CVS-192366-move-models

[pgladkows]

🛠 Summary

JIRA/Issue if applicable.
Describe the changes.

🧪 Checklist

• Unit tests added.
• The documentation updated.
• Change follows security best practices.
  ``

[Copilot]

Pull request overview

This PR refactors the functional-test pipeline/model scaffolding to reduce hard-coded model dependencies by removing several pipeline/custom-node helpers and switching some pipelines/MediaPipes to receive models via constructor arguments.

Changes:

• Removed unused custom node wrappers and multiple pipeline classes from the functional test suite.
• Updated several pipelines/MediaPipes to require explicit model(s) passed in, instead of constructing default models internally.
• Refactored MultiLevelPipeline and ImageClassificationMediaPipe to use injected model(s) rather than importing specific model classes.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
tests/functional/object_model/custom_node.py	Removes CustomNodeDemultiply and CustomNodeElastic1T helper classes.
tests/functional/constants/pipelines.py	Removes multiple pipeline helpers and updates pipeline/MediaPipe constructors to accept injected models; adjusts multi-level and image-classification pipeline wiring.

Suppressed comments (1)

tests/functional/constants/pipelines.py:882

• MultiLevelPipeline._create_nodes() reuses and mutates the same model instance (self.model) for every node (updates name and shapes). Because each Node holds a reference to the model, all nodes end up sharing the last mutated state, which breaks node naming/output naming and shape expectations.

            model = self.model
            model.name = f"{model.name}_{idx}"
            model.update_shapes(shape)
            model.set_input_shape_for_ovms(shape)
            model_nodes.append(Node(f"model_{idx}", model))

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.