def create_plot(depth, level_width, acc, x, y, data_source, instance,
node_list):
circle_radius = 5
# title = "Decision Tree " + ("\t\t\t\tAccuracy (%): " + str(round(acc * 100, 1))) \
# + " (Test Size: " + str('not available') + ", " \
# + "Train Size: " + str('not available') + ")"
hover = HoverTool(names=["circles"])
wheel = WheelZoomTool()
wheel.zoom_on_axis = False
p = figure(width=900,
tools=[hover, wheel, ResetTool(),
PanTool()],
x_range=x,
y_range=list(y),
tooltips=TOOLTIPS)
arrow_dc, label, _ = draw_arrow(depth, level_width, x, y, data_source.data,
p, instance, node_list)
p.add_layout(label)
p.circle("y",
"x",
radius=circle_radius,
radius_units='screen',
source=data_source,
name="circles",
legend="attribute_type",
color=factor_cmap('attribute_type',
palette=get_all_colors(),
factors=instance['all_attr_list']))
# Final settings
p.axis.visible = False
p.toolbar.active_scroll = wheel
p.outline_line_color = "white"
p.grid.grid_line_color = None
p.axis.axis_line_color = None
return p, arrow_dc
def create_plot(depth, level_width, acc, x, y, data_source, instance):
''' create glyphs, text and arrows and insert them into the figures
Args:
depth: maximum width of tree
level_width: list of width of each level
acc: accuracy of tree
y: y coordinates of nodes
x: x coordinates of nodes
data_source: data source to be used
instance: data class
Returns:
plot p
'''
circle_radius = 5
test_size = int(len(instance.data) * instance.test_percentage / 100)
title = "Decision Tree " + ("\t\t\t\tAccuracy (%): " + str(round(acc * 100, 1))) \
+ " (Test Size: " + str(test_size) + ", " \
+ "Train Size: " + str(int(len(instance.data) - test_size)) + ")"
hover = HoverTool(names=["circles"])
wheel = WheelZoomTool()
wheel.zoom_on_axis = False
p = figure(title=title,
toolbar_location=None,
tools=[hover, wheel, ResetTool(),
PanTool()],
x_range=x,
y_range=list(y),
tooltips=TOOLTIPS)
_, label = draw_arrow(depth, level_width, x, y, data_source.data, p,
instance)
p.add_layout(label)
p.circle("y",
"x",
radius=circle_radius,
radius_units='screen',
source=data_source,
name="circles",
legend="attribute_type",
color=factor_cmap('attribute_type',
palette=get_all_colors(),
factors=instance.all_attr_list))
# Final settings
p.axis.visible = False
p.toolbar.active_scroll = wheel
p.outline_line_color = "white"
p.grid.grid_line_color = None
p.axis.axis_line_color = None
return p
def lower_plot(df, p1):
# configure wheelzoom tool
wz = WheelZoomTool()
wz.maintain_focus = False
wz.dimensions = 'width'
wz.zoom_on_axis = True
p2_tools = 'box_zoom,pan,save, hover, reset'
p2 = figure(width=fwidth,
height=fhgt + 35,
x_axis_type="datetime",
tools=p2_tools,
x_range=p1.x_range)
p2.add_tools(wz)
p2.min_border_top = fborder
p2.min_border_bottom = fborder
# pressure
h_line = p2.line(x='time',
y='p',
source=df,
line_width=4,
color=pcol,
legend='Pressure')
p2.y_range = Range1d(df['p'].min() - 5, df['p'].max() + 5)
p2.yaxis.axis_label = 'Pressure (hPa)'
# wind
p2.extra_y_ranges = {
"winddir": Range1d(start=0, end=360),
"windspd": Range1d(start=0, end=df['ff'].max() + df['ff'].max() * 0.1)
}
p2.add_layout(
LinearAxis(y_range_name='winddir',
ticker=SingleIntervalTicker(interval=45,
num_minor_ticks=3)), 'right')
