python bokeh legend out of the plot size

I'm new in python, and somebody help me with this code, but I want to change some parameter:

First the size of the legend out of the plot, some time the legend are to big (example: D_0__Bacteria;D_1__Firmicutes;D_2__Clostridia;D_3__Clostridiales;D_4__Peptostreptococcaceae;D_5__Acetoanaerobium), and some time it is short (Acetoanaerobium), so I just want to make the legend auto fix the size (in the figure the legend are not complete)!!!.

Second, the label that appear when the pointer get hover the zone of the bar, show the name and the value of the data that correspond, (hover.tooltips = [('Taxon','example: Acetoanaerobium'),('Value','the corresponding value example: 99')])

third: the position of the plot (figure), in the middle

#!/usr/bin/env python

import pandas as pd
from bokeh.io import show, output_file
from bokeh.models import ColumnDataSource
from bokeh.plotting import figure
from bokeh.core.properties import value
from bokeh.palettes import Spectral
from bokeh.models import HoverTool
#from bokeh.plotting import figure, output_file, show, ColumnDataSource
import itertools
import sys

data_in = sys.argv[1]
data_out = sys.argv[2]

output_file(data_out + ".html")

df = pd.read_csv(data_in, sep='\t')
df.set_index('#OTU_ID', inplace=True)

#print(df)
s_data = df.columns.values # linia de samples !!!
t_data = df.index.values    #columna de datos

#print(s_data)
#print(t_data)

# You have two rows with 'uncultured' data. I added these together.
# This may or may not be what you want.
df = df.groupby('#OTU_ID')[s_data].transform('sum')

#grouped = df.groupby(["columnA", "columnB"], as_index=False).count()

#print(grouped)

# create a color iterator
# See https://stackoverflow.com/q/39839409/50065
# choose an appropriate pallete from
# https://docs.bokeh.org/en/latest/docs/reference/palettes.html
# if you have a large number of organisms
color_iter = itertools.cycle(Spectral[5])    
colors = [next(color_iter) for organism in t_data]

# create a ColumnDataSource
data = {'xs': list(s_data)}

for organism in t_data:
    data[organism] = list(df.loc[organism])
source = ColumnDataSource(data=data)


#print(organism)
# create our plot
plotX = figure(x_range=s_data, plot_height=500, title="Relative Abundance",
           toolbar_location=None, tools="hover")

plotX.vbar_stack(t_data, x='xs', width=0.93, source=source,
            legend=[value(x) for x in t_data], color=colors)

plotX.xaxis.axis_label = 'Sample'
plotX.yaxis.axis_label = 'Percent (%)'
plotX.legend.location = "bottom_left"
plotX.legend.orientation = "vertical"

# Position the legend outside the plot area
# https://stackoverflow.com/questions/48240867/how-can-i-make-legend-outside-plot-area-with-stacked-bar
new_legend = plotX.legend[0]
plotX.legend[0].plot = None
plotX.add_layout(new_legend, 'below')

hover = plotX.select(dict(type=HoverTool))
hover.tooltips = [('Taxon','unknow_var'),('Value','unknow_var')]
# I don't know what variable to addd in unknow_var

show(plotX)

the in file is a file.txt, tab delimited file like:

#OTU_ID columnA columnB columnC columnD columnN
D_0__Bacteria;D_1__Actinobacteria;D_2__Acidimicrobiia;D_3__Acidimicrobiales;D_4__uncultured;D_5__uncultured_bacterium   1   3   7   0.9 2
D_0__Bacteria;D_1__Acidobacteria;D_2__Subgroup_25;D_3__uncultured_Acidobacteria_bacterium;D_0__Bacteria;D_1__Actinobacteria;D_2__Actinobacteria;D_3__Streptomycetales;D_4__Streptomycetaceae;D_5__Kitasatospora 5   3   13  7   5
D_0__Bacteria;D_1__Firmicutes;D_2__Bacilli;D_3__Bacillales;D_4__Bacillaceae;D_5__Anoxybacillus  0.1 0.8 7   1   0.4
D_0__Bacteria;D_1__Firmicutes;D_2__Bacilli;D_3__Lactobacillales;D_4__Carnobacteriaceae;D_5__Carnobacterium  3   7   9   16  11
D_0__Bacteria;D_1__Firmicutes;D_2__Bacilli;D_3__Bacillales;D_4__Bacillaceae;D_5__Oceanobacillus 5   2   15  1   7
D_0__Bacteria;D_1__Firmicutes;D_2__Clostridia;D_3__Clostridiales;D_4__Family_XII;D_5__Fusibacter    8   9   0   11  22
D_0__Bacteria;D_1__Firmicutes;D_2__Clostridia;D_3__Clostridiales;D_4__Peptostreptococcaceae;D_5__Acetoanaerobium    99  3   12  1   3
D_4__Clostridiaceae_2;D_5__Alkaliphilus 33  45  1   0   9
D_4__Peptococcaceae;D_5__uncultured 0   3   9   10  11

in this example the value are not in % as the y-legend say, the values are just an example !!!

thanks so much !!!

Bokeh legends do not auto-size (there is no option to make them do so). You will need to set the legend width to be wide enough to cover any label you might have. Additionally, since they are drawn on the same canvas as the plot, you will need to make the plot wider, to accommodate the width you set on the legend. If you don't want the central plot area to get bigger, you can set the various min_border, min_border_left values on the plot to make more space around the inner plot area.

Alternatively, instead of resizing the plot and legend, you could consider making the legend font size smaller.

p.legend.label_text_font_size = "8px"