Better model.info for graphical model all fitted simultaneously

[Jammy2211]

The following script composes a graphical model using AnalysisFactor's and a FactorGraphModel (but fits them not using expectation propagation, but instead as just one big model):

https://github.com/Jammy2211/autofit_workspace/blob/release/scripts/howtofit/chapter_graphical_models/tutorial_2_graphical_model.py

The model.info file appears as follows:

Total Free Parameters = 7 

model                                                                                     CollectionPriorModel (N=7)
    0                                                                                     Gaussian (N=3)
    1                                                                                     Gaussian (N=3)
    2                                                                                     Gaussian (N=3) 

0
    centre                                                                                UniformPrior, lower_limit = 0.0, upper_limit = 100.0
    normalization                                                                         UniformPrior, lower_limit = 0.0, upper_limit = 100.0
    sigma                                                                                 UniformPrior, lower_limit = 0.0, upper_limit = 25.0
1
    centre                                                                                UniformPrior, lower_limit = 0.0, upper_limit = 100.0
    normalization                                                                         UniformPrior, lower_limit = 0.0, upper_limit = 100.0
    sigma                                                                                 UniformPrior, lower_limit = 0.0, upper_limit = 25.0
2
    centre                                                                                UniformPrior, lower_limit = 0.0, upper_limit = 100.0
    normalization                                                                         UniformPrior, lower_limit = 0.0, upper_limit = 100.0
    sigma                                                                                 UniformPrior, lower_limit = 0.0, upper_limit = 25.0

The model.results file is:

Bayesian Evidence                                                                         527.75383275
Maximum Log Likelihood                                                                    551.00961857
Maximum Log Posterior                                                                     551.00961857

model                                                                                     CollectionPriorModel (N=7)
    0                                                                                     Gaussian (N=3)
    1                                                                                     Gaussian (N=3)
    2                                                                                     Gaussian (N=3)

Maximum Log Likelihood Model:

2
    centre                                                                                50.329
    normalization                                                                         4.623
    sigma                                                                                 13.379
0
    normalization                                                                         1.903
    sigma                                                                                 7.046
1
    normalization                                                                         1.882
    sigma                                                                                 8.156


Summary (3.0 sigma limits):

2
    centre                                                                                50.32 (48.09, 52.19)
    normalization                                                                         4.69 (3.88, 5.71)
    sigma                                                                                 13.58 (11.28, 17.43)
0
    normalization                                                                         1.77 (1.24, 2.38)
    sigma                                                                                 6.91 (4.47, 10.28)
1
    normalization                                                                         1.97 (1.27, 2.82)
    sigma                                                                                 8.59 (5.41, 13.20)


Summary (1.0 sigma limits):

2
    centre                                                                                50.32 (49.69, 50.98)
    normalization                                                                         4.69 (4.37, 4.99)
    sigma                                                                                 13.58 (12.83, 14.50)
0
    normalization                                                                         1.77 (1.58, 2.00)
    sigma                                                                                 6.91 (6.10, 7.92)
1
    normalization                                                                         1.97 (1.76, 2.18)
    sigma                                                                                 8.59 (7.57, 9.75)

instances

Following https://github.com/rhayes777/PyAutoFit/issues/418, are there ways that we can use the same functionality to produce files that better describe the graph's structure?

[rhayes777]

This is a bit tricky because at the moment the thing being optimised is just a classical AutoFit model.

I could probably change the model.info quite easily but the model.results is constructed in a very different way for factor graphs