Fix order of columns after `TSDataset.to_flatten`

[Mr-Geekman]

🚀 Feature Request

It can be useful to have a more strict rule about order of columns after TSDataset.to_flatten.

Proposal

I propose the following order of columns:

• timestamp,
• segment,
• target,
• other columns in alphabetical order.

How it can be done:

1. Find TSDataset.to_flatten method.
2. Move line df_dict["segment"] = np.repeat(segments, len(df.index)) after the line df_dict["timestamp"] = np.tile(df.index, len(segments)).
3. Before for column in columns cycle:
   • Move value "target" at the beginning of columns list,
   • Order other values in alphabetical order.

Test cases

1. Fix doctest of TSDataset.to_flatten.
2. Fix docstes of TSDataset.to_pandas.
3. Make sure current test pass.
4. Add test on order of columns into test_dataset. test_to_flatten_with_exog .
   • The expected_df should be changed to have a correct order of columns.
   • Instead of 3 asserts at the end use pd.testing.assert_frame_equal(obtained_df, expectged_df).

Additional context

No response