def show_indexed_timeseries_plotly(timeseries: TimeSeries,
istart=None,
istop=None,
fig: go.FigureWidget = None,
col=None,
row=None,
zero_start=False,
xlabel='time (s)',
ylabel=None,
title=None,
neurodata_vis_spec=None,
**kwargs):
if ylabel is None and timeseries.unit:
ylabel = timeseries.unit
tt = get_timeseries_tt(timeseries, istart=istart, istop=istop)
if zero_start:
tt = tt - tt[0]
data, unit = get_timeseries_in_units(timeseries,
istart=istart,
istop=istop)
trace_kwargs = dict()
if col is not None or row is not None:
trace_kwargs.update(row=row, col=col)
fig.add_trace(x=tt, y=data, **trace_kwargs, **kwargs)
layout_kwargs = dict(xaxis_title=xlabel)
if ylabel is not None:
layout_kwargs.update(yaxis_title=ylabel)
if title is not None:
layout_kwargs.update(title=title)
fig.update_layout(**layout_kwargs)
def update(
self,
index: int,
start_label: str = "start_time",
before: float = 0.0,
after: float = 1.0,
order=None,
group_inds=None,
labels=None,
align_to_zero=False,
fig: go.FigureWidget = None,
):
data, time_ts_aligned = self.align_data(start_label, before, after,
index)
if group_inds is None:
group_inds = np.zeros(len(self.trials), dtype=np.int)
if align_to_zero:
for trial_no in order:
data_zero_id = bisect(time_ts_aligned[trial_no], 0)
data[trial_no] -= data[trial_no][data_zero_id]
fig = fig if fig is not None else go.FigureWidget()
fig.data = []
fig.layout = {}
return self.plot_group(group_inds, data, time_ts_aligned, fig, order)
def addMesh(fig: go.FigureWidget,
faces,
props=dict(color='white',
opacity=1.0,
facecolor=None,
lighting=None)):
(verts, facesIndices) = getVertsAndFaces(faces)
xVert, yVert, zVert, _ = np.array(verts).T
xFace, yFace, zFace = np.array(facesIndices).T
return fig.add_mesh3d(x=xVert,
y=yVert,
z=zVert,
i=xFace,
j=yFace,
k=zFace,
lighting=props["lighting"],
facecolor=props["facecolor"],
color=props["color"],
opacity=props["opacity"])
def createFigureWidget(self):
x_id = self.dimensions[0]
d_val = self.data[x_id]
if hasattr(self.data[x_id], 'codes'):
d_val = self.data[x_id].codes
hist, bin_edges = np.histogram(d_val.flatten(), bins='auto')
xedges = []
for i in range(len(bin_edges) - 1):
xedges.append((bin_edges[i + 1] + bin_edges[i]) / 2)
traces = []
trace = {
'type': "bar",
'name': self.data.label,
'marker': {
'color': 'rgba(0, 0, 0, 0.1)'
},
'x': xedges,
'y': hist,
}
if self.only_subsets == False:
traces.append(trace)
for sset in self.data.subsets:
if hasattr(sset, "disabled") == False or sset.disabled == False:
s_val = sset[x_id]
if hasattr(sset[x_id], 'codes'):
s_val = sset[x_id].codes
hist, bin_edges2 = np.histogram(
s_val.flatten(),
range=(bin_edges[0], bin_edges[len(bin_edges) - 1]),
bins=(len(bin_edges) - 1))
color = sset.style.color
trace = {
'type': "bar",
'name': sset.label,
'marker': {
'color': color
},
'x': xedges,
'y': hist,
}
traces.append(trace)
layout = {
'title': self.options['title'].value,
'margin': {
'l': 50,
'r': 0,
'b': 50,
't': 30
},
'xaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['xaxis'].value,
'linecolor': self.data.get_component(x_id).color,
'tickcolor': self.data.get_component(x_id).color,
'ticklen': 4,
'linewidth': 4,
},
'yaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['yaxis'].value,
},
'legend': {
'orientation': self.margins['legend_orientation'].value,
'x': self.margins['legend_xpos'].value,
'y': self.margins['legend_ypos'].value
},
'showlegend': self.margins['showlegend'].value,
'barmode': 'overlay',
}
return FigureWidget({'data': traces, 'layout': layout})
def createFigureWidget(self, x_id, y_id_list):
traces = []
alpha_min, alpha_max, alpha_delta = self.getDeltaFunction(
len(y_id_list))
alpha_val = alpha_max
self.polyfun = {}
for step, y_id in enumerate(y_id_list):
z = np.polyfit(self.data[x_id].flatten().astype('float'),
self.data[y_id].flatten().astype('float'),
self.options['fit_degree'].value)
f = np.poly1d(z)
color = "#444444"
color = 'rgba' + str(self.getDeltaColor(color, alpha_val, step))
trace = {
'type':
"scattergl",
'mode':
"markers",
'name':
self.data.label + "_" + y_id,
'marker':
dict({
'symbol': 'circle',
'size': self.options['marker_size'].value,
'color': color,
'line': {
'width': self.options['line_width'].value,
'color': color
}
}),
'x':
self.data[x_id].flatten(),
'y':
self.data[y_id].flatten(),
}
if self.only_subsets == False:
traces.append(trace)
x_new = self.data[x_id].flatten().astype('float').tolist()
x_new.sort()
y_new = f(x_new)
trace = {
'type': "scattergl",
'mode': "lines",
'name': self.data.label + "_fit_" + y_id,
'line': {
'width': self.options['line_width'].value,
'color': color
},
'x': x_new,
'y': y_new,
'showlegend': False
}
if self.only_subsets == False:
traces.append(trace)
alpha_val = alpha_val - alpha_delta
self.polyfun[y_id] = f
for sset in self.data.subsets:
if hasattr(sset, "disabled") == False or sset.disabled == False:
alpha_val = alpha_max
for step, y_id in enumerate(y_id_list):
z = np.polyfit(sset[x_id].flatten().astype('float'),
sset[y_id].flatten().astype('float'),
self.options['fit_degree'].value)
f = np.poly1d(z)
color = sset.style.color
color = 'rgba' + str(
self.getDeltaColor(color, alpha_val, step))
trace = {
'type':
"scattergl",
'mode':
"markers",
'name':
sset.label + "_" + y_id,
'marker':
dict({
'symbol': 'circle',
'size': self.options['marker_size'].value,
'color': color,
'line': {
'width': self.options['line_width'].value,
'color': color
}
}),
'x':
sset[x_id].flatten(),
'y':
sset[y_id].flatten(),
}
traces.append(trace)
x_new = sset[x_id].flatten().astype('float').tolist()
x_new.sort()
y_new = f(x_new)
trace = {
'type': "scattergl",
'mode': "lines",
'name': sset.label + "_fit_" + y_id,
'line': {
'width': self.options['line_width'].value,
'color': color
},
'x': x_new,
'y': y_new,
'showlegend': False
}
traces.append(trace)
alpha_val = alpha_val - alpha_delta
self.polyfun[sset.label + "_" + y_id] = f
y_color = 'rgb(0,0,0)'
if len(y_id_list) == 1:
y_color = self.data.get_component(y_id_list[0]).color
layout = {
'title': self.options['title'].value,
'margin': {
'l': 50,
'r': 0,
'b': 50,
't': 30
},
'xaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['xaxis'].value,
'linecolor': self.data.get_component(x_id).color,
'tickcolor': self.data.get_component(x_id).color,
'ticklen': 4,
'linewidth': 4,
},
'yaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['yaxis'].value,
'linecolor': y_color,
'tickcolor': y_color,
'ticklen': 4,
'linewidth': 4,
},
'showlegend': self.margins['showlegend'].value,
'legend': {
'orientation': self.margins['legend_orientation'].value,
'x': self.margins['legend_xpos'].value,
'y': self.margins['legend_ypos'].value
},
}
return FigureWidget({'data': traces, 'layout': layout})
def _init_widgets(self):
"""Initializes all configuration widgets. Possible image config parameters are:"""
layout = Layout(width="200px", max_width="200px")
self._alpha = FloatSlider(
value=1,
min=0,
max=1.0,
step=0.1,
description="alpha:",
description_tooltip="Overlay image alpha level (0 to 1).",
disabled=False,
continuous_update=True,
orientation="horizontal",
readout=True,
readout_format=".1f",
layout=layout,
)
self._lut = Dropdown(
options=PapayaConfigWidget.lut_options,
value="Red Overlay",
description="lut:",
description_tooltip="The color table name.",
layout=layout,
)
self._nlut = Dropdown(
options=PapayaConfigWidget.lut_options,
value="Red Overlay",
description="negative-lut:",
description_tooltip=
"The color table name used by the negative side of the parametric pair.",
layout=layout,
)
self._min = FloatText(
value=None,
description="min:",
description_tooltip="The display range minimum.",
step=0.01,
continuous_update=True,
disabled=False,
layout=layout,
)
self._minp = BoundedFloatText(
value=None,
min=0,
max=100,
step=1,
continuous_update=True,
description="min %:",
description_tooltip=
"The display range minimum as a percentage of image max.",
disabled=False,
layout=layout,
)
self._max = FloatText(
value=None,
description="max:",
description_tooltip="The display range maximum.",
step=0.01,
continuous_update=True,
disabled=False,
layout=layout,
)
self._maxp = BoundedFloatText(
value=None,
min=0,
max=100,
step=1,
continuous_update=True,
description="max %:",
description_tooltip=
"The display range minimum as a percentage of image max.",
disabled=False,
layout=layout,
)
self._sym = Checkbox(
value=False,
description="symmetric",
description_tooltip=
"When selected, sets the negative range of a parametric pair to the same size as the positive range.",
disabled=False,
layout=layout,
)
# figure to display histogram of image data
fig = Figure()
fig.update_layout(
height=300,
margin=dict(l=15, t=15, b=15, r=15, pad=4),
showlegend=True,
legend_orientation="h",
)
self._hist = FigureWidget(fig)
self._hist.add_trace(
Histogram(x=[], name="All image data", visible="legendonly"))
self._hist.add_trace(Histogram(x=[], name="Image data without 0s"))
self._handlers = defaultdict()
def update(
self,
index: int,
start_label: str = "start_time",
before: float = 0.0,
after: float = 1.0,
order=None,
group_inds=None,
labels=None,
align_to_zero=False,
sem=False,
fig: go.FigureWidget = None,
):
data, time_ts_aligned = self.align_data(start_label, before, after,
index)
if group_inds is None:
group_inds = np.zeros(len(self.trials), dtype=np.int)
if align_to_zero:
for trial_no in order:
data_zero_id = bisect(time_ts_aligned[trial_no], 0)
data[trial_no] -= data[trial_no][data_zero_id]
fig = go.FigureWidget() if fig is None else fig
fig.data = []
fig.layout = {}
if sem:
group_stats = []
for group in np.unique(group_inds):
this_mean = np.nanmean(data[group_inds == group, :], axis=0)
err = scipy.stats.sem(data[group_inds == group, :],
axis=0,
nan_policy="omit")
group_stats.append(
dict(
mean=this_mean,
lower=this_mean - 2 * err,
upper=this_mean + 2 * err,
group=group,
))
for stats in group_stats:
plot_kwargs = dict()
color = color_wheel[stats["group"]]
if labels is not None:
plot_kwargs.update(text=labels[stats["group"]])
fig.add_scattergl(x=time_ts_aligned[0],
y=stats["lower"],
line_color=color)
fig.add_scattergl(x=time_ts_aligned[0],
y=stats["upper"],
line_color=color,
fill='tonexty',
opacity=0.2)
fig.add_scattergl(x=time_ts_aligned[0],
y=stats["mean"],
line_color=color,
**plot_kwargs)
else:
fig = self.plot_group(group_inds, data, time_ts_aligned, fig,
order)
return fig
def player_strength_by_horizon(player_eps: DF, players_gw_eps: DF, horizon: str, position: str, team: str, player: str, ctx: Context):
"""
Returns a plotly chart with expected points as the y-axis and cost on the x-axis for a specific time horizon. This chart can be displayed in the Jupyter notebook.
