ENH: Add initial/mean point to chains to fix (#249)

* Add initial/mean point to chains (closes #243)
* A update/fix for the test suite

espei/optimizers/opt_mcmc.py

@@ -96,6 +96,7 @@ def initialize_new_chains(params, chains_per_parameter, std_deviation, determini
         # apply a Gaussian random to each parameter with std dev of std_deviation*parameter
         tiled_parameters = np.tile(params, (nchains, 1))
         chains = rng.normal(tiled_parameters, np.abs(tiled_parameters * std_deviation))
+        chains[0] = params #Ensure the initial guess is always included in the set, as the std_deviation may be too large to generate feasible points around a "good" initial point.
         return chains
 
     @staticmethod

tests/test_mcmc.py

@@ -129,11 +129,14 @@ def test_parameter_initialization():
     initial_parameters = np.array([1, 10, 100, 1000])
     opt = EmceeOptimizer(Database())
     deterministic_params = opt.initialize_new_chains(initial_parameters, 1, 0.10, deterministic=True)
+    print(repr(deterministic_params))
     expected_parameters = np.array([
-        [9.81708401e-01, 9.39027722e+00, 1.08016748e+02, 9.13512881e+02],
-        [1.03116874, 9.01412995, 112.79594345, 916.44725799],
+        list(initial_parameters), # The first element is always a start at the initial parameters
+        #These values are known due to deterministic=True above
+        [1.03116874e+00, 9.01412995e+00, 1.12795943e+02, 9.16447258e+02],
         [1.00664662e+00, 1.07178898e+01, 9.63696718e+01, 1.36872292e+03],
-        [1.07642366e+00, 1.16413520e+01, 8.71742457e+01, 9.61836382e+02]])
+        [1.07642366e+00, 1.16413520e+01, 8.71742457e+01, 9.61836382e+02],        
+    ])
     assert np.all(np.isclose(deterministic_params, expected_parameters))