IPTW restrict outlier weights

[pzivich]

Summary:
In addition of the support of bounds, I should add a restriction (drop observations with extreme weights). This addition adds greater flexibility and only requires some minor modifications to the IPTW procedure, predominantly during the fit() function.

What this adds:
This addition allows users to have more control over the estimation process. Particularly how outliers are handled based on their weights. This will allow for better sensitivity analyses to be coded easily.

Implementation plan:
This would be accomplished by adding an optional restrict statement in the fit() function. Behind the scenes, it would filter out the observations that are outside the specified criteria.

If this option is used, summary() will additionally provide a warning to the user that the target population has changed

Lastly, I should add a function called describe_restrictions() which summarizes the attributes of the restricted population. This is helpful information for potentially understanding the target population. I should also allow access to the ID's of the excluded units or a summary data set of them for users to further analyze themselves