Args:
player_eps: The data frame with data to chart.
horizon: The time horizon of the chart, e.g. Next GW, Next 8 GWs, etc.
Returns:
The plotly chart.
"""
def if_in_cols(df: DF, col: str, other):
return df[col].fillna(other) if col in df.columns else other
def in_team_trace(player_eps: DF) -> Scatter:
return Scatter(
x=player_eps['Current Cost'],
y=player_eps['Expected Points ' + horizon],
mode='markers',
marker={'size': MAX_POINT_SIZE, 'color': 'white', 'line': {'width': 1}},
name='In Team',
text=player_eps['Label'])
def position_traces(player_eps: DF, ctx: Context) -> list:
def trace(player_eps: DF, position: str) -> Scatter:
return Scatter(
x=player_eps['Current Cost'],
y=player_eps['Expected Points ' + horizon],
name=position,
mode='markers',
marker={'color': POSITION_COLORS[position],
'opacity': player_eps['Opacity'],
'size': player_eps['Size']},
text=player_eps['Label'])
return [trace(player_eps[player_eps['Field Position'] == position], position) for position in ctx.position_by_type.values()]
def event_capture_trace(player_eps: DF) -> Scatter:
return Scatter(
x=player_eps['Current Cost'],
y=player_eps['Expected Points ' + horizon],
mode='markers',
opacity=0,
marker={'size': player_eps['Size'], 'color': 'white', 'opacity': 1.0},
name='Player',
text=player_eps['Label'])
def player_clicked(trace, points, selector):
message.value = ''
try:
player = None
player_code = None
for ind in points.point_inds:
player = player_eps_chart.iloc[ind]
player_code = player_eps_chart.index[ind]
if player is not None:
player_gw_eps = players_gw_eps.loc[player_code]
detail.children = tuple([display_player(player, player_gw_eps, horizon, ctx)])
else:
detail.children = tuple([])
# Make sure exceptions are displayed in footer because they are swallowed otherwise.
except Exception as e:
message.value = traceback.format_exc()
def break_text(text: str, max_num_per_line: int, sep=','):
lines = []
items = text.split(sep)
while len(items) > 0:
lines += [sep.join(items[:max_num_per_line])]
items = items[max_num_per_line:]
return (sep + '<br>').join(lines)
def get_player_eps_chart(player_eps: DF) -> DF:
return (player_eps[
['Name', 'Name and Short Team', 'Long Name', 'Team Name', 'Field Position', 'Current Cost', 'Total Points', 'Minutes Percent',
'News And Date', 'Minutes Played', 'Total Points Recent Fixtures', 'Avg Total Points Recent Fixtures', 'Avg ICT Index Recent Fixtures',
'Avg Influence Recent Fixtures', 'Avg Creativity Recent Fixtures', 'Avg Threat Recent Fixtures', 'ICT Index', 'Influence', 'Creativity', 'Threat',
'Chance Avail This GW', 'Stats Completeness', 'Stats Completeness Percent', 'Profile Picture', 'Team Last Updated', 'Player Last Updated']
+ [col for col in player_eps.columns if col.startswith('Expected Points ') or col.startswith('Fixtures ')]
+ (['In Team?'] if 'In Team?' in player_eps.columns else [])]
# Add visualisation columns that need to be calculated on the unfiltered set.
.assign(**{'Opacity': lambda df: if_in_cols(df, 'Stats Completeness Percent', 100) * (1 - MIN_OPACITY) / 100 + MIN_OPACITY})
.assign(**{'Size': lambda df: np.maximum(np.where(df['Field Position'] != 'GK',
df.groupby('Field Position')['Avg Threat Recent Fixtures'].transform(lambda x: x / x.max() * MAX_POINT_SIZE),
df['Avg Influence Recent Fixtures'] / df['Avg Influence Recent Fixtures'].max() * MAX_POINT_SIZE), MIN_POINT_SIZE)}))
def get_player_eps_formatted(player_eps_chart: DF) -> DF:
return (player_eps_chart
.pipe(ctx.dd.format)
[player_eps_chart.columns]
.astype(str))
def get_labels(player_eps_chart: DF) -> S:
player_eps_formatted = player_eps_chart.pipe(get_player_eps_formatted)
return (player_eps_formatted['Name and Short Team']
+ ', ' + player_eps_formatted['Field Position']
+ ', Cost: ' + player_eps_formatted['Current Cost']
+ '<br>Exp. Points: ' + player_eps_formatted[f'Expected Points {horizon}']
+ ', Total Points: ' + player_eps_formatted['Total Points']
+ '<br>Minutes Percent: ' + player_eps_formatted['Minutes Percent']
+ f', Stats Completeness (Recent {ctx.player_fixtures_look_back} Fixtures): ' + player_eps_formatted['Stats Completeness']
+ np.where(player_eps_formatted['Field Position'] != 'GK', f'<br>Average Threat (Recent {ctx.player_fixtures_look_back} Fixtures): ' + player_eps_formatted['Avg Threat Recent Fixtures'],
f'<br>Avg Influence (Recent {ctx.player_fixtures_look_back} Fixtures): ' + player_eps_formatted['Avg Influence Recent Fixtures'])
+ ', ICT: ' + player_eps_formatted['ICT Index']
+ '<br>Next: ' + player_eps_formatted[f'Fixtures Next 8 GWs'].map(lambda v: break_text(v, 4))
+ '<br>News: ' + player_eps_formatted['News And Date']
)
player_eps_chart = player_eps.pipe(get_player_eps_chart)
pad = 0.5
min_cost, max_cost = (player_eps_chart['Current Cost'].min() - pad, player_eps_chart['Current Cost'].max() + pad)
min_eps, max_eps = (player_eps_chart[f'Expected Points {horizon}'].min() - pad, player_eps_chart[f'Expected Points {horizon}'].max() + pad)
last_updated = player_eps_chart['Player Last Updated'].min()
if position != SEL_ALL:
player_eps_chart = player_eps_chart[lambda df: df['Field Position'] == position]
if team != SEL_ALL:
player_eps_chart = player_eps_chart[lambda df: df['Team Name'] == team]
if player is not None and player != '':
player_eps_chart = player_eps_chart[lambda df: df['Name'].str.lower().str.contains(player.lower())]
# Add labels
player_eps_chart = player_eps_chart.assign(**{'Label': lambda df: df.pipe(get_labels)})
traces = []
if 'In Team?' in player_eps_chart.columns:
traces += [in_team_trace(player_eps_chart[lambda df: df['In Team?'] == True])]
traces += position_traces(player_eps_chart, ctx)
traces += [event_capture_trace(player_eps_chart)]
chart = FigureWidget(
traces,
layout={
'title': f'Expected Points from Game Week {ctx.next_gw} for {horizon}' if horizon != 'Next GW' else f'Expected Points for Game Week {ctx.next_gw}',
'xaxis': dict(title='Current Cost (lower is better)', showspikes=True, range=[min_cost, max_cost]),
'yaxis': dict(title=f'Expected Points {horizon} (higher is better)', showspikes=True, range=[min_eps, max_eps]),
'hovermode': 'closest',
'legend': dict(itemsizing='constant')
},
)
# Register the click event handler. Unfortunately, this cannot be done on the scatter trace itself at creation time for some reason.
