#!/usr/bin/env python
# -*- coding: utf-8 -*-
from json import JSONDecoder
import random
import pygal
from pygal.style import Style
import pandas
from datetime import datetime
from datetime import timedelta
from dateutil.parser import parse
import numpy
from scipy import stats
from pytaxonomies import Taxonomies
import re
import matplotlib.pyplot as plt
from matplotlib import pylab
import os
class DateError(Exception):
def __init__(self, value):
self.value = value
def __str__(self):
return repr(self.value)
# ############### Date Tools ################
def dateInRange(datetimeTested, begin=None, end=None):
if begin is None:
begin = datetime(1970, 1, 1)
if end is None:
end = datetime.now()
return begin <= datetimeTested <= end
def toDatetime(date):
return parse(date)
def checkDateConsistancy(begindate, enddate, lastdate):
if begindate is not None and enddate is not None:
if begindate > enddate:
raise DateError('begindate ({}) cannot be after enddate ({})'.format(begindate, enddate))
if enddate is not None:
if toDatetime(enddate) < lastdate:
raise DateError('enddate ({}) cannot be before lastdate ({})'.format(enddate, lastdate))
if begindate is not None:
if toDatetime(begindate) > datetime.now():
raise DateError('begindate ({}) cannot be after today ({})'.format(begindate, datetime.now().date()))
def setBegindate(begindate, lastdate):
return max(begindate, lastdate)
def setEnddate(enddate):
return min(enddate, datetime.now())
def getLastdate(last):
return (datetime.now() - timedelta(days=int(last))).replace(hour=0, minute=0, second=0, microsecond=0)
def getNDaysBefore(date, days):
return (date - timedelta(days=days)).replace(hour=0, minute=0, second=0, microsecond=0)
def getToday():
return (datetime.now()).replace(hour=0, minute=0, second=0, microsecond=0)
# ############### Tools ################
def getTaxonomies(dataframe):
taxonomies = Taxonomies()
taxonomies = list(taxonomies.keys())
notInTaxo = []
count = 0
for taxonomy in taxonomies:
empty = True
for it in dataframe.iterrows():
if it[0].startswith(taxonomy):
empty = False
dataframe = dataframe.drop([it[0]])
count = count + 1
if empty is True:
notInTaxo.append(taxonomy)
if dataframe.empty:
emptyOther = True
else:
emptyOther = False
for taxonomy in notInTaxo:
taxonomies.remove(taxonomy)
return taxonomies, emptyOther
def buildDoubleIndex(index1, index2, datatype):
it = -1
newindex1 = []
for index in index2:
if index == 0:
it += 1
newindex1.append(index1[it])
arrays = [newindex1, index2]
tuples = list(zip(*arrays))
return pandas.MultiIndex.from_tuples(tuples, names=['event', datatype])
def buildNewColumn(index2, column):
it = -1
newcolumn = []
for index in index2:
if index == 0:
it += 1
newcolumn.append(column[it])
return newcolumn
def addColumn(dataframe, columnList, columnName):
dataframe.loc[:, columnName] = pandas.Series(columnList, index=dataframe.index)
def concat(data):
return pandas.concat(data, axis=1)
def createFakeEmptyTagsSeries():
return pandas.Series({'Faketag': 0})
def removeFaketagRow(dataframe):
return dataframe.drop(['Faketag'])
def getCopyDataframe(dataframe):
return dataframe.copy()
def createDictTagsColour(colourDict, tags):
temp = tags.groupby(['name', 'colour']).count()['id']
levels_name = temp.index.levels[0]
levels_colour = temp.index.levels[1]
labels_name = temp.index.labels[0]
labels_colour = temp.index.labels[1]
for i in range(len(labels_name)):
