chore : remove pyproject.toml, add library namespace

__init__.py

@@ -3,9 +3,9 @@
 A collection of reusable components and pipelines for Kubeflow Pipelines.
 
 Usage:
-    from kfp_components import components, pipelines
-    from kfp_components.components import training
-    from kfp_components.pipelines import evaluation
+    from library import components, pipelines
+    from library.components import training
+    from library.pipelines import evaluation
 """
 
 # Import submodules to enable the convenient import patterns shown above
@@ -17,5 +17,5 @@
     # Fallback to absolute imports (works during testing with sys.path modification)
     import components  # noqa: F401
     import pipelines  # noqa: F401
-
-__all__ = ["components", "pipelines"]
+#since the 
+__all__ = ["pipelines", "components"]

library/__init__.py

@@ -0,0 +1 @@
+"""Toplevel package for reusable library components and pipelines."""

library/components/__init__.py

@@ -0,0 +1,5 @@
+"""Convenience imports for the reusable component categories in ``library``."""
+
+from . import data_processing, deployment, evaluation, training
+
+__all__ = ["data_processing", "deployment", "evaluation", "training"]

library/components/data_processing/__init__.py

@@ -0,0 +1,4 @@
+"""Data processing components exposed under ``library.components.data_processing``."""
+
+# Auto-generated imports will be added here by scripts/update_init_imports.py
+# Components will be imported dynamically based on subdirectories

components/data_processing/sdg/component.py → library/components/data_processing/sdg/component.py

@@ -4,7 +4,6 @@
 within Kubeflow Pipelines.
 """
 
-import kfp.compiler
 from kfp import dsl
 
 
@@ -262,7 +261,9 @@ def sdg(
 
 
 if __name__ == "__main__":
-    kfp.compiler.Compiler().compile(
+    from kfp import compiler
+
+    compiler.Compiler().compile(
         sdg,
         package_path=__file__.replace(".py", "_component.yaml"),
     )

components/data_processing/yoda_data_processor/README.md → library/components/data_processing/yoda_data_processor/README.md

@@ -20,26 +20,7 @@ Downloads the yoda_sentences dataset from HuggingFace, renames columns to match
 
 ## Metadata 🗂️
 
-- **Name**: yoda_data_processor
-- **Stability**: alpha
-- **Dependencies**:
-  - Kubeflow:
-    - Name: Pipelines, Version: >=2.15.2
-  - External Services:
-    - Name: HuggingFace Datasets, Version: >=4.4.2
-- **Tags**:
-  - data_processing
-  - dataset_preparation
-  - text_processing
-  - yoda_speak
-  - translation
-- **Last Verified**: 2025-12-19 11:30:16+00:00
-- **Owners**:
-  - Approvers:
-    - mprahl
-    - nsingla
-  - Reviewers:
-    - HumairAK
+See [metadata.yaml](metadata.yaml) for the component's tags, dependencies, owners, and last verification date.
 
 ## Additional Resources 📚
 

components/data_processing/yoda_data_processor/component.py → library/components/data_processing/yoda_data_processor/component.py

@@ -1,7 +1,13 @@
-import kfp.compiler
 from kfp import dsl
 
 
+def add_yoda_prefix(example: dict[str, str]) -> dict[str, str]:
+    """Prefix a prompt with the Yoda translation instruction."""
+    updated_example = dict(example)
+    updated_example["prompt"] = "Translate the following to Yoda speak: " + updated_example["prompt"]
+    return updated_example
+
+
 @dsl.component(
     packages_to_install=["datasets"],
 )
@@ -35,10 +41,6 @@ def prepare_yoda_dataset(
     # Add prefix to prompts
     print("Adding Yoda speak prefix to prompts")
 
-    def add_yoda_prefix(example):
-        example["prompt"] = "Translate the following to Yoda speak: " + example["prompt"]
-        return example
-
     dataset = dataset.map(add_yoda_prefix)
 
