fastp will overwrite -o without checking if it is the same as -i. Worth adding a check for this scenario.

While attempting to fill an incomplete batch, no attempt is made to follow the strategy specified by step_to_index_fn in choosing the next batch/samples from the next batch as of now.

Please note that if punkt resource is not found on NLTK, your code would not work.
I suggest to add a check somewhere in the code for the resource, or add to documentation that users should perform nltk.download('punkt') before using this code.

Hi,

I was wondering if there's a mistake in the FAQ section of the documentation ("Should I apply ICA first or autoreject first?").
According to the MNE docs the reject parameter in ica.fit() only applies to an instance of Raw, not Epochs.
However, here ica.fit() is used with reject on epochs several times, e.g.:

>>> reject = get_rejection_threshold(epochs)
>>> ica.fit(epo

If you accidentally provide a ill-formatted configuration file (such as forgetting to specify the config and resulting in the first file being considered the config), you get a confusing error message:

input/sample.fq.tagged.1.tophat/accepted_hits.bam(1): expected <

An message like 'Error processing configuration file', would be must more useful.

Dear @olivierhagolle, I would like to try to run MAJA using a JobOrder configuration file instead of the command-line mode.

I saw the example of a JobOrder file in the user manual (MU), but is there more documentation on it somewhere? In particular, what values do I need to change for processing Sentinel 2 scenes? The example seems to be for another input data type (judging from the file n

We will probably have to do this sooner or later, although the style may not fit well to document individual PipeOps.