Fix input resolution for steps with dynamic artifact names (#3228)

* Fix input resolution for steps with dynamic artifact names

* Improve logic

* Linting

* Add test

* Fix variable access

* Really fix test

* Rename

src/zenml/models/v2/core/pipeline_run.py

@@ -550,6 +550,15 @@ def is_templatable(self) -> bool:
         """
         return self.get_metadata().is_templatable
 
+    @property
+    def step_substitutions(self) -> Dict[str, Dict[str, str]]:
+        """The `step_substitutions` property.
+
+        Returns:
+            the value of the property.
+        """
+        return self.get_metadata().step_substitutions
+
     @property
     def model_version(self) -> Optional[ModelVersionResponse]:
         """The `model_version` property.

src/zenml/orchestrators/input_utils.py

@@ -20,7 +20,7 @@
 from zenml.config.step_configurations import Step
 from zenml.enums import ArtifactSaveType, StepRunInputArtifactType
 from zenml.exceptions import InputResolutionError
-from zenml.utils import pagination_utils
+from zenml.utils import pagination_utils, string_utils
 
 if TYPE_CHECKING:
     from zenml.models import PipelineRunResponse
@@ -53,7 +53,8 @@ def resolve_step_inputs(
     current_run_steps = {
         run_step.name: run_step
         for run_step in pagination_utils.depaginate(
-            Client().list_run_steps, pipeline_run_id=pipeline_run.id
+            Client().list_run_steps,
+            pipeline_run_id=pipeline_run.id,
         )
     }
 
@@ -66,11 +67,23 @@ def resolve_step_inputs(
                 f"No step `{input_.step_name}` found in current run."
             )
 
+        # Try to get the substitutions from the pipeline run first, as we
+        # already have a hydrated version of that. In the unlikely case
+        # that the pipeline run is outdated, we fetch it from the step
+        # run instead which will costs us one hydration call.
+        substitutions = (
+            pipeline_run.step_substitutions.get(step_run.name)
+            or step_run.config.substitutions
+        )
+        output_name = string_utils.format_name_template(
+            input_.output_name, substitutions=substitutions
+        )
+
         try:
-            outputs = step_run.outputs[input_.output_name]
+            outputs = step_run.outputs[output_name]
         except KeyError:
             raise InputResolutionError(
-                f"No step output `{input_.output_name}` found for step "
+                f"No step output `{output_name}` found for step "
                 f"`{input_.step_name}`."
             )
 
@@ -83,12 +96,12 @@ def resolve_step_inputs(
             # This should never happen, there can only be a single regular step
             # output for a name
             raise InputResolutionError(
-                f"Too many step outputs for output `{input_.output_name}` of "
+                f"Too many step outputs for output `{output_name}` of "
                 f"step `{input_.step_name}`."
             )
         elif len(step_outputs) == 0:
             raise InputResolutionError(
-                f"No step output `{input_.output_name}` found for step "
+                f"No step output `{output_name}` found for step "
                 f"`{input_.step_name}`."
             )
 

src/zenml/orchestrators/step_run_utils.py

@@ -309,6 +309,9 @@ def create_cached_step_runs(
         for invocation_id in cache_candidates:
             visited_invocations.add(invocation_id)
 
+            # Make sure the request factory has the most up to date pipeline
+            # run to avoid hydration calls
+            request_factory.pipeline_run = pipeline_run
             try:
                 step_run_request = request_factory.create_request(
                     invocation_id

tests/integration/functional/steps/test_step_naming.py → tests/integration/functional/steps/test_dynamic_artifact_names.py

@@ -12,6 +12,7 @@
 #  or implied. See the License for the specific language governing
 #  permissions and limitations under the License.
 
+from contextlib import ExitStack as does_not_raise
 from typing import Callable, Tuple
 
 import pytest
@@ -122,6 +123,11 @@ def mixed_with_unannotated_returns() -> (
     )
 
 
+@step
+def step_with_string_input(input_: str) -> None:
+    pass
+
+
 @pytest.mark.parametrize(
     "step",
     [
@@ -362,3 +368,17 @@ def _inner(pass_to_step: str = ""):
     assert p2_step_subs["date"] == "step_level"
     assert p1_step_subs["funny_name"] == "pipeline_level"
     assert p2_step_subs["funny_name"] == "step_level"
+
+
+def test_dynamically_named_artifacts_in_downstream_steps(
+    clean_client: "Client",
+):
+    """Test that dynamically named artifacts can be used in downstream steps."""
+
+    @pipeline(enable_cache=False)
+    def _inner(ret: str):
+        artifact = dynamic_single_string_standard()
+        step_with_string_input(artifact)
+
+    with does_not_raise():
+        _inner("output_1")