for trace in chart.data:
trace.on_click(player_clicked, True)
data_quality = HTML(f'''<p><center>Color: field position -
Size: threat (recent fixtures) for non-goalies, influence (recent fixtures) for goalies -
Opacity: stats completeness (recent fixtures)<br>
Data last updated: {last_updated.strftime("%d %b %Y %H:%M:%S")}</center></p>''')
message = HTML('<p><center>Click on player for more detail!</center></p>')
out = Output()
detail = VBox([])
chart_and_detail = VBox([message, out, chart, data_quality, detail])
return chart_and_detail
def createFigureWidget(self):
dimensions = self.dimensions
data_lines = []
i=0
if self.only_subsets == False:
colors = [i for r in range(self.data.size)]
colorscale = [[0,'#EEEEEE']]
else:
colors = []
colorscale = []
i=-1
for sset in self.data.subsets:
if hasattr(sset,"disabled") == False or sset.disabled == False:
i = i+1
tmplist = [i for r in range(len(sset.to_index_list()))]
colors.extend(tmplist)
colorscale.append([i,sset.style.color])
t_color = len(colorscale)
if (t_color > 1):
for c in colorscale:
c[0] = c[0]/(t_color-1)
else:
colorscale = [[0,'#EEEEEE'], [1,'#EEEEEE']]
traces = []
for dimension in dimensions:
line={}
if hasattr(self.data[dimension].flatten(), 'codes'):
line['values'] = self.data[dimension].flatten().codes.tolist()
tickvals, tickmask = np.unique(self.data[dimension].flatten().codes, return_index=True)
ticktext = self.data[dimension][tickmask]
line['tickvals'] = tickvals.tolist()
line['ticktext'] = ticktext.tolist()
else:
line['values'] = self.data[dimension].flatten().tolist()
line['label'] = dimension
for sset in self.data.subsets:
if hasattr(sset,"disabled") == False or sset.disabled == False:
if hasattr(sset[dimension].flatten(), 'codes'):
tmplist = sset[dimension].codes.tolist()
else:
tmplist = sset[dimension].tolist()
line['values'].extend(tmplist)
data_lines.append(line);
trace = {
'type' : 'parcoords',
'line' : {
'color' : colors,
'colorscale' : colorscale
},
'dimensions' : data_lines,
}
traces.append(trace)
for sset in self.data.subsets:
if hasattr(sset,"disabled") == False or sset.disabled == False:
color = sset.style.color
trace = {
'type': "scatter",
'name' : sset.label,
'textposition' : 'middle right',
'x' : [-1000],
'y' : [i],
'mode' : 'markers',
'marker': {
'color' : color,
'size' : 20,
'line': {
'width': 1,
'color' : 'light grey'
},
'symbol' : 'square'
}
}
traces.append(trace)
layout = {
'title' : self.options['title'].value,
'xaxis': {
'title' : self.options['xaxis'].value,
'range' : [0,1],
'showgrid':False,
'showline':False,
'showticklabels':False,
'zeroline':False
},
'yaxis': {
'title' : self.options['yaxis'].value,
'range' : [0,1],
'showgrid':False,
'showline':False,
'showticklabels':False,
'zeroline':False
},
'showlegend': self.margins['showlegend'].value,
'legend' : {
'orientation' : self.margins['legend_orientation'].value,
'x' : self.margins['legend_xpos'].value,
'y' : self.margins['legend_ypos'].value
},
}
return FigureWidget(data = traces, layout = layout)
def createFigureWidget(self, x_id, y_id, z_id):
traces = []
color = "#444444"
color = 'rgba' + str(self.getDeltaColor(color, .8))
trace = {
'type':
"scatter3d",
'mode':
"markers",
'name':
self.data.label,
'marker':
dict({
'symbol': 'circle',
'size': self.options['marker_size'].value,
'color': color,
}),
'x':
self.data[x_id].flatten(),
'y':
self.data[y_id].flatten(),
'z':
self.data[z_id].flatten(),
}
if self.only_subsets == False:
traces.append(trace)
for sset in self.data.subsets:
if hasattr(sset, "disabled") == False or sset.disabled == False:
color = sset.style.color
color = 'rgba' + str(self.getDeltaColor(color, .8))
trace = {
'type':
"scatter3d",
'mode':
"markers",
'name':
sset.label,
'marker':
dict({
'symbol': 'circle',
'size': self.options['marker_size'].value,
'color': color,
}),
'x':
sset[x_id].flatten(),
'y':
sset[y_id].flatten(),
'z':
sset[z_id].flatten(),
}
traces.append(trace)
layout = {
'margin': {
'l': 0,
'r': 0,
'b': 0,
't': 30
},
'scene': {
'xaxis': {
'title': self.options['xaxis'].value,
'linecolor': self.data.get_component(x_id).color,
'tickcolor': self.data.get_component(x_id).color,
'ticklen': 4,
'linewidth': 4,
},
'yaxis': {
'title': self.options['yaxis'].value,
'linecolor': self.data.get_component(y_id).color,
'tickcolor': self.data.get_component(y_id).color,
'ticklen': 4,
'linewidth': 4,
},
'zaxis': {
'title': self.options['zaxis'].value,
'linecolor': self.data.get_component(z_id).color,
'tickcolor': self.data.get_component(z_id).color,
'ticklen': 4,
'linewidth': 4,
}
},
'showlegend': self.margins['showlegend'].value,
'legend': {
'orientation': self.margins['legend_orientation'].value,
'x': self.margins['legend_xpos'].value,
'y': self.margins['legend_ypos'].value
}
}
return FigureWidget({'data': traces, 'layout': layout})
def createFigureWidget(self):
traces = []
Xpca = []
for x_id in self.dimensions:
if hasattr(self.data[x_id].flatten(), 'codes'):
Xpca.append(self.data[x_id].flatten().codes.tolist())
else:
Xpca.append(self.data[x_id].flatten())
Xpca = np.array(Xpca).transpose()
self.pca.fit(Xpca)
Xt = self.pca.transform(Xpca)
data_l = [self.data[col_id].astype(str) for col_id in self.dimensions]
data_l = zip(*data_l)
t_val = [' - '.join( dat ) for dat in data_l]
x_val = Xt[:,0]
y_val = Xt[:,1]
color = "#444444"
trace = {
'type': "scattergl", 'mode': "markers", 'name': self.data.label,
'marker': dict({
'symbol':'circle', 'size': self.options['marker_size'].value, 'color': color,
'line' : { 'width' : self.options['line_width'].value, 'color' : color }
}),
'x': x_val.tolist(),
'y': y_val.tolist(),
'text': t_val,
}
#print(trace)
if self.only_subsets == False:
traces.append(trace)
for sset in self.data.subsets:
if hasattr(sset,"disabled") == False or sset.disabled == False:
sset_mask = sset.to_mask()
color = sset.style.color
trace = {
'type': "scattergl", 'mode': "markers", 'name': sset.label,
'marker': dict({
'symbol':'circle', 'size': self.focused_size_marker, 'color': color,
'line' : { 'width' : self.options['line_width'].value, 'color' : color}
}),
'x': (x_val[sset_mask]).tolist(),
'y': (y_val[sset_mask]).tolist(),
}
traces.append(trace)
layout = {
'title' : self.options['title'].value,
'margin' : {
'l':self.margins['left'].value,
'r':self.margins['right'].value,
'b':self.margins['bottom'].value,
't':self.margins['top'].value
},
'xaxis': { 'autorange' : True, 'zeroline': True,
'title' : self.options['xaxis'].value,
'type' : self.options['xscale'].value,
},
'yaxis': { 'autorange':True, 'zeroline': True,
'title' : self.options['yaxis'].value,
'type' : self.options['yscale'].value
},
'showlegend': self.margins['showlegend'].value,
'legend' : {
'orientation' : self.margins['legend_orientation'].value,
'x' : self.margins['legend_xpos'].value,
'y' : self.margins['legend_ypos'].value
}
}
return FigureWidget({
'data': traces,
'layout': layout
})
def createFigureWidget(self):
traces = []
params = []
for param in self.dimensions:
if hasattr(self.data[param].flatten(), 'codes'):
params.append(self.data[param].flatten().codes.tolist())
else:
params.append(self.data[param].flatten())
self.mat = np.corrcoef(params)
self.mat = np.nan_to_num(self.mat)
self.mat = np.flip(self.mat, axis=1)
trace = {
'type': 'heatmap',
'x': self.dimensions[::-1],
'y': self.dimensions,
'z': self.mat,
'colorscale': 'RdBu',
'autocolorscale': False,
'reversescale': True,
'zauto': False,
'zmin': -1,
'zmax': 1,
}
layout = {
'title': self.options['title'].value,
'margin': {
'l': self.margins['left'].value,
'r': self.margins['right'].value,
'b': self.margins['bottom'].value,
't': self.margins['top'].value
},
'showlegend': True,
'xaxis': {
'side': 'top'
},
'annotations': []
}
if self.only_subsets == False:
traces.append(trace)
for i in range(len(self.dimensions)):
for j in range(len(self.dimensions)):
currentValue = self.mat[i][j]
if abs(currentValue) > 0.6:
textColor = 'white'
else:
textColor = 'black'
result = {
'xref': 'x1',
'yref': 'y1',
'x': j,
'y': i,
'text': round(currentValue, 2),
'showarrow': False,
'font': {
'color': textColor
}
}
layout['annotations'].append(result)
for sset in self.data.subsets:
if hasattr(sset, "disabled") == False or sset.disabled == False:
params = []
for param in self.dimensions:
if hasattr(sset[param].flatten(), 'codes'):
params.append(sset[param].flatten().codes.tolist())
else:
params.append(sset[param].flatten())
mat = np.corrcoef(params)
mat = np.nan_to_num(mat, 1)
mat = np.flip(mat, axis=1)
trace = {
'type': 'heatmap',
'x': self.dimensions[::-1],
'y': self.dimensions,
'z': mat,
'colorscale': 'RdBu',
'autocolorscale': False,
'reversescale': True,
'zauto': False,
'zmin': -1,
'zmax': 1,
}
traces.append(trace)
for i in range(len(self.dimensions)):
for j in range(len(self.dimensions)):
currentValue = mat[i][j]
if abs(currentValue) > 0.6:
textColor = 'white'
else:
textColor = 'black'
result = {
'xref': 'x1',
'yref': 'y1',
'x': j,
'y': i,
'text': round(currentValue, 2),
'showarrow': False,
'font': {
'color': textColor
}
}
layout['annotations'].append(result)
return FigureWidget({'data': traces, 'layout': layout})
def createFigureWidget(self):
x_id = self.dimensions[0]
y_id_list = [
self.dimensions[i] for i in range(1, len(self.dimensions))
]
d_val = self.data[x_id]
if hasattr(self.data[x_id], 'codes'):
d_val = self.data[x_id].codes.flatten()
hist, bin_edges = np.histogram(d_val.flatten(), bins='auto')
bin_list = bin_edges.searchsorted(d_val.flatten(), 'right')
xedges = []
for i in range(len(bin_edges) - 1):
xedges.append((bin_edges[i + 1] + bin_edges[i]) / 2)
traces = []
alpha_min, alpha_max, alpha_delta = self.getDeltaFunction(
len(y_id_list), 0.1, 0.5)
alpha_val = alpha_max
for y_id in y_id_list:
d_val = self.data[y_id]
if hasattr(self.data[y_id], 'codes'):
d_val = self.data[y_id].codes
y_mean = []
y_std_u = []
y_std_l = []
y_min = []
y_max = []
for i in range(1, len(bin_edges)):
d_col = d_val[(bin_list == i)].flatten()
if len(d_col) > 0:
i_mean = d_col.mean()
i_min = d_col.min()
i_max = d_col.max()
i_std = d_col.std()
else:
i_mean = 0
i_std = 0
i_min = 0
i_max = 0
y_mean.append(i_mean)
y_std_u.append(i_mean + i_std)
y_std_l.append(i_mean - i_std)
y_min.append(i_min)
y_max.append(i_max)
color = "#444444"
color = 'rgba' + str(self.getDeltaColor(color, alpha_val))
trace = {
'type': "scatter",
'name': '-',
'marker': {
'color': color
},
'line': {
'width': 0
},
'x': xedges,
'y': y_min,
'mode': 'lines',
'showlegend': False
}
if self.only_subsets == False:
traces.append(trace)
trace = {
'type': "scatter",
'name': self.data.label + "_" + y_id,
'marker': {
'color': color
},
'fillcolor': color,
'fill': 'tonexty',
'line': {
'color': color
},
'x': xedges,
'y': y_mean,
'mode': 'lines',
}
if self.only_subsets == False:
traces.append(trace)
trace = {
'type': "scatter",
'name': '+',
'marker': {
'color': color
},
'fillcolor': color,
'fill': 'tonexty',
'line': {
'width': 0
},
'x': xedges,
'y': y_max,
'mode': 'lines',
'showlegend': False
}
if self.only_subsets == False:
traces.append(trace)
alpha_val = alpha_val - alpha_delta
for sset in self.data.subsets:
if hasattr(sset, "disabled") == False or sset.disabled == False:
s_val = sset[x_id]
if hasattr(sset[x_id], 'codes'):
s_val = sset[x_id].codes
bin_list = bin_edges.searchsorted(s_val, 'right')
alpha_val = alpha_max
for y_id in y_id_list:
s_val = sset[y_id]
if hasattr(sset[y_id], 'codes'):
s_val = sset[y_id].codes
y_s_mean = []
y_s_std_u = []
y_s_std_l = []
for i in range(1, len(bin_edges)):
s_col = s_val[(bin_list == i)]
if len(s_col) > 0:
i_mean = s_col.mean()
i_std = s_col.std()
i_min = s_col.min()
i_max = s_col.max()
else:
i_mean = 0
i_std = 0
i_min = 0
i_max = 0
y_s_mean.append(i_mean)
y_s_std_u.append(i_mean + i_std)
y_s_std_l.append(i_mean - i_std)
color = sset.style.color
color = 'rgba' + str(self.getDeltaColor(color, alpha_val))
trace = {
'type': "scatter",
'name': '-',
'marker': {
'color': color
},
'line': {
'width': 0
},
'x': xedges,
'y': y_s_std_l,
'mode': 'lines',
'showlegend': False
}
traces.append(trace)
trace = {
'type': "scatter",
'name': sset.label + "_" + y_id,
'marker': {
'color': color
},
'fillcolor': color,
'fill': 'tonexty',
'line': {
'color': color
},
'x': xedges,
'y': y_s_mean,
'mode': 'lines',
}
traces.append(trace)
trace = {
'type': "scatter",
'name': '+',
'marker': {
'color': color
},
'fillcolor': color,
'fill': 'tonexty',
'line': {
'width': 0
},
'x': xedges,
'y': y_s_std_u,
'mode': 'lines',
'showlegend': False
}
traces.append(trace)
alpha_val = alpha_val - alpha_delta
y_color = 'rgb(0,0,0)'
if len(y_id_list) == 1:
y_color = self.data.get_component(y_id_list[0]).color
layout = {
'title': self.options['title'].value,
'margin': {
'l': 50,
'r': 0,
'b': 50,
't': 30
},
'xaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['xaxis'].value,
'linecolor': self.data.get_component(x_id).color,
'tickcolor': self.data.get_component(x_id).color,
'ticklen': 4,
'linewidth': 4,
},
'yaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['yaxis'].value,
'linecolor': y_color,
'tickcolor': y_color,
'ticklen': 4,
'linewidth': 4,
},
'showlegend': self.margins['showlegend'].value,
'barmode': 'overlay',
'legend': {
'orientation': self.margins['legend_orientation'].value,
'x': self.margins['legend_xpos'].value,
'y': self.margins['legend_ypos'].value
},
}
return FigureWidget({'data': traces, 'layout': layout})
class PapayaConfigWidget(VBox):
"""A widget that displays widgets to adjust NLPapayaViewer image parameters."""
