The pretty-print of a dimension 1-dimension Distribution does not show its parameters

[mbaudin47]

What is the idea?

When we create a Distribution, we can pretty-print it. This was first developed at #2282, #2447 for ComposedDistribution. But when we print a distribution in dimension 1, then the pretty-print does not show the parameters of the distribution.

Why is this needed?

The parameters are an important part of the distribution. When hidden, lots of informations are lost.

Additional Context

Here is an example:

import openturns as ot

d = ot.Gumbel()
d

This prints:

---

Gumbel

• name=Gumbel
• dimension=1
• weight=1
• range=]-inf (-3.47309), (32.2362) +inf[
• description=[X0]
• isParallel=true
• isCopula=false

---

We can see the type of distribution, but not its parameters.

Instead, the not-so-pretty print works well:

print(d)

This prints:

Gumbel(beta = 1, gamma = 0)

Analysis

The pretty-print feature of a Distribution is available of the code in DistributionImplementation::repr_html(). The limitation is because none of the implementations overload that pretty-print with their own parameters. This implicit choice was made because the current pretty-print shows the attributes of the class: the parameters are not attributes of this generic class.