Python pyplot.boxplot函数代码示例

def plot_difficulties(difficulties, bins=10):
    # Data   
    plot_data = []
    names = []
    for y_true, c_val in [(0,0), (0,1), (1,0), (1,1)]:
        diff_yc = difficulties[2*y_true+c_val]
        plot_data.append(diff_yc)
        names.append('y=%d, c=%d' %(y_true, c_val))
        print("y=%d, c=%d, mean=%.5f, std=%.5f" % (y_true, c_val, np.mean(diff_yc), np.std(diff_yc)))

    # Boxplots
    fig, axes = plt.subplots()
    plt.boxplot(plot_data)
    xtickNames = plt.setp(axes, xticklabels=names)
    axes.set_ylim([-.01, 1.01])
    axes.set_ylabel('Difficulty')
    plt.show()

    # Histogram
    fig, axes = plt.subplots()
    plt.yscale('log', nonposy='clip')
    hist = plt.hist(plot_data, label=names, bins=bins)
    plt.legend()
    axes.set_xlabel('Difficulty')
    axes.set_ylabel('Count (log-scale)')
    plt.show()

def bivariate_analysis_catg_cont(catg_cont_list,df,target_name,sub_len,COUNTER,PLOT_ROW_SIZE,PLOT_COLUMNS_SIZE):

    # No need to remove string varible as they are handled by chi2 function of sklearn.
    # clean_catg_cont_list = clean_str_list(df,catg_cont_list)
    clean_catg_cont_list = catg_cont_list
    clean_df = df.dropna()

    for col in clean_catg_cont_list:

        col_classes =df[target_name].unique()

        summary = clean_df[col].describe()
        count = summary[0]
        mean = summary[1]
        std = summary[2]

        plt.subplot(PLOT_ROW_SIZE,PLOT_COLUMNS_SIZE,COUNTER)
        plt.title("mean "+str(np.float32(mean))+" std "+str(np.float32(std)),fontsize=10)

        x = [np.array(clean_df[clean_df[target_name]==i][col]) for i in col_classes]
        y = clean_df[target_name]

        f_value,p_val = evaluate_anova(np.array(clean_df[col]).reshape(-1,1),y)

        plt.xlabel(target_name+"\n f_value: "+str(np.float32(f_value[0]))+" / p_val: "+str(p_val[0]), fontsize=10)
        plt.ylabel(col, fontsize=10)
        plt.boxplot(x)

        print (col+" vs "+target_name+" plotted....")

        COUNTER +=1

    return plt,COUNTER

def plot(lookup):
        data = []
        for iiDiameter in sorted(lookup.keys()):
                data.append(lookup[iiDiameter])
        plt.boxplot(data, sym='')
        plt.setp(plt.gca(),'xticklabels',sorted(lookup.keys()))
        plt.show()

    def handle(self, *args, **options):
        fs = 10  # fontsize
        
        versions = models.SourceLine.objects.filter(
            project__startswith='django-').order_by(
            'project').values_list(
            'project', 'progradon__complexity')
        for vers, complexity_iter in itertools.groupby(
            versions, key=operator.itemgetter(1)):
            print vers, ':'
            print '-', ', '.join(str(x) for x in complexity_iter)
        data = models.SourceLine.objects.filter(
            project='django-1.0.1').values_list(
            'progradon__complexity', flat=True)
        plt.boxplot(data) # , labels=labels)
        
        plt.show()

        # xs, ys, areas = zip(*data)
        # ys = areas
        # colors = np.random.rand(len(xs))
        # plt.scatter(xs, ys, c=colors) # s=areas)
        # plt.xlabel('file index')
        # plt.ylabel('version index')
        plt.savefig('z.png')

    def visualize_performance(self):
        intra = self._intra
        inter = self._inter

        labels = [1]*len(intra) + [-1]*len(inter)
        scores = intra+inter

        self._common_visualize_performance( labels, scores)

        plt.figure()
        plt.boxplot([intra, inter])
        plt.xticks([1, 2], ['intra', 'inter'])
        plt.title('Distribution of scores')
        plt.savefig('comparison_score_distribution.pdf')


        plt.figure()
        start = np.min(np.min(intra), np.min(inter))
        end = np.max(np.max(intra), np.max(inter))
        intra_hist, intra_bin = np.histogram(intra,50, (start, end))
        inter_hist, inter_bin = np.histogram(inter,50, (start, end))


        plt.plot(intra_bin[:-1], intra_hist/float(intra_hist.sum()), label='intra', color='blue')
        plt.plot(inter_bin[:-1], inter_hist/float(inter_hist.sum()), label='inter', color='red')
        plt.legend()
        plt.xlabel('Comparison scores')
        plt.ylabel('Probability')
        plt.title('Score distribution')

