#!/usr/bin/env python
# coding: utf-8

# # Loading libraries and data 

# In[124]:


import pandas as pd
import altair as alt

from github import Github
import configparser


# In[176]:


config = configparser.ConfigParser()
config.read('secrets.ini');


# In[177]:


g = Github(config['corona']['api_key'])


# In[209]:


START_CASES_INFECTION = 100
START_CASES_DEATH = 10


# In[210]:


URL = "https://covid.ourworldindata.org/data/ecdc/full_data.csv"


# In[211]:


df = pd.read_csv(URL)


# In[212]:


df.date = pd.to_datetime(df.date)


# In[213]:


df = df.sort_values(['location','date'], ascending=True)


# In[214]:


df.shape[0]


# In[215]:


df.head()


# In[216]:


df.info()


# # Date of data

# In[217]:


df.date.max()


# # Creating data for graphs

# In[218]:


df = df.drop(columns=['new_cases', 'new_deaths'])


# In[219]:


df_infection = df[df.total_cases >= START_CASES_INFECTION].copy()
df_death = df[df.total_deaths >= START_CASES_DEATH].copy()


# In[220]:


df_infection.head()


# In[221]:


df_infection[df_infection.location == 'Austria'].tail(3)


# In[222]:


df_death.head()


# In[223]:


df_infection['total_cases_normalized'] = df_infection.groupby(
    "location")[['total_cases']].transform(lambda x: x / x.min() * START_CASES_INFECTION)
df_infection.total_cases_normalized = df_infection.total_cases_normalized.astype("int")

df_death['total_deaths_normalized'] = df_death.groupby(
    "location")[['total_deaths']].transform(lambda x: x / x.min() * START_CASES_DEATH)
df_death.total_deaths_normalized = df_death.total_deaths_normalized.astype("int")


# In[224]:


df_infection["days_after"] = df_infection.groupby("location").cumcount()
df_death["days_after"] = df_death.groupby("location").cumcount()


# In[225]:


df_infection.head()


# In[226]:


df_death.head()


# In[227]:


df_summary = df[df.date == df.date.max()].drop(columns='date')
df_summary['cases'] = df_summary.total_cases
df_summary = df_summary.drop(['total_cases', 'total_deaths'], axis=1)


# In[228]:


df_summary.head()


# In[229]:


df_infection.location.unique()


# In[230]:


df_death.location.unique()


# In[231]:


COUNTRIES = set(df[(df.date == df.date.max())].nlargest(columns='total_cases', n=34).sort_values('location').location)
COUNTRIES.remove('World')


# # Altair

# ## Storing data locally 

# In[232]:


url_infection = 'data_infection.json'
url_death = 'data_death.json'
url_summary = 'data_summary.json'

df_infection[df_infection.location.isin(COUNTRIES)].to_json(url_infection, orient='records')
df_death[df_death.location.isin(COUNTRIES)].to_json(url_death, orient='records')
df_summary[df_summary.location.isin(COUNTRIES)].to_json(url_summary, orient='records')


# ## Uploading data to github

# In[202]:


repo = g.get_repo("Datenspieler/notebooks_for_blog")


# In[203]:


FILELIST = ['data_infection.json', 'data_death.json', 'data_summary.json']


# In[236]:


for file_for_upload in FILELIST:
    print('Uploading', file_for_upload, end=' - ')
    with open(file_for_upload) as f:
        data = f.read()
    contents = repo.get_contents("2020-corona/" + file_for_upload)
    repo.update_file(contents.path, "Update data as of " + df.date.max().strftime('%Y-%m-%d'), data, 
                     contents.sha, branch="master")
    
    contents = repo.get_contents("2020-corona/" + file_for_upload)

    print(contents.last_modified)


# ## Preparing link to data in github 

# In[60]:


# In future version files should be pushed to git automatically, this time I uploaded them 

BASEURL = 'https://raw.githubusercontent.com/Datenspieler/notebooks_for_blog/master/2020-corona/'

url_infection = BASEURL + url_infection
url_death = BASEURL + url_death
url_summary = BASEURL + url_summary


# ## Alternatively local version

# In[233]:


url_infection = df_infection[df_infection.location.isin(COUNTRIES)]
url_death = df_death[df_death.location.isin(COUNTRIES)]
url_summary = df_summary[df_summary.location.isin(COUNTRIES)]


# ## Plotting data 

# In[235]:


highlight = alt.selection(type='single', on='mouseover',
                          fields=['location'], nearest=True)

base_infection = alt.Chart(url_infection).encode(
    alt.X('days_after', type='quantitative', scale=alt.Scale(domain=[0,30], type='ordinal'),
         title='Days since the 100th confirmed infection'),
    alt.Y('total_cases_normalized', type='quantitative', scale=alt.Scale(type='log', base=10), 
          title='Total confirmed infections of COVID-19, normalized'),
    alt.Color('location:N', title="Country", legend=None),
    alt.Tooltip(['location:N', 'total_cases:Q', 'date:T'])
).properties(
    title='Development of Corona infections',
)

base_death = alt.Chart(url_death).encode(
    alt.X('days_after', type='quantitative',  scale=alt.Scale(domain=[0,30], type='ordinal'),
         title='Days since the 10th confirmed death'),
    alt.Y('total_deaths_normalized', type='quantitative', scale=alt.Scale(type='log', base=10), 
          title='Total confirmed deaths of COVID-19, normalized'),
    alt.Color('location:N', title="Country", legend=None),
    alt.Tooltip(['location:N', 'total_deaths:Q', 'date:T'])
).properties(
    title='Development of Corona deaths',
)

base_summary = alt.Chart(url_summary).mark_bar().encode(
    x = alt.X('cases', type='quantitative', title='Confirmed cases', scale=alt.Scale(type='linear')),
    y = alt.Y('location:N', title='Country'),
    color = alt.condition(highlight, alt.Color('location:N', title="Country", legend=None), alt.ColorValue("grey")),
    tooltip = alt.Tooltip(['location:N', 'cases:Q'])
).add_selection(
    highlight
).properties(
    width=200,
    title='Corona cases by country'
)


points_infection = base_infection.mark_circle().encode(
    opacity=alt.value(0)
).add_selection(
    highlight
).properties(
    width=600
).interactive(
    bind_y = False
)

points_death = base_death.mark_circle().encode(
    opacity=alt.value(0)
).add_selection(
    highlight
).properties(
    width=600
).interactive(
    bind_y = False
)


lines_infection = base_infection.mark_line().encode(
    size=alt.condition(~highlight, alt.value(1), alt.value(5), legend=None)
)

lines_death = base_death.mark_line().encode(
    size=alt.condition(~highlight, alt.value(1), alt.value(5), legend=None)
)

chart = alt.vconcat(points_infection + lines_infection, points_death + lines_death)

chart = alt.hconcat(base_summary, chart)


#chart.save('corona.html')
chart.save('corona.json')

chart


# In[ ]: