Can someone help me determine why these images are coming out different from one another?
I am trying to re-generate a plot I made with plotnine. Unfortunately, running the same code seems to generate a different output. Here's the original output I was getting:
Here is the new output I am getting:
But I have not changed the code at all.
Here is the code:
from dfply import select, X
from matplotlib.font_manager import FontProperties
from mizani.breaks import log_breaks, mpl_breaks
from mizani.formatters import currency_format, custom_format, mpl_format,scientific_format
from mizani.transforms import log10_trans, identity_trans
from pandas import DataFrame, read_csv
from pathlib import Path
from plotnine import (
aes,
element_blank,
element_line,
element_rect,
element_text,
facet_wrap,
geom_blank,
geom_line,
geom_point,
geom_vline,
ggplot,
labeller,
scale_x_continuous,
scale_y_continuous,
scale_y_log10,
theme)
from sys import path
from time import asctime, localtime, time
module_path = Path(__file__).resolve()
analysis_folder = module_path.parent
scripts_folder = analysis_folder.parent
operations_folder = str(scripts_folder) + '/operations'
project_folder = scripts_folder.parent
output_folder = str(project_folder) + '/output_analysis'
path.insert(0, str(operations_folder))
path.insert(0, str(scripts_folder))
from fixed_precision import fixed_precision as fixed_precision
from label_title import label_title as label_title
from revised_phase1_y_breaks import revised_phase1_y_breaks as revised_phase1_y_breaks
from revised_phase1_y_format import revised_phase1_y_format as revised_phase1_y_format
font_path = str(project_folder) + '/resources/fonts/Arial-Unicode-Regular.ttf'
plot_title_properties = FontProperties(fname=font_path, size=12)
axis_title_properties = FontProperties(fname=font_path, size=11)
legend_title_properties = FontProperties(fname=font_path, size=11)
strip_text_properties = FontProperties(fname=font_path, size=11)
text_properties = FontProperties(fname=font_path, size=10)
phase1_dataframe = read_csv(
str(output_folder)
+ '/phase1_dataframe.csv')
def create_phase1_plot():
"""
Create the Phase 1 plot.
:return: 'Date, time, and year: date_time_year
Module: module_path
Saved image: phase1_plot_location'
:rtype: str
"""
phase1_dataframe_combined = phase1_dataframe >> select (
X.series,
X.trait,
X.time,
X.ancestor_seed,
X.ancestor_value,
X.ancestor_value_log10,
X.evolved_population_average_value,
X.evolved_population_average_value_log10)
# Use ancestor data to plot ancestor values.
fitness_ancestor_value = (
phase1_dataframe_combined.query(
"trait == 'fitness'")
# ['ancestor_value_log10']
['ancestor_value']
.mean())
genome_length_ancestor_value = (
phase1_dataframe_combined.query(
"trait == 'genome_length'")
# ['ancestor_value_log10']
['ancestor_value']
.mean())
trait_names = ["fitness", "genome_length",]
value = list([fitness_ancestor_value, genome_length_ancestor_value])
ancestor_values = {'trait': trait_names, 'ancestor_value': value}
ancestor_data = DataFrame(ancestor_values)
# Generate fake data for framing.
trait_names = ["fitness", "fitness", "genome_length", "genome_length"]
# trait_names = ["fitness", "fitness"]
value = [0.001, 10000000000, 70, 130]
# value = [-0.0000000001, 10000000000]
frame_nums = {'trait': trait_names, 'value': value}
frame_data = DataFrame(frame_nums)
phase1_plot = (
ggplot(
data=phase1_dataframe_combined,
mapping=aes(
x='time',
y='evolved_population_average_value',
color= 'factor(ancestor_seed)',
group= 'series'))
+ geom_line(size= .5)
+ geom_vline(xintercept=20000)
+ geom_vline(xintercept=100000)
+ geom_vline(xintercept=500000)
+ geom_point(
data= ancestor_data,
mapping=aes(
x=0,
y='ancestor_value'
),
shape= '*',
size= 4,
inherit_aes= False)
+ geom_blank(
data=frame_data,
mapping=aes(y='value'),
inherit_aes=False)
+ facet_wrap(
'~ trait',
nrow=2,
scales='free_y',
labeller= labeller(trait=label_title))
+ scale_x_continuous(
name='Time (updates)',
labels = currency_format(
prefix='',
suffix='',
digits=0,
big_mark=','))
+ scale_y_log10(name='Average value',
breaks=revised_phase1_y_breaks,
labels=revised_phase1_y_format
)
+ theme(
figure_size=(5.2, 4.375),
legend_title_align= 'center',
text=element_text(fontproperties=text_properties),
title=element_text(
margin= {'b': 20},
fontproperties=plot_title_properties),
legend_position= 'none',
plot_background=element_rect(fill='white'),
axis_text_x=element_text(
margin={'t':6},
color='black',
size=8,
fontproperties=text_properties),
axis_text_y=element_text(
margin={'r':6},
color='black',
size=8,
fontproperties=text_properties),
axis_title_x=element_text(
margin= {'t': 12},
fontproperties=axis_title_properties),
axis_title_y=element_text(
margin= {'r': 18},
fontproperties=axis_title_properties),
axis_ticks_major_x=element_blank(),
axis_ticks_major_y=element_blank(),
rect=element_rect(
color='white',
size=3,
fill='white'),
panel_grid_major_y=element_line(
linetype='solid',
color='gray',
size=.5),
strip_text_x = element_text(
fontproperties=strip_text_properties),
panel_spacing = .3))
phase1_plot_name= 'revised_phase1_2.pdf'
phase1_plot.save(
filename = phase1_plot_name,
path = (
output_folder
+ '/masked/phase1_plot'))
date_time_year = asctime(localtime(time()))
return (
'\n\n'
'Date, time, and year: %s \n'
'Module: %s \n'
'Saved image: %s'
% (
date_time_year,
module_path, (
output_folder
+ '/masked/phase1_plot/'
+ phase1_plot_name)))
if __name__ == '__main__':
print(create_phase1_plot())
You do not need
theme(panel_spacing=*)unless you want wider spacing.Your first result is from earlier version (<0.12.0) of
plotnine.plotnine==0.12.0got a layout manager so your axis text will not overlap content from other panels.