colourDict[levels_name[labels_name[i]]] = levels_colour[labels_colour[i]]
def createTagsPlotStyle(dataframe, colourDict, taxonomy=None):
colours = []
if taxonomy is not None:
for it in dataframe.iterrows():
if it[0].startswith(taxonomy):
colours.append(colourDict[it[0]])
else:
for it in dataframe.iterrows():
colours.append(colourDict[it[0]])
style = Style(background='transparent',
plot_background='#eeeeee',
foreground='#111111',
foreground_strong='#111111',
foreground_subtle='#111111',
opacity='.6',
opacity_hover='.9',
transition='400ms ease-in',
colors=tuple(colours))
return style
# ############### Formatting ################
def eventsListBuildFromList(filename):
with open(filename, 'r') as myfile:
s = myfile.read().replace('\n', '')
decoder = JSONDecoder()
s_len = len(s)
Events = []
end = 0
while end != s_len:
Event, end = decoder.raw_decode(s, idx=end)
Events.append(Event)
data = []
for e in Events:
data.append(pandas.DataFrame.from_dict(e, orient='index'))
Events = pandas.concat(data)
for it in range(Events['attribute_count'].size):
if Events['attribute_count'][it] is None:
Events['attribute_count'][it] = '0'
else:
Events['attribute_count'][it] = int(Events['attribute_count'][it])
Events = Events.set_index('id')
return Events
def eventsListBuildFromArray(jdata):
'''
returns a structure listing all primary events in the sample
'''
data = [pandas.DataFrame.from_dict(e, orient='index') for e in jdata['response']]
events = pandas.concat(data)
events = events.set_index(['id'])
return events
def attributesListBuild(events):
attributes = [pandas.DataFrame(attribute) for attribute in events['Attribute']]
return pandas.concat(attributes)
def tagsListBuild(Events):
Tags = []
if 'Tag' in Events.columns:
for Tag in Events['Tag']:
if type(Tag) is not list:
continue
Tags.append(pandas.DataFrame(Tag))
if Tags:
Tags = pandas.concat(Tags)
columnDate = buildNewColumn(Tags.index, Events['date'])
addColumn(Tags, columnDate, 'date')
index = buildDoubleIndex(Events.index, Tags.index, 'tag')
Tags = Tags.set_index(index)
else:
Tags = None
return Tags
def selectInRange(Events, begin=None, end=None):
inRange = []
for i, Event in Events.iterrows():
if dateInRange(parse(Event['date']), begin, end):
inRange.append(Event.tolist())
inRange = pandas.DataFrame(inRange)
temp = Events.columns.tolist()
if inRange.empty:
return None
inRange.columns = temp
return inRange
def isTagIn(dataframe, tag):
temp = dataframe[dataframe['name'].str.contains(tag)].index.tolist()
index = []
for i in range(len(temp)):
if temp[i][0] not in index:
index.append(temp[i][0])
return index
def renameColumns(dataframe, namelist):
dataframe.columns = namelist
return dataframe
def replaceNaN(dataframe, value):
return dataframe.fillna(value)
# ############### Basic Stats ################
def getNbitems(dataframe):
return len(dataframe.index)
def getNbAttributePerEventCategoryType(attributes):
return attributes.groupby(['event_id', 'category', 'type']).count()['id']
def getNbOccurenceTags(Tags):
return Tags.groupby('name').count()['id']
# ############### Charts ################
def createTable(colors, categ_types_hash, tablename='attribute_table.html'):
with open(tablename, 'w') as target:
target.write('\n\n
\n\n\n')
for categ_name, types in categ_types_hash.items():
table = pygal.Treemap(pretty_print=True)
target.write('\n
{}
\n'.format(colors[categ_name], categ_name))
