Annotations mismatch when dataset has neutral loss

[aeisenbarth]

When a dataset has been processed with chemical modifications or neutral loss, the annotations dataframe does not match the number of ion images.

That is because dataset.all_annotation_images() literally downloads all images, where as dataset.results() excludes chemical modifications by default. Thus there are more ion images than dataframe rows and the assertion/assumption fails.

There are two extra arguments include_chem_mods=True, include_neutral_losses=True that need to be passed so that the dataframe includes all annotations.