p2.add_layout(LinearAxis(y_range_name="windspd"), 'right')
p2.circle(x='time',
y='dd',
source=df,
line_width=4,
color='black',
y_range_name='winddir',
legend='Wind Direction')
p2.line(x='time',
y='ff',
source=df,
line_width=4,
color=ffcol,
y_range_name='windspd',
legend='Wind speed')
p2.yaxis[0].axis_label = 'Pressure (hPa)'
p2.yaxis[1].axis_label = 'Wind direction (°)'
p2.yaxis[2].axis_label = 'Wind speed (m s⁻¹)'
p2.xaxis[0].axis_label = 'Time (UTC)'
p2.yaxis[2].major_label_text_color = ffcol
p2.yaxis[2].axis_label_text_color = ffcol
p2.yaxis[2].minor_tick_line_color = ffcol
p2.yaxis[2].major_tick_line_color = ffcol
p2.yaxis[2].axis_line_color = ffcol
p2.yaxis[0].major_label_text_color = pcol
p2.yaxis[0].axis_label_text_color = pcol
p2.yaxis[0].minor_tick_line_color = pcol
p2.yaxis[0].major_tick_line_color = pcol
p2.yaxis[0].axis_line_color = pcol
# hover
hover_p2 = p2.select(dict(type=HoverTool))
hover_p2.tooltips = [("Timestamp", "@time{%d %b %Y %H:%M} UTC"),
('Pressure', '@p{f0.0} hPa'),
('Winddirection', '@dd{int} °'),
('Windspeed', '@ff{f0.0} (m s⁻¹)')]
hover_p2.formatters = {"time": "datetime"}
hover_p2.mode = 'vline'
hover_p2.renderers = [h_line] #### to fix if missing value
# legend
p2.legend.location = (0, 15) # above plot
p2.legend.orientation = 'horizontal'
p2.legend.click_policy = "hide"
p2.legend.label_text_font_size = font_size_legend
p2.add_layout(p2.legend[0], 'above')
# font style
p2 = set_font_style_axis(p2)
#set boarders for zoom
p2.x_range.max_interval = timedelta(7.5)
return p2
def upper_plot(df):
# configure wheelzoom tool
wz = WheelZoomTool()
wz.maintain_focus = False
wz.dimensions = 'width'
wz.zoom_on_axis = True
p1_tools = 'box_zoom,pan,save, hover, reset' #, xwheel_zoom' # zoom bounds auto?
p1 = figure(width=fwidth,
height=fhgt,
x_axis_type="datetime",
tools=p1_tools)
p1.x_range = Range1d(start=df.index[-1] - timedelta(days=1),
end=df.index[-1],
bounds=(df.index[0], df.index[-1]))
p1.add_tools(wz)
p1.min_border_top = fborder
p1.min_border_bottom = fborder
# hover for temp, dew point and rel. humidity
hover_p1 = p1.select(dict(type=HoverTool))
hover_p1.tooltips = [("Timestamp", "@time{%d %b %Y %H:%M} UTC"),
('Temperature', "@tl{f0.0} °C")] #
if 'tp' in df.columns:
hover_p1[0].tooltips.append(('Dewpoint', '@tp{f0.0} °C'))
else:
hover_p1[0].tooltips.append(('Relative Humidity', '@rf{f0.0} %'))
# sunshine duration
if 'so' in df.columns:
if ssdcum:
varso = 'ssd_cum' # 'ssd_cum' or 'so'
unitso = 'h'
else:
varso = 'so' # 'ssd_cum' or 'so'
unitso = 'min'
if ssdcum and df[varso].sum(
) > 0: #axis would disappear when there was no rain measured
p1.extra_y_ranges[varso] = Range1d(start=0,
end=(df[varso].max() +
df[varso].max() * 0.1))
else:
p1.extra_y_ranges[varso] = Range1d(start=0, end=10)
p1.add_layout(LinearAxis(y_range_name=varso), 'right')
if ssdcum:
p1.line(x='time',
y=varso,
source=df,
line_width=4,
color=socol,
y_range_name=varso,
legend='Sunshine duration (24h)')
else:
p1.vbar(top=varso,
x='time',
source=df,
width=get_width(),
fill_color=socol,
line_alpha=0,
line_width=0,
fill_alpha=0.5,
y_range_name=varso,
legend='Sunshine duration')
p1.yaxis[1].axis_label = 'Sunshine duration (' + unitso + ')'
