Error computing test_value for transformed dist

[junpenglao]

import pymc3 as pm
with pm.Model() as model:
    a1 = pm.Uniform('a1', lower=0., upper=1.)
    b1 = pm.Uniform('b1', lower=0., upper=1. - a1)
    pm.sample(1000)

raise:

Traceback (most recent call last):

  File "<ipython-input-1-e7f2b743f1a1>", line 5, in <module>
    pm.sample(1000)

  File "/usr/local/lib/python3.5/dist-packages/pymc3/sampling.py", line 273, in sample
    return sample_func(**sample_args)[discard:]

  File "/usr/local/lib/python3.5/dist-packages/pymc3/sampling.py", line 288, in _sample
    for it, strace in enumerate(sampling):

  File "/usr/local/lib/python3.5/dist-packages/tqdm/_tqdm.py", line 862, in __iter__
    for obj in iterable:

  File "/usr/local/lib/python3.5/dist-packages/pymc3/sampling.py", line 367, in _iter_sample
    _update_start_vals(start, model.test_point, model)

  File "/usr/local/lib/python3.5/dist-packages/pymc3/sampling.py", line 483, in _update_start_vals
    b[tname] = transform_func[0].forward(a[name]).eval()

  File "/usr/local/lib/python3.5/dist-packages/theano/gof/graph.py", line 516, in eval
    self._fn_cache[inputs] = theano.function(inputs, self)

  File "/usr/local/lib/python3.5/dist-packages/theano/compile/function.py", line 326, in function
    output_keys=output_keys)

  File "/usr/local/lib/python3.5/dist-packages/theano/compile/pfunc.py", line 486, in pfunc
    output_keys=output_keys)

  File "/usr/local/lib/python3.5/dist-packages/theano/compile/function_module.py", line 1807, in orig_function
    output_keys=output_keys).create(

  File "/usr/local/lib/python3.5/dist-packages/theano/compile/function_module.py", line 1446, in __init__
    accept_inplace)

  File "/usr/local/lib/python3.5/dist-packages/theano/compile/function_module.py", line 177, in std_fgraph
    update_mapping=update_mapping)

  File "/usr/local/lib/python3.5/dist-packages/theano/gof/fg.py", line 174, in __init__
    self.__import_r__(output, reason="init")

  File "/usr/local/lib/python3.5/dist-packages/theano/gof/fg.py", line 345, in __import_r__
    self.__import__(variable.owner, reason=reason)

  File "/usr/local/lib/python3.5/dist-packages/theano/gof/fg.py", line 390, in __import__
    raise MissingInputError(error_msg, variable=r)

MissingInputError: Input 0 of the graph (indices start from 0), used to compute sigmoid(a1_interval__), was not provided and not given a value. Use the Theano flag exception_verbosity='high', for more information on this error.

Diagnostic:
Error is from the https://github.com/pymc-devs/pymc3/blob/master/pymc3/sampling.py#L483

for example:

from pymc3.util import is_transformed_name, get_untransformed_name
def update_start_vals(a, b, model):
    """Update a with b, without overwriting existing keys. Values specified for
    transformed variables on the original scale are also transformed and inserted.
    """
    for name in a:
        for tname in b:
            if is_transformed_name(tname) and get_untransformed_name(tname) == name:
                transform_func = [d.transformation for d in model.deterministics if d.name == name]
                if transform_func:
                    b[tname] = transform_func[0].forward(a[name]).eval()

    a.update({k: v for k, v in b.items() if k not in a})

a = {'a1': .5, 'b1':.7}
b = model.test_point
update_start_vals(a, b, model)

gives the same error.

[junpenglao]

Also related to below:
#2109 (comment)