`ValueError: `bins` must increase monotonically, when an array` when propagating uncertainty

[andreaegg]

When using the FitResult.propagate() function with a parametric model I get the following error:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
~\AppData\Local\Temp\ipykernel_12144\449243555.py in <module>
     26         Pfcn = lambda V0,sigma: mqr.rbd_gauss(V3,V0=V0,sigma=sigma)
     27         Pfit = resultlocal.evaluate(Pfcn)
---> 28         PfitUncert = resultlocal.propagate(Pfcn)
     29         Vfit = resultlocal.model
     30         VfitUncert = resultlocal.modelUncert

c:\users\andre\documents\doktorat\programs\deerlab\deerlab\model.py in propagate(model, lb, ub, *constants)
   1102         subset = [param_idx[np.where(np.asarray(_paramlist)==param)[0][0]] for param in modelparam]
   1103         # Propagate the uncertainty from that subset to the model
-> 1104         modeluq = fitresults.paramUncert.propagate(lambda param: model(*constants,*[param[s] for s in subset]),lb,ub)
   1105         return modeluq
   1106     # ----------------------------------------------------------------------------

c:\users\andre\documents\doktorat\programs\deerlab\deerlab\classes.py in propagate(self, model, lb, ub, samples)
    532 
    533             # Construct new uncertainty object
--> 534             return  UQResult('bootstrap',data=sampled_model,lb=lb,ub=ub)
    535 
    536     #--------------------------------------------------------------------------------

c:\users\andre\documents\doktorat\programs\deerlab\deerlab\classes.py in __init__(self, uqtype, data, covmat, lb, ub, threshold, profiles, noiselvl)
    170             covmat = np.squeeze(samples).T@np.squeeze(samples)/np.shape(samples)[0] - means*means.T
    171             self.mean = means
--> 172             self.median = self.percentile(50)
    173             self.median = np.array([nth_samples[0] if np.all(nth_samples==nth_samples[0]) else self.median[n] for n,nth_samples in enumerate(samples.T)])
    174             self.std = np.squeeze(np.std(samples,0))

c:\users\andre\documents\doktorat\programs\deerlab\deerlab\classes.py in percentile(self, p)
    350         for n in range(self.nparam):
    351             # Get parameter PDF
--> 352             values,pdf = self.pardist(n)
    353             # Compute corresponding CDF
    354             cdf = np.cumsum(pdf)

c:\users\andre\documents\doktorat\programs\deerlab\deerlab\classes.py in pardist(self, n)
    278                 minbin = np.minimum(np.min(samplen),np.mean(samplen)-3*sigma)
    279                 bins = np.linspace(minbin,maxbin, 2**10 + 1)
--> 280                 count, edges = np.histogram(samplen, bins=bins)
    281 
    282                 # Generate kernel

<__array_function__ internals> in histogram(*args, **kwargs)

c:\Users\andre\AppData\Local\Programs\Python\Python37\lib\site-packages\numpy\lib\histograms.py in histogram(a, bins, range, normed, weights, density)
    791     a, weights = _ravel_and_check_weights(a, weights)
    792 
--> 793     bin_edges, uniform_bins = _get_bin_edges(a, bins, range, weights)
    794 
    795     # Histogram is an integer or a float array depending on the weights.

c:\Users\andre\AppData\Local\Programs\Python\Python37\lib\site-packages\numpy\lib\histograms.py in _get_bin_edges(a, bins, range, weights)
    430         if np.any(bin_edges[:-1] > bin_edges[1:]):
    431             raise ValueError(
--> 432                 '`bins` must increase monotonically, when an array')
    433 
    434     else:

ValueError: `bins` must increase monotonically, when an array