def test_add_subtract(self):
'''Test add and subtract on some subclasses'''
bar = vincent.Bar()
area = vincent.Area()
area + ({
'value': 'basis'
}, 'marks', 0, 'properties', 'enter', 'interpolate')
bar + ('red', 'marks', 0, 'properties', 'hover', 'fill', 'value')
assert area.marks[0]['properties']['enter'].has_key('interpolate')
assert bar.marks[0]['properties']['hover']['fill']['value'] == 'red'
bar - ('domain', 'scales', 1)
bar -= ('name', 'scales', 1)
area - ('scale', 'axes', 0)
area -= ('type', 'axes', 1)
assert bar.scales[1] == {'nice': True, 'range': 'height'}
assert area.axes == [{'type': 'x'}, {'scale': 'y'}]
def api_score_timeline(player_id):
score_series = common.scores.get_score_timeline(
player_id,
int(time.time()) - SECONDS_IN_3_WEEKS)
vis = vincent.Area()
vis.tabular_data(score_series, axis_time='day')
vis += ({'labels': {'angle': {'value': 25}}}, 'axes', 0, 'properties')
vis += ({'value': 22}, 'axes', 0, 'properties', 'labels', 'dx')
vis.update_vis(padding={'bottom': 50, 'left': 60, 'right': 40, 'top': 10})
vis.update_vis(width=700)
# here we go..
# determine the bottom limit - pad below by a bit, and truncate to mult of 10
domain_min = round((score_series.min() - 10) / 10.0) * 10
vis += (False, 'scales', 1, 'zero')
vis += (domain_min, 'scales', 1, 'domainMin')
vis += (domain_min, 'marks', 0, 'properties', 'enter', 'y2', 'value')
# colors..
# remove the hover color change, it is stupid
vis -= ('hover', 'marks', 0, 'properties')
# and change the main color to red because I wanna
vis += ({'value': '#c1251f'}, 'marks', 0, 'properties', 'update', 'fill')
return json.dumps(vis.vega)
def __email_summary(self):
email_interval = self.__config['email']['interval']
email_subject = '服务器监控数据'
while True:
time.sleep(email_interval)
now = datetime.now().strftime('%Y-%m-%d %H:%M:%S')
stat_data = {
'cpu_stat': {
'used_percent': [],
'created_at': []
},
'mem_stat': {
'used_percent': [],
'created_at': []
}
}
for table_name in stat_data.keys():
for_select = 'SELECT used_percent, created_at FROM %s WHERE created_at < "%s" ORDER BY created_at' \
% (table_name, now)
for row in self.__db.query(for_select):
stat_data[table_name]['used_percent'].append(row[0])
stat_data[table_name]['created_at'].append(row[1])
for_delete = 'DELETE FROM %s WHERE created_at < "%s"' % (
table_name, now)
self.__db.execute(for_delete)
# cpu
cpu_stat_data = self.__data_preprocess(stat_data['cpu_stat'])
# memory
mem_stat_data = self.__data_preprocess(stat_data['mem_stat'])
cpu_data_count = len(cpu_stat_data['created_at'])
mem_data_count = len(mem_stat_data['created_at'])
if self.__config['email_type'] == 'advanced':
# 生成绘图JSON数据的细节还得仔细研究研究
for index in xrange(cpu_data_count):
cpu_stat_data['created_at'][index] = datetime.strptime(
cpu_stat_data['created_at'][index],
'%Y-%m-%d %H:%M:%S')
series = pandas.Series(cpu_stat_data['used_percent'],
index=cpu_stat_data['created_at'])
cpu_graph = vincent.Area(series)
cpu_graph.axis_titles(x=u'Time', y=u'Usage (%)')
ax = AxisProperties(labels=PropertySet(angle=ValueRef(
value=-30)))
cpu_graph.axes[0].properties = ax
cpu_graph_json = cpu_graph.to_json()
for index in xrange(mem_data_count):
mem_stat_data['created_at'][index] = datetime.strptime(
mem_stat_data['created_at'][index],
'%Y-%m-%d %H:%M:%S')
series = pandas.Series(mem_stat_data['used_percent'],
index=mem_stat_data['created_at'])
mem_graph = vincent.Area(series)
mem_graph.axis_titles(x=u'Time', y=u'Usage (%)')
ax = AxisProperties(labels=PropertySet(angle=ValueRef(
value=-30)))
mem_graph.axes[0].properties = ax
mem_graph_json = mem_graph.to_json()
email_content = self.__render_content(
template_name='templates/monitor_stat.html',