lut_options = [
"Grayscale",
"Red Overlay",
"Green Overlay",
"Blue Overlay",
"Gold",
"Spectrum",
"Overlay (Positives)",
"Overlay (Negatives)",
]
def __init__(self, viewer, *args, **kwargs):
"""
Parameters
----------
viewer: NlPapayaViewer
associated viewer.
"""
super().__init__(*args, **kwargs)
self._viewer = viewer
self._init_widgets()
self.children = [
VBox([
VBox(
[self._hist],
layout=Layout(
height="auto",
margin="0px 0px 0px 0px",
padding="5px 5px 5px 5px",
),
),
VBox(
[
self._alpha,
self._lut,
self._nlut,
self._min,
self._minp,
self._max,
self._maxp,
self._sym,
],
layout=Layout(width="230px"),
),
])
]
def _init_widgets(self):
"""Initializes all configuration widgets. Possible image config parameters are:"""
layout = Layout(width="200px", max_width="200px")
self._alpha = FloatSlider(
value=1,
min=0,
max=1.0,
step=0.1,
description="alpha:",
description_tooltip="Overlay image alpha level (0 to 1).",
disabled=False,
continuous_update=True,
orientation="horizontal",
readout=True,
readout_format=".1f",
layout=layout,
)
self._lut = Dropdown(
options=PapayaConfigWidget.lut_options,
value="Red Overlay",
description="lut:",
description_tooltip="The color table name.",
layout=layout,
)
self._nlut = Dropdown(
options=PapayaConfigWidget.lut_options,
value="Red Overlay",
description="negative-lut:",
description_tooltip=
"The color table name used by the negative side of the parametric pair.",
layout=layout,
)
self._min = FloatText(
value=None,
description="min:",
description_tooltip="The display range minimum.",
step=0.01,
continuous_update=True,
disabled=False,
layout=layout,
)
self._minp = BoundedFloatText(
value=None,
min=0,
max=100,
step=1,
continuous_update=True,
description="min %:",
description_tooltip=
"The display range minimum as a percentage of image max.",
disabled=False,
layout=layout,
)
self._max = FloatText(
value=None,
description="max:",
description_tooltip="The display range maximum.",
step=0.01,
continuous_update=True,
disabled=False,
layout=layout,
)
self._maxp = BoundedFloatText(
value=None,
min=0,
max=100,
step=1,
continuous_update=True,
description="max %:",
description_tooltip=
"The display range minimum as a percentage of image max.",
disabled=False,
layout=layout,
)
self._sym = Checkbox(
value=False,
description="symmetric",
description_tooltip=
"When selected, sets the negative range of a parametric pair to the same size as the positive range.",
disabled=False,
layout=layout,
)
# figure to display histogram of image data
fig = Figure()
fig.update_layout(
height=300,
margin=dict(l=15, t=15, b=15, r=15, pad=4),
showlegend=True,
legend_orientation="h",
)
self._hist = FigureWidget(fig)
self._hist.add_trace(
Histogram(x=[], name="All image data", visible="legendonly"))
self._hist.add_trace(Histogram(x=[], name="Image data without 0s"))
self._handlers = defaultdict()
def _set_values(self, config, range, data):
"""Sets config values from the specified `config` and creates histogram for `data`.
Parameters
----------
config : dict
configuration parameters for the image. Possible keywords are:
alpha : int
the overlay image alpha level (0 to 1).
lut : str
the color table name.
negative_lut : str
the color table name used by the negative side of the parametric pair.
max : int
the display range maximum.
maxPercent : int
the display range maximum as a percentage of image max.
min : int
the display range minimum.
minPercent : int
the display range minimum as a percentage of image min.
symmetric : bool
if true, sets the negative range of a parametric pair to the same size as the positive range.
range: float
range of image values.
data: []
flattened image data.
"""
self._alpha.value = config.get("alpha", 1)
self._lut.value = config.get("lut", PapayaConfigWidget.lut_options[1])
self._nlut.value = config.get("negative_lut",
PapayaConfigWidget.lut_options[1])
self._min.value = config.get("min", 0)
self._minp.value = self._get_per_from_value(range,
config.get("min", 0))
self._max.value = config.get("max", 0.1)
self._maxp.value = self._get_per_from_value(range,
config.get("max", 0.1))
self._sym.value = config.get("symmetric", "false") == "true"
# set histogram data
self._hist.data[0].x = data
# leave out 0 values
self._hist.data[1].x = [] if (data == []
or data is None) else data[data != 0]
def _add_handlers(self, image):
"""Add config widget event handlers to change the config values for the specified `image`.
Parameters
----------
image: neurolang_ipywidgets.PapayaImage
image whose config values will be viewed/modified using this config widget.
"""
# Dropdown does not support resetting event handlers after Dropdown.unobserve_all is called
# So handlers are stored to be removed individually
# github issue https://github.com/jupyter-widgets/ipywidgets/issues/1868
self._handlers["alpha"] = partial(self._config_changed,
image=image,
name="alpha")
self._handlers["lut"] = partial(self._config_changed,
image=image,
name="lut")
self._handlers["nlut"] = partial(self._config_changed,
image=image,
name="negative_lut")
self._handlers["min"] = partial(self._config_changed,
image=image,
name="min")
self._handlers["minp"] = partial(self._set_min_max,
image=image,
name="minPercent")
self._handlers["max"] = partial(self._config_changed,
image=image,
name="max")
self._handlers["maxp"] = partial(self._set_min_max,
image=image,
name="maxPercent")
self._handlers["sym"] = partial(self._config_bool_changed,
image=image,
name="symmetric")
self._alpha.observe(self._handlers["alpha"], names="value")
self._lut.observe(self._handlers["lut"], names="value")
self._nlut.observe(self._handlers["nlut"], names="value")
self._min.observe(self._handlers["min"], names="value")
self._minp.observe(self._handlers["minp"], names="value")
self._max.observe(self._handlers["max"], names="value")
self._maxp.observe(self._handlers["maxp"], names="value")
self._sym.observe(self._handlers["sym"], names="value")
def _remove_handlers(self):
"""Removes all event handlers set for the config widgets."""
if len(self._handlers):
self._alpha.unobserve(self._handlers["alpha"], names="value")
self._lut.unobserve(self._handlers["lut"], names="value")
self._nlut.unobserve(self._handlers["nlut"], names="value")
self._min.unobserve(self._handlers["min"], names="value")
self._minp.unobserve(self._handlers["minp"], names="value")
self._max.unobserve(self._handlers["max"], names="value")
self._maxp.unobserve(self._handlers["maxp"], names="value")
self._sym.unobserve(self._handlers["sym"], names="value")
self._handlers = defaultdict()
@debounce(0.5)
def _config_changed(self, change, image, name):
if name == "min":
self._minp.unobserve(self._handlers["minp"], names="value")
self._minp.value = self._get_per_from_value(
image.range, change.new)
self._minp.observe(self._handlers["minp"], names="value")
elif name == "max":
self._maxp.unobserve(self._handlers["maxp"], names="value")
self._maxp.value = self._get_per_from_value(
image.range, change.new)
self._maxp.observe(self._handlers["maxp"], names="value")
self._set_config(image, name, change.new)
@debounce(0.5)
def _set_min_max(self, change, image, name):
if name == "minPercent":
self._min.unobserve(self._handlers["min"], names="value")
self._min.value = self._get_value_from_per(image.range, change.new)
self._set_config(image, "min", self._min.value)
self._min.observe(self._handlers["min"], names="value")
elif name == "maxPercent":
self._max.unobserve(self._handlers["max"], names="value")
self._max.value = self._get_value_from_per(image.range, change.new)
self._set_config(image, "max", self._max.value)
self._max.observe(self._handlers["max"], names="value")
def _config_bool_changed(self, change, image, name):
value = "false"
if change.new:
value = "true"
self._set_config(image, name, value)
def _set_config(self, image, key, value):
image.config[key] = value
self._viewer.set_images()
def _get_per_from_value(self, range, value):
return round(value * 100 / range, 0)
def _get_value_from_per(self, range, per):
return round(per * range / 100, 2)
def set_image(self, image):
"""Sets the image whose config values will be viewed/modified using this config widget.
If image is `None`, all config values are reset.
Parameters
----------
image: neurolang_ipywidgets.PapayaImage
image whose config values will be viewed/modified using this config widget.
