def create_plot(title, xlabel, ylabel, source, lines, yformatter=None):
fig = plotting.figure(
plot_width=PLOT_WIDTH,
plot_height=PLOT_HEIGHT,
tools=TOOLS
)
fig.title.text = title
fig.xaxis.axis_label = xlabel
fig.yaxis.axis_label = ylabel
fig.xaxis.formatter = models.NumeralTickFormatter(format='0,0')
if yformatter is not None:
fig.yaxis.formatter = yformatter
is_legend = False
for x in lines:
if x.legend != '':
is_legend = True
fig.line(x='Time', y=x.col, color=x.color, legend_label=x.legend, source=source)
else:
fig.line(x='Time', y=x.col, color=x.color, source=source)
if is_legend:
fig.legend.click_policy="hide"
return fig
def create_packets_plot(source, is_sender):
side_name = 'Sender' if is_sender else 'Receiver'
if is_sender:
lines = [
linedesc('pktSent', 'Sent', 'green'),
linedesc('pktSndLoss', 'Lost', 'orange'),
linedesc('pktRetrans', 'Retransmitted', 'blue'),
linedesc('pktSndDrop', 'Dropped', 'red'),
linedesc('pktFlightSize', 'On Flight', 'black'),
]
else:
lines = [
linedesc('pktRecv', 'Received', 'green'),
linedesc('pktRcvLoss', 'Lost', 'orange'),
linedesc('pktRcvRetrans', 'Retransmitted', 'blue'),
linedesc('pktRcvBelated', 'Belated', 'grey'),
linedesc('pktRcvDrop', 'Dropped', 'red'),
]
return create_plot(
'Packets (' + side_name + ' Side)',
'Time (ms)',
'Number of Packets',
source,
lines,
models.NumeralTickFormatter(format='0,0')
)
def figure_histogram(hist, edges, title: str, normalized=False, x_range=(0, 5000)):
""" Create and return `bokeh` histogram figure. """
fig = plotting.figure(
plot_height = 400,
plot_width=800,
tools=TOOLS,
background_fill_color='#fafafa',
x_range=x_range
)
fig.quad(
top=hist,
bottom=0,
left=edges[:-1],
right=edges[1:],
fill_color='navy',
line_color='white',
alpha=0.5
)
fig.title.text = title
fig.xaxis.axis_label = 'x, us'
fig.y_range.start = 0
fig.grid.grid_line_color='white'
if normalized:
fig.yaxis.axis_label = 'f(x)'
else:
fig.yaxis.axis_label = 'f(x), packets'
fig.yaxis.formatter = models.NumeralTickFormatter(format='0,0')
return fig
def _prepare_plot(self):
mean = np.array(self.mean)
deviation = np.array(self.deviation)
self._figure = bp.figure(title=self._title, plot_width=1200, plot_height=600)
self._figure.title.text_font_size = '20pt'
self._figure.xaxis.axis_label = "Base length"
self._figure.xaxis.axis_label_text_font_size = '16pt'
self._figure.xaxis.major_label_text_font_size = '12pt'
self._figure.yaxis[0].formatter = bm.NumeralTickFormatter(format="0.00%")
self._figure.yaxis[0].major_label_text_font_size = '12pt'
self._figure.line(self.scales, mean, legend="Error mean", line_color="red", line_width=2)
self._figure.circle(self.scales, mean, legend="Error mean", line_color="red", fill_color="red")
self._figure.square(self.scales, deviation, legend="Deviation", fill_color=None, line_color="green", line_width=2)
self._figure.line(self.scales, deviation, legend="Deviation", line_color="green")
self._figure.ygrid.minor_grid_line_color = '#e5e5e5'
self._figure.ygrid.minor_grid_line_dash = [6, 4]
self._figure.ygrid.minor_grid_line_alpha = 0.7
self._figure.legend.location = "top_right"
self._figure.legend.label_text_font_size = '16pt'
def new_plot(self, title, units, *, line_width=None, precision=2):
"""
Creates a blank line plot for with timestamps on the x-axis and
a line for each data series on the y-axis.
