"""

Plotting horizontal, terminal based histograms

"""

from __future__ import print_function

import math

from .utils.helpers import *

from .utils.commandhelp import hist

def plot_horiz_hist(f, width=20, bincount=None, binwidth=None, pch="o", title="", ylab=False, show_summary=False, regular=False):

"""

Make a histogram

Arguments:

width -- the width of the histogram in # of lines

bincount -- number of bins in the histogram

binwidth -- width of bins in the histogram

pch -- shape of the bars in the plot

colour -- colour of the bars in the terminal

title -- title at the top of the plot

ylab -- boolen value for whether or not to display x-axis labels

show_summary -- boolean value for whether or not to display a summary

regular -- boolean value for whether or not to start y-labels at 0

"""

if pch is None:

pch = "o"

if isinstance(f, str):

with open(f) as fh:

f = fh.readlines()

min_val, max_val = None, None

n, mean, sd = 0.0, 0.0, 0.0

for number in read_numbers(f):

n += 1

if min_val is None or number < min_val:

min_val = number

if max_val is None or number > max_val:

max_val = number

mean += number

mean /= n

for number in read_numbers(f):

sd += (mean - number)**2

sd /= (n - 1)

sd **= 0.5

bins = list(calc_bins(n, min_val, max_val, bincount, binwidth))

hist = dict((i, 0) for i in range(len(bins)))

for number in read_numbers(f):

for i, b in enumerate(bins):

if number <= b:

hist[i] += 1

break

if number == max_val and max_val > bins[len(bins) - 1]:

hist[len(hist) - 1] += 1

min_count = min(hist.values())

max_count = max(hist.values())

# `min_display_count` and `max_display_count` are the min/max

# counts that will be displayed on the x-axis of our

# graph. If the user sets the `regular` argument to True,

# we use a `min_display_count` of 0.

if regular:

min_display_count = 0

else:

min_display_count = min_count

max_display_count = max_count + 1

if width is None:

width = int(max_display_count - min_display_count)

if width > 40:

width = 40

# Calculate how many counts each horizontal unit (square)

# represents. This will be useful for knowing how long

# each of our bars should be.

counts_per_horizontal_unit = float(max_display_count - min_display_count) / width

# If we need to display y-labels, use `bins` to generate

# them. `ylabels_width` represents the width of the

# y-labels "column" so that we can add the appropriate

# amount of padding in the rest of our graph.

if ylab:

ylabels = [str(b) for b in bins]

ylabels_width = max(len(l) for l in ylabels) + 1

else:

ylabels_width = 0

# Print the title, as per usual

if title:

print(box_text(title, width*2 + ylabels_width))

print()

# Print the guts of the graph!

for bin_n, count in hist.iteritems():

line = ""

if ylab:

line += ylabels[bin_n].ljust(ylabels_width)

line += "|"

# This is why we calculated `counts_per_horizontal_unit`

# earlier.

n_squares = int((count - min_display_count) / counts_per_horizontal_unit) + 1

line += (" " + pch) * n_squares

print(line)

print(" " * ylabels_width + "+" + "-" * width * 2)

# Printing the x-labels is quite difficult. We only want to print

# a label for a square if it different to the previous label. This

# is because we don't want to print labels that look like

# "1 1 1 1 2 2 2 2...".

#

# We therefore generate a list of "candidate" x-labels, which are

# the labels we would show if we didn't care about repetition. We

# use this list of candidates to generate a list of deduplicated

# labels.

#

# First we generate the candidates

candidate_xlabels = [str(int(l)) for l in list(drange(

min_display_count,

max_display_count,

float(max_display_count - min_display_count) / width))]

# Then we deduplicate `candidate_xlabels`

xlabels = []

for cand in candidate_xlabels:

if cand not in xlabels:

xlabels.append(cand)

else:

xlabels.append("")

# Print the labels vertically by printing a row with all

# the first characters in each label, then a row with all

# the second characters, etc.

max_xlabel_len = max(len(l) for l in xlabels)

for row_n in range(0, max_xlabel_len):

row = ""

for label in xlabels:

if len(label) > row_n:

row += label[row_n]

else:

row += " "

row += " "

print(" " * (ylabels_width + 2) + row)

# Finally, print the summary statistics, as per usual

if show_summary:

center = max(map(len, map(str, [n, min_val, mean, max_val])))

center += 15

print()

print("-" * (2 + center))

print("|" + "Summary".center(center) + "|")

print("-" * (2 + center))

summary = "|" + ("observations: %d" % n).center(center) + "|\n"

summary += "|" + ("min value: %f" % min_val).center(center) + "|\n"

summary += "|" + ("mean : %f" % mean).center(center) + "|\n"

summary += "|" + ("std dev : %f" % sd).center(center) + "|\n"

summary += "|" + ("max value: %f" % max_val).center(center) + "|\n"

summary += "-" * (2 + center)

print(summary)

def calc_bins(n, min_val, max_val, h=None, binwidth=None):

"""

Calculate number of bins for the histogram

"""

if not h:

h = max(10, math.log(n + 1, 2))

if binwidth == 0:

binwidth = 0.1

if binwidth is None:

binwidth = (max_val - min_val) / h

for b in drange(min_val, max_val, step=binwidth, include_stop=True):

if b.is_integer():

yield int(b)

else:

yield b

def read_numbers(numbers):

"""

Read the input data in the most optimal way

"""

if isiterable(numbers):

for number in numbers:

yield float(str(number).strip())

else:

with open(numbers) as fh:

for number in fh:

yield float(number.strip())