"""
if image:
self._remove_handlers()
self._set_values(image.config, image.range,
image.image.get_fdata().flatten())
self._add_handlers(image)
else:
self.reset()
def reset(self):
"""Resets values for all config widgets."""
self._remove_handlers()
self._set_values({}, 100, [])
self.layout.visibility = "hidden"
def createFigureWidget(self, x_id, y_id_list):
traces = []
alpha_min, alpha_max, alpha_delta = self.getDeltaFunction(len(y_id_list))
alpha_val = alpha_max
x_values = self.data[x_id].flatten()
x_sort = x_values.argsort()
for y_id in y_id_list:
y_values = self.data[y_id].flatten()
color = "#444444"
color = 'rgba'+str(self.getDeltaColor(color, alpha_val))
trace = {
'type': "scattergl", 'mode': self.options['marker_type'].value, 'name': self.data.label + "_" + y_id,
'line' : { 'width' : self.options['line_width'].value, 'color' : color },
'x': x_values[x_sort],
'y': y_values[x_sort],
}
if self.only_subsets == False:
traces.append(trace)
alpha_val = alpha_val - alpha_delta
for sset in self.data.subsets:
if hasattr(sset,"disabled") == False or sset.disabled == False:
alpha_val = alpha_max
x_values = sset[x_id].flatten()
x_sort = x_values.argsort()
for step, y_id in enumerate(y_id_list):
y_values = sset[y_id].flatten()
color = sset.style.color
color = 'rgba'+str(self.getDeltaColor(color, alpha_val, step))
trace = {
'type': "scattergl", 'mode': self.options['marker_type'].value, 'name': sset.label + "_" + y_id,
'line' : { 'width' : self.options['line_width'].value, 'color' : color},
'x': x_values[x_sort],
'y': y_values[x_sort],
}
traces.append(trace)
alpha_val = alpha_val - alpha_delta
y_color = 'rgb(0,0,0)'
if len(y_id_list) == 1:
y_color = self.data.get_component(y_id_list[0]).color
layout = {
'margin' : {'l':50,'r':0,'b':50,'t':30 },
'xaxis': { 'autorange' : True, 'zeroline': True,
'title' : self.options['xaxis'].value,
'type' : self.options['xscale'].value ,
'linecolor' : self.data.get_component(x_id).color,
'tickcolor' : self.data.get_component(x_id).color,
'ticklen' : 4,
'linewidth' : 4,
},
'yaxis': { 'autorange':True, 'zeroline': True,
'title' : self.options['yaxis'].value,
'type' : self.options['yscale'].value,
'linecolor' : y_color,
'tickcolor' : y_color,
'ticklen' : 4,
'linewidth' : 4,
},
'showlegend': self.margins['showlegend'].value,
'legend' : {
'orientation' : self.margins['legend_orientation'].value,
'x' : self.margins['legend_xpos'].value,
'y' : self.margins['legend_ypos'].value
},
}
return FigureWidget({
'data': traces,
'layout': layout
})
def createFigureWidget(self, x_id, y_id):
traces = []
color = "#444444"
df = self.data.to_dataframe()
df_merge = df.merge(df, on=y_id)
dfo = pd.crosstab(df_merge[x_id + '_x'], df_merge[x_id + '_y'])
idx = dfo.columns.union(dfo.index)
dfo = dfo.reindex(index=idx, columns=idx, fill_value=0)
self.G = nx.convert_matrix.from_pandas_adjacency(dfo)
if self.options['layout_type'].value == 'bipartite_layout':
pos = nx.bipartite_layout(self.G)
elif self.options['layout_type'].value == 'circular_layout':
pos = nx.circular_layout(self.G)
elif self.options['layout_type'].value == 'kamada_kawai_layout':
pos = nx.kamada_kawai_layout(self.G)
elif self.options['layout_type'].value == 'planar_layout':
pos = nx.planar_layout(self.G)
elif self.options['layout_type'].value == 'random_layout':
pos = nx.random_layout(self.G)
elif self.options['layout_type'].value == 'shell_layout':
pos = nx.shell_layout(self.G)
elif self.options['layout_type'].value == 'spectral_layout':
pos = nx.spectral_layout(self.G)
elif self.options['layout_type'].value == 'shell_layout':
pos = nx.shell_layout(self.G)
else:
pos = nx.spring_layout(
self.G,
k=self.options['layout_k'].value,
iterations=self.options['layout_iter'].value)
degree = dict(nx.degree(self.G))
n_x = [0 for i in range(len(self.G.nodes()))]
n_y = [0 for i in range(len(self.G.nodes()))]
n_t = ['' for i in range(len(self.G.nodes()))]
n_s = [0 for i in range(len(self.G.nodes()))]
for i, node in enumerate(self.G.nodes()):
n_x[i] = pos[node][0]
n_y[i] = pos[node][1]
n_t[i] = node
n_s[i] = degree[node]
max_degree = max(n_s)
min_degree = min(n_s)
n_s = [
math.ceil(
(n - min_degree) / (max_degree - min_degree) *
self.focused_size_marker * self.options['marker_size'].value) +
self.options['marker_size'].value for n in n_s
]
node_trace = {
'type': "scatter",
'name': self.data.label + "_" + x_id,
'x': n_x,
'y': n_y,
'text': n_t,
'mode': 'markers',
'hoverinfo': 'text',
'marker': {
'size': n_s,
'color': self.data.get_component(x_id).color,
'line': {
'width': self.options['line_width'].value,
'color': self.data.get_component(x_id).color,
}
},
'showlegend': True,
}
e_x = [0 for i in range(2 * len(self.G.edges()))]
e_y = [0 for i in range(2 * len(self.G.edges()))]
for i, edge in enumerate(self.G.edges()):
e_x[2 * i] = pos[edge[0]][0]
e_y[2 * i] = pos[edge[0]][1]
e_x[2 * i + 1] = pos[edge[1]][0]
e_y[2 * i + 1] = pos[edge[1]][1]
color = self.data.get_component(y_id).color
color = re.match(r"rgb\((\d+),\s*(\d+),\s*(\d+)\)", color)
color = (int(color[1]), int(color[2]), int(color[3]))
color = '#%02x%02x%02x' % color
color = 'rgba' + str(self.getDeltaColor(color, 0.2))
edge_trace = {
'type': "scatter",
'name': self.data.label + "_" + y_id,
'x': e_x,
'y': e_y,
'line': {
'width': self.options['line_width'].value,
'color': color,
},
'hoverinfo': 'none',
'mode': 'lines',
'showlegend': True,
}
if self.only_subsets == False:
traces.append(edge_trace)
for sset in self.data.subsets:
if hasattr(sset, "disabled") == False or sset.disabled == False:
color = sset.style.color
dfs = df[sset.to_mask()]
df_merge = dfs.merge(dfs, on=y_id)
dfs = pd.crosstab(df_merge[x_id + '_x'], df_merge[x_id + '_y'])
idx = dfs.columns.union(dfs.index)
dfs = dfs.reindex(index=idx, columns=idx, fill_value=0)
G = nx.convert_matrix.from_pandas_adjacency(dfs)
e_x = [0 for i in range(2 * len(G.edges()))]
e_y = [0 for i in range(2 * len(G.edges()))]
for i, edge in enumerate(G.edges()):
e_x[2 * i] = pos[edge[0]][0]
e_y[2 * i] = pos[edge[0]][1]
e_x[2 * i + 1] = pos[edge[1]][0]
e_y[2 * i + 1] = pos[edge[1]][1]
trace = {
'type': "scatter",
'name': str(sset.label) + "_" + str(y_id),
'x': e_x,
'y': e_y,
'line': {
'width': self.options['line_width'].value,
'color': color
},
'hoverinfo': 'none',
'mode': 'lines',
'showlegend': True,
}
traces.append(trace)
traces.append(node_trace)
layout = {
'title': self.options['title'].value,
'margin': {
'l': self.margins['left'].value,
'r': self.margins['right'].value,
'b': self.margins['bottom'].value,
't': self.margins['top'].value
},
'xaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['xaxis'].value,
'autorange': True,
'zeroline': False,
'showticklabels': False,
'showline': True,
'showgrid': False,
},
'yaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['yaxis'].value,
'autorange': True,
'zeroline': False,
'showticklabels': False,
'showline': True,
'showgrid': False,
},
'showlegend': self.margins['showlegend'].value,
'legend': {
'orientation': self.margins['legend_orientation'].value,
'x': self.margins['legend_xpos'].value,
'y': self.margins['legend_ypos'].value
}
}
return FigureWidget({'data': traces, 'layout': layout})
def createFigureWidget(self):
traces = []
xtickmode = "auto"
xtickvals = None
xticktext = None
dimensions = []
color = "#444444"
for dimension in self.dimensions:
value = {'label': dimension}
if hasattr(self.data[dimension].flatten(), 'codes'):
val = self.data[dimension].flatten().codes.tolist()
else:
val = self.data[dimension].flatten()
value["values"] = val
dimensions.append(value)
trace = {
'type':
"splom",
'name':
self.data.label,
'dimensions':
dimensions,
'marker':
dict({
'symbol': 'circle',
'size': self.options['marker_size'].value,
'color': color,
'line': {
'width': self.options['line_width'].value,
'color': color
}
}),
'showupperhalf':
False,
'diagonal': {
'visible': False
},
}
if self.only_subsets == False:
traces.append(trace)
for sset in self.data.subsets:
if hasattr(sset, "disabled") == False or sset.disabled == False:
color = sset.style.color
dimensions = []
for dimension in self.dimensions:
value = {'label': dimension}
if hasattr(sset[dimension].flatten(), 'codes'):
val = sset[dimension].flatten().codes.tolist()
else:
val = sset[dimension].flatten()
value["values"] = val
dimensions.append(value)
trace = {
'type':
"splom",
'name':
sset.label,
'dimensions':
dimensions,
'marker':
dict({
'symbol': 'circle',
'size': self.options['marker_size'].value,
'color': color,
'line': {
'width': self.options['line_width'].value,
'color': color
}
}),
'showupperhalf':
False,
'diagonal': {
'visible': False
},
}
traces.append(trace)
layout = {
'title': self.options['title'].value,
'margin': {
'l': self.margins['left'].value,
'r': self.margins['right'].value,
'b': self.margins['bottom'].value,
't': self.margins['top'].value
},
'showlegend': self.margins['showlegend'].value,
'legend': {
'orientation': self.margins['legend_orientation'].value,
'x': self.margins['legend_xpos'].value,
'y': self.margins['legend_ypos'].value
},
'plot_bgcolor': 'rgba(245,245,245, 0.95)',
}
for i, dimension in enumerate(self.dimensions):
layout['xaxis' + str(i + 1)] = {
'autorange': True,
'zeroline': False,
'showline': True,
#'mirror' : 'ticks',
'linecolor': self.data.get_component(dimension).color,
'tickcolor': self.data.get_component(dimension).color,
