mport pandas as pd
data = pd.read_csv('assets/gapminder.csv', thousands= ',', index_col= 'Year')

from bokeh.io import curdoc

from bokeh.models import LinearInterpolator, CategoricalColorMapper, ColumnDataSource, HoverTool

from bokeh.palettes import Spectral6

from bokeh.plotting import figure
PLOT_OPTS = dict(
    height=400, x_axis_type='log',
    x_range=(100, 100000), y_range=(0, 100)
)

source = ColumnDataSource(dict(
x=data.loc[2010].income,
y=data.loc[2010].life,
country=data.loc[2010].Country,
population=data.loc[2010].population,
region=data.loc[2010].region
))

size_mapper = LinearInterpolator(
    x=[data.population.min(), data.population.max()],
    y=[3, 60]
)

color_mapper = CategoricalColorMapper(
   factors=list(data.region.unique()),
   palette=Spectral6,
)

p = figure(
    title=str(2010),
    toolbar_location = 'above',
    tools=[HoverTool(tooltips='@country', show_arrow=False)],**PLOT_OPTS)

p.circle(
    x='x', y='y',
    size={'field':'population', 'transform':size_mapper},
    color={'field':'region', 'transform':color_mapper},
    alpha=0.6,
    source=source,
    legend='region'
)
p.legend.border_line_color = None
p.legend.location = (0,-50)
p.right.append(p.legend[0])

from bokeh.models.widgets import Slider

def update(attr, old, new):
    year = new
    new_data = dict(
        x=data.loc[2010].income,
        y=data.loc[2010].life,
        country=data.loc[2010].Country,
        region=data.loc[2010].region,
        population=data.loc[2010].population,
    )
    source.data = new_data
    p.title.text = str(year)

slider = Slider(start=1800, end=2010, step=1, title='Year')
slider.on_change('value', update)

from bokeh.layouts import column
layout = column(p, slider)
curdoc().add_root(layout)