Self containment metric

[carolineychen8]

No description provided.

[paco-barreras]

Let's assume this information is in the "location_id" column. Which is passed as kwargs or in traj_cols similar to home_attribution.

home_activity_type can be "home_id", defaulting to home, but can take any location id.

[paco-barreras]

If there is NO activity, then np.nan is appropriate. But if there are stops at home and no non-home, then the containment should be 1.0

[paco-barreras]

Overly-defensive and fails silently. Dist from home should never be np.nan for any of the stops and, thus, we should just let it fail.

[paco-barreras]

If none of the stops correspond to home_id, maybe raise a warning when the WHOLE stop table does not contain a home.

[paco-barreras]

Self containment is expected to be 1.0 not na

[andresmondragont]

Failing Test:

• test_from_df_values_types_na

===================================================================== FAILURES ======================================================================
___________________________________________________________ test_from_df_values_types_na ____________________________________________________________

value_na_input_df =   user_id     timestamp  latitude  longitude              datetime
0       A  1672560000.0   40.7128    -74.006  2023-...2617600.0   36.1699  -115.1398  2023-01-01T18:00:00Z
4       C  1672531200.0   40.7128    -74.006  2023-01-01T00:00:00Z
expected_value_na_output_df =   user_id   timestamp  latitude  longitude            datetime  tz_offset
0       A  1672560000   40.7128    -74.006 2...699  -115.1398 2023-01-01 18:00:00          0
4       C  1672531200   40.7128    -74.006 2023-01-01 00:00:00          0

    def test_from_df_values_types_na(value_na_input_df, expected_value_na_output_df):
        """Tests from_df for correct values, type casting, and NA handling."""
        traj_cols = {
            "user_id": "user_id", "timestamp": "timestamp", "latitude": "latitude",
            "longitude": "longitude", "datetime": "datetime",
        }
        mixed_tz = 'naive'
    
        result = loader.from_df(value_na_input_df.copy(), traj_cols=traj_cols,
                                parse_dates=True, mixed_timezone_behavior=mixed_tz, sort_times=False)
    
>       pd.testing.assert_frame_equal(result, expected_value_na_output_df, check_dtype=True)
E       AssertionError: Attributes of DataFrame.iloc[:, 4] (column name="datetime") are different
E       
E       Attribute "dtype" are different
E       [left]:  datetime64[us]
E       [right]: datetime64[ns]

nomad/tests/test_io.py:249: AssertionError

[andresmondragont]

fix failing test: test_from_df_values_types_na