def make_histogram(image, filename, numbins=32):
width, height = image.size
data = []
for i in range(width):
for j in range(height):
pixel = image.getpixel((i,j))
data.append(pixel)
p = Plot()
hist = Histogram(data, numbins)
p.append(hist)
p.write(filename)
def plotmixing(samples, param_name, filename=None):
p = Plot(title='%s mixing plot' % param_name)
points = Points([(i,s[param_name]) for i,s in enumerate(samples)],
style='lines')
points.linewidth=1
p.append(points)
p.show()
if filename:
p.write(filename)
def do_iris_HAC():
numvars, names, data, labels = read_data('iris.cluster.txt')
DE = HAC(data, .1)
data = []
for grouping in DE:
if grouping[0] != 0:
K = grouping[2]
K.compute_means()
data.append((grouping[1], K.error()))
print grouping[0], grouping[1], K.error()
for cluster in K.clusters:
cluster.indices.sort()
print cluster.indices
plot = Plot()
plot.title = 'SSE as a function of number of clusters in HAC'
plot.append(Points(data, style='lines'))
plot.write('hac_clusters.gpi')
from psodata import PSOData
from evilplot import Plot, Points, RawData
MAX_BAR_SAMPLES = 40
MAX_SAMPLES = 500
parser = optparse.OptionParser()
parser.add_option('--print', dest='print_page', action='store_true',
help='Send to the printer')
opts, args = parser.parse_args()
if not args:
parser.error('Log file not specified.')
plot = Plot()
plot.ylogscale = 10
plot.xlabel = 'Function Evaluations'
plot.ylabel = 'Best Function Value'
for filename in args:
data = PSOData(open(filename))
trim = int(len(data) / 10)
points = []
bars = []
iterations = len(data[0])
samples_step = int(math.ceil(iterations / MAX_SAMPLES))
bar_samples_step = int(math.ceil(iterations / MAX_BAR_SAMPLES))
for iteration in islice(data[0], 0, None, samples_step):
points.append((iteration, data.average(iteration)))
for iteration in islice(data[0], 0, None, bar_samples_step):
def plotposterior(samples, param_name, min=None, max=None, prior=None,
filename=None):
p = Plot(title='Posterior of %s' % param_name)
density = Density([s[param_name] for s in samples], title='Posterior')
p.append(density)
if prior:
priord = Function(prior, title='Prior')
p.append(priord)
if min is not None:
p.xmin = min
if max is not None:
p.xmax = max
p.show()
if filename:
p.write(filename)
def main():
tm_results = parse_file('with_topic_model.log')
ntm_results = parse_file('without_topic_model.log')
p = Plot(title="Precision/Recall for Labeled Mentions")
p.xmin = 0
p.xmax = 1
p.ymin = 0
p.ymax = 1
p.append(Points(tm_results[0].prec_rec,
style='lines', title='With topic model'))
p.append(Points(ntm_results[4].prec_rec,
style='lines', title='Without topic model'))
p.append(Points(tm_results[14].prec_rec,
style='lines', title='With topic model after sampling'))
p.append(Points(ntm_results[14].prec_rec,
style='lines', title='Without topic model after sampling'))
p.append(Points([(tm_results[0].baseline_rec,
tm_results[0].baseline_prec)],
title='Baseline performance'))
p.write('prec_rec.gpi')
p.show()