Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add support for importable histogram print function #43

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Empty file added data_hacks/__init__.py
Empty file.
50 changes: 30 additions & 20 deletions data_hacks/histogram.py
Original file line number Diff line number Diff line change
Expand Up @@ -121,27 +121,29 @@ def test_median():
assert "4.50" == "%.2f" % median([4.0, 5, 2, 1, 9, 10])


def histogram(stream, options):
def _histogram(stream, minimum=None, maximum=None, num_buckets=None, logscale=False,
custbuckets=None, calc_mvsd=True,
bucket_format='%10.4f', calc_percentage=False, dot='∎'):
"""
Loop over the stream and add each entry to the dataset, printing out at the
end.
stream yields Decimal()
"""
if not options.min or not options.max:
if not minimum or not maximum:
# glob the iterator here so we can do min/max on it
data = list(stream)
else:
data = stream
bucket_scale = 1

if options.min:
min_v = Decimal(options.min)
if minimum:
min_v = Decimal(minimum)
else:
min_v = min(data, key=lambda x: x.value)
min_v = min_v.value
if options.max:
max_v = Decimal(options.max)
if maximum:
max_v = Decimal(maximum)
else:
max_v = max(data, key=lambda x: x.value)
max_v = max_v.value
Expand All @@ -151,11 +153,9 @@ def histogram(stream, options):
diff = max_v - min_v

boundaries = []
bucket_counts = []
buckets = 0

if options.custbuckets:
bound = options.custbuckets.split(',')
if custbuckets:
bound = custbuckets.split(',')
bound_sort = sorted(map(Decimal, bound))

# if the last value is smaller than the maximum, replace it
Expand All @@ -174,8 +174,8 @@ def histogram(stream, options):
# so no need to do a -1!
bucket_counts = [0 for x in range(len(boundaries))]
buckets = len(boundaries)
elif options.logscale:
buckets = options.buckets and int(options.buckets) or 10
elif logscale:
buckets = num_buckets and int(num_buckets) or 10
if buckets <= 0:
raise ValueError('# of buckets must be > 0')

Expand All @@ -202,7 +202,7 @@ def log_steps(k, n):
for step in log_steps(buckets, diff):
boundaries.append(min_v + step)
else:
buckets = options.buckets and int(options.buckets) or 10
buckets = num_buckets and int(num_buckets) or 10
if buckets <= 0:
raise ValueError('# of buckets must be > 0')
step = diff / buckets
Expand All @@ -216,7 +216,7 @@ def log_steps(k, n):
accepted_data = []
for record in data:
samples += record.count
if options.mvsd:
if calc_mvsd:
mvsd.add(record.value, record.count)
accepted_data.append(record)
# find the bucket this goes in
Expand All @@ -237,29 +237,39 @@ def log_steps(k, n):
if skipped:
print("# %d value%s outside of min/max" %
(skipped, skipped > 1 and 's' or ''))
if options.mvsd:
if calc_mvsd:
print("# Mean = %f; Variance = %f; SD = %f; Median %f" %
(mvsd.mean(), mvsd.var(), mvsd.sd(),
median(accepted_data, key=lambda x: x.value)))
print "# each " + options.dot + " represents a count of %d" % bucket_scale
bucket_min = min_v
print "# each " + dot + " represents a count of %d" % bucket_scale
bucket_max = min_v
percentage = ""
format_string = options.format + ' - ' + options.format + ' [%6d]: %s%s'
format_string = bucket_format + ' - ' + bucket_format + ' [%6d]: %s%s'
for bucket in range(buckets):
bucket_min = bucket_max
bucket_max = boundaries[bucket]
bucket_count = bucket_counts[bucket]
star_count = 0
if bucket_count:
star_count = bucket_count / bucket_scale
if options.percentage:
if calc_percentage:
percentage = " (%0.2f%%)" % (100 * Decimal(bucket_count) /
Decimal(samples))
print format_string % (bucket_min, bucket_max, bucket_count, options.dot *
print format_string % (bucket_min, bucket_max, bucket_count, dot *
star_count, percentage)


def histogram(stream, options):
_histogram(stream, options.min, options.max, options.buckets, options.logscale,
options.custbuckets, options.mvsd, options.format, options.percentage,
options.dot)


def print_histogram(samples, **kwargs):
stream = [str(x) for x in samples]
_histogram(load_stream(stream, False, False), **kwargs)


if __name__ == "__main__":
parser = OptionParser()
parser.usage = "cat data | %prog [options]"
Expand Down