def plotTrace(paramY, paramX=None, label='', history=100, tracelen=None, name=None, group=None):
"""Plotting a trace.
Args:
:paramY(Record): The data for the ordinate is paramY.data.ravel()
Kwargs:
:paramX(Record): The data for the abscissa is paramX.data.ravel() if paramX is not None
:label(str): Label for param
:history(int): Length of history buffer
"""
if paramY is None:
return
if name is None:
name = "Trace(%s)" %paramY.name
if(not paramY.name in traces):
ipc.broadcast.init_data(name, data_type='vector', xlabel=label, history_length=history, group=group)
histograms[paramY.name] = True
if paramX is None:
ipc.new_data(name, data_y=paramY.data.ravel())
else:
x = paramX.data.ravel()
y = paramY.data.ravel()
if tracelen is not None:
x = x[:tracelen]
y = y[:tracelen]
if x.size != y.size:
logging.warning("For %s x- and y-dimension do not match (%i, %i). Cannot plot trace." % (name,x.size,y.size))
return
ipc.new_data(name, data_y=np.array([x,y], copy=False))
def plotHistogram(param, hmin=None, hmax=None, bins=100, label='', density=False, vline=None, history=10000, mask=None, log10=False, name_extension="", name=None, group=None):
"""Plotting a histogram.
Args:
:param(Record): The histogram is based on param.data.flat
Kwargs:
:hmin(float): Minimum, default = record.data.min()
:hmax(float): Maximum, default = record.data.max()
:bins(int): Nr. of bins, default = 100
:label(str): Label for param
:density(bool): If True, the integral of the histogram is 1
:history(int): Length of history buffer
"""
if param is None:
return
if name is None:
name = "Histogram(%s)%s" % (param.name, name_extension)
if(not param.name in histograms):
ipc.broadcast.init_data(name, data_type='vector', xlabel=label, vline=vline, history_length=history, group=group)
histograms[param.name] = True
data = param.data
if mask is not None:
data = data[mask]
if log10:
data=np.log10(data)
if hmin is None: hmin = data.min()
if hmax is None: hmax = data.max()
H,B = np.histogram(data.flat, range=(hmin, hmax), bins=bins, density=density)
ipc.new_data(name, H, xmin=B.min(), xmax=B.max(), vline=vline)
def plotTrace(paramY, paramX=None, label='', history=10000, tracelen=None, name=None, group=None):
"""Plotting a trace.
Args:
:paramY(Record): The data for the ordinate is paramY.data.ravel()
Kwargs:
:paramX(Record): The data for the abscissa is paramX.data.ravel() if paramX is not None
:label(str): Label for param
:history(int): Length of history buffer
"""
if paramY is None:
return
if name is None:
name = "Trace(%s)" %paramY.name
if(not paramY.name in traces):
ipc.broadcast.init_data(name, data_type='vector', xlabel=label, history_length=history, group=group)
histograms[paramY.name] = True
if paramX is None:
ipc.new_data(name, data_y=paramY.data.ravel())
else:
x = paramX.data.ravel()
y = paramY.data.ravel()
if tracelen is not None:
x = x[:tracelen]
y = y[:tracelen]
if x.size != y.size:
logging.warning("For %s x- and y-dimension do not match (%i, %i). Cannot plot trace." % (name,x.size,y.size))
return
ipc.new_data(name, data_y=np.array([x,y], copy=False))
def plotScatter(X,
Y,
name=None,
history=10000,
xlabel=None,
ylabel=None,
group=None):
"""Plotting the scatter of two parameters X and Y.
(No buffer in the backend).
Args:
:X(Record): An event parameter, e.g. injector position in x
:Y(Record): An event parameter, e.g. injector position in y
Kwargs:
:name(str): The key that appears in the interface (default = MeanMap(X.name, Y.name))
:xlabel(str):
:ylabel(str):
"""
if name is None:
name = "Scatter(%s,%s)" % (X.name, Y.name)
if (not name in _existingPlots):
if xlabel is None: xlabel = X.name
if ylabel is None: ylabel = Y.name
ipc.broadcast.init_data(name,
data_type='tuple',
history_length=history,
xlabel=xlabel,
ylabel=ylabel,
group=group)
_existingPlots[name] = True
ipc.new_data(name, np.array([X.data, Y.data]))
def plotMeanMap(X,
Y,
Z,
xmin=0,
xmax=10,
xbins=10,
ymin=0,
ymax=10,
ybins=10,
xlabel=None,
ylabel=None,
msg='',
dynamic_extent=False,
initial_reset=False,
name=None,
group=None):
"""Plotting the meanmap of Z as a function of two parameters X and Y.