def stats_fn(data_frame):
    global scene
    stat_file = open("Stat_tests_" + scene[:-4] + ".txt", "w")
    seen_pairs = []
    for algorithm in data_frame:
        for algorithm2 in data_frame:
            if (algorithm != algorithm2) and ((algorithm, algorithm2) not in seen_pairs):
                seen_pairs.append((algorithm, algorithm2))
                seen_pairs.append((algorithm2, algorithm))
                statistical_significance = stats.wilcoxon(data_frame[algorithm], data_frame[algorithm2])
                print >> stat_file, algorithm, " VS ", algorithm2, " -->", statistical_significance
                print >> stat_file, algorithm, " median = ", np.median(data_frame[algorithm])
                print >> stat_file, algorithm2, " median = ", np.median(data_frame[algorithm2])
                print >> stat_file, "----------------------------------------------------------"
    # # This part is for drawing the different boxplots
    figure_name = scene + "_.png"
    current_path = os.getcwd()
    os.chdir("/home/omohamme/INRIA/experiments/moop_sim_comparison/boxplots/" + scene[:-4] + "/")
    plt.figure(figsize=(15.0, 11.0))
    plt.boxplot(data_frame.values())
    plt.xticks(range(1, len(data_frame.keys()) + 1), data_frame.keys())
    plt.title(figure_name)
    plt.savefig(figure_name)
    os.chdir(current_path)

    stat_file.close()

def create_boxplot(data, save_dir, correct_entropy=1):
    """
    data_file - path file containing entropy values for the lines added by the mutant files
    save_directory - directory to save the plot in, not including the name of the plot itself
    correct_entropy - the entropy of the lines added by the repair program
    """
    print "CREATE BOXPLOT"
    # fid = open(data_file,'r')
    # data=[float(l.strip()) for l in fid.readlines()]
    print data
    assert len(data) > 0
    # plot mutant entropy
    plt.boxplot(data)
    # plot correct entropy
    p1 = plt.plot([0, 2], [correct_entropy, correct_entropy], color="g")
    # label the repaired program
    l1 = plt.legend([p1], ["repaired program"])

    # annotate the plot
    plt.ylabel("Entropy (bits)")
    plt.title("Entropy of lines added in mutant programs")

    # generate a random number as the name of the plot
    name = str(random.randint(0, sys.maxint))
    plt.savefig(os.path.join(save_dir, name + ".png"), bbox_inches=0)
    print os.path.join(save_dir, name + ".png")
    return name

def plot(revisions, benchmarks, subdir='.', baseurl='https://github.com/idaholab/moose/commit/'):
    data = []
    labels = []
    for rev, bench in zip(revisions, benchmarks):
        data.append(bench.realruns)
        labels.append(rev[:7])

    median = sorted(data[0])[int(len(data[0])/2)]
    plt.axhline(y=median*1.05, linestyle='--', linewidth=2, color='red', alpha=.5, label='+5%')
    plt.axhline(y=median*1.01, linestyle=':', linewidth=2, color='red', label='+1%')
    plt.axhline(y=median, dashes=[48, 4, 12, 4], color='black', alpha=.5)
    plt.axhline(y=median*.99, linestyle=':', linewidth=2, color='green', label='-1%')
    plt.axhline(y=median*.95, linestyle='--', linewidth=2, color='green', alpha=.5, label='-5%')

    plt.boxplot(data, labels=labels, whis=1.5)
    plt.xticks(rotation=90)
    plt.ylabel('Time (seconds)')

    fig = plt.gcf()

    ax = fig.axes[0]
    labels = ax.get_xticklabels()
    for label in labels:
        label.set_url(urlparse.urljoin(baseurl, label.get_text()))

    legend = ax.legend(loc='upper right')

    fig.subplots_adjust(bottom=.15)
    fig.savefig(os.path.join(subdir, benchmarks[0].name + '.svg'))
    plt.clf()

def plot(work_time_deltas_hours):
 
    # 45 minutes break is assumed    
    work_overtime = sum([w - 8.75 for w in work_time_deltas_hours ])
 
    plt.boxplot(work_time_deltas_hours)
    plt.ylabel("Working Hours")
        
    plt.xticks([0,1,2],())    
        
    yvalues = numpy.arange(numpy.floor(numpy.min(work_time_deltas_hours)),numpy.ceil(numpy.max(work_time_deltas_hours)),0.25)    
    plt.yticks(yvalues,[ str(math.floor(x)) + "h " + str(int((x % 1.0) * 60)) +"min" for x  in yvalues],rotation=0)
  