p1.yaxis[1].axis_label_text_font_size = font_size_label
p1.yaxis[1].major_label_text_color = socol
p1.yaxis[1].axis_label_text_color = socol
p1.yaxis[1].minor_tick_line_color = socol
p1.yaxis[1].major_tick_line_color = socol
p1.yaxis[1].axis_line_color = socol
hover_p1[0].tooltips.append(
('Sunshine duration', '@so{int} min per 10 min'))
hover_p1[0].tooltips.append(
('Cumulated sunshine duration', '@ssd_cum{f0.0} h'))
# temperature
h_line = p1.line(x='time',
y='tl',
source=df,
line_width=4,
color=tcol,
legend='Temperature')
p1.yaxis[0].axis_label = 'Temperature (°C)'
p1.yaxis[0].major_label_text_color = tcol
p1.yaxis[0].axis_label_text_color = tcol
p1.yaxis[0].minor_tick_line_color = tcol
p1.yaxis[0].major_tick_line_color = tcol
p1.yaxis[0].axis_line_color = tcol
p1.yaxis[0].axis_label_text_font_style = "normal"
# dew point
if 'tp' in df.columns:
p1.y_range = Range1d(df['tp'].min() - 2, df['tl'].max() + 2)
p1.line(x='time',
y='tp',
source=df,
line_width=4,
color=hcol,
legend='Dewpoint')
else:
# relative humidity
p1.y_range = Range1d(df['tl'].min() - 2, df['tl'].max() + 2)
p1.extra_y_ranges = {'rf': Range1d(start=0, end=100)}
p1.add_layout(LinearAxis(y_range_name='rf'), 'right')
p1.line(x='time',
y='rf',
source=df,
line_width=4,
color=hcol,
legend='Relative Humidity',
y_range_name='rf')
p1.yaxis[1].axis_label = 'Relative humidity (%)'
p1.yaxis[1].major_label_text_color = hcol
p1.yaxis[1].axis_label_text_color = hcol
p1.yaxis[1].minor_tick_line_color = hcol
p1.yaxis[1].major_tick_line_color = hcol
p1.yaxis[1].axis_line_color = hcol
# precipitation (3 h sums)
if 'rrsum' in df.columns:
if df['rrsum'].sum(
) > 0: #axis would disappear when there was no rain measured
p1.extra_y_ranges['rrsum'] = Range1d(start=0,
end=(df['rrsum'].max() * 2))
else:
p1.extra_y_ranges['rrsum'] = Range1d(start=0, end=10)
p1.add_layout(LinearAxis(y_range_name='rrsum'), 'right')
timeoffset = 0 # timeoffset: dodge to correctly bin bar in time
if rrsum_period > 10: timeoffset = -60 * rrsum_period / 2 * 1000
rr = p1.vbar(x=dodge('time', timeoffset, range=p1.x_range),
top='rrsum',
width=get_width() * rrsum_period / 10,
source=df,
fill_color=pcol,
line_alpha=0,
line_width=0,
fill_alpha=0.5,
legend='Precipitation',
y_range_name='rrsum')
rr.level = 'underlay'
if rrsum_period >= 60:
rr_period = str(round_dec(rrsum_period / 60, decimals=1)) + ' h'
else:
rr_period = str(rrsum_period) + ' min'
hover_p1[0].tooltips.append(
('Precipitation', '@rrsumm{f0.0} mm in ' + rr_period))
hover_p1[0].tooltips.append(
('Cumulated precipitation', '@rr_cum{f0.0} mm'))
p1.yaxis[2].major_label_text_color = pcol
p1.yaxis[2].axis_label_text_color = pcol
p1.yaxis[2].minor_tick_line_color = pcol
p1.yaxis[2].major_tick_line_color = pcol
p1.yaxis[2].axis_line_color = pcol
p1.yaxis[2].axis_label = 'Precipitation (mm)'
# hover
hover_p1.formatters = {"time": "datetime"}
hover_p1.mode = 'vline'
hover_p1.renderers = [h_line] #### to fix if missing value
# legend
p1.legend.location = (0, 15) # above plot
p1.legend.orientation = 'horizontal'
p1.legend.click_policy = "hide"
p1.legend.label_text_font_size = font_size_legend
p1.add_layout(p1.legend[0], 'above')
# font style
p1 = set_font_style_axis(p1)
return p1