     # Split the dataset into train and eval sets
@@ -63,7 +65,9 @@ def add_yoda_prefix(example):
 
 
 if __name__ == "__main__":
-    kfp.compiler.Compiler().compile(
+    from kfp import compiler
+
+    compiler.Compiler().compile(
         prepare_yoda_dataset,
         package_path=__file__.replace(".py", "_component.yaml"),
     )

library/components/data_processing/yoda_data_processor/tests/support.py

@@ -0,0 +1,108 @@
+"""Shared test helpers for yoda_data_processor tests."""
+
+import json
+import random
+from pathlib import Path
+
+
+class InMemoryDataset:
+    """A tiny dataset double that preserves transformation behavior."""
+
+    def __init__(self, rows: list[dict[str, str]]):
+        """Initialize the dataset with a copy of the provided rows."""
+        self.rows = [dict(row) for row in rows]
+
+    def __len__(self) -> int:
+        """Return the number of rows in the dataset."""
+        return len(self.rows)
+
+    def rename_column(self, old_name: str, new_name: str) -> "InMemoryDataset":
+        """Return a dataset with one column renamed."""
+        return InMemoryDataset(
+            [
+                {
+                    (new_name if key == old_name else key): value
+                    for key, value in row.items()
+                }
+                for row in self.rows
+            ]
+        )
+
+    def remove_columns(self, columns: list[str]) -> "InMemoryDataset":
+        """Return a dataset without the requested columns."""
+        columns_to_remove = set(columns)
+        return InMemoryDataset(
+            [
+                {
+                    key: value
+                    for key, value in row.items()
+                    if key not in columns_to_remove
+                }
+                for row in self.rows
+            ]
+        )
+
+    def map(self, transform) -> "InMemoryDataset":
+        """Return a dataset with the transform applied to each row."""
+        return InMemoryDataset([transform(dict(row)) for row in self.rows])
+
+    def train_test_split(self, test_size: float, seed: int) -> dict[str, "InMemoryDataset"]:
+        """Split the dataset deterministically into train and test subsets."""
+        shuffled_rows = [dict(row) for row in self.rows]
+        random.Random(seed).shuffle(shuffled_rows)
+
+        test_count = min(len(shuffled_rows), max(1, int(round(len(shuffled_rows) * test_size))))
+        split_index = len(shuffled_rows) - test_count
+        return {
+            "train": InMemoryDataset(shuffled_rows[:split_index]),
+            "test": InMemoryDataset(shuffled_rows[split_index:]),
+        }
+
+    def save_to_disk(self, path: str) -> None:
+        """Persist rows in a simple JSONL file within the output directory."""
+        output_dir = Path(path)
+        output_dir.mkdir(parents=True, exist_ok=True)
+        with output_dir.joinpath("data.jsonl").open("w", encoding="utf-8") as handle:
+            for row in self.rows:
+                handle.write(json.dumps(row) + "\n")
+
+
+def sample_rows() -> list[dict[str, str]]:
+    """Return a representative input dataset for Yoda processor tests."""
+    return [
+        {
+            "sentence": "Train yourself to let go of everything you fear to lose.",
+            "translation_extra": "Let go of everything you fear to lose, train yourself to.",
+            "translation": "unused-1",
+        },
+        {
+            "sentence": "Do or do not. There is no try.",
+            "translation_extra": "Do or do not. Try, there is not.",
+            "translation": "unused-2",
+        },
+        {
+            "sentence": "Named must your fear be before banish it you can.",
+            "translation_extra": "Before banish it you can, named must your fear be.",
+            "translation": "unused-3",
+        },
+        {
+            "sentence": "Wars not make one great.",
+            "translation_extra": "Great, wars make one not.",
+            "translation": "unused-4",
+        },
+        {
+            "sentence": "Pass on what you have learned.",
+            "translation_extra": "What you have learned, pass on.",
+            "translation": "unused-5",
+        },
+    ]
+
+
+def read_saved_rows(path: Path) -> list[dict[str, str]]:
+    """Load the saved JSONL rows from a component output directory."""
+    data_file = path / "data.jsonl"
+    return [
+        json.loads(line)
+        for line in data_file.read_text(encoding="utf-8").splitlines()
+        if line
+    ]