Open
Description
The prior, curve, plot workflow fails with scalar dims and extra dims specified on the LinearTrend component.
from pymc_marketing.mmm import LinearTrend
trend = LinearTrend(n_changepoints=4, dims=("geo", "product"))
# This is fine
coords = dict(geo=["A", "B"], product=["X", "Y", "Z"])
prior = trend.sample_prior(coords=coords)
# This fails...
curve = trend.sample_curve(prior)
trend.plot_curve(curve)Here is the traceback:
Trackback
---------------------------------------------------------------
ValueError Traceback (most recent call last)
Cell In[7], line 2
1 # This fails...
----> 2 curve = trend.sample_curve(prior)
File ~/GitHub/pymc-eco/pymc-marketing/pymc_marketing/mmm/linear_trend.py:400, in LinearTrend.sample_curve(self, parameters, max_value)
398 with pm.Model(coords=coords):
399 name = "trend"
--> 400 pm.Deterministic(
401 name,
402 self.apply(t),
403 dims=("t", *cast(Dims, self.dims)),
404 )
406 return pm.sample_posterior_predictive(
407 parameters,
408 var_names=[name],
409 ).posterior_predictive[name]
File ~/mamba/envs/pymc-marketing-dev/lib/python3.10/site-packages/pymc/model/core.py:2263, in Deterministic(name, var, model, dims)
2261 var = var.copy(model.name_for(name))
2262 model.deterministics.append(var)
-> 2263 model.add_named_variable(var, dims)
2265 from pymc.printing import str_for_potential_or_deterministic
2267 var.str_repr = types.MethodType(
2268 functools.partial(str_for_potential_or_deterministic, dist_name="Deterministic"), var
2269 )
File ~/mamba/envs/pymc-marketing-dev/lib/python3.10/site-packages/pymc/model/core.py:1476, in Model.add_named_variable(self, var, dims)
1474 for dim in dims:
1475 if dim not in self.coords and dim is not None:
-> 1476 raise ValueError(f"Dimension {dim} is not specified in `coords`.")
1477 if any(var.name == dim for dim in dims if dim is not None):
1478 raise ValueError(f"Variable `{var.name}` has the same name as its dimension label.")
ValueError: Dimension geo is not specified in `coords`.