[Fix] Days argument in forecast - doesn't correspond to the number of prediction days

[Inglezos]

Summary

For example if I use ita_scenario.fit_predict(oxcgrt_data=oxcgrt_data, name="Forecast", days=[1]) and then ita_sim_df = ita_scenario.simulate(variables=["Confirmed", "Infected", "Fatal", "Recovered"], color_dict={"Confirmed": "blue", "Infected": "orange", "Fatal": "red", "Recovered": "green"}, name="Forecast"), I see that in the plot but mostly in the ita_sim_df, the future days are up to 14Apr21, not only up to one day into the future (that would be 11 or 12 Apr21) as I would expect by using days. Is this considered normal? Because I feel that a value of around 2-3 days are added beforehand to the output days.

Codes

import covsirphy as cs

data_loader = cs.DataLoader(directory="kaggle/input")
jhu_data = data_loader.jhu()
population_data = data_loader.population()
oxcgrt_data = data_loader.oxcgrt()

country = "Italy"
ita_scenario = cs.Scenario(jhu_data, population_data, country)
ita_scenario.register(extras=[oxcgrt_data])
ita_delay = ita_scenario.estimate_delay(use_difference=False)
ita_scenario.records(variables=["Confirmed", "Infected", "Fatal", "Recovered"], color_dict={"Confirmed": "blue", "Infected": "orange", "Fatal": "red", "Recovered": "green"})

_ = ita_scenario.trend()
ita_scenario.estimate(cs.SIRF)
ita_scenario.add(name="Main", days=30)

ita_scenario.fit_predict(oxcgrt_data=oxcgrt_data, name="Forecast", days=[1]).summary(name="Forecast")
ita_sim_df = ita_scenario.simulate(variables=["Confirmed", "Infected", "Fatal", "Recovered"],
                                   color_dict={"Confirmed": "blue", "Infected": "orange", "Fatal": "red", "Recovered": "green"},
                                   name="Forecast")

Environment

• CovsirPhy version: v2.19.0
• Python version: 3.8.5
• Installation: conda/pip
• System: Windows

[lisphilar]

Yes, this could be fixed.

Counts of days starts from 0 at this time and confusing. So, days=[1] means that the end date of the future phase is 13Apr when today is 11Apr with the latest version.

Scenario.simulate() simulates to the next date (14Apr) of the end date (13Apr). This behavior is from PhaseUnit.simulate(). PhaseUnit.simulate() needs this begavior internally because values on 14Apr will be used as initial values when the next phase (14Apr - ) is available. However, yes, this is a confusing behavior for users of Scenario.simulate(). The last date (14Apr) should be removed from Scenario.simulate() output.

[lisphilar]

The last date of simulation was changed from the next date (14Apr in the above example) to the exact end date (13Apr) at version 2.19.1-zeta-fu2 with #718.

Next, I will fix the count of days in prediction.

[lisphilar]

I think this was fixed with #756 (version 2.19.1-theta).

[Inglezos]

I think so too, we can close it.