Heston model has accurate density approximations for European option prices, which are of interest.

The module implementing this method should live under tf_quant_finance/volatility/heston_approximation.py. It should support both European option puts and calls approximations. Tests should be in heston_approximation_test.py in the same folder.

A point load can be mathematically represented as a distribution, e.g., a Dirac delta. It breaks the Gridap flow, since one cannot use Gauss quadratures and numerical integration (what we usually do in FEM) to compute the integral of f*v in that case.

I don't want to consider hacks, e.g., touch the vector entry in a particular node in which you want to put the force (assuming the force is on

Code coverage should be added to the Travis config (perhaps the cabal file and/or the stack options need to be changed in order to account for hpc coverage generation); currently in Travis there is only a project key.
This tool uploads hpc coverage reports to codecov.io .

We should try to improve the fallback operations for RandomVariables in probnum.random_variables._arithmetic and the unary fallback operations in RandomVariable itself.
Ideas include:

• Generic random variables can always be multiplied by Diracs which provides a closed form expression for the mean by leveraging linearity E[aX] = a E[X] if a ~ Dirac
• One should be able to defin