else:
newgrids = grids
tempgdict = newgrids[list(grids.keys())[0]]['grid'].getGeoDict()
# Upsample layers to same as topofile if desired for better looking hillshades
if upsample is True and topofile is not None:
try:
topodict = GDALGrid.getFileGeoDict(topofile)
if topodict.dx >= tempgdict.dx or topodict.dy >= tempgdict.dy:
print('Upsampling not possible, resolution of results already smaller than DEM')
pass
else:
tempgdict1 = GeoDict({'xmin': tempgdict.xmin-xbuff,
'ymin': tempgdict.ymin-ybuff,
'xmax': tempgdict.xmax+xbuff,
'ymax': tempgdict.ymax+ybuff,
'dx': topodict.dx,
'dy': topodict.dy,
'nx': topodict.nx,
'ny': topodict.ny},
adjust='res')
tempgdict2 = tempgdict1.getBoundsWithin(tempgdict)
for k, layer in enumerate(plotorder):
newgrids[layer]['grid'] = newgrids[layer]['grid'].subdivide(tempgdict2)
except:
print('Upsampling failed, continuing')
# Downsample all of them for plotting, if needed, and replace them in
# grids (to save memory)
tempgrid = newgrids[list(grids.keys())[0]]['grid']
xsize = tempgrid.getGeoDict().nx
ysize = tempgrid.getGeoDict().ny
inchesx, inchesy = fig.get_size_inches()
divx = int(np.round(xsize/(500.*inchesx)))
divy = int(np.round(ysize/(500.*inchesy)))
xmin, xmax, ymin, ymax = tempgrid.getBounds()
gdict = tempgrid.getGeoDict() # Will be replaced if downsampled
del tempgrid
gc.collect()
if divx <= 1:
divx = 1
if divy <= 1:
divy = 1
if (divx > 1. or divy > 1.) and ds:
if dstype == 'max':
func = np.nanmax
elif dstype == 'min':
func = np.nanmin
elif dstype == 'med':
func = np.nanmedian
else:
func = np.nanmean
for k, layer in enumerate(plotorder):
layergrid = newgrids[layer]['grid']
dat = block_reduce(layergrid.getData().copy(),
block_size=(divy, divx),
cval=float('nan'),
func=func)
if k == 0:
lons = block_reduce(np.linspace(xmin, xmax, layergrid.getGeoDict().nx),
block_size=(divx,),
func=np.mean,
cval=float('nan'))
if math.isnan(lons[-1]):
lons[-1] = lons[-2] + (lons[1]-lons[0])
lats = block_reduce(np.linspace(ymax, ymin, layergrid.getGeoDict().ny),
block_size=(divy,),
func=np.mean,
cval=float('nan'))
if math.isnan(lats[-1]):
lats[-1] = lats[-2] + (lats[1]-lats[0])
gdict = GeoDict({'xmin': lons.min(),
'xmax': lons.max(),
'ymin': lats.min(),
'ymax': lats.max(),
'dx': np.abs(lons[1]-lons[0]),
'dy': np.abs(lats[1]-lats[0]),
'nx': len(lons),
'ny': len(lats)},
adjust='res')
newgrids[layer]['grid'] = Grid2D(dat, gdict)
del layergrid, dat
else:
lons = np.linspace(xmin, xmax, xsize)
lats = np.linspace(ymax, ymin, ysize) # backwards so it plots right side up
#make meshgrid
llons1, llats1 = np.meshgrid(lons, lats)
# See if there is an oceanfile for masking
bbox = PolygonSH(((cutxmin, cutymin), (cutxmin, cutymax), (cutxmax, cutymax), (cutxmax, cutymin)))
if oceanfile is not None:
try:
f = fiona.open(oceanfile)
oc = next(f)
f.close
shapes = shape(oc['geometry'])
# make boundaries into a shape
ocean = shapes.intersection(bbox)
except:
print('Not able to read specified ocean file, will use default ocean masking')
oceanfile = None
if inventory_shapefile is not None:
try:
f = fiona.open(inventory_shapefile)
invshp = list(f.items(bbox=(bxmin, bymin, bxmax, bymax)))
f.close()
inventory = [shape(inv[1]['geometry']) for inv in invshp]
except:
print('unable to read inventory shapefile specified, will not plot inventory')
