The default kwargs for epoch_ch_sd (which are both referenced in flag_ch_sd and passed into marks_array_2_flags as kwargs) are below. Note that the default value for init_method is null (which translates to None in python).
In flag_ch_sd, if init_method is None, then an empty list is passed into flag_sd_t_inds, and nothing else is done. See line 501 in the code block below. Then see line 511, flag_sd_t_inds (which is an empty list in this case) is passed into self.flagged_epochs.add_flag_cat. This is why no bad_pylossless_ch_sdannotations are set at this step in the pipeline.
IF you instead set init_method to q, for example, then you'll hit the code block starting in line 505 , and many flagged_epoch annotations will be created (I tested this on one file so far, and 24 bad_pylossless_ch_sdannotations were set).
3 questions:
is setting [epoch_ch_sd][init_method] to null/None the default value we want for the default config?
Do we even want to allow this value to be None?
If the answer to 2 is Yes, should we at least provide a logger.warn to the end-user so they are aware that there will be no assessment for outlying epochs in flag_ch_sd? Because it took me an hour of code introspection to realize what was going on here.
The default kwargs for
epoch_ch_sd(which are both referenced inflag_ch_sdand passed intomarks_array_2_flagsas kwargs) are below. Note that the default value forinit_methodisnull(which translates toNonein python).https://github.com/lina-usc/pylossless/blob/c444b0af2d7bb502e23d82d10dde095d540c71ae/assets/ll_default_config.yaml#L129-L135
In
flag_ch_sd, ifinit_methodisNone, then an empty list is passed intoflag_sd_t_inds, and nothing else is done. See line501in the code block below. Then see line511,flag_sd_t_inds(which is an empty list in this case) is passed intoself.flagged_epochs.add_flag_cat. This is why nobad_pylossless_ch_sdannotationsare set at this step in the pipeline.https://github.com/lina-usc/pylossless/blob/c444b0af2d7bb502e23d82d10dde095d540c71ae/pylossless/pipeline.py#L495-L511
IF you instead set
init_methodtoq, for example, then you'll hit the code block starting in line505, and many flagged_epoch annotations will be created (I tested this on one file so far, and 24bad_pylossless_ch_sdannotationswere set).3 questions:
epoch_ch_sd][init_method] tonull/Nonethe default value we want for thedefault config?None?logger.warnto the end-user so they are aware that there will be no assessment for outlying epochs inflag_ch_sd? Because it took me an hour of code introspection to realize what was going on here.