Introduce revisited valuation semantics

[matteolimberto-da]

This PR introduces a revisited fundamental asset pricing formula.

The existing one did the following:

• map a claim to an expression representing a stochastic process
• make a model assumption, using geometric brownian motion as an elementary process
• serialise the resulting formula to text

The revisited FAPF does only the first step: map a claim to a stochastic process representing its value at time t. The process primitives that are used are the ones outlined in the PJ paper: exch, disc, observable (in the future also snell and absorb.

Moreover, we now handle the propagation of acquisition time, such that

• scale nodes that are evaluated before today are replaced with their numerical value
• scale nodes that are unknown are mapped to the corresponding stoch. process

There are still a couple of things to iron out, specifically

• implementation of snell and absorb primitive processes to handle anytime and until nodes (I think we can come to a better expression than what there is in the paper)
• simplification functions

I would like to hear your feedback on this and also discuss whether we need support for printing out the formula in e.g. MathML like it's currently done for the existing fapf.

The tests can give you an idea of the output which is currently generated.

[matteolimberto-da]

A nice POC would be to implement a bond pricer in e.g. Typescript that

• maps the claim to a stochastic process (using CC)
• applies model assumptions (e.g. deterministic interest rates)
• evaluates the resulting expression to get the price of the bond

[matteolimberto-da]

Will close this one and open in Daml Finance