inventory_shapefile = None
# # Find cities that will be plotted
if mapcities is True and cityfile is not None:
try:
mycity = BasemapCities.loadFromGeoNames(cityfile=cityfile)
bcities = mycity.limitByBounds((bxmin, bxmax, bymin, bymax))
#bcities = bcities.limitByPopulation(40000)
bcities = bcities.limitByGrid(nx=4, ny=4, cities_per_grid=2)
except:
print('Could not read in cityfile, not plotting cities')
mapcities = False
cityfile = None
# Load in topofile
if topofile is not None:
try:
topomap = GDALGrid.load(topofile, resample=True, method='linear', samplegeodict=gdict)
except:
topomap = GMTGrid.load(topofile, resample=True, method='linear', samplegeodict=gdict)
topodata = topomap.getData().copy()
# mask oceans if don't have ocean shapefile
if oceanfile is None:
topodata = maskoceans(llons1, llats1, topodata, resolution='h', grid=1.25, inlands=True)
else:
print('no hillshade is possible\n')
topomap = None
topodata = None
# Load in roads, if needed
if maproads is True and roadfolder is not None:
try:
roadslist = []
for folder in os.listdir(roadfolder):
road1 = os.path.join(roadfolder, folder)
shpfiles = glob.glob(os.path.join(road1, '*.shp'))
if len(shpfiles):
shpfile = shpfiles[0]
f = fiona.open(shpfile)
shapes = list(f.items(bbox=(bxmin, bymin, bxmax, bymax)))
for shapeid, shapedict in shapes:
roadslist.append(shapedict)
f.close()
except:
print('Not able to plot roads')
roadslist = None
val = 1
for k, layer in enumerate(plotorder):
layergrid = newgrids[layer]['grid']
if 'label' in list(grids[layer].keys()):
label1 = grids[layer]['label']
else:
label1 = layer
try:
sref = grids[layer]['description']['name']
except:
sref = None
ax = fig.add_subplot(rowpan, colpan, val)
val += 1
clat = bymin + (bymax-bymin)/2.0
clon = bxmin + (bxmax-bxmin)/2.0
# setup of basemap ('lcc' = lambert conformal conic).
# use major and minor sphere radii from WGS84 ellipsoid.
m = Basemap(llcrnrlon=bxmin, llcrnrlat=bymin, urcrnrlon=bxmax, urcrnrlat=bymax,
rsphere=(6378137.00, 6356752.3142),
resolution='l', area_thresh=1000., projection='lcc',
lat_1=clat, lon_0=clon, ax=ax)
x1, y1 = m(llons1, llats1) # get projection coordinates
axsize = ax.get_window_extent().transformed(fig.dpi_scale_trans.inverted())
if k == 0:
wid, ht = axsize.width, axsize.height
if colormaps is not None and \
len(colormaps) == len(newgrids) and \
colormaps[k] is not None:
palette = colormaps[k]
else: # Find preferred default color map for each type of layer
if 'prob' in layer.lower() or 'pga' in layer.lower() or \
'pgv' in layer.lower() or 'cohesion' in layer.lower() or \
'friction' in layer.lower() or 'fs' in layer.lower():
palette = cm.jet
elif 'slope' in layer.lower():
palette = cm.gnuplot2
elif 'precip' in layer.lower():
palette = cm2.s3pcpn
else:
palette = defaultcolormap
if topodata is not None:
if k == 0:
ptopo = m.transform_scalar(
np.flipud(topodata), lons+0.5*gdict.dx,
lats[::-1]-0.5*gdict.dy, np.round(300.*wid),
np.round(300.*ht), returnxy=False, checkbounds=False,
order=1, masked=False)
#use lightsource class to make our shaded topography
ls = LightSource(azdeg=135, altdeg=45)
ls1 = LightSource(azdeg=120, altdeg=45)
ls2 = LightSource(azdeg=225, altdeg=45)
intensity1 = ls1.hillshade(ptopo, fraction=0.25, vert_exag=1.)
intensity2 = ls2.hillshade(ptopo, fraction=0.25, vert_exag=1.)