(No buffer in the backend).
Args:
:X(Record,float): An event parameter, e.g. injector position in x
:Y(Record,float): An event parameter, e.g. injector position in y
:Z(Record,float): Some metric, e.g. hit rate, size, etc...
Kwargs:
:xmin(int): default = 0
:xmax(int): default = 10
:xbins(int): default = 10
:ymin(int): default = 0
:ymax(int): default = 10
:ybins(int): default = 10
:name(str): The key that appears in the interface (default = MeanMap(X.name, Y.name))
:xlabel(str):
:ylabel(str):
:msg(msg): Any message to be displayed in the plot
"""
if name is None:
name = "MeanMap(%s,%s,%s)" % (X.name, Y.name, Z.name)
if (not name in _existingPlots):
if xlabel is None: xlabel = X.name
if ylabel is None: ylabel = Y.name
ipc.broadcast.init_data(name,
data_type='triple',
history_length=1,
xmin=xmin,
xmax=xmax,
ymin=ymin,
ymax=ymax,
xbins=xbins,
ybins=ybins,
xlabel=xlabel,
ylabel=ylabel,
flipy=True,
dynamic_extent=dynamic_extent,
initial_reset=initial_reset,
group=group)
_existingPlots[name] = True
x = X if not isinstance(X, Record) else X.data
y = Y if not isinstance(Y, Record) else Y.data
z = Z if not isinstance(Z, Record) else Z.data
ipc.new_data(name, np.array([x, y, z]), msg=msg)
def plotNormalizedHistogram(value,
weight,
hmin=0,
hmax=10,
bins=10,
name=None,
group=None,
buffer_length=100):
if name is None:
if hasattr(value, "name"):
name = "Normalized histogram of {0}".format(value.name)
else:
name = "Normalize histogram"
if name not in normalized_histograms:
ipc.broadcast.init_data(name,
data_type='normalized_histogram',
history_length=buffer_length,
hmin=hmin,
hmax=hmax,
bins=bins,
group=group)
normalized_histograms[name] = True
value = value if not isinstance(value, Record) else value.data
weight = weight if not isinstance(weight, Record) else weight.data
ipc.new_data(name,
np.array([value, weight]),
data_type="normalized_histogram")
def plotScatterBg(X,
Y,
name=None,
history=10000,
xlabel=None,
ylabel=None,
bg_filename=None,
bg_xmin=0.,
bg_xmax=1.,
bg_ymin=0.,
bg_ymax=0.,
bg_angle=0.,
group=None):
"""Plotting the scatter of two parameters X and Y.
"""
if name is None:
name = "ScatterBg(%s,%s)" % (X.name, Y.name)
if (not name in _existingPlots):
if xlabel is None: xlabel = X.name
if ylabel is None: ylabel = Y.name
ipc.broadcast.init_data(name,
data_type='tuple',
history_length=history,
xlabel=xlabel,
ylabel=ylabel,
bg_filename=bg_filename,
bg_xmin=bg_xmin,
bg_xmax=bg_xmax,
bg_ymin=bg_ymin,
bg_ymax=bg_ymax,
bg_angle=bg_angle,
group=group)
_existingPlots[name] = True
ipc.new_data(name, np.array([X.data, Y.data]))
def plotMeanMap(X,Y,Z, xmin=0, xmax=10, xbins=10, ymin=0, ymax=10, ybins=10, plotid=None, xlabel=None, ylabel=None, msg=''):
"""Plotting the meanmap of Z as a function of two parameters X and Y.
(No buffer in the backend).
Args:
:X(Record,float): An event parameter, e.g. injector position in x
:Y(Record,float): An event parameter, e.g. injector position in y
:Z(Record,float): Some metric, e.g. hit rate, size, etc...