    # Debug
    print("Mean: "+str(numpy.mean(work_time_deltas_hours))) 
    print("Min: "+str(numpy.min(work_time_deltas_hours)))
    print("Max: "+str(numpy.max(work_time_deltas_hours)))
    print("Median: "+str(numpy.median(work_time_deltas_hours)))
    print("Work overtime: "+ str(work_overtime))
    print("Days tracked: "+str(len(work_time_deltas_hours)))
     
    plt.text(1.35,10,"Mean: " + str(math.floor(numpy.mean(work_time_deltas_hours))) + "h " + str(int((numpy.mean(work_time_deltas_hours) % 1.0) * 60)) + "min"
             "\nMax: " + str(math.floor(numpy.max(work_time_deltas_hours))) + "h " + str(int((numpy.max(work_time_deltas_hours) % 1.0) * 60)) + "min"
             "\nMin: "+ str(math.floor(numpy.min(work_time_deltas_hours))) + "h " + str(int((numpy.min(work_time_deltas_hours) % 1.0) * 60)) + "min"
             "\nMedian: "+ str(math.floor(numpy.median(work_time_deltas_hours))) + "h " + str(int((numpy.median(work_time_deltas_hours) % 1.0) * 60)) + "min"+
             "\nOvertime: " + str(math.floor(work_overtime)) +"h "+ str(int((work_overtime % 1.0) * 60)) + "min" +
             "\nDays: " + str(len(work_time_deltas_hours)),
             bbox=dict(boxstyle='round', facecolor='white', alpha=0.5))
    
    plt.title("Working Hours Boxplot")
    plt.show()   

    def sale_price_per_sq_foot_boxplot(self, groupby, title):
        """Boxplot of sale price per square foot, grouped by a groupby variable

        title is the plot title"""
        fig = init_fig()

        # This figure needs to be extra wide
        fig.set_size_inches(10, 4)

        # Remove missings and restrict to the columns we need
        data = self.data[[groupby, "sale_price_per_sqft"]].dropna()

        # The boxplot function takes a list of Series, so we make one Series for each
        # group, and append them all into a list
        groups = list()
        values = data[groupby].value_counts().index  # All the levels of the groupby variable

        for value in values:
            groups.append(data.loc[data[groupby] == value, "sale_price_per_sqft"])

        # Now make the plot. The empty string means we don't want the outliers, since
        # they will mess up the axis scale
        plt.boxplot(groups, 0, "")

        plt.ylabel("Sale Price per Sq. Ft.")
        plt.title(title)
        plt.xticks(np.arange(len(values)) + 1, values)

        return fig_to_svg(fig)

def distance_distribution_plot(learner,box_kwargs=None,**kwargs):
    """
    plots the distribution of distances to/from predicted events from/to
    actual events, dependning on kwargs
    
    Args:
        learner: the learner object to use
        kwargs: passed to event_distance_distribution (ie: to_true=T/F)
    """
    train_scores = learner._scores_by_params(train=True)
    valid_scores = learner._scores_by_params(train=False)
    if (box_kwargs is None):
        box_kwargs = dict(whis=[5,95])
    name = learner.description.lower()
    x_values = learner.param_values()
    train_dist = Learning.event_distance_distribution(train_scores,**kwargs)
    valid_dist = Learning.event_distance_distribution(valid_scores,**kwargs)
    dist_plot = lambda x: [v for v in x]
    train_plot = dist_plot(train_dist)
    valid_plot = dist_plot(valid_dist)
    plt.boxplot(x=train_plot,**box_kwargs)
    plt.boxplot(x=valid_plot,**box_kwargs)
    plt.gca().set_yscale('log')
    PlotUtilities.lazyLabel("Tuning parameter","Distance Distribution (idx)",
                            "Event distributions for {:s}".format(name),
                            frameon=False)

def boxplot_by_pft(var, timestep, cmtnum, stages, ref_veg_map, ref_run_status):
  '''
  Work in progress...
  '''

  data, units = stitch_stages(var, timestep, stages)
  print "data size:", data.size
  print data.shape

  d2 = data
  # d2 = sum_across_compartments(data)
  # print "data size after summing compartments:", d2.size

  d3 = mask_by_cmt(d2, cmtnum, ref_veg_map)
  print "data size after masking cmt:", d3.count()

  d3 = mask_by_failed_run_status(d3, ref_run_status)
  print "data count after masking run status:", d3.count()

  pft0avg = np.ma.average(d3, axis=(2,3))
  #plt.plot(pft0avg) # Line plot
  plt.boxplot(
      pft0avg,
      labels = ["PFT {}".format(i) for i in range(0, 10)],
      whis='range',
      showfliers=False,
      patch_artist=True,
      boxprops=dict(color='blue', alpha=0.25),
      whiskerprops=dict(color='red'),
      capprops=dict(color='blue'),
  )
  plt.ylabel(units)
  plt.show(block=True)

def make_plot_lfw_reorder_other(save=False):
    conn = pm.Connection()
    db = conn['hyperopt']
    Jobs = db['jobs']
    
    exp_key = 'thor_model_exploration.model_exploration_bandits.LFWBanditModelExplorationOther/hyperopt.Random'