"""
plot = plotting.figure(title=title, tools=[])
self.active_plot = plot
self.plots.append(plot)
self.colors = list(colors)
self.units = units
self.line_width = line_width or self.line_width
plot.plot_width = self.width
plot.plot_height = self.height
plot.x_range = self.x_range
datetime_tick_formats = {
key: ["%a %b %d %H:%M:%S"]
for key in ("seconds", "minsec", "minutes", "hourmin", "hours", "days")
}
plot.xaxis.formatter = models.DatetimeTickFormatter(**datetime_tick_formats)
# https://bokeh.pydata.org/en/latest/docs/reference/models/formatters.html
# plot.yaxis.formatter = models.NumeralTickFormatter(format="0a")
plot.yaxis.formatter = models.NumeralTickFormatter(format="0,0.00 a")
# With default precision level 2 (decimal places)
# The units_formats = "@y{0,0.00}"
# and the value would look like 1,234,567.89
units_formats = f"@y{{0,0.{'0' * precision}}}"
hover = models.HoverTool(
# mode = vline would be nice to use,
# but then separate hovers block each when lines are too close.
# Would like a single hover box with time once, and a value per line
# perhaps this will help acheive that:
# https://stackoverflow.com/questions/29435200/bokeh-plotting-enable-tooltips-for-only-some-glyphs
mode="mouse", # other optins: vline
line_policy="nearest", # other optins: prev, next, nearest, interp, none
tooltips=[
("Name", "$name"),
("Time", "@x{%a %m/%d %H:%M:%S}"),
(self.units, units_formats),
],
formatters={"x": "datetime", "Time": "datetime", "@x": "datetime"},
)
plot.add_tools(hover)
plot.add_tools(models.BoxZoomTool())
plot.add_tools(models.HelpTool())
plot.add_tools(models.PanTool())
plot.add_tools(models.WheelZoomTool(dimensions="width"))
plot.toolbar.active_scroll = plot.select_one(models.WheelZoomTool)
plot.add_tools(models.WheelZoomTool(dimensions="height"))
plot.add_tools(models.UndoTool())
plot.add_tools(models.RedoTool())
plot.add_tools(models.ResetTool())
plot.add_tools(models.SaveTool())
def plot_percents(data: list) -> plotting.Figure:
"""Return the bokeh of temperature data."""
plot = make_timeseries_plot()
plot.yaxis.formatter = models.NumeralTickFormatter(format="0%")
plot.y_range = models.Range1d(0, 1)
plot = make_fact_lines(plot, data)
plot.legend.location = "top_left" # because the legend requires glyphs
return plot
def create_bandwidth_plot(source):
lines = [linedesc('mbpsBandwidth', '', 'green')]
return create_plot(
'Bandwith',
'Time (ms)',
'Bandwith (Mbps)',
source,
lines,
models.NumeralTickFormatter(format='0,0')
)
def create_window_size_plot(source):
lines = [
linedesc('pktFlowWindow', 'Flow Window', 'green'),
linedesc('pktCongestionWindow', 'Congestion Window', 'red'),
]
return create_plot(
'Window Size',
'Time (ms)',
'Number of Packets',
source,
lines,
models.NumeralTickFormatter(format='0,0')
)
def create_avail_buffer_plot(source, is_sender):
side_name = 'Sending' if is_sender else 'Receiving'
if is_sender:
if not 'byteAvailSndBuf' in source.column_names:
return None
lines = [linedesc('MBAvailSndBuf', '', 'green')]
else:
if not 'byteAvailRcvBuf' in source.column_names:
return None
lines = [linedesc('MBAvailRcvBuf', '', 'green')]
return create_plot(
'Available ' + side_name + ' Buffer Size',
'Time (ms)',
'MB',
source,
lines,
models.NumeralTickFormatter(format='0,0.00')
)
def create_rate_plot(source, is_sender):
side_name = 'Sending' if is_sender else 'Receiving'
if is_sender:
lines = [
linedesc('mbpsSendRate', 'Sendrate', 'green'),
# Please don't delete the following line.
# It is used in some cases.
#linedesc('mbpsMaxBW', 'Bandwidth Limit', 'black'),
]
else:
lines = [linedesc('mbpsRecvRate', '', 'green')]
return create_plot(
side_name + ' Rate',
'Time (ms)',
'Rate (Mbps)',
source,
lines,
models.NumeralTickFormatter(format='0,0.00')
)
def create_bytes_plot(source, is_sender):
side_name = 'Sender' if is_sender else 'Receiver'
if is_sender:
lines = [
linedesc('MBSent', 'Sent', 'green'),
linedesc('MBSndDrop', 'Dropped', 'red')
]
else:
lines = [
linedesc('MBRecv', 'Received', 'green'),
linedesc('MBRcvDrop', 'Dropped', 'red'),
]
return create_plot(
'Megabytes (' + side_name + ' Side)',
'Time (ms)',
'MB',
source,
lines,
models.NumeralTickFormatter(format='0,0.00')
)
def make_color_bar(cmap, formatting):
""" Generates color bar from make_color_mapper()
:param (Bokeh object) cmap: colormapper from make_color_mapper()
:param (dict) formatting: see DEFAULTFORMAT from params.py
:return: None (adds to Bokeh object) """
color_bar = models.ColorBar(color_mapper=cmap, label_standoff=10, location='bottom_right',
height=formatting['cbar_height'], background_fill_color=None,
major_label_text_font_size=formatting['cbar_fontsize'],
major_label_text_font=formatting['font'],
major_tick_line_color=formatting['cbar_tick_color'],
major_tick_line_alpha=formatting['cbar_tick_alpha'],
title=formatting['cbar_title'],
title_text_font_size=formatting['cbar_fontsize'],
title_text_font=formatting['font'],
title_text_align=formatting['cbar_title_align'],
title_text_font_style=formatting['cbar_style'],
title_standoff=int(formatting['width'] * \
formatting['cbar_title_standoff_ratio']))
if formatting['cbar_textfmt']:
color_bar.formatter = models.NumeralTickFormatter(
format=formatting['cbar_textfmt'])
return color_bar
def __init__(self, readers, rule):
super().__init__(readers, rule)
self.figure.yaxis[0].formatter = bom.NumeralTickFormatter(format="0%")
def set_yaxis_cash(plot):
plot.yaxis.axis_label = "Vuositulot"
plot.yaxis[0].formatter = bm.NumeralTickFormatter(format="€0")
def panel_eda(data: pd.DataFrame):
"""
The main panel with exploratory data analysis consisting of:
- packet timestamp vs packet inter-arrival time plot,
- basic statistics table,
- bar chart and table with sliced into intervals packet inter-arrival times,
- histograms regular and normalized, 100us bins,
- histograms regular and normalized, 10us bins,
- ECDF.
Attributes:
data:
`pd.DataFrame` with packet inter-arrival times data preliminary
cleaned up and consisting of the two columns `ws.time` and `ws.iat.us`,
where `ws.time` is the packet timestamp in seconds and `ws.iat.us` is
the corresponding inter-arrival time from the previous data packet
in microseconds.
Returns:
`models.widgets.Panel` bokeh panel.
"""
# Figure: Packet timestamp vs Inter-arrival packet time
source_data = models.ColumnDataSource(data)
fig_iat = plotting.figure(
plot_height=400,
plot_width=1000,
x_range=(0, 10),
tools=TOOLS
)
fig_iat.title.text = 'Inter-arrival packet time'
fig_iat.xaxis.axis_label = 'Time, s'
fig_iat.yaxis.axis_label = 'IAT, us'
fig_iat.yaxis.formatter = models.NumeralTickFormatter(format='0,0')
fig_iat.line(x='ws.time', y='ws.iat.us', source=source_data)
# Table: Statistics
stats = {}
stats['stats'] = [
'25th percentile (Q1), us',
'50th percentile (Median, Q2), us',
'75th percentile (Q3), us',
'90th percentile, us',
'95th percentile, us',
'99th percentile, us',
'Interquartile range (IQR, Q3 - Q1), us',
'Mean, us',
'Standard deviation, us',
'Min, us',
'Max, us',
'Packets',
]
stats['value'] = get_stats(data['ws.iat.us'])
source_stats = models.ColumnDataSource(pd.DataFrame(stats))
columns = [
models.widgets.TableColumn(field='stats', title='Statistic'),