Kwargs:
:xmin(int): default = 0
:xmax(int): default = 10
:xbins(int): default = 10
:ymin(int): default = 0
:ymax(int): default = 10
:ybins(int): default = 10
:plotid(str): The key that appears in the interface (default = MeanMap(X.name, Y.name))
:xlabel(str):
:ylabel(str):
:msg(msg): Any message to be displayed in the plot
"""
if plotid is None:
plotid = "MeanMap(%s,%s)" %(X.name, Y.name)
if (not plotid in meanMaps):
if xlabel is None: xlabel = X.name
if ylabel is None: ylabel = Y.name
ipc.broadcast.init_data(plotid, data_type='triple', history_length=1,
xmin=xmin, xmax=xmax, ymin=ymin, ymax=ymax,
xbins=xbins, ybins=ybins,
xlabel=xlabel, ylabel=ylabel, flipy=True)
x = X if not isinstance(X, Record) else X.data
y = Y if not isinstance(Y, Record) else Y.data
z = Z if not isinstance(Z, Record) else Z.data
ipc.new_data(plotid, np.array([x, y, z]), msg=msg)
def plotScatterColor(X,Y,Z, plotid=None, history=100, xlabel=None, ylabel=None, zlabel=None, vmin=None, vmax=None):
"""Plotting the scatter of two parameters X and Y and use Z for color.
(No buffer in the backend).
Args:
:X(Record): An event parameter, e.g. injector position in x
:Y(Record): An event parameter, e.g. injector position in y
:Z(Record): An event parameter, e.g. injector position in z
Kwargs:
:plotid(str): The key that appears in the interface (default = MeanMap(X.name, Y.name))
:xlabel(str):
:ylabel(str):
"""
if plotid is None:
plotid = "ScatterColor(%s,%s)" %(X.name, Y.name)
if (not plotid in scatterPlots):
if xlabel is None: xlabel = X.name
if ylabel is None: ylabel = Y.name
if zlabel is None: zlabel = Z.name
if vmin is None: vmin = 0
if vmax is None: vmax = 1
ipc.broadcast.init_data(plotid, data_type='triple', history_length=history,
xlabel=xlabel, ylabel=ylabel, zlabel=zlabel,
vmin=vmin, vmax=vmax)
ipc.new_data(plotid, np.array([X.data, Y.data, Z.data]))
def plotScatter(X, Y, plotid=None, history=100, xlabel=None, ylabel=None):
"""Plotting the scatter of two parameters X and Y.
(No buffer in the backend).
Args:
:X(Record): An event parameter, e.g. injector position in x
:Y(Record): An event parameter, e.g. injector position in y
Kwargs:
:plotid(str): The key that appears in the interface (default = MeanMap(X.name, Y.name))
:xlabel(str):
:ylabel(str):
"""
if plotid is None:
plotid = "Scatter(%s,%s)" % (X.name, Y.name)
if (not plotid in scatterPlots):
if xlabel is None: xlabel = X.name
if ylabel is None: ylabel = Y.name
ipc.broadcast.init_data(
plotid,
data_type='tuple',
history_length=history,
xlabel=xlabel,
ylabel=ylabel)
ipc.new_data(plotid, np.array([X.data, Y.data]))
def plotImage(record, history=10, vmin=None, vmax=None, log=False, mask=None, msg=None, alert=False, name=None):
"""Plotting an image.
Args:
:record(Record): record.data is plotted as an image
Kwargs:
:history(int): Length of history buffer
:vmin(float): Minimum value
:vmax(float): Maximum value
:log(boolean): Plot image in log scale (needs restart of GUI, only works with grayscale colormap)
:mask(boolean or int): Multiply image with mask
"""
if record is None:
return
if name is None:
n = record.name
else:
n = name
if(not n in images):
ipc.broadcast.init_data(n, data_type='image', history_length=history, vmin=vmin, vmax=vmax, log=log)
images[n] = True
image = record.data
sh = image.shape
if (image.ndim == 3):
image = image.reshape(sh[0]*sh[2], sh[1])
if mask is None:
mask = np.ones_like(image)
ipc.new_data(n, image*mask, msg=msg, alert=alert)
def plotHistogram(param, hmin=None, hmax=None, bins=100, label='', density=False, vline=None, history=100, mask=None, log10=False, name_extension="", name=None, group=None):
"""Plotting a histogram.