intensity = intensity1*0.5 + intensity2*0.5
#hillshm_im = m.transform_scalar(np.flipud(hillshm), lons, lats[::-1], np.round(300.*wid), np.round(300.*ht), returnxy=False, checkbounds=False, order=0, masked=False)
#m.imshow(hillshm_im, cmap='Greys', vmin=0., vmax=3., zorder=1, interpolation='none') # vmax = 3 to soften colors to light gray
#m.pcolormesh(x1, y1, hillshm, cmap='Greys', linewidth=0., rasterized=True, vmin=0., vmax=3., edgecolors='none', zorder=1);
# plt.draw()
# Get the data
dat = layergrid.getData().copy()
# mask out anything below any specified thresholds
# Might need to move this up to before downsampling...might give illusion of no hazard in places where there is some that just got averaged out
if maskthreshes is not None and len(maskthreshes) == len(newgrids):
if maskthreshes[k] is not None:
dat[dat <= maskthreshes[k]] = float('NaN')
dat = np.ma.array(dat, mask=np.isnan(dat))
if logscale is not False and len(logscale) == len(newgrids):
if logscale[k] is True:
dat = np.log10(dat)
label1 = r'$log_{10}$(' + label1 + ')'
if scaletype.lower() == 'binned':
# Find order of range to know how to scale
order = np.round(np.log(np.nanmax(dat) - np.nanmin(dat)))
if order < 1.:
scal = 10**-order
else:
scal = 1.
if lims is None or len(lims) != len(newgrids):
clev = (np.linspace(np.floor(scal*np.nanmin(dat)), np.ceil(scal*np.nanmax(dat)), 10))/scal
else:
if lims[k] is None:
clev = (np.linspace(np.floor(scal*np.nanmin(dat)), np.ceil(scal*np.nanmax(dat)), 10))/scal
else:
clev = lims[k]
# Adjust to colorbar levels
dat[dat < clev[0]] = clev[0]
for j, level in enumerate(clev[:-1]):
dat[(dat >= clev[j]) & (dat < clev[j+1])] = clev[j]
# So colorbar saturates at top
dat[dat > clev[-1]] = clev[-1]
#panelhandle = m.contourf(x1, y1, datm, clev, cmap=palette, linewidth=0., alpha=ALPHA, rasterized=True)
vmin = clev[0]
vmax = clev[-1]
else:
if lims is not None and len(lims) == len(newgrids):
if lims[k] is None:
vmin = np.nanmin(dat)
vmax = np.nanmax(dat)
else:
vmin = lims[k][0]
vmax = lims[k][-1]
else:
vmin = np.nanmin(dat)
vmax = np.nanmax(dat)
# Mask out cells overlying oceans or block with a shapefile if available
if oceanfile is None:
dat = maskoceans(llons1, llats1, dat, resolution='h', grid=1.25, inlands=True)
else:
#patches = []
if type(ocean) is PolygonSH:
ocean = [ocean]
for oc in ocean:
patch = getProjectedPatch(oc, m, edgecolor="#006280", facecolor=watercolor, lw=0.5, zorder=4.)
#x, y = m(oc.exterior.xy[0], oc.exterior.xy[1])
#xy = zip(x, y)
#patch = Polygon(xy, facecolor=watercolor, edgecolor="#006280", lw=0.5, zorder=4.)
##patches.append(Polygon(xy, facecolor=watercolor, edgecolor=watercolor, zorder=500.))
ax.add_patch(patch)
##ax.add_collection(PatchCollection(patches))
if inventory_shapefile is not None:
for in1 in inventory:
if 'point' in str(type(in1)):
x, y = in1.xy
x = x[0]
y = y[0]
m.scatter(x, y, c='m', s=50, latlon=True, marker='^',
zorder=100001)
else:
x, y = m(in1.exterior.xy[0], in1.exterior.xy[1])
xy = list(zip(x, y))
patch = Polygon(xy, facecolor='none', edgecolor='k', lw=0.5, zorder=10.)
#patches.append(Polygon(xy, facecolor=watercolor, edgecolor=watercolor, zorder=500.))
ax.add_patch(patch)
palette.set_bad(clear_color, alpha=0.0)
# Plot it up
dat_im = m.transform_scalar(
np.flipud(dat), lons+0.5*gdict.dx, lats[::-1]-0.5*gdict.dy,
np.round(300.*wid), np.round(300.*ht), returnxy=False,
checkbounds=False, order=0, masked=True)
if topodata is not None: # Drape over hillshade
#turn data into an RGBA image
cmap = palette
#adjust data so scaled between vmin and vmax and between 0 and 1
dat1 = dat_im.copy()
dat1[dat1 < vmin] = vmin
dat1[dat1 > vmax] = vmax
dat1 = (dat1 - vmin)/(vmax-vmin)
rgba_img = cmap(dat1)
maskvals = np.dstack((dat1.mask, dat1.mask, dat1.mask))
rgb = np.squeeze(rgba_img[:, :, 0:3])
rgb[maskvals] = 1.