models.widgets.TableColumn(field='value', title='Value'),
]
table_stats = models.widgets.DataTable(columns=columns, source=source_stats)
# Bar chart, table: Sliced into intervals inter-arrival packet times
fig_slicing, table_slicing = perform_slicing(data['ws.iat.us'])
# Histograms regular and normalized, 100us bins
bins_100 = [n * 100 for n in range(0, 5001)]
hist_100, edges_100 = np.histogram(data['ws.iat.us'], bins=bins_100)
fig_hist_100 = figure_histogram(hist_100, edges_100, 'Histogram of inter-arrival packet time, 100us bins')
hist_norm_100, edges_norm_100 = np.histogram(data['ws.iat.us'], bins=bins_100, density=True)
fig_hist_norm_100 = figure_histogram(hist_norm_100, edges_norm_100, 'Normalized histogram of inter-arrival packet time, 100us bins', True)
# Histograms regular and normalized, 100us bins
bins_10 = [n * 10 for n in range(0, 50001)]
hist_10, edges_10 = np.histogram(data['ws.iat.us'], bins=bins_10)
fig_hist_10 = figure_histogram(hist_10, edges_10, 'Histogram of inter-arrival packet time, 10us bins')
hist_norm_10, edges_norm_10 = np.histogram(data['ws.iat.us'], bins=bins_10, density=True)
fig_hist_norm_10 = figure_histogram(hist_norm_10, edges_norm_10, 'Normalized histogram of inter-arrival packet time, 10us bins', True)
# Synchronize x axeses of the histograms
fig_hist_100.x_range = \
fig_hist_norm_100.x_range = \
fig_hist_10.x_range = \
fig_hist_norm_10.x_range
# ECDF
x, y = ecdf(data['ws.iat.us'])
fig_ecdf = figure_ecdf(x, y)
# Create grid
grid = layouts.gridplot(
[
[fig_iat, table_stats],
[fig_slicing, table_slicing],
[fig_hist_100, fig_hist_norm_100],
[fig_hist_10, fig_hist_norm_10],
[None, fig_ecdf],
]
)
# Create panel
panel = models.widgets.Panel(child=grid, title='Exploratory Data Analysis')
return panel
def perform_slicing(s: pd.Series):
"""
Perform slicing of the packet IAT series `s` into following bins:
'0 - 10',
'10 - 100',
'100 - 500',
'500 - 1,000',
'1,000 - 5,000',
'5,000 - 10,000',
'10,000 - 50,000',
'50,000 - 100,000',
'100,000 - 500,000',
where each diapason is specified in microseconds (us).
Attributes:
s: `
pd.Series` of packet inter-arrival time in microseconds (us).
Returns:
`bokeh` bar chart figure and table with data.
"""
bins = [0, 10, 100, 500, 1000, 5000, 10000, 50000, 100000, 500000]
hist, edges = np.histogram(s, bins=bins)
bins_str = [
'0 - 10',
'10 - 100',
'100 - 500',
'500 - 1,000',
'1,000 - 5,000',
'5,000 - 10,000',
'10,000 - 50,000',
'50,000 - 100,000',
'100,000 - 500,000'
]
d = {}
d['edges.us'] = bins_str
d['packets'] = hist
df = pd.DataFrame(d)
n = df['packets'].sum()
df['packets_cumsum'] = df['packets'].cumsum()
df['percentage'] = df['packets'] * 100 / n
df['percentage_cumsum'] = round(df['percentage'].cumsum(), 4)
df['percentage'] = round(df['percentage'], 4)
# Figure
fig = plotting.figure(
plot_height=300,
plot_width=1000,
x_range=bins_str,
tools=TOOLS
)
fig.title.text = 'Inter-arrival packet time diapason vs Packets'
fig.xaxis.axis_label = 'IAT, us'
fig.yaxis.axis_label = 'Packets'
fig.yaxis.formatter = models.NumeralTickFormatter(format='0,0')
# fig.xaxis.major_label_orientation = math.pi/4
fig.vbar(x=bins_str, top=hist, width=0.9)
# Table
source = models.ColumnDataSource(df)
columns = [
models.widgets.TableColumn(field='edges.us', title='IAT, us'),
models.widgets.TableColumn(field='packets', title='Packets', formatter=models.NumberFormatter(format='0,0')),
models.widgets.TableColumn(field='packets_cumsum', title='Packets cumsum', formatter=models.NumberFormatter(format='0,0')),
models.widgets.TableColumn(field='percentage', title='Packets, %'),
models.widgets.TableColumn(field='percentage_cumsum', title='Packets cumsum, %'),
]
table = models.widgets.DataTable(columns=columns, source=source)
return fig, table