def overall_collection_stats():
doc_lengths = []
for prefix in all_prefixes:
doc_lengths += raw_doc_lengths(prefix).values()
print "All documents length mean: %.2f" % n.mean(doc_lengths)
print "All documents length std: %.2f" % n.std(doc_lengths)
print "All documents length median: %.2f" % n.median(doc_lengths)
medlist, meanlist, stdlist = zip(*map(length_stats, all_prefixes))
print "Mean of means: %.2f" % n.mean(meanlist)
print "Mean of std: %.2f" % n.mean(stdlist)
print "Mean of median: %.2f" % n.mean(medlist)
def overall_collection_stats():
doc_lengths = []
for prefix in all_prefixes:
doc_lengths += raw_doc_lengths(prefix).values()
print "All documents length mean: %.2f" % n.mean(doc_lengths)
print "All documents length std: %.2f" % n.std(doc_lengths)
print "All documents length median: %.2f" % n.median(doc_lengths)
medlist, meanlist, stdlist = zip(*map(length_stats,all_prefixes))
print "Mean of means: %.2f" % n.mean(meanlist)
print "Mean of std: %.2f" % n.mean(stdlist)
print "Mean of median: %.2f" % n.mean(medlist)
def doclength_histogram(path, prefix):
values = p.array(raw_doc_lengths(prefix).values())
num_bins = 1000
bin_upper_limit = p.mean(values) + 3 * p.std(values)
print "UL: "+ str(bin_upper_limit)
bins = p.array(range(1,1001)) * (bin_upper_limit/1000.0)
p.hist(values, bins)
p.xlabel('Document size (unicode codepoints)')
p.ylabel('Number of documents')
p.title('Document Size Histogram for %s' % prefix)
p.savefig(path, dpi=72)
p.close()
def doclength_histogram(path, prefix):
values = p.array(raw_doc_lengths(prefix).values())
num_bins = 1000
bin_upper_limit = p.mean(values) + 3 * p.std(values)
print "UL: " + str(bin_upper_limit)
bins = p.array(range(1, 1001)) * (bin_upper_limit / 1000.0)
p.hist(values, bins)
p.xlabel('Document size (unicode codepoints)')
p.ylabel('Number of documents')
p.title('Document Size Histogram for %s' % prefix)
p.savefig(path, dpi=72)
p.close()
def length_stats(prefix): doc_lengths = raw_doc_lengths(prefix).values() median = n.median(doc_lengths) mean = n.mean(doc_lengths) std = n.std(doc_lengths) return median, mean, std
def docs_under_thresh(prefix, thresh):
doc_lengths = raw_doc_lengths(prefix).values()
return len(filter(lambda x: x < thresh, doc_lengths))
def indices_under_thresh(prefix, thresh):
doc_lengths = raw_doc_lengths(prefix).iteritems()
return [id for (id, len) in doc_lengths if len < thresh]
def length_stats(prefix):
doc_lengths = raw_doc_lengths(prefix).values()
median = n.median(doc_lengths)
mean = n.mean(doc_lengths)
std = n.std(doc_lengths)
return median, mean, std
