def get_pies(df_m, df_m_trend):
"""Get pies info"""
return {
"all": serie_to_dict(to_percentages(df_m.mean())),
"month": serie_to_dict(to_percentages(df_m.tail(1).T.iloc[:, 0])),
"year": serie_to_dict(to_percentages(df_m.iloc[-12:].mean())),
}
def get_year_data(dfi):
df = dfi.pivot_table(values=c.COL_PAGES,
index=c.COL_DATE,
columns=c.COL_LANGUAGE,
aggfunc="sum")
df = to_year_start(df)
out = {x: serie_to_dict(df[x]) for x in df.columns}
out["Total"] = serie_to_dict(to_year_start(dfi)[c.COL_PAGES])
return out
def get_month_data(dfi):
out = {
i: serie_to_dict(dfa.resample("MS")[c.COL_PAGES].sum())
for i, dfa in dfi.set_index(c.COL_DATE).groupby(c.COL_LANGUAGE)
}
out["Total"] = serie_to_dict(
dfi.set_index(c.COL_DATE).resample("MS")[c.COL_PAGES].sum())
for name, data in {
**out
}.items(): # The ** is to avoid problems while mutating the dict
serie = smooth_serie(pd.Series(data))
out[f"{name}_trend"] = serie_to_dict(serie[6:])
return out
def get_dashboard(dfi):
out = serie_to_dict(dfi.groupby(c.COL_LANGUAGE)[c.COL_PAGES].sum())
out["Total"] = int(dfi[c.COL_PAGES].sum())
out["Years"] = int(dfi[c.COL_DATE].dt.year.nunique())
return out
def get_top(dfi, groupby, top_n=20):
df = (dfi.groupby(groupby).agg({
c.COL_PAGES: "sum"
}).sort_values(c.COL_PAGES, ascending=False)[:top_n])
return serie_to_dict(df[c.COL_PAGES])
def get_year_percent(data, cumsum=True):
df = pd.DataFrame(data)
if cumsum:
df = df.cumsum()
# Get percentatges
df = 100 * df.div(df["Total"], axis=0).fillna(0)
return {x: serie_to_dict(df[x]) for x in df.columns if x != "Total"}
def extract_data(vdp, df, export=False):
"""Extract data from the dataframe"""
calendars = read_calendars()
df_w_trend = df.resample("W-MON").sum().apply(smooth_serie)
df_m = df.resample("MS").sum()
df_m_trend = df_m.apply(smooth_serie)
to_dict_reversed = lambda df: serie_to_dict()
out = {
"week_trend": serie_to_dict(df_w_trend),
"month": serie_to_dict(df_m),
"month_percent": serie_to_dict(to_percentages(df_m)),
"month_trend": serie_to_dict(df_m_trend),
"month_trend_percent": serie_to_dict(to_percentages(df_m_trend)),
"pies": get_pies(df_m, df_m_trend),
"colors": {name: data["color"] for name, data in calendars.items()},
}
out["cards"] = get_cards(out, calendars)
if export:
vdp.write_yaml(out, f"I will fail")
return out