model_to_graphviz displays incomplete graph

[JackCaster]

model_to_graphviz is awesome!
However, I have an example in which the graph seems to be incomplete: see this notebook. There, I am modelling a beta distribution. Instead of modelling directly the parameters alpha and beta, I transform them to work with central tendency and precision.

In the graph above I would expect the blocks for alpha and beta to be connected to the likelihood block.

OS: Windows 10
PyMC3 version: updated to master
Other packages:

# Name                    Version                   Build  Channel
appdirs                   1.4.3            py36h28b3542_0
asn1crypto                0.24.0                   py36_0
attrs                     18.1.0                   py36_0
automat                   0.7.0                    py36_0
backcall                  0.1.0                    py36_0
blas                      1.0                         mkl
bleach                    2.1.3                    py36_0
ca-certificates           2018.03.07                    0
certifi                   2018.8.13                py36_0
cffi                      1.11.5           py36h945400d_0
colorama                  0.3.9            py36h029ae33_0
constantly                15.1.0           py36h28b3542_0
cryptography              2.3              py36h74b6da3_0
cryptography-vectors      2.3                      py36_0
cycler                    0.10.0           py36h009560c_0
cython                    0.28.5           py36h6538335_0
decorator                 4.3.0                    py36_0
entrypoints               0.2.3                    py36_2
freetype                  2.9.1                ha9979f8_1
graphviz                  2.38                 hfd603c8_2
h5py                      2.8.0            py36h3bdd7fb_1
hdf5                      1.10.2               hac2f561_1
html5lib                  1.0.1                    py36_0
hyperlink                 18.0.0                   py36_0
icc_rt                    2017.0.4             h97af966_0
icu                       58.2                 ha66f8fd_1
idna                      2.7                      py36_0
incremental               17.5.0                   py36_0
intel-openmp              2018.0.3                      0
ipykernel                 4.8.2                    py36_0
ipython                   6.5.0                    py36_0
ipython_genutils          0.2.0            py36h3c5d0ee_0
ipywidgets                7.4.0                    py36_0
jedi                      0.12.1                   py36_0
jinja2                    2.10                     py36_0
joblib                    0.12.2                   py36_0
jpeg                      9b                   hb83a4c4_2
jsonschema                2.6.0            py36h7636477_0
jupyter                   1.0.0                    py36_5
jupyter_client            5.2.3                    py36_0
jupyter_console           5.2.0                    py36_1
jupyter_core              4.4.0                    py36_0
kiwisolver                1.0.1            py36h6538335_0
libgpuarray               0.7.6                hfa6e2cd_0
libpng                    1.6.34               h79bbb47_0
libpython                 2.1                      py36_0
libsodium                 1.0.16               h9d3ae62_0
m2w64-binutils            2.25.1                        5
m2w64-bzip2               1.0.6                         6
m2w64-crt-git             5.0.0.4636.2595836               2
m2w64-gcc                 5.3.0                         6
m2w64-gcc-ada             5.3.0                         6
m2w64-gcc-fortran         5.3.0                         6
m2w64-gcc-libgfortran     5.3.0                         6
m2w64-gcc-libs            5.3.0                         7
m2w64-gcc-libs-core       5.3.0                         7
m2w64-gcc-objc            5.3.0                         6
m2w64-gmp                 6.1.0                         2
m2w64-headers-git         5.0.0.4636.c0ad18a               2
m2w64-isl                 0.16.1                        2
m2w64-libiconv            1.14                          6
m2w64-libmangle-git       5.0.0.4509.2e5a9a2               2
m2w64-libwinpthread-git   5.0.0.4634.697f757               2
m2w64-make                4.1.2351.a80a8b8               2
m2w64-mpc                 1.0.3                         3
m2w64-mpfr                3.1.4                         4
m2w64-pkg-config          0.29.1                        2
m2w64-toolchain           5.3.0                         7
m2w64-tools-git           5.0.0.4592.90b8472               2
m2w64-windows-default-manifest 6.4                           3
m2w64-winpthreads-git     5.0.0.4634.697f757               2
m2w64-zlib                1.2.8                        10
mako                      1.0.7                    py36_0
markupsafe                1.0              py36hfa6e2cd_1
matplotlib                2.2.3            py36hd159220_0
mistune                   0.8.3            py36hfa6e2cd_1
mkl                       2018.0.3                      1
mkl-service               1.1.2            py36hb217b18_4
mkl_fft                   1.0.4            py36h1e22a9b_1
mkl_random                1.0.1            py36h77b88f5_1
msys2-conda-epoch         20160418                      1
nbconvert                 5.3.1                    py36_0
nbformat                  4.4.0            py36h3a5bc1b_0
notebook                  5.6.0                    py36_0
numpy                     1.15.0           py36h9fa60d3_0
numpy-base                1.15.0           py36h4a99626_0
openssl                   1.0.2p               hfa6e2cd_0
pandas                    0.23.4           py36h830ac7b_0
pandoc                    2.2.1                h1a437c5_0
pandocfilters             1.4.2                    py36_1
parso                     0.3.1                    py36_0
patsy                     0.5.0                    py36_0
pickleshare               0.7.4            py36h9de030f_0
pip                       10.0.1                   py36_0
prometheus_client         0.3.1                    py36_0
prompt_toolkit            1.0.15           py36h60b8f86_0
pyasn1                    0.4.4                    py36_0
pyasn1-modules            0.2.2                    py36_0
pycparser                 2.18                     py36_1
pygments                  2.2.0            py36hb010967_0
pygpu                     0.7.6            py36h452e1ab_0
pyhamcrest                1.9.0                      py_2    conda-forge
pymc3                     3.5                       <pip>
pyopenssl                 18.0.0                   py36_0
pyparsing                 2.2.0                    py36_1
pyqt                      5.9.2            py36ha878b3d_0
python                    3.6.6                hea74fb7_0
python-dateutil           2.7.3                    py36_0
python-graphviz           0.8.4                    py36_1
pytz                      2018.5                   py36_0
pywin32                   223              py36hfa6e2cd_1
pywinpty                  0.5.4                    py36_0
pyzmq                     17.1.2           py36hfa6e2cd_0
qt                        5.9.6            vc14h62aca36_0
qtconsole                 4.4.0                    py36_0
scipy                     1.1.0            py36h672f292_0
seaborn                   0.9.0                    py36_0
send2trash                1.5.0                    py36_0
service_identity          17.0.0           py36h28b3542_0
setuptools                40.0.0                   py36_0
simplegeneric             0.8.1                    py36_2
sip                       4.19.8           py36h6538335_0
six                       1.11.0                   py36_1
sqlite                    3.24.0               h7602738_0
statsmodels               0.9.0            py36h452e1ab_0
terminado                 0.8.1                    py36_1
testpath                  0.3.1            py36h2698cfe_0
theano                    1.0.2            py36he980bc4_0
tornado                   5.1              py36hfa6e2cd_0
tqdm                      4.24.0                   py36_0
traitlets                 4.3.2            py36h096827d_0
twisted                   18.7.0           py36hfa6e2cd_1
vc                        14.1                 h0510ff6_3
vs2015_runtime            15.5.2                        3
wcwidth                   0.1.7            py36h3d5aa90_0
webencodings              0.5.1                    py36_1
wheel                     0.31.1                   py36_0
widgetsnbextension        3.4.0                    py36_0
wincertstore              0.2              py36h7fe50ca_0
winpty                    0.4.3                         4
zeromq                    4.2.5                he025d50_1
zlib                      1.2.11               h8395fce_2
zope                      1.0                      py36_0
zope.interface            4.5.0            py36hfa6e2cd_0