    H = Jobs.group(['spec.order'],
                   {'exp_key': exp_key, 'state':2, 
                    'spec.preproc.size.0':250
                   },
                   {'losses': []},
                   'function(d, o){o.losses.push(d.result.loss);}')
        
    order_choices = params.order_choices
    ords = pluck(H, 'spec.order')
    reinds = [ords.index(_o) for _o in order_choices]
    H = [H[_r] for _r in reinds]

    od = {'lpool': 'p', 'activ': 'a', 'lnorm': 'n'}
    order_labels = [','.join([od[b] for b in Before]) + '|' + ','.join([od[b] for b in After]) for (Before, After) in order_choices]
 
    import matplotlib.pyplot as plt
    fig = plt.figure(figsize=(18,8))
    plt.boxplot([1-np.array(h['losses']) for h in H])
    means = [1-np.array(h['losses']).mean() for h in H]
    plt.plot(range(1,len(H)+1), means, color='green')
    plt.scatter(range(1,len(H)+1), means)
    
    plt.xticks(range(1,len(ords)+1),  order_labels, rotation=60)
    
    plt.ylabel('Absolute performance')
    plt.xlabel('Architecture tag')

def main():

    data = []
    data_month = []

    # Post to database
    con = mdb.connect(host='192.168.1.143', db='monitor', user='crblackw')
    
    #Format of data structure
    #[mm][dd][data]
    #mm:    This is the month of the dataset.  Keep in mind that it is indexed from zero.  So August (8) is actually 7.
    #dd:    This is the day within the month.
    #data:  This is an array of the the data from the day.  Each datapoint is a tuple of (datetime, value).
    
    with con:
        cur = con.cursor()
        #cur.execute("SELECT temp_actual FROM sensor1 GROUP BY HOUR(datetime) LIMIT 0, 30")
        for m in range(1,12):
            for d in range(1,31):
                cur.execute("SELECT datetime,temp_actual FROM sensor1 WHERE DAY(datetime) = %i AND MONTH(datetime) = %i" %(d,m))
                data_month.append(np.array(cur.fetchall()))  
            data.append(data_month)
            data_month = []
    con.close()
    
    plt.boxplot(data[7-1][11][:,1])
    plt.show()
    
    '''

def __create_num_threads_vs_jct_graph(num_threads_to_jcts, output_dir, phase):
  """
  Create a graph of num threads per disk vs. JCT for the specified phase, which must be either
  "write" or "read". num_threads_to_jcts should be a dictionary of the form:
    { num threads : ( list of write JCTs, list of read JCTs ) }
  """
  assert phase in ["write", "read"]

  num_ticks = len(num_threads_to_jcts) + 2
  xmax = num_ticks - 1
  max_jct = max([jct
    for write_jcts, read_jcts in num_threads_to_jcts.itervalues()
    for jct in (write_jcts if phase == "write" else read_jcts)])
  ymax = max_jct * 1.1
  pyplot.title("Num threads per disk vs. JCT ({} phase)".format(phase))
  pyplot.xlabel("Num threads per disk")
  pyplot.ylabel("JCT (s)")
  pyplot.grid(b=True)
  pyplot.xlim(xmin=0, xmax=xmax)
  pyplot.ylim(ymin=0, ymax=ymax)

  # Build a list of lists of JCTs, sorted by num threads per disk.
  all_jcts = [write_jcts if phase == "write" else read_jcts
    for _, (write_jcts, read_jcts) in sorted(num_threads_to_jcts.iteritems())]
  pyplot.boxplot(all_jcts, whis=[0, 100])

  # Replace the visually-correct x-axis values with the numerically correct values.
  pyplot.xticks(xrange(num_ticks), [""] + sorted(num_threads_to_jcts.keys()) + [""])

  # Save the graph as a PDF.
  output_filepath = path.join(output_dir, "{}_phase_num_threads_vs_jct.pdf".format(phase))
  with backend_pdf.PdfPages(output_filepath) as pdf:
    pdf.savefig()

  pyplot.close()