Args:
:param(Record): The histogram is based on param.data.flat
Kwargs:
:hmin(float): Minimum, default = record.data.min()
:hmax(float): Maximum, default = record.data.max()
:bins(int): Nr. of bins, default = 100
:label(str): Label for param
:density(bool): If True, the integral of the histogram is 1
:history(int): Length of history buffer
"""
if param is None:
return
if name is None:
name = "Histogram(%s)%s" % (param.name, name_extension)
if(not param.name in histograms):
ipc.broadcast.init_data(name, data_type='vector', xlabel=label, vline=vline, history_length=history, group=group)
histograms[param.name] = True
data = param.data
if mask is not None:
data = data[mask]
if log10:
data=np.log10(data)
if hmin is None: hmin = data.min()
if hmax is None: hmax = data.max()
H,B = np.histogram(data.flat, range=(hmin, hmax), bins=bins, density=density)
ipc.new_data(name, H, xmin=B.min(), xmax=B.max(), vline=vline)
def plotHistogram(value, hmin=0, hmax=10, bins=10, name=None, group=None, buffer_length=100):
if name is None:
name = "Histogram of {0}".format(value.name)
if (name not in histograms):
ipc.broadcast.init_data(name, data_type='histogram', history_length=buffer_length,
hmin=hmin, hmax=hmax, bins=bins, group=group)
histograms[name] = True
value = value if not isinstance(value, Record) else value.data
ipc.new_data(name, value)
def plotTimestamp(timestamp, name=None, group=None, **kwargs):
if name is None:
name = "History(Fiducial)"
if not name in histories:
ipc.broadcast.init_data(name,
data_type='scalar',
group=group,
**kwargs)
ipc.new_data(name, timestamp.fiducials, **kwargs)
def plotMeanMapDynamic(X, Y, Z, norm=1., msg='', update=100, xmin=0, xmax=100, ymin=0, ymax=100, step=10, \
localRadius=100, overviewStep=100, xlabel=None, ylabel=None, name=None, group=None):
"""Plotting the mean of parameter Z as a function of parameters X and Y.
(Using a buffer in the backend).
Args:
:X(Record): An event parameter e.g. Motor position in X
:Y(Record): An event parameter e.g. Motor position in Y
:Z(Record): An event parameter e.g. Intensity
Kwargs:
:norm(int): Z is normalized by a given value, e.g. gmd (default = 1)
:msg (str): A message to be displayed in the plot
:update (int): After how many new data points, an update is send to the frontend (default = 100)
:xmin (int): (default = 0)
:xmax (int): (default = 100)
:ymin (int): (default = 0)
:ymax (int): (default = 100)
:step (int): The resolution of the map in units of x,y (default = 10)
:xlabel (str): (default = X.name)
:ylabel (str): (default = Y.name)
:localRadius (int): The radius of a square neighborehood around the current position (X.data, Y.data) (default = 100)
:overviewStep (int): The resolution of the overiew map (default = 100)
"""
if name is None:
name = "%s(%s,%s)" % (Z.name, X.name, Y.name)
if (not name in _existingPlots):
if xlabel is None: xlabel = X.name
if ylabel is None: ylabel = Y.name
_existingPlots[name] = _MeanMap(name,
xmin,
xmax,
ymin,
ymax,
step,
localRadius,
overviewStep,
xlabel,
ylabel,
group=group)
m = existingPlots[name]
m.append(X, Y, Z, norm)
if (not m.counter % update):
m.gatherSumsAndNorms()
m.gatherOverview()
if (ipc.mpi.is_main_worker()):
m.updateCenter(X, Y)
m.updateLocalLimits()
m.updateLocalMap()
m.updateOverviewMap(X, Y)
print m.localXmin, m.localXmax, m.localYmin, m.localYmax
ipc.new_data(name+' -> Local', m.localMap, msg=msg, \
xmin=m.localXmin, xmax=m.localXmax, ymin=m.localYmin, ymax=m.localYmax)
ipc.new_data(name + ' -> Overview', m.overviewMap)
def plotImage(record,
history=10,
vmin=None,
vmax=None,
log=False,
mask=None,
msg=None,
alert=False,
name=None,
group=None,
send_rate=None,
roi_center=None,
roi_diameters=None):
"""Plotting an image.
Args:
:record(Record): record.data is plotted as an image
Kwargs:
:history(int): Length of history buffer
:vmin(float): Minimum value
:vmax(float): Maximum value
:log(boolean): Plot image in log scale (needs restart of GUI, only works with grayscale colormap)
:mask(boolean or int): Multiply image with mask
"""
if record is None:
return
if name is None:
n = record.name
else:
n = name
if (not n in images):
ipc.broadcast.init_data(n,
data_type='image',
history_length=history,
vmin=vmin,
vmax=vmax,
log=log,
group=group)
images[n] = True
image = record.data
sh = image.shape
if (image.ndim == 3):
image = image.reshape(sh[0] * sh[2], sh[1])
if mask is None:
mask = np.ones_like(image)
ipc.new_data(n,
image * mask,
msg=msg,
alert=alert,
send_rate=send_rate,
center=roi_center,
diameters=roi_diameters)
def plotHistory(param,
label='',
history=100,
hline=None,
runningHistogram=False,
window=20,
bins=100,
hmin=0,
hmax=100,
name_extension="",
name=None,
group=None,
**kwargs):
"""Plotting history of a parameter.
