Selecting Observations¶

When using SBART we can reject observations in two different ways:

• globally - they are effectivelly discarded for all SBART operations
• temporarily -Only disable them for the creation of the Stellar Templates

Even though the two rejection types have completely different effects, we always use the same format

Based on Header values¶

• Allows to define bounds for the "valid" values that a given header value can take.
• The bounds argument takes a list of list, with each entry defining a "region" that the value can take.
• If we want to place no upper/lower limits, we can place a None

    # Only use observations that can have AIRMASS smaller than 1.5 or larger than 1.6
    valid_AIRM_condition = KEYWORD_condition(KW = "airmass",
                                          bounds=[[None, 1.5], [1.6, None]]
                                         )

    # Only select observations that have a previous RV error (the one from the CCF) smaller than 50 meter_second
    # Note1: the conditions for 'RV related' values (previousRV, drift, BERV) must be in meter_second OR kilometer_second
    # Note2: the units will be converted later on, so either "unit" can be chosen

    valid_error_condition = KEYWORD_condition(KW="previousRV_ERR",
                                              bounds = [0, 50*meter_second])

Rejection based on filenames¶

Provide a list of bad filenames that will be rejected.

Note: Filename only, do NOT pass a path

    bad_filename_condition = FNAME_condition(filename_list=["filename.fits"])

Rejecting based on the "subInstrument"¶

Reject all observations from ESPRESSO21

    bad_subInst_condition = SubInstrument_condition("ESPRESSO21")

Rejecting based on Warning Flags¶

Reject observation if a given warning flag was set when loading the spectra

    bad_subInst_condition = WarningFlag_Notset("HIERARCH ESO QC SCIRED DRIFT CHECK")

Combining conditions¶

To combine multiple conditions, sum them together. The observations will be rejected if any Condition decides to reject it

    full_conditions = valid_AIRM_condition + valid_error_condition + bad_filename_condition + bad_subInst_condition

Applying the conditions¶

For a DataClass object, we can use its reject_observations method

    data.reject_observations(full_conditions)

Otherwise, other data-aggregation methods usually allow to provide a condition model to help specify the observations of interest.