Skip to content

Check for observed variables in the trace #7641

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Jan 20, 2025
Merged

Check for observed variables in the trace #7641

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
a592208
Check for observed variables in the trace as well as the model
zaxtax Jan 12, 2025
e33e517
Bugfix
zaxtax Jan 12, 2025
e895a5c
Update tests
zaxtax Jan 19, 2025
d813da1
Add logic to handle conditional nodes for observed variables
zaxtax Jan 19, 2025
2fcf395
Remove redundant test
zaxtax Jan 20, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 13 additions & 3 deletions pymc/sampling/forward.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -345,10 +345,13 @@ def draw(
return [np.stack(v) for v in drawn_values]


def observed_dependent_deterministics(model: Model):
def observed_dependent_deterministics(model: Model, extra_observeds=None):
"""Find deterministics that depend directly on observed variables."""
if extra_observeds is None:
extra_observeds = []

deterministics = model.deterministics
observed_rvs = set(model.observed_RVs)
observed_rvs = set(model.observed_RVs + extra_observeds)
blockers = model.basic_RVs
return [
deterministic
Expand Down Expand Up @@ -767,13 +770,15 @@ def sample_posterior_predictive(
if "coords" not in idata_kwargs:
idata_kwargs["coords"] = {}
idata: InferenceData | None = None
observed_data = None
stacked_dims = None
if isinstance(trace, InferenceData):
_constant_data = getattr(trace, "constant_data", None)
if _constant_data is not None:
trace_coords.update({str(k): v.data for k, v in _constant_data.coords.items()})
constant_data.update({str(k): v.data for k, v in _constant_data.items()})
idata = trace
observed_data = trace.get("observed_data", None)
trace = trace["posterior"]
if isinstance(trace, xarray.Dataset):
trace_coords.update({str(k): v.data for k, v in trace.coords.items()})
Expand Down Expand Up @@ -816,7 +821,12 @@ def sample_posterior_predictive(
if var_names is not None:
vars_ = [model[x] for x in var_names]
else:
vars_ = model.observed_RVs + observed_dependent_deterministics(model)
observed_vars = model.observed_RVs
if observed_data is not None:
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

BTW, the the observed_dependent_deterministics above is not going to work if these variables are not observed in the model.

That happens with auto-imputation models, which I assume the as_model wrapper won't handle correctly either because the models are different depending on whether you pass data or not.

Just something to keep in mind

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree with this. You'll have to adapt observed_dependent_deterministics to also accept a list of extra variables that will depend on your observed_data

observed_vars += [
model[x] for x in observed_data if x in model and x not in observed_vars
]
vars_ = observed_vars + observed_dependent_deterministics(model, observed_vars)

vars_to_sample = list(get_default_varnames(vars_, include_transformed=False))

Expand Down
18 changes: 18 additions & 0 deletions tests/sampling/test_forward.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -540,6 +540,24 @@ def test_normal_scalar_idata(self):
ppc = pm.sample_posterior_predictive(idata, return_inferencedata=False)
assert ppc["a"].shape == (nchains, ndraws)

def test_external_trace_det(self):
with pm.Model() as model:
mu = pm.Normal("mu", 0.0, 1.0)
a = pm.Normal("a", mu=mu, sigma=1, observed=0.0)
b = pm.Deterministic("b", a + 1)
trace = pm.sample(tune=50, draws=50, chains=1, compute_convergence_checks=False)

# test that trace is used in ppc
with pm.Model() as model_ppc:
mu = pm.Normal("mu", 0.0, 1.0)
a = pm.Normal("a", mu=mu, sigma=1)
c = pm.Deterministic("c", a + 1)

ppc = pm.sample_posterior_predictive(
trace=trace, model=model_ppc, return_inferencedata=False
)
assert list(ppc.keys()) == ["a", "c"]

def test_normal_vector(self):
with pm.Model() as model:
mu = pm.Normal("mu", 0.0, 1.0)
Expand Down
Loading