Read structured dataset from a folder

plugins/flytekit-huggingface/flytekitplugins/huggingface/sd_transformers.py

@@ -1,3 +1,4 @@
+import os
 import typing
 
 import datasets
@@ -59,12 +60,11 @@ def decode(
     ) -> datasets.Dataset:
         local_dir = ctx.file_access.get_random_local_directory()
         ctx.file_access.get_data(flyte_value.uri, local_dir, is_multipart=True)
-        path = f"{local_dir}/00000"
-
+        files = [item.path for item in os.scandir(local_dir)]
         if current_task_metadata.structured_dataset_type and current_task_metadata.structured_dataset_type.columns:
             columns = [c.name for c in current_task_metadata.structured_dataset_type.columns]
-            return datasets.Dataset.from_parquet(path, columns=columns)
-        return datasets.Dataset.from_parquet(path)
+            return datasets.Dataset.from_parquet(files, columns=columns)
+        return datasets.Dataset.from_parquet(files)
 
 
 StructuredDatasetTransformerEngine.register(HuggingFaceDatasetToParquetEncodingHandler())

plugins/flytekit-huggingface/setup.py

@@ -38,4 +38,5 @@
         "Topic :: Software Development :: Libraries",
         "Topic :: Software Development :: Libraries :: Python Modules",
     ],
+    entry_points={"flytekit.plugins": [f"{PLUGIN_NAME}=flytekitplugins.{PLUGIN_NAME}"]},
 )

plugins/flytekit-huggingface/tests/test_huggingface_plugin_sd.py

@@ -68,3 +68,15 @@ def test_datasets_renderer():
     df = pd.DataFrame({"col1": [1, 3, 2], "col2": list("abc")})
     dataset = datasets.Dataset.from_pandas(df)
     assert HuggingFaceDatasetRenderer().to_html(dataset) == str(dataset).replace("\n", "<br>")
+
+
+def test_parquet_to_datasets():
+    df = pd.DataFrame({"name": ["Alice"], "age": [10]})
+
+    @task
+    def create_sd() -> StructuredDataset:
+        return StructuredDataset(dataframe=df)
+
+    sd = create_sd()
+    dataset = sd.open(datasets.Dataset).all()
+    assert dataset.data == datasets.Dataset.from_pandas(df).data

plugins/flytekit-polars/flytekitplugins/polars/sd_transformers.py

@@ -64,11 +64,10 @@ def decode(
     ) -> pl.DataFrame:
         local_dir = ctx.file_access.get_random_local_directory()
         ctx.file_access.get_data(flyte_value.uri, local_dir, is_multipart=True)
-        path = f"{local_dir}/00000"
         if current_task_metadata.structured_dataset_type and current_task_metadata.structured_dataset_type.columns:
             columns = [c.name for c in current_task_metadata.structured_dataset_type.columns]
-            return pl.read_parquet(path, columns=columns)
-        return pl.read_parquet(path)
+            return pl.read_parquet(local_dir, columns=columns)
+        return pl.read_parquet(local_dir, use_pyarrow=True)
 
 
 StructuredDatasetTransformerEngine.register(PolarsDataFrameToParquetEncodingHandler())

plugins/flytekit-polars/tests/test_polars_plugin_sd.py

@@ -66,3 +66,16 @@ def test_polars_renderer():
     assert PolarsDataFrameRenderer().to_html(df) == pd.DataFrame(
         df.describe().transpose(), columns=df.describe().columns
     ).to_html(index=False)
+
+
+def test_parquet_to_polars():
+    data = {"name": ["Alice"], "age": [5]}
+
+    @task
+    def create_sd() -> StructuredDataset:
+        df = pd.DataFrame(data=data)
+        return StructuredDataset(dataframe=df)
+
+    sd = create_sd()
+    polars_df = sd.open(pl.DataFrame).all()
+    assert polars_df == pl.DataFrame(data)

plugins/flytekit-spark/flytekitplugins/spark/sd_transformers.py

@@ -1,6 +1,7 @@
 import typing
 
 import pandas as pd
+import pyspark as _pyspark
 from pyspark.sql.dataframe import DataFrame
 
 from flytekit import FlyteContext
@@ -38,7 +39,10 @@ def encode(
     ) -> literals.StructuredDataset:
         path = typing.cast(str, structured_dataset.uri) or ctx.file_access.get_random_remote_directory()
         df = typing.cast(DataFrame, structured_dataset.dataframe)
-        df.write.mode("overwrite").parquet(path)
+        ss = _pyspark.sql.SparkSession.builder.getOrCreate()
+        # Avoid generating SUCCESS files
+        ss.conf.set("mapreduce.fileoutputcommitter.marksuccessfuljobs", "false")
+        df.write.mode("overwrite").parquet(path=path)
         return literals.StructuredDataset(uri=path, metadata=StructuredDatasetMetadata(structured_dataset_type))
 
 

plugins/flytekit-spark/setup.py

@@ -34,4 +34,5 @@
         "Topic :: Software Development :: Libraries :: Python Modules",
     ],
     scripts=["scripts/flytekit_install_spark3.sh"],
+    entry_points={"flytekit.plugins": [f"{PLUGIN_NAME}=flytekitplugins.{PLUGIN_NAME}"]},
 )