Fix CI

.pre-commit-config.yaml

@@ -14,7 +14,7 @@ repos:
         exclude: ^requirements-dev\.txt$
     -   id: trailing-whitespace
 - repo: https://github.com/pre-commit/mirrors-mypy
-  rev: v0.941
+  rev: v0.942
   hooks:
     - id: mypy
       name: Run static type checks
@@ -42,11 +42,11 @@ repos:
     - id: pyupgrade
       args: [--py37-plus]
 - repo: https://github.com/psf/black
-  rev: 22.1.0
+  rev: 22.3.0
   hooks:
     - id: black
 - repo: https://github.com/PyCQA/pylint
-  rev: v2.12.2
+  rev: v2.13.2
   hooks:
     - id: pylint
       args: [--rcfile=.pylintrc]

pymc/backends/arviz.py

@@ -9,6 +9,7 @@
     Iterable,
     Mapping,
     Optional,
+    Sequence,
     Tuple,
     Union,
 )
@@ -178,7 +179,10 @@ def __init__(
                 " one of trace, prior, posterior_predictive or predictions."
             )
 
-        untyped_coords = {**self.model.coords, **(coords or {})}
+        # Make coord types more rigid
+        untyped_coords: Dict[str, Optional[Sequence[Any]]] = {**self.model.coords}
+        if coords:
+            untyped_coords.update(coords)
         self.coords = {
             cname: np.array(cvals) if isinstance(cvals, tuple) else cvals
             for cname, cvals in untyped_coords.items()
@@ -649,8 +653,8 @@ def predictions_to_inference_data(
     )
     if hasattr(idata_orig, "posterior"):
         assert idata_orig is not None
-        converter.nchains = idata_orig.posterior.dims["chain"]
-        converter.ndraws = idata_orig.posterior.dims["draw"]
+        converter.nchains = idata_orig["posterior"].dims["chain"]
+        converter.ndraws = idata_orig["posterior"].dims["draw"]
     else:
         aelem = next(iter(predictions.values()))
         converter.nchains, converter.ndraws = aelem.shape[:2]

pymc/backends/report.py

@@ -115,7 +115,7 @@ def _run_convergence_checks(self, idata: arviz.InferenceData, model):
             self._add_warnings([warn])
             return
 
-        if idata.posterior.sizes["chain"] == 1:
+        if idata["posterior"].sizes["chain"] == 1:
             msg = (
                 "Only one chain was sampled, this makes it impossible to "
                 "run some convergence checks"
@@ -124,7 +124,7 @@ def _run_convergence_checks(self, idata: arviz.InferenceData, model):
             self._add_warnings([warn])
             return
 
-        elif idata.posterior.sizes["chain"] < 4:
+        elif idata["posterior"].sizes["chain"] < 4:
             msg = (
                 "We recommend running at least 4 chains for robust computation of "
                 "convergence diagnostics"
@@ -140,7 +140,7 @@ def _run_convergence_checks(self, idata: arviz.InferenceData, model):
             if is_transformed_name(rv_name):
                 rv_name2 = get_untransformed_name(rv_name)
                 rv_name = rv_name2 if rv_name2 in valid_name else rv_name
-            if rv_name in idata.posterior:
+            if rv_name in idata["posterior"]:
                 varnames.append(rv_name)
 
         self._ess = ess = arviz.ess(idata, var_names=varnames)
@@ -158,7 +158,7 @@ def _run_convergence_checks(self, idata: arviz.InferenceData, model):
             warnings.append(warn)
 
         eff_min = min(val.min() for val in ess.values())
-        eff_per_chain = eff_min / idata.posterior.sizes["chain"]
+        eff_per_chain = eff_min / idata["posterior"].sizes["chain"]
         if eff_per_chain < 100:
             msg = (
                 "The effective sample size per chain is smaller than 100 for some parameters. "

pymc/sampling.py

@@ -614,6 +614,7 @@ def sample(
         f"({n_tune*n_chains:_d} + {n_draws*n_chains:_d} draws total) "
         f"took {mtrace.report.t_sampling:.0f} seconds."
     )
+    mtrace.report._log_summary()
 
     idata = None
     if compute_convergence_checks or return_inferencedata:
@@ -622,19 +623,18 @@ def sample(
             ikwargs.update(idata_kwargs)
         idata = pm.to_inference_data(mtrace, **ikwargs)
 
-    if compute_convergence_checks:
-        if draws - tune < 100:
-            warnings.warn(
-                "The number of samples is too small to check convergence reliably.", stacklevel=2
-            )
-        else:
-            mtrace.report._run_convergence_checks(idata, model)
-    mtrace.report._log_summary()
+        if compute_convergence_checks:
+            if draws - tune < 100:
+                warnings.warn(
+                    "The number of samples is too small to check convergence reliably.",
+                    stacklevel=2,
+                )
+            else:
+                mtrace.report._run_convergence_checks(idata, model)
 
-    if return_inferencedata:
-        return idata
-    else:
-        return mtrace
+        if return_inferencedata:
+            return idata
+    return mtrace
 
 
 def _check_start_shape(model, start: PointType):
@@ -1621,7 +1621,7 @@ def sample_posterior_predictive(
     _trace: Union[MultiTrace, PointList]
     nchain: int
     if isinstance(trace, InferenceData):
-        _trace = dataset_to_point_list(trace.posterior)
+        _trace = dataset_to_point_list(trace["posterior"])
         nchain, len_trace = chains_and_samples(trace)
     elif isinstance(trace, xarray.Dataset):
         _trace = dataset_to_point_list(trace)
@@ -1704,7 +1704,7 @@ def sample_posterior_predictive(
 
     if not vars_to_sample:
         if return_inferencedata and not extend_inferencedata:
-            return None
+            return InferenceData()
         elif return_inferencedata and extend_inferencedata:
             return trace
         return {}

pymc/util.py

@@ -248,7 +248,7 @@ def chains_and_samples(data: Union[xarray.Dataset, arviz.InferenceData]) -> Tupl
     if isinstance(data, xarray.Dataset):
         dataset = data
     elif isinstance(data, arviz.InferenceData):
-        dataset = data.posterior
+        dataset = data["posterior"]
     else:
         raise ValueError(
             "Argument must be xarray Dataset or arviz InferenceData. Got %s",