Args:
:param(Record): The history is based on param.data
Kwargs:
:label(str): Label for param
:history(int): Length of history buffer
"""
if param is None:
return
if name is None:
name = "History(%s)%s" % (param.name, name_extension)
if (not param.name in histories):
if runningHistogram:
data_type = 'running_hist'
ipc.broadcast.init_data(name,
data_type=data_type,
ylabel=label,
history_length=history,
window=window,
bins=bins,
hmin=hmin,
hmax=hmax,
group=group,
**kwargs)
else:
data_type = 'scalar'
ipc.broadcast.init_data(name,
data_type=data_type,
ylabel=label,
history_length=history,
hline=hline,
group=group,
**kwargs)
histories[param.name] = True
ipc.new_data(name, param.data, hline=hline, **kwargs)
def plotHeatmap(X,
Y,
xmin=0,
xmax=1,
xbins=10,
ymin=0,
ymax=1,
ybins=10,
name=None,
group=None):
"""Plotting the heatmap of two parameters X and Y. Has been tested in MPI mode.
(Using a buffer in the backend).
Args:
:X(Record): An event parameter, e.g. hit rate
:Y(Record): An event parameter, e.g. some motor position
Kwargs:
:xmin(int): default = 0
:xmax(int): default = 1
:xbins(int): default = 10
:ymin(int): default = 0
:ymax(int): default = 1
:ybins(int): default = 10
"""
if name is None:
name = "Heatmap(%s,%s)" % (X.name, Y.name)
if not (name in _existingPlots):
# initiate (y, x) in 2D array to get correct orientation of image
_existingPlots[name] = np.zeros((ybins, xbins), dtype=int)
ipc.broadcast.init_data(name, data_type="image", group=group)
deltaX = (xmax - float(xmin)) / xbins
deltaY = (ymax - float(ymin)) / ybins
nx = np.ceil((X.data - xmin) / deltaX)
if (nx < 0):
nx = 0
elif (nx >= xbins):
nx = xbins - 1
ny = np.ceil((Y.data - ymin) / deltaY)
if (ny < 0):
ny = 0
elif (ny >= ybins):
ny = ybins - 1
# assign y to row and x to col in 2D array
_existingPlots[name][ny, nx] += 1
current_heatmap = np.copy(heatmaps[name])
ipc.mpi.sum(current_heatmap)
if ipc.mpi.is_main_event_reader():
ipc.new_data(name, current_heatmap[()])
def plotScatterBg(X,Y, name=None, history=100, xlabel=None, ylabel=None, bg_filename=None, bg_xmin=0., bg_xmax=1., bg_ymin=0., bg_ymax=0., bg_angle=0., group=None):
"""Plotting the scatter of two parameters X and Y.
"""
if name is None:
name = "ScatterBg(%s,%s)" %(X.name, Y.name)
if (not name in _existingPlots):
if xlabel is None: xlabel = X.name
if ylabel is None: ylabel = Y.name
ipc.broadcast.init_data(name, data_type='tuple', history_length=history,
xlabel=xlabel, ylabel=ylabel,
bg_filename=bg_filename,
bg_xmin=bg_xmin, bg_xmax=bg_xmax,
bg_ymin=bg_ymin, bg_ymax=bg_ymax,
bg_angle=bg_angle, group=group)
_existingPlots[name] = True
ipc.new_data(name, np.array([X.data, Y.data]))
def plotMeanMapDynamic(X, Y, Z, norm=1., msg='', update=100, xmin=0, xmax=100, ymin=0, ymax=100, step=10, \
localRadius=100, overviewStep=100, xlabel=None, ylabel=None, plotid=None):
"""Plotting the mean of parameter Z as a function of parameters X and Y.
(Using a buffer in the backend).