draped_hsv = ls.blend_hsv(rgb, np.expand_dims(intensity, 2))
m.imshow(draped_hsv, zorder=3., interpolation='none')
# This is just a dummy layer that will be deleted to make the
# colorbar look right
panelhandle = m.imshow(dat_im, cmap=palette, zorder=0.,
vmin=vmin, vmax=vmax)
else:
panelhandle = m.imshow(dat_im, cmap=palette, zorder=3.,
vmin=vmin, vmax=vmax, interpolation='none')
#panelhandle = m.pcolormesh(x1, y1, dat, linewidth=0., cmap=palette, vmin=vmin, vmax=vmax, alpha=ALPHA, rasterized=True, zorder=2.);
#panelhandle.set_edgecolors('face')
# add colorbar
cbfmt = '%1.1f'
if vmax is not None and vmin is not None:
if (vmax - vmin) < 1.:
cbfmt = '%1.2f'
elif vmax > 5.: # (vmax - vmin) > len(clev):
cbfmt = '%1.0f'
#norm = mpl.colors.Normalize(vmin=vmin, vmax=vmax)
if scaletype.lower() == 'binned':
cbar = fig.colorbar(panelhandle, spacing='proportional',
ticks=clev, boundaries=clev, fraction=0.036,
pad=0.04, format=cbfmt, extend='both')
#cbar1 = ColorbarBase(cbar.ax, cmap=palette, norm=norm, spacing='proportional', ticks=clev, boundaries=clev, fraction=0.036, pad=0.04, format=cbfmt, extend='both', extendfrac='auto')
else:
cbar = fig.colorbar(panelhandle, fraction=0.036, pad=0.04,
extend='both', format=cbfmt)
#cbar1 = ColorbarBase(cbar.ax, cmap=palette, norm=norm, fraction=0.036, pad=0.04, extend='both', extendfrac='auto', format=cbfmt)
if topodata is not None:
panelhandle.remove()
cbar.set_label(label1, fontsize=10)
cbar.ax.tick_params(labelsize=8)
parallels = m.drawparallels(getMapLines(bymin, bymax, 3),
labels=[1, 0, 0, 0], linewidth=0.5,
labelstyle='+/-', fontsize=9, xoffset=-0.8,
color='gray', zorder=100.)
m.drawmeridians(getMapLines(bxmin, bxmax, 3), labels=[0, 0, 0, 1],
linewidth=0.5, labelstyle='+/-', fontsize=9,
color='gray', zorder=100.)
for par in parallels:
try:
parallels[par][1][0].set_rotation(90)
except:
pass
#draw roads on the map, if they were provided to us
if maproads is True and roadslist is not None:
try:
for road in roadslist:
try:
xy = list(road['geometry']['coordinates'])
roadx, roady = list(zip(*xy))
mapx, mapy = m(roadx, roady)
m.plot(mapx, mapy, roadcolor, lw=0.5, zorder=9)
except:
continue
except Exception as e:
print(('Failed to plot roads, %s' % e))
#add city names to map
if mapcities is True and cityfile is not None:
try:
fontname = 'Arial'
fontsize = 8
if k == 0: # Only need to choose cities first time and then apply to rest
fcities = bcities.limitByMapCollision(
m, fontname=fontname, fontsize=fontsize)
ctlats, ctlons, names = fcities.getCities()
cxis, cyis = m(ctlons, ctlats)
for ctlat, ctlon, cxi, cyi, name in zip(ctlats, ctlons, cxis, cyis, names):
m.scatter(ctlon, ctlat, c='k', latlon=True, marker='.',
zorder=100000)
ax.text(cxi, cyi, name, fontname=fontname,
fontsize=fontsize, zorder=100000)
except Exception as e:
print('Failed to plot cities, %s' % e)
#draw star at epicenter
plt.sca(ax)
if edict is not None:
elat, elon = edict['lat'], edict['lon']
ex, ey = m(elon, elat)
plt.plot(ex, ey, '*', markeredgecolor='k', mfc='None', mew=1.0,
ms=15, zorder=10000.)