[twiecki]

Seems like a problem with tracing the nodes through Determinsitics.

[ColCarroll]

Funny! Should be able to take a look at it this week. Deterministics are hard to "see" when tracing the graph.

In the mean time, the function does return a graphviz graph, so you can add the connections manually. I'll post that snippet when I start looking if it is easy.

[ColCarroll]

well, this is funny for any number of reasons:

[junpenglao]

Well, it's not wrong about that link.

[twiecki]

Should be fixed by #3170. Reopen if not.

[twiecki]

Looking at it again I might be wrong. @ColCarroll?

[ColCarroll]

@JackCaster 's example is still a bug:

mu_real = 0.87
phi_real = 10

n_samples = 500

alpha_real = mu_real * phi_real
beta_real = (1-mu_real) * phi_real
Y = np.random.beta(a=alpha_real, b=beta_real, size=n_samples)

with pm.Model() as model:
    
    # Prior on 'mu'
    mu = pm.Beta('mu', alpha=1, beta=1)
    
    # Prior on 'phi'
    phi = pm.HalfNormal('phi', sd=25)
    
    # Parametrize in terms of a, b
    alpha = pm.Deterministic('alpha', mu*phi)
    beta = pm.Deterministic('beta', (1-mu)*phi)
    
    #alpha = pm.HalfNormal('alpha', sd=25)
    #beta = pm.HalfNormal('beta', sd=25)    
    
    # Likelihood
    likelihood = pm.Beta('likelihood', alpha=alpha, beta=beta, observed=Y)

pm.model_to_graphviz(model)

My example is a lot better than previously, but still draws extra lines.

with pm.Model() as model:
    w = pm.Normal('w')
    x = pm.Deterministic('x', w)
    y = pm.Deterministic('y', x)
    u = pm.Deterministic('u', y)
    z = pm.Normal('z', u)
pm.model_to_graphviz(model)