for d in types:
table.add(d['label'], d['value'])
target.write(table.render_table(transpose=True))
target.write('\n\n')
def createTreemap(data, title, treename='attribute_treemap.svg', tablename='attribute_table.html'):
labels_categ = data.index.labels[0]
labels_types = data.index.labels[1]
names_categ = data.index.levels[0]
names_types = data.index.levels[1]
categ_types_hash = {}
for categ_id, type_val, total in zip(labels_categ, labels_types, data):
if not categ_types_hash.get(names_categ[categ_id]):
categ_types_hash[names_categ[categ_id]] = []
dict_to_print = {'label': names_types[type_val], 'value': total}
categ_types_hash[names_categ[categ_id]].append(dict_to_print)
colors = {categ: "#%06X" % random.randint(0, 0xFFFFFF) for categ in categ_types_hash.keys()}
style = Style(background='transparent',
plot_background='#FFFFFF',
foreground='#111111',
foreground_strong='#111111',
foreground_subtle='#111111',
opacity='.6',
opacity_hover='.9',
transition='400ms ease-in',
colors=tuple(colors.values()))
treemap = pygal.Treemap(pretty_print=True, legend_at_bottom=True, style=style)
treemap.title = title
treemap.print_values = True
treemap.print_labels = True
for categ_name, types in categ_types_hash.items():
treemap.add(categ_name, types)
createTable(colors, categ_types_hash)
treemap.render_to_file(treename)
def tagsToLineChart(dataframe, title, dates, colourDict):
style = createTagsPlotStyle(dataframe, colourDict)
line_chart = pygal.Line(x_label_rotation=20, style=style, show_legend=False)
line_chart.title = title
line_chart.x_labels = dates
for it in dataframe.iterrows():
line_chart.add(it[0], it[1].tolist())
line_chart.render_to_file('tags_repartition_plot.svg')
def tagstrendToLineChart(dataframe, title, dates, split, colourDict):
style = createTagsPlotStyle(dataframe, colourDict)
line_chart = pygal.Line(x_label_rotation=20, style=style, show_legend=False)
line_chart.title = title
line_chart.x_labels = dates
xi = numpy.arange(split)
for it in dataframe.iterrows():
slope, intercept, r_value, p_value, std_err = stats.linregress(xi, it[1])
line = slope * xi + intercept
line_chart.add(it[0], line, show_dots=False)
line_chart.render_to_file('tags_repartition_trend_plot.svg')
def tagsToTaxoLineChart(dataframe, title, dates, colourDict, taxonomies, emptyOther):
style = createTagsPlotStyle(dataframe, colourDict)
line_chart = pygal.Line(x_label_rotation=20, style=style)
line_chart.title = title
line_chart.x_labels = dates
for taxonomy in taxonomies:
taxoStyle = createTagsPlotStyle(dataframe, colourDict, taxonomy)
taxo_line_chart = pygal.Line(x_label_rotation=20, style=taxoStyle)
taxo_line_chart.title = title + ': ' + taxonomy
taxo_line_chart.x_labels = dates
for it in dataframe.iterrows():
if it[0].startswith(taxonomy):
taxo_line_chart.add(re.sub(taxonomy + ':', '', it[0]), it[1].tolist())
dataframe = dataframe.drop([it[0]])
taxo_line_chart.render_to_file('plot/' + taxonomy + '.svg')
if not emptyOther:
taxoStyle = createTagsPlotStyle(dataframe, colourDict)
taxo_line_chart = pygal.Line(x_label_rotation=20, style=taxoStyle)
taxo_line_chart.title = title + ': other'
taxo_line_chart.x_labels = dates
for it in dataframe.iterrows():
taxo_line_chart.add(it[0], it[1].tolist())
taxo_line_chart.render_to_file('plot/other.svg')
def tagstrendToTaxoLineChart(dataframe, title, dates, split, colourDict, taxonomies, emptyOther):
