Sampling constraints

[kahaaga]

We need sampling constraints for both UncertainIndexValueDataset and UncertainValueDataset.

In particular, we need the following SamplingConstraints for regular UncertainDatasets.

• [x] NoConstraint. Sample the full distributions.
• [x] TruncateLowerQuantile. Truncate the lower end of the distributions at some quantile.
• [x] TruncateUpperQuantile. Truncate the upper end of the distributions at some quantile.
• [x] TruncateQuantiles. Truncate both lower and upper end of the distributions at some quantiles.
• [x] TruncateMinimum. Truncate the lower end of the distributions at some set minimum value.
• [x] TruncateMaximum. Truncate the upper end of the distributions at some set maximumvalue.
• [x] TruncateRange. Truncate both lower and upper end of the distributions at some set minimum and maximum values.
• [x] TruncateStd. Truncate the furnishing distributions at some multiple of the standard deviation around the mean.
• [ ] TruncateMAD. Truncate at the mean absolute deviation (more robust to outliers than TruncateStd).
• [ ] TruncateGMAD. Truncate at the geometric mean absolute deviation.
• [ ] TruncateIQR. Truncate at the interquartile range.
• [ ] TruncateQ1. Truncate at the first quartile.
• [ ] TruncateQ2. Truncate at the second quartile.
• [ ] TruncateQ3. Truncate at the third quartile.
• [ ] TruncateIDR. Truncate at the interdecile range.

For UncertainIndexDatasets, we additionally need:

• [x] StrictlyIncreasing. Sample strictly increasing values. Done in #38.
• [ ] StrictlyIncreasingViolationsAllowed. Sample strictly increasing values with some fraction of violations of that criterion allowed.
• [x] StrictlyDecreasing. Sample strictly decreasing values. Done in #38.
• [ ] StrictlyDecreasingViolationsAllowed. Sample strictly decreasing values with some fraction of violations of that criterion allowed.

It should be possible to apply multiple constraints to a dataset.