def main():
# prepare to read data and apply basic filtering
rows = common.rawData()
rows = (row for row in rows if row['SamplingRate'] == 1)
rows = (row for row in rows if row['Effort'] == 5)
percents = []
time_aggressive = []
time_only_metrics = []
for row in rows:
if row['conj_total'] == None:
continue
pruned = row['compl_cr_pruned_conj'] + row['compl_cr_pruned_disj']
total = row['conj_total'] + row['disj_total']
percents.append(float(pruned) / total)
time_aggressive.append(row['time_compl_cr'])
time_only_metrics.append(row['time'])
print 'Aggressive pruning efficiency: %lf of total' % common.average(
percents)
print '\nTimes:'
print 'When everything is computed (no pruning, no metrics): %d seconds' % average_time(
'data/times-compute-all.txt')
print 'Only metrics are applied: %d seconds' % common.average(
time_only_metrics)
print 'Moderate pruning is applied: %d seconds' % average_time(
'data/times-moderate-pruning.txt')
print 'Aggressive pruning is applied: %d seconds' % common.average(
time_aggressive)
def main():
[outfile, kind] = sys.argv[1:]
# prepare to read data and apply basic filtering
rows = common.rawData()
rows = (row for row in rows if row['SamplingRate'] == 1)
# collect and index individual data points of interest
data = {}
for row in rows:
coord = (row['Application'], row['Effort'])
value = row[kind + '_interesting']
if value == None: value = 0
data.setdefault(coord, []).append(value)
# summarize by averaging values at each (app, effort) coordinate
for coord, counts in data.iteritems():
data[coord] = common.average(counts)
# extract effort levels to serve as X axis category labels
efforts = (coord[1] for coord in data.iterkeys())
efforts = set(efforts)
efforts = sorted(efforts)
categories = [(effort, ) for effort in efforts]
# tick marks, in same order as common.apps
# line styles rotate automatically
ticks = [
tick_mark.blackcircle3, tick_mark.blacksquare3, tick_mark.blacktri3,
tick_mark.blackdtri3, tick_mark.plus3, tick_mark.x3
]
# create plot area
theme.output_file = outfile
common.setTheme()
x_coord = category_coord.T(categories, 0)
x_axis = axis.X(label='/ieffort', format='%d%%')
y_coord = log_coord.T()
y_axis = axis.Y(label='Number of interesting predicates',
format=common.format_count)
loc = {'conj': (100, 20), 'disj': (130, 7)}[kind]
leg = legend.T(loc=loc)
ar = area.T(x_axis=x_axis,
x_coord=x_coord,
y_axis=y_axis,
y_coord=y_coord,
legend=leg)
# add one line plot for each application
for (app, tick) in izip(common.apps, ticks):
slice = [(effort, data[app, effort]) for effort in efforts]
line = line_plot.T(label='/C/6' + app, data=slice, tick_mark=tick)
ar.add_plot(line)
# save the rendered result
ar.draw()
def plots():
prior = None
for fate in fates:
series = [(app, data[app, fate]) for app in common.apps]
overall = common.average([point[1] for point in series])
series.append(('Overall', overall))
bars = bar_plot.T(data=series, stack_on=prior, label=fate)
prior = bars
yield bars
def basic_suite(datapoints):
basic_list = []
basic_list.append(('Average', float_str_format(average(datapoints))))
basic_list.append(('Median', median(datapoints)))
basic_list.append(('Mode', mode(datapoints)))
basic_list.append(('Range', data_range(datapoints)))
basic_list.append(('Variance', float_str_format(variance(datapoints))))
basic_list.append(('Standard Deviation', float_str_format(standard_deviation(datapoints))))
return basic_list
def get_average_by_manufacturer(table, manufacturer):
"""Answer the question: What is the average amount of games in stock of
a given manufacturer?? Args: table with all games. Returns: number of games"""
stock_amount = [int(game[4]) for game in table if game[2] == manufacturer]
while not stock_amount:
ui.print_error_message("There is no such manufacturer.")