'ticklen': 4,
'linewidth': 4,
}
layout['yaxis' + str(i + 1)] = {
'autorange': True,
'zeroline': False,
'showline': True,
#'mirror' : 'ticks',
'linecolor': self.data.get_component(dimension).color,
'tickcolor': self.data.get_component(dimension).color,
'ticklen': 4,
'linewidth': 4,
}
return FigureWidget({'data': traces, 'layout': layout})
def createFigureWidget(self):
dimensions = self.dimensions
data_lines = []
nodes = [{} for dim in dimensions]
values = []
colors = []
traces = []
if self.options['colorscale'].value == GlueParallelCategoriesPlotly.default_color:
color_set = [self.data.get_component(dim).color for dim in dimensions]
else:
color_set = cl.scales['8']['qual'][self.options['colorscale'].value]
if (len(self.dimensions) > 8):
color_set = cl.interp( color_set, len(dimensions) )
color_set = cl.to_rgb(color_set)
for i, dimension in enumerate(dimensions):
dvalues = np.unique(self.data[dimension].flatten())
if len(dvalues) < self.options['grouping_limit'].value or hasattr(self.data[dimension].flatten(), 'codes'):
nodes[i]['values'] = np.unique(self.data[dimension].flatten())
nodes[i]['masks'] = []
for val in nodes[i]['values']:
nodes[i]['masks'].append(self.data[dimension] == val)
else:
hist, bin_edges = np.histogram(self.data[dimension].flatten(), bins='auto')
if (len(bin_edges) > self.options['grouping_limit'].value):
hist, bin_edges = np.histogram(self.data[dimension].flatten(), bins=self.options['grouping_limit'].value)
nodes[i]['values'] = []
for edge in range(len(bin_edges)-1):
nodes[i]['values'].append( "{:.1f}".format(bin_edges[edge]) + " - " + "{:.1f}".format(bin_edges[edge+1]))
nodes[i]['masks'] = []
for edge in range(len(bin_edges)-1):
if edge == 0:
nodes[i]['masks'].append((self.data[dimension] >= bin_edges[edge]) & (self.data[dimension] <= bin_edges[edge+1]))
else :
nodes[i]['masks'].append((self.data[dimension] > bin_edges[edge]) & (self.data[dimension] <= bin_edges[edge+1]))
line = {}
line['values'] = np.array(['' for i in range(self.data.size)], dtype = 'object')
colorv = np.array(['#EEEEEE' for i in range(self.data.size)])
for k, value in enumerate(nodes[i]['values']):
mask = nodes[i]['masks'][k]
line['values'][mask] = value
for sset in self.data.subsets:
if hasattr(sset,"disabled") == False or sset.disabled == False:
sset_mask = sset.to_mask()
color = sset.style.color
colorv[mask & sset_mask] = color
mask = mask & ~sset_mask
colors.extend(colorv)
line['values'] = line['values'].tolist()
line['label'] = dimension
data_lines.append(line);
parcats = {
'type' : 'parcats',
'dimensions' : data_lines,
'line' : {
'color' : colors,
'shape': 'hspline'
}
}
traces.append(parcats)
layout = {
'title' : self.options['title'].value,
'xaxis': {
'title' : self.options['xaxis'].value,
'range' : [0,1],
'showgrid':False,
'showline':False,
'showticklabels':False,
'zeroline':False
},
'yaxis': {
'title' : self.options['yaxis'].value,
'range' : [0,1],
'showgrid':False,
'showline':False,
'showticklabels':False,
'zeroline':False
},
'showlegend': self.margins['showlegend'].value,
'legend' : {
'orientation' : self.margins['legend_orientation'].value,
'x' : self.margins['legend_xpos'].value,
'y' : self.margins['legend_ypos'].value
}
}
if self.only_subsets == False:
trace = {
'type': "scatter",
'name' : self.data.label,
'textposition' : 'middle right',
'x' : [-1000],
'y' : [i],
'mode' : 'markers',
'marker': {
'color' : "#EEEEEE",
'size' : 20,
'line': {
'width': 0,
},
'symbol' : 'square'
}
}
traces.append(trace)
for sset in self.data.subsets:
color = sset.style.color
trace = {
'type': "scatter",
'name' : sset.label,
'textposition' : 'middle right',
'x' : [-1000],
'y' : [i],
'mode' : 'markers',
'marker': {
'color' : color,
'size' : 20,
'line': {
'width': 0,
},
'symbol' : 'square'
}
}
traces.append(trace)
data = traces
FigureWidget(data = data, layout = layout)
return FigureWidget(data = data, layout = layout)
def createFigureWidget(self, x_id, y_id_list):
traces = []
alpha_min, alpha_max, alpha_delta = self.getDeltaFunction(
len(y_id_list))
alpha_val = alpha_max
xtickmode = "auto"
xtickvals = None
xticktext = None
if hasattr(self.data[x_id].flatten(), 'codes'):
x_val = self.data[x_id].flatten().codes.tolist()
tickvals, tickmask = np.unique(self.data[x_id].flatten().codes,
return_index=True)
ticktext = self.data[x_id][tickmask]
xtickmode = "array"
xtickvals = tickvals.tolist()
xticktext = ticktext.tolist()
else:
x_val = self.data[x_id].flatten()
for step, y_id in enumerate(y_id_list):
color = "#444444"
color = 'rgba' + str(self.getDeltaColor(color, alpha_val, step))
if hasattr(self.data[y_id].flatten(), 'codes'):
y_val = self.data[y_id].flatten().codes.tolist()
else:
y_val = self.data[y_id].flatten()
trace = {
'type': "scattergl",
'mode': "markers",
'name': self.data.label + "_" + y_id,
'marker': {
'symbol': 'circle',
'size': self.options['marker_size'].value,
'color': color,
'line': {
'width': self.options['line_width'].value,
'color': color
}
},
'x': x_val,
'y': y_val,
}
if self.only_subsets == False:
traces.append(trace)
alpha_val = alpha_val - alpha_delta
for sset in self.data.subsets:
if hasattr(sset, "disabled") == False or sset.disabled == False:
alpha_val = alpha_max
if hasattr(sset[x_id].flatten(), 'codes'):
x_val = sset[x_id].flatten().codes.tolist()
else:
x_val = sset[x_id].flatten()
for step, y_id in enumerate(y_id_list):
if hasattr(sset[y_id].flatten(), 'codes'):
y_val = sset[y_id].flatten().codes.tolist()
else:
y_val = sset[y_id].flatten().astype('float')
color = sset.style.color
color = 'rgba' + str(
self.getDeltaColor(color, alpha_val, step))
trace = {
'type': "scattergl",
'mode': "markers",
'name': sset.label + "_" + y_id,
'marker': {
'symbol': 'circle',
'size': self.options['marker_size'].value,
'color': color,
'line': {
'width': self.options['line_width'].value,
'color': color
}
},
'x': x_val,
'y': y_val,
}
traces.append(trace)
alpha_val = alpha_val - alpha_delta
y_fid = y_id_list[0]
layout = {
'title': self.options['title'].value,
'margin': {
'l': self.margins['left'].value,
'r': self.margins['right'].value,
'b': self.margins['bottom'].value,
't': self.margins['top'].value
},
'xaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['xaxis'].value,
'type': self.options['xscale'].value,
'tickmode': xtickmode,
'tickvals': xtickvals,
'ticktext': xticktext,
'linecolor': self.data.get_component(x_id).color,
'tickcolor': self.data.get_component(x_id).color,
'ticklen': 4,
'linewidth': 4,
},
'yaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['yaxis'].value,
'type': self.options['yscale'].value,
'linecolor': self.data.get_component(y_fid).color,
'tickcolor': self.data.get_component(y_fid).color,
'ticklen': 4,
'linewidth': 4,
},
'showlegend': self.margins['showlegend'].value,
'legend': {
'orientation': self.margins['legend_orientation'].value,
'x': self.margins['legend_xpos'].value,
'y': self.margins['legend_ypos'].value
}
}
return FigureWidget({'data': traces, 'layout': layout})
def display_eps(player_gw_eps: DF) -> Widget:
data = (player_gw_eps
.reset_index()
.sort_values(['Season', 'Game Week']))
data_formatted = data.pipe(ctx.dd.format)
data = (data.assign(**{'Label': 'FDR: ' + data_formatted['Fixture Short Name FDR'] + ', '
+ 'Rel. Strength: ' + data_formatted['Rel Strength'] + ', '
+ 'Cost: ' + data_formatted['Fixture Cost']})
.assign(**{'Game Week': lambda df: 'GW ' + df['Game Week'].apply('{:.0f}'.format)})
.assign(**{'Season Game Week': lambda df: df['Season'] + ', GW ' + df['Game Week']}))
x_axis = [data['Season'], data['Game Week']]
eps_trace = Scatter(x=x_axis,
y=data['Expected Points'],
name='Expected Points',
line=dict(color='rgb(255, 127, 14)'),
mode='lines')
ftp_trace = Scatter(x=x_axis,
y=data['Fixture Total Points'],
name='Actual Points',
line=dict(color='rgba(44, 160, 44, 0.3)'),
mode='lines')
ftpr_trace = Scatter(x=x_axis,
y=data['Rolling Avg Game Points'],
name='Rolling Actual Points',
line=dict(color='rgb(44, 160, 44)'),
line_shape='spline',
mode='lines')
fs_trace = Scatter(x=x_axis,
y=data['Expected Points'],
name='Rel. Strength',
mode='markers',
marker=dict(color=(data['Team FDR'].fillna(3)), colorscale=FDR_COLOR_SCALE),
text=data['Label'])
last_gw = [ctx.current_season, f'GW {ctx.next_gw - 1}']
first_gw = [ctx.current_season, 'GW 1']
last_season_gws_color = 'rgb(230, 230, 230)'
past_gws_color = 'rgb(240, 240, 240)'
last_season_shape = Shape(type='rect', yref='paper', x0=-6, x1=first_gw, y0=0, y1=1, fillcolor=last_season_gws_color, layer='below', line_width=0, opacity=0.5)
past_shape = Shape(type='rect', yref='paper', x0=first_gw, x1=last_gw, y0=0, y1=1, fillcolor=past_gws_color, layer='below', line_width=0, opacity=0.5)
start_shape = Shape(type='line', yref='paper', x0=first_gw, x1=first_gw, y0=0, y1=1, line=dict(width=2, color='DarkGrey'), layer='below')
current_gw_shape = Shape(type='line', yref='paper', x0=last_gw, x1=last_gw, y0=0, y1=1, line=dict(width=2, color='DarkGrey'), layer='below')
max_points = max(max(data['Expected Points'].max(), data['Fixture Total Points'].max()) + 1, 15)
min_points = min(data['Expected Points'].min(), data['Fixture Total Points'].min()) - 1
layout = dict(
yaxis=dict(title=f'Points', showspikes=True, range=[min_points, max_points]),
xaxis=dict(tickfont=dict(size=8)),