data={
'cpu_stat': {
'data': cpu_graph_json,
'target_file_name': 'cpu_graph.png'
},
'mem_stat': {
'data': mem_graph_json,
'target_file_name': 'mem_graph.png'
}
})
if email_content is None:
print u'模板渲染失败！'
break
print self.__email_it(subject=email_subject,
content=email_content,
attach_files=[{
'file_name': 'cpu_graph.png',
'file_id': 'cpu_stat'
}, {
'file_name': 'mem_graph.png',
'file_id': 'mem_stat'
}])
else:
# 娶最大的5个
max_n = 5
cpu_data_tuples = [(cpu_stat_data['created_at'][index],
cpu_stat_data['used_percent'][index])
for index in xrange(cpu_data_count)]
cpu_data_sorted = sorted(cpu_data_tuples,
key=lambda item: item[1])
cpu_sorted_max = []
if cpu_data_count >= max_n:
cpu_sorted_max.extend(cpu_data_sorted[0 - max_n:])
mem_data_tuples = [(mem_stat_data['created_at'][index],
mem_stat_data['used_percent'][index])
for index in xrange(mem_data_count)]
mem_data_sorted = sorted(mem_data_tuples,
key=lambda item: item[1])
mem_sorted_max = []
if mem_data_count >= max_n:
mem_sorted_max.extend(mem_data_sorted[0 - max_n:])
email_content = self.__render_content(
template_name='templates/monitor_stat.html',
data={
'max_n': max_n,
'cpu_stat': {
'max_n_data': cpu_sorted_max
},
'mem_stat': {
'max_n_data': mem_sorted_max
}
})
if email_content is None:
print u'模板渲染失败！'
break
print self.__email_it(subject=email_subject,
content=email_content)
# -*- coding: utf-8 -*-
'''
Builds a Vega grammar specification from vincent.Area()
'''
import vincent
import random
vis = vincent.Area()
vis.tabular_data([random.randint(10, 100) for x in range(0, 16, 1)])
#Generate both the Vega JSON and a data JSON.
path = r'/vega.json'
vis.to_json(path, split_data=True, html=True)
#Lets add a data interpolation parameter and resave the JSON
vis + ({'value': 'basis'}, 'marks', 0, 'properties', 'enter', 'interpolate')
vis.to_json(path, split_data=True, html=True)
import pandas as p
import vincent
from pandas.tseries.resample import TimeGrouper
from pandas.tseries.offsets import DateOffset
tweets = p.read_csv('./tweets.csv')
tweets['created_at'] = p.to_datetime(p.Series(tweets['created_at']))
# set index to 'created_at'
tweets.set_index('created_at', drop=False, inplace=True)
tweets.index = tweets.index.tz_localize('GMT').tz_convert('EST')
# convert to 12 hour format
tweets.index = tweets.index - DateOffset(hours=12)
# created_at index is formatted to per minute
tweets_pm = tweets['created_at'].resample('1t', how='count')
# create time series graph via Vincent
vincent.core.initialize_notebook()
area = vincent.Area(tweets_pm)
area.colors(brew='Spectral')
area.display()
# <codecell> price['YHOO'] # <codecell> line = vincent.Line(price[['FB', 'YHOO']]) line.axis_titles(x='Date', y='Price') line.legend(title='FB vs YHOO') line.width = 600 line.height = 400 line.display() # <codecell> area = vincent.Area(list_data) area.width = 600 area.height = 400 area.display() # <codecell> stacked = vincent.StackedArea(multi_iter1, iter_idx='index') stacked.axis_titles(x='Index', y='Value') stacked.legend(title='Categories') stacked.height = 300 stacked.width = 600 stacked.display() # <codecell>
#Daily data
dates = pd.date_range('4/1/2013 00:00:00', periods=1441, freq='T')
data = [random.randint(20, 100) for x in range(len(dates))]
series = pd.Series(data, index=dates)
vis = vincent.Line()
vis.tabular_data(series, axis_time='day')
vis += ({'value': 'basis'}, 'marks', 0, 'properties', 'enter', 'interpolate')
vis.update_vis(width=800)
vis.axis_label(x_label='Time', y_label='Data')
vis.to_json(path, html=True)
#Resample to hourly, which can take a lambda function in addition to the
#standard mean, max, min, etc.