m.drawmapboundary(fill_color=watercolor)
m.fillcontinents(color=clear_color, lake_color=watercolor)
m.drawrivers(color=watercolor)
##m.drawcoastlines()
#draw country boundaries
m.drawcountries(color=countrycolor, linewidth=1.0)
#add map scale
m.drawmapscale((bxmax+bxmin)/2., (bymin+(bymax-bymin)/9.), clon, clat, np.round((((bxmax-bxmin)*111)/5)/10.)*10, barstyle='fancy', zorder=10)
# Add border
autoAxis = ax.axis()
rec = Rectangle((autoAxis[0]-0.7, autoAxis[2]-0.2), (autoAxis[1]-autoAxis[0])+1, (autoAxis[3]-autoAxis[2])+0.4, fill=False, lw=1, zorder=1e8)
rec = ax.add_patch(rec)
rec.set_clip_on(False)
plt.draw()
if sref is not None:
label2 = '%s\nsource: %s' % (label1, sref) # '%s\n' % label1 + r'{\fontsize{10pt}{3em}\selectfont{}%s}' % sref #
else:
label2 = label1
plt.title(label2, axes=ax, fontsize=fontsizesub)
#draw scenario watermark, if scenario
if isScenario:
plt.sca(ax)
cx, cy = m(clon, clat)
plt.text(cx, cy, 'SCENARIO', rotation=45, alpha=0.10, size=72, ha='center', va='center', color='red')
#if ds: # Could add this to print "downsampled" on map
# plt.text()
if k == 1 and rowpan == 1:
# adjust single level plot
axsize = ax.get_window_extent().transformed(fig.dpi_scale_trans.inverted())
ht2 = axsize.height
fig.set_figheight(ht2*1.6)
else:
plt.tight_layout()
# Make room for suptitle - tight layout doesn't account for it
plt.subplots_adjust(top=0.92)
if printparam is True:
try:
fig = plt.gcf()
dictionary = grids['model']['description']['parameters']
paramstring = 'Model parameters: '
halfway = np.ceil(len(dictionary)/2.)
for i, key in enumerate(dictionary):
if i == halfway and colpan == 1:
paramstring += '\n'
paramstring += ('%s = %s; ' % (key, dictionary[key]))
print(paramstring)
fig.text(0.01, 0.015, paramstring, fontsize=fontsizesmallest)
plt.draw()
except:
print('Could not display model parameters')
if edict is not None:
eventid = edict['eventid']
else:
eventid = ''
time1 = datetime.datetime.utcnow().strftime('%d%b%Y_%H%M')
outfile = os.path.join(outfolder, '%s_%s_%s.pdf' % (eventid, suptitle, time1))
pngfile = os.path.join(outfolder, '%s_%s_%s.png' % (eventid, suptitle, time1))
if savepdf is True:
print('Saving map output to %s' % outfile)
plt.savefig(outfile, dpi=300)
if savepng is True:
print('Saving map output to %s' % pngfile)
plt.savefig(pngfile)
if showplots is True:
plt.show()
else:
plt.close(fig)
return newgrids
def plot_3d_wavefunction(
ax,
intensity,
phase,
cmap=mpl.cm.gist_rainbow,
azdeg=315,
altdeg=45,
satmin=0.25,
satmax=0.75,
alpha=0.9,
flatenning=5,
):
"""On axis ax, make a 3d colorplot of a wave function with intensity and
phase in numpy arrays.
Parameters
ax - axis on which to plot the image
intensity - Intensity of the complex wave function
phase - Phase of complex wave function
cmap - color map for plotting phase
azdeg - azimuthal angle of light source
altdeg - angle of light source from z axis
alpha - transperancy of plot
"""
from matplotlib.colors import LightSource
def renormalize(array, newmin=0, newmax=1):
oldmin, oldmax = [np.amin(array), np.amax(array)]
return (array - oldmin) * (newmax - newmin) / (oldmax - oldmin)
# Get grids of x and y
X, Y = np.meshgrid(*[np.arange(intensity.shape[i]) for i in [0, 1]])
# Height is intensity
Z = renormalize(intensity)
# Setup light source
ls = LightSource(azdeg, altdeg)
# rgb =
# C = rgb
C = ls.blend_hsv(
cmap(phase % (2 * np.pi) / (2 * np.pi)),
renormalize(intensity).reshape([*inten.shape, 1]),
hsv_min_sat=1.0,
hsv_max_sat=1.0,
hsv_min_val=satmin,
hsv_max_val=satmax,
)
ax.plot_surface(
X,
Y,
Z,
facecolors=C,
rcount=X.shape[1],
ccount=X.shape[0],
alpha=alpha,
shade=True,
)
ax.set_axis_off()
ax.set_zlim(0, flatenning)
