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

# In[91]:


# read my matomo api token
with open("matomo-token") as f:
    token = f.read().strip()


# In[18]:


url = "https://mybinder.org/matomo/index.php"


# In[92]:


import requests

def api_request(method, **params):
    """Make a matomo API request"""
    data={
        "token_auth": token,
        "format": "JSON",
        "module": "API",
        "method": method,
        "idSite": "1",
    }
    data.update(params)
    r = requests.post(url, data=data)
    try:
        r.raise_for_status()
    except requests.HTTPError as e:
        print(e)
        if r.status_code >= 500:
            # monthly queries seem to time out a lot
            # eventually, the result is cached and it responds quickly
            print("retrying")
            time.sleep(1)
            return api_request(method, **params)
        else:
            raise
    return r.json()

api_request("API.getMatomoVersion")
            


# In[70]:


import pandas as pd

months = {}


# retry a bunch because it seems to time out the first time we request each month,
# but eventually succeed

# In[93]:


for month in range(1, 11):
    if month not in months:
        date = f"2021-{month:02}-01"
        print(f"Fetching {date}")
        records = api_request(
            "UserCountry.getContinent",
            period="month",
            date=date,
            showColumns="label,nb_visits,sum_daily_nb_uniq_visitors",
        )
        for record in records:
            record["date"] = date
        months[month] = records


# In[102]:


df.tail(10)


# In[99]:


from itertools import chain

import altair as alt

df = pd.DataFrame(chain(*months.values()))
df.head()

alt.Chart(df).mark_line().encode(
    color=alt.Color(
        "label", title="Continent", sort={"encoding": "y", "order": "descending"}
    ),
    x=alt.Y("date", title="Month"),
    y=alt.Y(
        "nb_visits",
        title="Monthly visits",
        scale=alt.Scale(type="log"),
    ),
).interactive()


# And again, summing all non-NA-EU together,
# ignoring "Unknown"

# In[117]:


gross = df[df.label != "Unknown"]
gross["Gross Region"] = gross.label

gross.loc[~gross.label.isin(["North America", "Europe"]), "Gross Region"] = "Rest of World"

# regroup sum by gross region
gross = gross.groupby(["date", "Gross Region"]).nb_visits.sum().reset_index()
gross


# In[118]:


alt.Chart(gross).mark_line().encode(
    color=alt.Color(
        "Gross Region", sort={"encoding": "y", "order": "descending"}
    ),
    x=alt.Y("date", title="Month"),
    y=alt.Y(
        "nb_visits",
        title="Monthly visits",
        scale=alt.Scale(type="log"),
    ),
    tooltip=[
        "Gross Region",
        "date",
        "nb_visits",
    ],
).interactive()