Make print_report in filtering functions False by default

[niksirbi]

Our filtering functions accept a print_report boolean argument, which is True by default.
When True, the function prints the percentage of NaNs for each keypoint, before and after the filter.
This is a very useful feature for seeing the effects of filtering on missing values, but it gets quite annoying when you have to do chain multiple filters or to loop a single filter over multiple datasets (as one would typically do in a preprocessing pipeline).
In such cases, the printed output grows quite fast and is very verbose.

Therefore I propose making the default value to be False. This functionality will be still documented in the examples, so people will be aware of its existence should they need it.