shapes=[last_season_shape, start_shape, past_shape, current_gw_shape],
hovermode='closest',
legend=dict(yanchor='top', xanchor='left', x=0, y=1, bgcolor='rgba(0,0,0,0)'),
height=300,
margin=dict(l=20, r=0, t=5, b=20, pad=0),
annotations=[
dict(x=last_gw, y=0.9, yref='paper', text='Last Game Week', ax=80, ay=0),
dict(x=first_gw, y=0.9, yref='paper', text='Start of Season', ax=-80, ay=0)
]
)
return VBox([HTML('<h3>Expected Points vs Actual Points</h3>'),
FigureWidget([ftp_trace, ftpr_trace, eps_trace, fs_trace], layout=layout)])
def createFigureWidget(self):
data_lines = []
traces = []
i=0
if self.only_subsets == False:
colors = [i for r in range(self.data.size)]
colorscale = [[0,'#EEEEEE']]
else:
colors = []
colorscale = []
i=-1
values = []
labels = []
nodes_id = []
parents = []
self.masks = []
colors = []
df = self.data.to_dataframe()
ids = 1
queue = deque()
if self.options['colorscale'].value == GlueSunburstPlotly.default_color:
color_set = [self.data.get_component(dim).color for dim in self.dimensions]
else:
color_set = cl.scales['8']['qual'][self.options['colorscale'].value]
if (len(self.dimensions) > 8):
color_set = cl.interp( color_set, len(dimensions) )
color_set = cl.to_rgb(color_set)
color_set.insert(0,'rgb(255,255,255)')
queue.append({'id':ids,'dimension':0, 'label':"Data", 'mask':[True for i in range(self.data.size)], 'parent':''})
while len(queue) > 0:
toprocess = queue.popleft()
mask = toprocess['mask']
id = toprocess['id']
label = toprocess['label']
value = np.count_nonzero(mask)
if value > 0:
labels.append(label)
values.append(value)
parents.append(toprocess['parent'])
nodes_id.append(id)
self.masks.append(mask)
colors.append(color_set[toprocess['dimension']])
data = df[mask]
if toprocess['dimension'] < len(self.dimensions):
dimension = self.dimensions[toprocess['dimension']]
dvalues = np.unique(data[dimension].ravel())
if hasattr(self.data[dimension].flatten(), 'codes'):
for val in dvalues:
ids += 1
process = {'id':ids,'dimension':toprocess['dimension']+1, 'label':str(val), 'parent':toprocess['id']}
process['mask'] = (mask & (self.data[dimension] == val))
queue.append(process)
elif len(dvalues) < self.options['grouping_limit'].value:
for val in dvalues:
ids += 1
process = {'id':ids,'dimension':toprocess['dimension']+1, 'label':str(val), 'parent':toprocess['id']}
process['mask'] = (mask & (self.data[dimension] == val))
queue.append(process)
else:
hist, bin_edges = np.histogram(self.data[dimension].flatten(), bins='auto')
if (len(bin_edges) > self.options['grouping_limit'].value):
hist, bin_edges = np.histogram(self.data[dimension].flatten(), bins=self.options['grouping_limit'].value)
for edge in range(len(bin_edges)-1):
ids += 1
label = "{:.1f}".format(bin_edges[edge]) + " - " + "{:.1f}".format(bin_edges[edge+1])
process = {'id':ids,'dimension':toprocess['dimension']+1, 'label':label, 'parent':toprocess['id']}
if edge == 0:
process['mask'] = (mask & ((self.data[dimension] >= bin_edges[edge]) & (self.data[dimension] <= bin_edges[edge+1])))
else :
process['mask'] = (mask & ((self.data[dimension] > bin_edges[edge]) & (self.data[dimension] <= bin_edges[edge+1])))
queue.append(process)
trace = {
'type': "sunburst",
'ids': nodes_id,
'labels': labels,
'parents': parents,
'values': values,
'branchvalues' : 'total',
'maxdepth' : 4,
'outsidetextfont': {
'size': 20,
'color': "#377eb8"
},
'leaf': {
'opacity' : 1
},
'marker': {
'line': {
'width': 2
},
'colors' : colors
},
'domain':{
'x': [0, 0.9],
'y': [0, 1],
},
}
traces.append(trace)
layout = {
'title' : self.options['title'].value,
'margin' : {
'l':self.margins['left'].value,
'r':self.margins['right'].value,
'b':self.margins['bottom'].value,
't':self.margins['top'].value
},
'showlegend': True,
'xaxis': {
'title' : self.options['xaxis'].value,
'range' : [0,1],
'showgrid':False,
'showline':False,
'showticklabels':False,
'zeroline':False
},
'yaxis': {
'title' : self.options['yaxis'].value,
'range' : [0,1],
'showgrid':False,
'showline':False,
'showticklabels':False,
'zeroline':False
},
'showlegend': self.margins['showlegend'].value,
'legend' : {
'orientation' : self.margins['legend_orientation'].value,
'x' : self.margins['legend_xpos'].value,
'y' : self.margins['legend_ypos'].value
}
}
for i in range(len(self.dimensions)):
dimension = self.dimensions[i]
trace = {
'type': "scatter",
'name' : dimension,
'textposition' : 'middle right',
'x' : [-1000],
'y' : [i],
'mode' : 'markers',
'marker': {
'color' : color_set[i+1],
'size' : 20,
'line': {
'width': 0,
},
'symbol' : 'square'
}
}
traces.append(trace)
return FigureWidget({
'data': traces,
'layout': layout
})
def show_indexed_timeseries_plotly(
timeseries: TimeSeries,
istart: int = 0,
istop: int = None,
time_window: list = None,
trace_range: list = None,
offsets=None,
fig: go.FigureWidget = None,
col=None,
row=None,
zero_start=False,
scatter_kwargs: dict = None,
figure_kwargs: dict = None,
):
if istart != 0 or istop is not None:
if time_window is not None:
raise ValueError(
"input either time window or istart/stop but not both")
if not (0 <= istart < timeseries.data.shape[0] and
(istop is None or 0 < istop <= timeseries.data.shape[0])):
raise ValueError("enter correct istart/stop values")
t_istart = istart
t_istop = istop
elif time_window is not None:
t_istart = timeseries_time_to_ind(timeseries, time_window[0])
t_istop = timeseries_time_to_ind(timeseries, time_window[1])
else:
t_istart = istart
t_istop = istop
tt = get_timeseries_tt(timeseries, istart=t_istart, istop=t_istop)
data, unit = get_timeseries_in_units(timeseries,
istart=t_istart,
istop=t_istop)
if len(data.shape) == 1:
data = data[:, np.newaxis]
if trace_range is not None:
if not (0 <= trace_range[0] < data.shape[1]
and 0 < trace_range[1] <= data.shape[1]):
raise ValueError("enter correct trace range")
trace_istart = trace_range[0]
trace_istop = trace_range[1]
else:
trace_istart = 0
trace_istop = data.shape[1]
if offsets is None:
offsets = np.zeros(trace_istop - trace_istart)
if zero_start:
tt = tt - tt[0]
scatter_kwargs = dict() if scatter_kwargs is None else scatter_kwargs
if fig is None:
fig = go.FigureWidget(make_subplots(rows=1, cols=1))
row = 1 if row is None else row
col = 1 if col is None else col
for i, trace_id in enumerate(range(trace_istart, trace_istop)):
fig.add_trace(
go.Scattergl(x=tt,
y=data[:, trace_id] + offsets[i],
mode="lines",
**scatter_kwargs),
row=row,
col=col,
)
input_figure_kwargs = dict(
xaxis=dict(title_text="time (s)", range=[tt[0], tt[-1]]),
yaxis=dict(title_text=unit if unit is not None else None),
title=timeseries.name,
)
if figure_kwargs is None:
figure_kwargs = dict()
input_figure_kwargs.update(figure_kwargs)
fig.update_xaxes(input_figure_kwargs.pop("xaxis"), row=row, col=col)
fig.update_yaxes(input_figure_kwargs.pop("yaxis"), row=row, col=col)
fig.update_layout(**input_figure_kwargs)
return fig
def createFigureWidget(self):
data_lines = []
traces = []
sources = []
targets = []
values = []
labels = []
parents = []
self.masks = []
colors = []
df = self.data.to_dataframe()
ids = 0
queue = deque()
if self.options[
'colorscale'].value == GlueSankeytreePlotly.default_color:
color_set = [
self.data.get_component(dim).color for dim in self.dimensions
]
else:
color_set = cl.scales['8']['qual'][
self.options['colorscale'].value]
if (len(self.dimensions) > 8):
color_set = cl.interp(color_set, len(self.dimensions))
color_set = cl.to_rgb(color_set)
queue.append({
'id': ids,
'dimension': 0,
'label': 'Data',
'mask': [True for i in range(self.data.size)],
'parent': ''
})
while len(queue) > 0:
toprocess = queue.popleft()
mask = toprocess['mask']
id = toprocess['id']
label = toprocess['label']
value = np.count_nonzero(mask)
if True: #value > 0:
labels.append(label)
sources.append(toprocess['parent'])
targets.append(id)
values.append(value)
parents.append(toprocess['parent'])
colors.append(color_set[toprocess['dimension'] - 1])
self.masks.append(mask)
data = df[mask]
if toprocess['dimension'] < len(self.dimensions):
dimension = self.dimensions[toprocess['dimension']]
dvalues = np.unique(data[dimension].ravel())
if hasattr(self.data[dimension].flatten(), 'codes'):
for val in dvalues:
ids += 1
process = {
'id': ids,
'dimension': toprocess['dimension'] + 1,
'label': str(val),
'parent': toprocess['id']
}
process['mask'] = (mask &
(self.data[dimension] == val))
queue.append(process)
elif len(dvalues) < self.options['grouping_limit'].value:
for val in dvalues:
ids += 1
process = {
'id': ids,
'dimension': toprocess['dimension'] + 1,
'label': str(val),
'parent': toprocess['id']
}
process['mask'] = (mask &
(self.data[dimension] == val))
queue.append(process)
else:
hist, bin_edges = np.histogram(
self.data[dimension].flatten(), bins='auto')
if (len(bin_edges) >
self.options['grouping_limit'].value):
hist, bin_edges = np.histogram(
self.data[dimension].flatten(),
bins=self.options['grouping_limit'].value)
for edge in range(len(bin_edges) - 1):
ids += 1
label = '{:.1f}'.format(
bin_edges[edge]) + ' - ' + '{:.1f}'.format(
bin_edges[edge + 1])
process = {
'id': ids,
'dimension': toprocess['dimension'] + 1,
'label': label,
'parent': toprocess['id']
}
if edge == 0:
process['mask'] = (mask & (
(self.data[dimension] >= bin_edges[edge]) &
(self.data[dimension] <=
bin_edges[edge + 1])))
else:
process['mask'] = (mask & (
(self.data[dimension] > bin_edges[edge]) &