Args:
:X(Record): An event parameter e.g. Motor position in X
:Y(Record): An event parameter e.g. Motor position in Y
:Z(Record): An event parameter e.g. Intensity
Kwargs:
:norm(int): Z is normalized by a given value, e.g. gmd (default = 1)
:msg (str): A message to be displayed in the plot
:update (int): After how many new data points, an update is send to the frontend (default = 100)
:xmin (int): (default = 0)
:xmax (int): (default = 100)
:ymin (int): (default = 0)
:ymax (int): (default = 100)
:step (int): The resolution of the map in units of x,y (default = 10)
:xlabel (str): (default = X.name)
:ylabel (str): (default = Y.name)
:localRadius (int): The radius of a square neighborehood around the current position (X.data, Y.data) (default = 100)
:overviewStep (int): The resolution of the overiew map (default = 100)
"""
if plotid is None:
plotid = "%s(%s,%s)" % (Z.name, X.name, Y.name)
if (not plotid in meanMaps):
if xlabel is None: xlabel = X.name
if ylabel is None: ylabel = Y.name
meanMaps[plotid] = _MeanMap(plotid, xmin, xmax, ymin, ymax, step,
localRadius, overviewStep, xlabel, ylabel)
m = meanMaps[plotid]
m.append(X, Y, Z, norm)
if (not m.counter % update):
m.gatherSumsAndNorms()
m.gatherOverview()
if (ipc.mpi.is_main_worker()):
m.updateCenter(X, Y)
m.updateLocalLimits()
m.updateLocalMap()
m.updateOverviewMap(X, Y)
print m.localXmin, m.localXmax, m.localYmin, m.localYmax
ipc.new_data(plotid+' -> Local', m.localMap, msg=msg, \
xmin=m.localXmin, xmax=m.localXmax, ymin=m.localYmin, ymax=m.localYmax)
ipc.new_data(plotid + ' -> Overview', m.overviewMap)
def plotCorrelation(X, Y, history=100):
"""Plotting the correlation of two parameters X and Y over time.
(Using a buffer in the backend).
Args:
:X(Record): An event parameter, e.g. hit rate
:Y(Record): An event parameter, e.g. some motor position
Kwargs:
:history(int): Buffer length
"""
plotid = "Corr(%s,%s)" %(X.name, Y.name)
if (not plotid in correlations):
ipc.broadcast.init_data(plotid, history_length=100)
correlations[plotid] = True
x,y = (X.data, Y.data)
xArray.append(x)
yArray.append(y)
correlation = x*y/(np.mean(xArray)*np.mean(yArray))
ipc.new_data(plotid, correlation)
def plotCorrelation(X, Y, history=10000, name=None, group=None):
"""Plotting the correlation of two parameters X and Y over time.
(Using a buffer in the backend).
Args:
:X(Record): An event parameter, e.g. hit rate
:Y(Record): An event parameter, e.g. some motor position
Kwargs:
:history(int): Buffer length
"""
if name is None:
name = "Corr(%s,%s)" % (X.name, Y.name)
if (not name in _existingPlots):
ipc.broadcast.init_data(name, history_length=100, group=group)
_existingPlots[name] = True
x, y = (X.data, Y.data)
xArray.append(x)
yArray.append(y)
correlation = x * y / (np.mean(xArray) * np.mean(yArray))
ipc.new_data(name, correlation)
def plotCorrelation(X, Y, history=100, name=None, group=None):
"""Plotting the correlation of two parameters X and Y over time.
(Using a buffer in the backend).
Args:
:X(Record): An event parameter, e.g. hit rate
:Y(Record): An event parameter, e.g. some motor position
Kwargs:
:history(int): Buffer length
"""
if name is None:
name = "Corr(%s,%s)" %(X.name, Y.name)
if (not name in _existingPlots):
ipc.broadcast.init_data(name, history_length=100, group=group)
_existingPlots[name] = True
x,y = (X.data, Y.data)
xArray.append(x)
yArray.append(y)
correlation = x*y/(np.mean(xArray)*np.mean(yArray))
ipc.new_data(name, correlation)
def plotHistogram(value,
hmin=0,
hmax=10,
bins=10,
name=None,
group=None,
buffer_length=100):
if name is None:
name = "Histogram of {0}".format(value.name)
if (name not in histograms):
ipc.broadcast.init_data(name,
data_type='histogram',
history_length=buffer_length,
hmin=hmin,
hmax=hmax,
bins=bins,
group=group)
histograms[name] = True
value = value if not isinstance(value, Record) else value.data
ipc.new_data(name, value)
def plotHeatmap(X, Y, xmin=0, xmax=1, xbins=10, ymin=0, ymax=1, ybins=10, name=None, group=None):
"""Plotting the heatmap of two parameters X and Y. Has been tested in MPI mode.
(Using a buffer in the backend).