tests/unit/conftest.py

@@ -11,6 +11,7 @@
 #  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
 #  or implied. See the License for the specific language governing
 #  permissions and limitations under the License.
+from collections import defaultdict
 from datetime import datetime
 from typing import Any, Callable, Dict, List, Optional
 from uuid import uuid4
@@ -416,6 +417,12 @@ def sample_pipeline_run(
     sample_workspace_model: WorkspaceResponse,
 ) -> PipelineRunResponse:
     """Return sample pipeline run view for testing purposes."""
+    now = datetime.utcnow()
+    substitutions = {
+        "date": now.strftime("%Y_%m_%d"),
+        "time": now.strftime("%H_%M_%S_%f"),
+    }
+
     return PipelineRunResponse(
         id=uuid4(),
         name="sample_run_name",
@@ -430,6 +437,7 @@ def sample_pipeline_run(
             workspace=sample_workspace_model,
             config=PipelineConfiguration(name="aria_pipeline"),
             is_templatable=False,
+            steps_substitutions=defaultdict(lambda: substitutions.copy()),
         ),
         resources=PipelineRunResponseResources(tags=[]),
     )
@@ -543,10 +551,15 @@ def f(
         spec = StepSpec.model_validate(
             {"source": "module.step_class", "upstream_steps": []}
         )
+        now = datetime.utcnow()
         config = StepConfiguration.model_validate(
             {
                 "name": step_name,
                 "outputs": outputs or {},
+                "substitutions": {
+                    "date": now.strftime("%Y_%m_%d"),
+                    "time": now.strftime("%H_%M_%S_%f"),
+                },
             }
         )
         return StepRunResponse(

tests/unit/orchestrators/test_input_utils.py

@@ -12,14 +12,13 @@
 #  or implied. See the License for the specific language governing
 #  permissions and limitations under the License.
 
-from uuid import uuid4
 
 import pytest
 
 from zenml.config.step_configurations import Step
 from zenml.enums import StepRunInputArtifactType
 from zenml.exceptions import InputResolutionError
-from zenml.models import Page, PipelineRunResponse
+from zenml.models import Page
 from zenml.models.v2.core.artifact_version import ArtifactVersionResponse
 from zenml.models.v2.core.step_run import StepRunInputResponse
 from zenml.orchestrators import input_utils
@@ -29,6 +28,7 @@ def test_input_resolution(
     mocker,
     sample_artifact_version_model: ArtifactVersionResponse,
     create_step_run,
+    sample_pipeline_run,
 ):
     """Tests that input resolution works if the correct models exist in the
     zen store."""
@@ -60,7 +60,7 @@ def test_input_resolution(
     )
 
     input_artifacts, parent_ids = input_utils.resolve_step_inputs(
-        step=step, pipeline_run=PipelineRunResponse(id=uuid4(), name="foo")
+        step=step, pipeline_run=sample_pipeline_run
     )
     assert input_artifacts == {
         "input_name": StepRunInputResponse(
@@ -71,7 +71,7 @@ def test_input_resolution(
     assert parent_ids == [step_run.id]
 
 
-def test_input_resolution_with_missing_step_run(mocker):
+def test_input_resolution_with_missing_step_run(mocker, sample_pipeline_run):
     """Tests that input resolution fails if the upstream step run is missing."""
     mocker.patch(
         "zenml.zen_stores.sql_zen_store.SqlZenStore.list_run_steps",
@@ -97,11 +97,13 @@ def test_input_resolution_with_missing_step_run(mocker):
 
     with pytest.raises(InputResolutionError):
         input_utils.resolve_step_inputs(
-            step=step, pipeline_run=PipelineRunResponse(id=uuid4(), name="foo")
+            step=step, pipeline_run=sample_pipeline_run
         )
 
 
-def test_input_resolution_with_missing_artifact(mocker, create_step_run):
+def test_input_resolution_with_missing_artifact(
+    mocker, create_step_run, sample_pipeline_run
+):
     """Tests that input resolution fails if the upstream step run output
     artifact is missing."""
     step_run = create_step_run(
@@ -132,12 +134,12 @@ def test_input_resolution_with_missing_artifact(mocker, create_step_run):
 
     with pytest.raises(InputResolutionError):
         input_utils.resolve_step_inputs(
-            step=step, pipeline_run=PipelineRunResponse(id=uuid4(), name="foo")
+            step=step, pipeline_run=sample_pipeline_run
         )
 
 
 def test_input_resolution_fetches_all_run_steps(
-    mocker, sample_artifact_version_model, create_step_run
+    mocker, sample_artifact_version_model, create_step_run, sample_pipeline_run
 ):
     """Tests that input resolution fetches all step runs of the pipeline run."""
     step_run = create_step_run(
@@ -178,7 +180,7 @@ def test_input_resolution_fetches_all_run_steps(
     )
 
     input_utils.resolve_step_inputs(
-        step=step, pipeline_run=PipelineRunResponse(id=uuid4(), name="foo")
+        step=step, pipeline_run=sample_pipeline_run
     )
 
     # `resolve_step_inputs(...)` depaginates the run steps so we fetch all