style = createTagsPlotStyle(dataframe, colourDict)
line_chart = pygal.Line(x_label_rotation=20, style=style)
line_chart.title = title
line_chart.x_labels = dates
xi = numpy.arange(split)
for taxonomy in taxonomies:
taxoStyle = createTagsPlotStyle(dataframe, colourDict, taxonomy)
taxo_line_chart = pygal.Line(x_label_rotation=20, style=taxoStyle)
taxo_line_chart.title = title + ': ' + taxonomy
taxo_line_chart.x_labels = dates
for it in dataframe.iterrows():
if it[0].startswith(taxonomy):
slope, intercept, r_value, p_value, std_err = stats.linregress(xi, it[1])
line = slope * xi + intercept
taxo_line_chart.add(re.sub(taxonomy + ':', '', it[0]), line, show_dots=False)
dataframe = dataframe.drop([it[0]])
taxo_line_chart.render_to_file('plot/' + taxonomy + '_trend.svg')
if not emptyOther:
taxoStyle = createTagsPlotStyle(dataframe, colourDict)
taxo_line_chart = pygal.Line(x_label_rotation=20, style=taxoStyle)
taxo_line_chart.title = title + ': other'
taxo_line_chart.x_labels = dates
for it in dataframe.iterrows():
slope, intercept, r_value, p_value, std_err = stats.linregress(xi, it[1])
line = slope * xi + intercept
taxo_line_chart.add(it[0], line, show_dots=False)
taxo_line_chart.render_to_file('plot/other_trend.svg')
def tagsToPolyChart(dataframe, split, colourDict, taxonomies, emptyOther, order):
for taxonomy in taxonomies:
for it in dataframe.iterrows():
if it[0].startswith(taxonomy):
points = []
for i in range(split):
points.append((i, it[1][i]))
color = colourDict[it[0]]
label = re.sub(taxonomy + ':', '', it[0])
points = numpy.array(points)
dataframe = dataframe.drop([it[0]])
# get x and y vectors
x = points[:, 0]
y = points[:, 1]
# calculate polynomial
z = numpy.polyfit(x, y, order)
f = numpy.poly1d(z)
# calculate new x's and y's
x_new = numpy.linspace(x[0], x[-1], 50)
y_new = f(x_new)
plt.plot(x, y, '.', color=color)
plt.plot(x_new, y_new, color=color, label=label + 'trend')
pylab.title('Polynomial Fit with Matplotlib: ' + taxonomy)
pylab.legend(loc='center left', bbox_to_anchor=(1, 0.5))
ax = plt.gca()
# ax.set_facecolor((0.898, 0.898, 0.898))
box = ax.get_position()
ax.set_position([box.x0 - 0.01, box.y0, box.width * 0.78, box.height])
fig = plt.gcf()
fig.set_size_inches(20, 15)
fig.savefig('plotlib/' + taxonomy + '.png')
fig.clf()
if not emptyOther:
for it in dataframe.iterrows():
points = []
for i in range(split):
points.append((i, it[1][i]))
color = colourDict[it[0]]
label = it[0]
points = numpy.array(points)
# get x and y vectors
x = points[:, 0]
y = points[:, 1]
# calculate polynomial
z = numpy.polyfit(x, y, order)
f = numpy.poly1d(z)
# calculate new x's and y's
x_new = numpy.linspace(x[0], x[-1], 50)
y_new = f(x_new)
plt.plot(x, y, '.', color=color, label=label)
plt.plot(x_new, y_new, color=color, label=label + 'trend')
pylab.title('Polynomial Fit with Matplotlib: other')
pylab.legend(loc='center left', bbox_to_anchor=(1, 0.5))
ax = plt.gca()
#cax.set_facecolor((0.898, 0.898, 0.898))
box = ax.get_position()
ax.set_position([box.x0 - 0.01, box.y0, box.width * 0.78, box.height])
fig = plt.gcf()
fig.set_size_inches(20, 15)
fig.savefig('plotlib/other.png')
def createVisualisation(taxonomies):
chain = '\n\n\t\n\t\t\n\t\n\t'
chain = chain + '
'
for taxonomy in taxonomies:
chain = chain + '
\n'
chain = chain + '
\n'
chain = chain + '
'
chain = chain + '\n\t\n'
with open('test_tags_trend.html', 'w') as target:
target.write(chain)