[Fix] ValueError when adding phases manually to fully-complemented data

[lisphilar]

Summary

ValueError was raised when adding phases manually to fully-complemented data.
Data ("cook county.csv") was provided by @AnujTiwari with #851.

Codes

import covsirphy as cs
country_data = cs.CountryData("input/cook county.csv", country="Cook")
country_data.set_variables(
    date="date", confirmed="confirmed", recovered="recovered", fatal="fatal")

country_data.register_total()
country_data.cleaned()

jhu_data_cook = cs.JHUData.from_dataframe(country_data.cleaned())
population_data_cook = cs.PopulationData()
population_data_cook.update(5150233, country="Cook")

snl = cs.Scenario(country="Cook")
snl.register(jhu_data_cook, population_data_cook)
snl.records()

snl.trend(algo="Pelt-rbf").summary()
snl.estimate(cs.SIRF)

snl.clear(include_past=True).summary()
snl.add(end_date="28Apr2020")

Outputs

ValueError: @start must be the same as/over 15Mar2020, but 27Feb2020 was applied.

Environment

• CovsirPhy version: 2.21.0-iota-fu4
• Python version: 3.9.5
• Installation: poetry
• System: WSL (Ubuntu)

[lisphilar]

Reopened with #851.
https://gist.github.com/lisphilar/1063a31147d50dff252e682673b1e600