(self.data[dimension] <=
bin_edges[edge + 1])))
queue.append(process)
trace = {
'type': 'sankey',
'node': {
'pad': 15,
'thickness': 20,
'line': {
'color': 'black',
'width': 0.5
},
'label': labels[1:],
'color': colors[1:],
},
'link': {
'source': [s - 1 for s in sources[1:]],
'target': [t - 1 for t in targets[1:]],
'value':
values[1:],
'hovertemplate':
'%{label}<br><b>Source: %{source.label}<br>Target: %{target.label}<br> %{flow.value}</b>'
},
}
traces.append(trace)
layout = {
'title': self.options['title'].value,
'margin': {
'l': self.margins['left'].value,
'r': self.margins['right'].value,
'b': self.margins['bottom'].value,
't': self.margins['top'].value
},
'showlegend': True,
'xaxis': {
'title': self.options['xaxis'].value,
'range': [0, 1],
'showgrid': False,
'showline': False,
'showticklabels': False,
'zeroline': False
},
'yaxis': {
'title': self.options['yaxis'].value,
'range': [0, 1],
'showgrid': False,
'showline': False,
'showticklabels': False,
'zeroline': False
},
'legend': {
'orientation': self.margins['legend_orientation'].value,
'x': self.margins['legend_xpos'].value,
'y': self.margins['legend_ypos'].value
}
}
for i in range(len(self.dimensions)):
dimension = self.dimensions[i]
trace = {
'type': 'scatter',
'name': dimension,
'textposition': 'middle right',
'x': [-1000],
'y': [i],
'mode': 'markers',
'marker': {
'color': color_set[i],
'size': 20,
'line': {
'width': 0,
},
'symbol': 'square'
}
}
traces.append(trace)
return FigureWidget({'data': traces, 'layout': layout})
def createFigureWidget(self):
dimensions = self.dimensions
data_lines = []
sources = []
targets = []
nodes = [{} for dim in dimensions]
nodest = [{} for dim in dimensions]
values = []
labels = []
colors = []
self.masks = []
nodes_id = []
if self.options['colorscale'].value == GlueSankeyPlotly.default_color:
color_set = [
self.data.get_component(dim).color for dim in dimensions
]
else:
color_set = cl.scales['8']['qual'][
self.options['colorscale'].value]
if (len(self.dimensions) > 8):
color_set = cl.interp(color_set, len(dimensions))
color_set = cl.to_rgb(color_set)
for i, dimension in enumerate(dimensions):
dvalues = np.unique(self.data[dimension].flatten())
if len(dvalues) < self.options['grouping_limit'].value or hasattr(
self.data[dimension].flatten(), 'codes'):
nodes[i]['values'] = np.unique(self.data[dimension].flatten())
nodes[i]['local_nodes'] = [
len(nodes_id) + cnt
for cnt in range(len(nodes[i]['values']))
]
nodes[i]['masks'] = []
for val in nodes[i]['values']:
nodes[i]['masks'].append(self.data[dimension] == val)
else:
hist, bin_edges = np.histogram(self.data[dimension].flatten(),
bins='auto')
if (len(bin_edges) > self.options['grouping_limit'].value):
hist, bin_edges = np.histogram(
self.data[dimension].flatten(),
bins=self.options['grouping_limit'].value)
nodes[i]['values'] = []
for edge in range(len(bin_edges) - 1):
nodes[i]['values'].append(
'{:.1f}'.format(bin_edges[edge]) + ' - ' +
'{:.1f}'.format(bin_edges[edge + 1]))
nodes[i]['local_nodes'] = [
len(nodes_id) + cnt
for cnt in range(len(nodes[i]['values']))
]
nodes[i]['masks'] = []
for edge in range(len(bin_edges) - 1):
if edge == 0:
nodes[i]['masks'].append(
(self.data[dimension] >= bin_edges[edge])
& (self.data[dimension] <= bin_edges[edge + 1]))
else:
nodes[i]['masks'].append(
(self.data[dimension] > bin_edges[edge])
& (self.data[dimension] <= bin_edges[edge + 1]))
nodes_id.extend(nodes[i]['local_nodes'])
labels.extend(nodes[i]['values'])
colors.extend([color_set[i] for val in nodes[i]['values']])
self.masks.extend(nodes[i]['masks'])
link_colors = []
for i in range(len(dimensions) - 1):
for j in range(len(nodes[i]['local_nodes'])):
for k in range(len(nodes[i + 1]['local_nodes'])):
total = np.count_nonzero(
(nodes[i]['masks'][j] & nodes[i + 1]['masks'][k]))
if total > 0:
mask = nodes[i]['masks'][j] & nodes[i + 1]['masks'][k]
for sset in self.data.subsets:
if hasattr(sset, "disabled"
) == False or sset.disabled == False:
sset_mask = sset.to_mask()
color = sset.style.color
settotal = np.count_nonzero(mask & sset_mask)
mask = mask & ~sset_mask
sources.append(nodes[i]['local_nodes'][j])
targets.append(nodes[i + 1]['local_nodes'][k])
values.append(settotal)
link_colors.append(color)
total = total - settotal
if self.only_subsets == False:
sources.append(nodes[i]['local_nodes'][j])
targets.append(nodes[i + 1]['local_nodes'][k])
values.append(total)
link_colors.append('rgba(238, 238, 238, 0.6)')
traces = []
sankey = {
'type': 'sankey',
#'arrangement' : 'perpendicular',
'node': {
'pad': 8,
'thickness': 15,
'line': {
'color': 'black',
'width': 0.5
},
'label': labels,
'color': colors
},
'link': {
'source': sources,
'target': targets,
'value': values,
'hovertemplate':
'%{label}<br><b>Source: %{source.label}<br>Target: %{target.label}<br> %{flow.value}</b>',
'color': link_colors,
'line': {
'color': 'lightgrey',
'width': 0.5
}
},
#'domain':{
# 'x': [0, 0.9],
# 'y': [0, 1],
#},
}
traces.append(sankey)
layout = {
'title': self.options['title'].value,
'xaxis': {
'title': self.options['xaxis'].value,
'range': [0, 1],
'showgrid': False,
'showline': False,
'showticklabels': False,
'zeroline': False
},
'yaxis': {
'title': self.options['yaxis'].value,
'range': [0, 1],
'showgrid': False,
'showline': False,
'showticklabels': False,
'zeroline': False
},
'showlegend': self.margins['showlegend'].value,
'legend': {
'orientation': self.margins['legend_orientation'].value,
'x': self.margins['legend_xpos'].value,
'y': self.margins['legend_ypos'].value
},
'font': {
'size': 10,
},
}
for i in range(len(self.dimensions)):
dimension = self.dimensions[i]
trace = {
'type': 'scatter',
'name': dimension,
'textposition': 'middle right',
'x': [-1000],
'y': [i],
'mode': 'markers',
'marker': {
'color': color_set[i],
'size': 20,
'line': {
'width': 0,
},
'symbol': 'square'
}
}
traces.append(trace)
if self.only_subsets == False:
trace = {
'type': "scatter",
'name': self.data.label,
'textposition': 'middle right',
'x': [-1000],
'y': [i],
'mode': 'markers',
'marker': {
'color': "#EEEEEE",
'size': 20,
'line': {
'width': 1,
'color': 'light grey'
},
'symbol': 'square'
}
}
traces.append(trace)
for sset in self.data.subsets:
if hasattr(sset, "disabled") == False or sset.disabled == False:
color = sset.style.color
trace = {
'type': "scatter",
'name': sset.label,
'textposition': 'middle right',
'x': [-1000],
'y': [i],
'mode': 'markers',
'marker': {
'color': color,
'size': 20,
'line': {
'width': 1,
'color': 'light grey'
},
'symbol': 'square'
}
}
traces.append(trace)
data = traces
FigureWidget(data=data, layout=layout)
return FigureWidget(data=data, layout=layout)
def createFigureWidget(self):
x_id = self.dimensions[0]
y_id = self.dimensions[1]
z_id = self.dimensions[2]
x_value = self.data[x_id].flatten().astype('float')
y_value = self.data[y_id].flatten().astype('float')
x_value, x_inv = np.unique(x_value, return_inverse=True)
y_value, y_inv = np.unique(y_value, return_inverse=True)
try:
z_value = np.reshape(self.data[z_id].flatten(),
(len(x_value), len(y_value)))
except:
t_value = self.data[z_id].flatten()
x_value = np.sort(x_value)
y_value = np.sort(y_value)
z_value = np.zeros((len(x_value), len(y_value)))
for i, value in enumerate(t_value):
z_value[x_inv[i]][y_inv[i]] = value
#with self.debug:
# print (x_value,y_value,z_value)
traces = []
trace = {
'type': "contour",
'colorscale': self.options['color_scale'].value,
'showlegend': False,
'autocontour': False,
'ncontours': 1,
'contours': {
'coloring': 'heatmap'
},
'x': x_value.tolist(),
'y': y_value.tolist(),
'z': z_value.tolist(),
'zauto': False,
'zmin': self.options['color_range_min'].value,
'zmax': self.options['color_range_max'].value,
}
if self.only_subsets == False:
traces.append(trace)
for sset in self.data.subsets:
if hasattr(sset, "disabled") == False or sset.disabled == False:
color = sset.style.color
color = 'rgba' + str(self.getDeltaColor(color, 0.6))
trace = {
'type':
"scattergl",
'mode':
"markers",
'name':
sset.label + "_" + y_id,
'marker':
dict({
'symbol': 'circle',
'size': self.options['marker_size'].value,
'color': color,
'line': {
'width': self.options['line_width'].value,
'color': color
}
}),
'selected': {
'marker': {
'color': color,
'size': self.options['marker_size'].value
}
},
'unselected': {
'marker': {
'color': color,
'size': self.options['marker_size'].value
}
},
'x':
sset[x_id].flatten(),
'y':
sset[y_id].flatten(),
}
traces.append(trace)
layout = {
'title': self.options['title'].value,
'margin': {
'l': self.margins['left'].value,
'r': self.margins['right'].value,
'b': self.margins['bottom'].value,
't': self.margins['top'].value
},
'xaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['xaxis'].value,
'linecolor': self.data.get_component(x_id).color,
'tickcolor': self.data.get_component(x_id).color,
'ticklen': 4,
'linewidth': 4,
},
'yaxis': {
'autorange': True,
'zeroline': True,
'title': self.options['yaxis'].value,
'linecolor': self.data.get_component(y_id).color,
'tickcolor': self.data.get_component(y_id).color,
'ticklen': 4,
'linewidth': 4,
},
'showlegend': self.margins['showlegend'].value,
'legend': {
'orientation': self.margins['legend_orientation'].value,
'x': self.margins['legend_xpos'].value,
'y': self.margins['legend_ypos'].value
},
}
return FigureWidget({'data': traces, 'layout': layout})