return
stock = common.average(stock_amount)
ui.print_result([stock], 'Amount of games by {}'.format(manufacturer))
return stock
def variance(datapoints):
if len(datapoints) == 0:
return None
mu = average(datapoints)
squaresum = 0
for dp in datapoints:
squaresum += (dp.as_float()-mu)*(dp.as_float()-mu)
sigma = squaresum/len(datapoints)
return sigma
def main():
subdir = dirname(__file__)
# prepare to read data and apply basic filtering
rows = common.rawData()
rows = (row for row in rows if row['Effort'] == 5)
# prepare master data table
densities = [1, 1.01, 2, 5, 10, 100, 1000]
kinds = ['Conjunctions', 'Simple', 'Disjunctions']
data = dict(((app, density, kind), []) for app in common.apps
for density in densities for kind in kinds)
# collect and index individual data points of interest
for row in rows:
app = row['Application']
density = row['SamplingRate']
def record(kind, column):
count = row[column]
if count != None:
data[app, density, kind].append(count)
record('Conjunctions', 'conj_interesting')
record('Disjunctions', 'disj_interesting')
record('Simple', 'simple_preds')
# summarize by averaging values at each (app, density, kind) coordinate
for coord, counts in data.iteritems():
data[coord] = common.average(counts)
if data[coord] == 0:
data[coord] = None
# create common plot area elements
common.setTheme(size=(120, 93))
densities = [1, 1.01, 2, 5, 10, 100, 1000]
categories = [(label, ) for label in densities]
x_coord = category_coord.T(categories, 0)
x_axis = axis.X(label='Sampling Rate', format=format_x)
y_coord = log_coord.T()
y_axis = axis.Y(label='Number of Predicates', format=common.format_count)
# create single-application plot
theme.reinitialize()
leg = legend.T(loc=(115, 10))
ar = area.T(x_axis=x_axis,
x_coord=x_coord,
y_axis=y_axis,
y_coord=y_coord,
legend=leg)
# one data series (line) per category of counted predicates
app = 'print_tokens2'
for series in kinds:
counts = [(density, data[app, density, series])
for density in densities]
ar.add_plot(line_plot.T(label=series, data=counts))
# render plots to disk
ar.draw(canvas.init('%s/sampling-%s.pdf' % (subdir, app)))
ar.draw(canvas.init('%s/sampling-%s.eps' % (subdir, app)))
# LaTeX fragment with table rows
latex = file(subdir + '/sampling.tex', 'w')
for app in common.apps:
protected = app.replace('_', '\\_')
print >> latex, '\\prog{%s}' % protected,
for series in ['Simple', 'Conjunctions', 'Disjunctions']:
for density in [1, 100, 1000]:
count = data[app, density, series]
if count == None:
cell = '-'
else:
cell = locale.format('%.0f', count, grouping=True)
print >> latex, '&', cell,
print >> latex, '\\\\'
def returnOnShe(netIncome, dividends, currentShe, prevShe):
return (netIncome - dividends) / common.average(currentShe, prevShe)
def returnOnAssets(netIncome, currentAssets, prevAssets):
return netIncome / common.average(currentAssets, prevAssets)
def assetTurnover(sales, currentAssets, prevAssets):
return sales / common.average(currentAssets, prevAssets)
def main():
[outfile, kind] = sys.argv[1:]
# prepare to read data and apply basic filtering
rows = common.rawData()
rows = (row for row in rows if row['SamplingRate'] == 1)
rows = (row for row in rows if row['Effort'] == 5)
# prepare master data table
data = dict(((app, fate), []) for app in common.apps for fate in fates)
# collect and index individual data points of interest
for row in rows:
app = row['Application']
total = row[kind + '_total']
if total != None:
compute = row[kind + '_computed']
retain = row[kind + '_interesting']
waste = compute - retain
bound = row[kind + '_ub_est_pruned']
effort = row[kind + '_pdg_metric_pruned']
assert retain + waste + bound + effort == total
total = float(total)
data[app, computeRetain].append(retain / total)
data[app, computeDiscard].append(waste / total)
data[app, pruneBound].append(bound / total)
data[app, pruneEffort].append(effort / total)
# summarize by averaging at each data point
for coord, values in data.iteritems():
data[coord] = common.average(values)
# create plot area
theme.output_file = outfile
common.setTheme()
x_coord = common.appsCoord(overall=True)
x_axis = common.appsAxisX()
y_axis = axis.Y(label='Fraction of complex predicates',
format=common.format_percent)
leg = legend.T(loc=(23, 100), shadow=(1, -1, fill_style.gray50))
ar = area.T(x_axis=x_axis,
x_coord=x_coord,
y_axis=y_axis,
y_range=(0, 1),
y_grid_interval=0.1,
legend=leg)
# add stacked bars in reverse order so legend looks right
def plots():
prior = None
for fate in fates:
series = [(app, data[app, fate]) for app in common.apps]
overall = common.average([point[1] for point in series])
series.append(('Overall', overall))
bars = bar_plot.T(data=series, stack_on=prior, label=fate)
prior = bars
yield bars
bars = reversed(list(plots()))
ar.add_plot(*bars)
# save rendered plot
ar.draw()
def integer_overview(query): datapoints = DataPoint.get_by_query(query) return 'Average: ' + float_str_format(average(datapoints))
def average_time(fname):
data = file(fname).readlines()[1:]
data = [line.split(' ')[-1] for line in data]
data = [int(line) for line in data]
return common.average(data)
def inventoryTurnover(Cogs, inventory1, inventory2):
return Cogs / common.average(inventory1, inventory2)
def receivablesTurnover(creditSales, grossAcctReceivable1,
grossAcctReceivable2):
return creditSales / common.average(grossAcctReceivable1,
grossAcctReceivable2)
recommenderFile = '../play/test-data/codechanges.txt.inferred_ast_operations.recommender' RANKING_INDEX = 1 TIME_INDEX = 3 header,results = common.readCSV(recommenderFile) nom_rows = len(results) num_col = len(results[0]) maxRankings = 0 rankings = [] maxTime = 0; time = [] for line in results: if (line[RANKING_INDEX] > maxRankings): maxRankings = line[RANKING_INDEX] if (line[TIME_INDEX] > maxTime): maxTime = line[TIME_INDEX] rankings.append(line[RANKING_INDEX]) time.append(line[TIME_INDEX]) rankings = [float(x) for x in rankings] time = [float(x) for x in time] print 'maximum ranking = ' + str(maxRankings) print 'Average ranking = ' + str(common.average(rankings)) print 'maximum time = ' + str(maxTime) print 'Average time = '+ str(common.average(time))