Args:
:X(Record): An event parameter, e.g. hit rate
:Y(Record): An event parameter, e.g. some motor position
Kwargs:
:xmin(int): default = 0
:xmax(int): default = 1
:xbins(int): default = 10
:ymin(int): default = 0
:ymax(int): default = 1
:ybins(int): default = 10
"""
if name is None:
name = "Heatmap(%s,%s)" %(X.name, Y.name)
if not(name in _existingPlots):
# initiate (y, x) in 2D array to get correct orientation of image
_existingPlots[name] = np.zeros((ybins, xbins), dtype=int)
ipc.broadcast.init_data(name, data_type="image", group=group)
deltaX = (xmax - float(xmin))/xbins
deltaY = (ymax - float(ymin))/ybins
nx = np.ceil((X.data - xmin)/deltaX)
if (nx < 0):
nx = 0
elif (nx >= xbins):
nx = xbins - 1
ny = np.ceil((Y.data - ymin)/deltaY)
if (ny < 0):
ny = 0
elif (ny >= ybins):
ny = ybins - 1
# assign y to row and x to col in 2D array
_existingPlots[name][ny, nx] += 1
current_heatmap = np.copy(heatmaps[name])
ipc.mpi.sum(current_heatmap)
if ipc.mpi.is_main_event_reader():
ipc.new_data(name, current_heatmap[()])
def plotHistory(param, label='', history=100, runningHistogram=False, window=20, bins=100, hmin=0, hmax=100, name_extension=""):
"""Plotting history of a parameter.
Args:
:param(Record): The history is based on param.data
Kwargs:
:label(str): Label for param
:history(int): Length of history buffer
"""
if param is None:
return
plotid = "History(%s)%s" % (param.name, name_extension)
if (not param.name in histories):
if runningHistogram:
data_type = 'running_hist'
ipc.broadcast.init_data(plotid, data_type=data_type, ylabel=label, history_length=history, window=window, bins=bins, hmin=hmin, hmax=hmax)
else:
data_type = 'scalar'
ipc.broadcast.init_data(plotid, data_type=data_type, ylabel=label, history_length=history)
histories[param.name] = True
ipc.new_data(plotid, param.data)
def plotTimestamp(timestamp, name=None, group=None, **kwargs):
if name is None:
name = "History(Fiducial)"
if not name in histories:
ipc.broadcast.init_data(name, data_type='scalar', group=group, **kwargs)
ipc.new_data(name, timestamp.fiducials, **kwargs)
def onEvent(evt):
# Processing rate
analysis.event.printProcessingRate()
# Detector statistics
analysis.pixel_detector.printStatistics(evt["photonPixelDetectors"])
# Count Nr. of Photons
analysis.pixel_detector.totalNrPhotons(evt, "photonPixelDetectors", "CCD")
plotting.line.plotHistory(evt["analysis"]["nrPhotons - CCD"], label="Nr of photons / frame", history=50)
# Simple hitfinding (Count Nr. of lit pixels)
analysis.hitfinding.countLitPixels(
evt, "photonPixelDetectors", "CCD", aduThreshold=25, hitscoreThreshold=1000, mask=mask
)
# Compute the hitrate
analysis.hitfinding.hitrate(evt, evt["analysis"]["isHit - CCD"], history=10000)
# Plot the hitscore
plotting.line.plotHistory(evt["analysis"]["hitscore - CCD"], label="Nr. of lit pixels")
# Plot the hitrate
plotting.line.plotHistory(evt["analysis"]["hitrate"], label="Hit rate [%]")
# Plot injector position in x
plotting.line.plotHistory(evt["parameters"]["injector_posx"], label="Position in X [um]")
# Plot injector position in y
plotting.line.plotHistory(evt["parameters"]["injector_posz"], label="Position in Z [um]")
# Perform sizing on hits
if not evt["analysis"]["isHit - CCD"]:
D.append(evt["photonPixelDetectors"]["CCD"].data)
global iD
global nD
global bg
iD += 1
if (iD % nD) == 0:
print "Calculating new background average."