half_day = series['4/1/2013 00:00:00':'4/1/2013 12:00:00']
hourly = half_day.resample('H', how=lambda x: x.mean() + random.randint(-30, 40))
area = vincent.Area()
area.tabular_data(hourly, axis_time='hour')
area += ({'value': 'basis'}, 'marks', 0, 'properties', 'enter', 'interpolate')
area.update_vis(width=800)
area.axis_label(x_label='Time (Hourly)', y_label='Data')
area.to_json(path, html=True)
#Subset of minute data
half_hour = series['4/1/2013 00:00:00':'4/1/2013 00:30:00']
vis.tabular_data(half_hour, axis_time='minute')
vis.axis_label(x_label='Time (Minutes)', title='Data vs. Time')
vis.to_json(path, html=True)
tweets.set_index('created_at', drop=False, inplace=True)
tweets.index = tweets.index.tz_localize('GMT').tz_convert('Europe/Athens')
tweets.index = tweets.index - DateOffset(hours = 24)
tweets.index
#created_at timeseries in a per minute minute format
tweets1m = tweets['created_at'].resample('1t').count()
#average tweets per minute
avg = tweets1m.mean()
#plot tweets timeseries
import vincent
vincent.core.initialize_notebook()
area = vincent.Area(tweets1m)
area.colors(brew='Spectral')
area.axis_titles(x='time', y='tweets')
area.display()
#find most frequent tokens with NLTK
#remove english stopwords
import nltk
from nltk.corpus import stopwords
from nltk import FreqDist
#added "greek" file at ../nltk_data/corpora/stopwords
#Combined the following files:
#file source #1 https://code.grnet.gr/projects/gss/repository/revisions/d59fbcd2f0cd/entry/solr/conf/stopwords.txt
#file source #2 http://grepcode.com/file/repo1.maven.org/maven2/gr.skroutz/elasticsearch-skroutz-greekstemmer/0.0.4/org/elasticsearch/index/analysis/stopwords.txt
#NOAA 46401 Wave Period
vis1 = vincent.Line(width=600)
vis1.tabular_data(NOAA_46041, columns=['dominant_wave_period (s)'],
axis_time='day')
vis1.axis_label(x_label='Time', y_label='Dominant Wave Period (s)')
vis1.to_json('vis1.json')
#NOAA 46050 Binned Wind Speed
vis2 = vincent.Bar(width=600)
vis2.tabular_data(ws_binned)
vis2.axis_label(x_label='Wind Speed (m/s)', y_label='# of Obs')
vis2.to_json('vis2.json')
#NOAA 46243 Wave Height
vis3 = vincent.Area(width=600)
vis3.tabular_data(NOAA_46243, columns=['significant_wave_height (m)'],
axis_time='day')
vis3.axis_label(x_label='Time', y_label='Significant Wave Height (m)')
vis3.to_json('vis3.json')
#Map all buoys
buoy_map = folium.Map(location=[46.3014, -123.7390], zoom_start=7,
tiles='Stamen Terrain')
buoy_map.polygon_marker(location=[47.3489, -124.708], fill_color='#43d9de',
radius=12, popup=(vis1, 'vis1.json'))
buoy_map.polygon_marker(location=[44.639, -124.5339], fill_color='#43d9de',
radius=12, popup=(vis2, 'vis2.json'))
buoy_map.polygon_marker(location=[46.216, -124.1280], fill_color='#43d9de',
radius=12, popup=(vis3, 'vis3.json'))
buoy_map.create_map(path='NOAA_buoys.html')
# NOAA 46401 Wave Period.
vis1 = vincent.Line(NOAA_46041['dominant_wave_period (s)'],
width=400,
height=200)
vis1.axis_titles(x='Time', y='Dominant Wave Period (s)')
vis1.to_json('vis1.json')
# NOAA 46050 Binned Wind Speed.
vis2 = vincent.Bar(ws_binned, width=400, height=200)
vis2.axis_titles(x='Wind Speed (m/s)', y='# of Obs')
vis2.to_json('vis2.json')
# NOAA 46243 Wave Height.
vis3 = vincent.Area(NOAA_46243['significant_wave_height (m)'],
width=400,
height=200)
vis3.axis_titles(x='Time', y='Significant Wave Height (m)')
vis3.to_json('vis3.json')
# Map all buoys.
buoy_map = folium.Map(location=[46.3014, -123.7390],
zoom_start=7,
tiles='Stamen Terrain')
buoy_map.polygon_marker(location=[47.3489, -124.708],
fill_color='#43d9de',
radius=12,
popup=(vis1, 'vis1.json'))
buoy_map.polygon_marker(location=[44.639, -124.5339],
fill_color='#43d9de',
radius=12,