bg = numpy.array(D).mean(axis=0)
ipc.new_data("background average", bg)
else:
print "It's a hit"
if mode in ["default", "radial"]:
ccd_type = "photonPixelDetectors"
ccd_key = "CCD"
m = mask
elif mode == "bin":
t0 = time.time()
analysis.pixel_detector.bin(evt, "photonPixelDetectors", "CCD", binning, mask=mask)
t_bin = time.time() - t0
binnedMask = evt["analysis"]["binned mask - CCD"].data
ccd_type = "analysis"
ccd_key = "binned image - CCD"
m = binnedMask
t0 = time.time()
# Find the center of diffraction
analysis.sizing.findCenter(evt, ccd_type, ccd_key, mask=m, maxshift=40 / binning, threshold=13, blur=4)
t_center = time.time() - t0
if mode in ["default", "bin"]:
# Fitting sphere model to get size and intensity
t0 = time.time()
analysis.sizing.fitSphere(evt, ccd_type, ccd_key, mask=m, **dict(modelParams, **sizingParams))
t_size = time.time() - t0
elif mode == "radial":
# Calculate radial average
t0 = time.time()
cx = evt["analysis"]["offCenterX"].data + (sim.nx - 1) / 2.0
cy = evt["analysis"]["offCenterY"].data + (sim.ny - 1) / 2.0
analysis.pixel_detector.radial(evt, ccd_type, ccd_key, mask=m, cx=cx, cy=cy)
# Fitting sphere model to get size and intensity
analysis.sizing.fitSphereRadial(
evt, "analysis", "radial distance - CCD", "radial average - CCD", **dict(modelParams, **sizingParams)
)
t_size = time.time() - t0
# Fitting model
analysis.sizing.sphereModel(
evt, "analysis", "offCenterX", "offCenterY", "diameter", "intensity", m.shape, poisson=True, **modelParams
)
if mode == "radial":
analysis.pixel_detector.radial(evt, "analysis", "fit", mask=m, cx=cx, cy=cy)
# 1D arrays have to have same length, otherwise histoty keeping gets messed up
rlen = 100
ipc.new_data(
"radial fit",
numpy.array(
[
evt["analysis"]["radial distance - fit"].data.ravel()[:rlen],
evt["analysis"]["radial average - fit"].data.ravel()[:rlen],
],
copy=False,
),
)
ipc.new_data(
"radial CCD",
numpy.array(
[
evt["analysis"]["radial distance - CCD"].data.ravel()[:rlen],
evt["analysis"]["radial average - CCD"].data.ravel()[:rlen],
],
copy=False,
),
msg="BLA",
)
t_all = t_center + t_size
print "Time: %e sec (center / size : %.2f%% / %.2f%%)" % (
t_all,
100.0 * t_center / t_all,
100.0 * t_size / t_all,
)
# Error parameter
diff = (evt[ccd_type][ccd_key].data - evt["analysis"]["fit"].data)[m].sum() / float(
evt[ccd_type][ccd_key].data[m].sum()
)
ipc.new_data("fit diff", diff)
ipc.new_data("fit error", evt["analysis"]["fit error"].data)
ipc.new_data("fit mean error 1", evt["analysis"]["fit mean error 1"].data)
ipc.new_data("fit mean error 3", evt["analysis"]["fit mean error 3"].data)
plotting.line.plotHistory(evt["analysis"]["offCenterX"])
plotting.line.plotHistory(evt["analysis"]["offCenterY"])
plotting.line.plotHistory(evt["analysis"]["diameter"])
plotting.line.plotHistory(evt["analysis"]["intensity"])
# plotting.line.plotHistory(evt["analysis"]["error"])
# Attach a message to the plots
s0 = evt["analysis"]["diameter"].data
I0 = evt["analysis"]["intensity"].data
msg_glo = "diameter = %.2f nm, \nintensity = %.2f mJ/um2" % (s0, I0)
msg_fit = "Fit result: \ndiameter = %.2f nm, \nintensity = %.2f mJ/um2 \nerr=%f, merr1=%f, merr3=%f" % (
s0,
I0,
evt["analysis"]["fit error"].data,
evt["analysis"]["fit mean error 1"].data / evt["analysis"]["radial average - CCD"].data.sum(),
evt["analysis"]["fit mean error 3"].data,
)
# Plot the glorious shots
plotting.image.plotImage(evt[ccd_type][ccd_key], msg=msg_glo, log=True, mask=m)
# Plot the fitted model
plotting.image.plotImage(evt["analysis"]["fit"], msg=msg_fit, log=True, mask=m)
# Plot heatmap of injector pos in x vs. diameter
plotting.correlation.plotHeatmap(
evt["parameters"]["injector_posx"], evt["analysis"]["diameter"], **heatmapInjector
)
# Plot heatmap of center positions
plotting.correlation.plotHeatmap(
evt["analysis"]["offCenterX"], evt["analysis"]["offCenterY"], **heatmapCenterPos
)
def plotTimestamp(timestamp):
ipc.new_data('History(Fiducial)', timestamp.fiducials)
def onEvent(evt):
analysis.event.printProcessingRate()
ipc.new_data("TOF", evt["ionTOFs"]["tof"].data)
if numpy.random.randint(100) == 0:
time.sleep(1)
