fix(sagemaker-templates): Fix batch inference parameter resolution

The batch inference template was passing ParameterString objects directly
to job_arguments, which caused them to be serialized as literal JSON strings
instead of being resolved at runtime.

Changes:
- Use ProcessingInput with parameter as source for dynamic S3 paths
- Pass S3 URL parameter to script so it can extract filename
- Script constructs local path from filename
- Remove S3 download logic (ProcessingInput handles it)

This fixes the issue where InputDataUrl parameter was passed as
'{"Get": "Parameters.InputDataUrl"}' instead of the actual S3 path,
causing preprocessing jobs to fail with IndexError.

Resolves parameter resolution for batch inference pipelines.

Author: newman911

modules/sagemaker/sagemaker-templates/templates/batch_inference/seed_code/build_app/ml_pipelines/transformer/pipeline.py

@@ -144,7 +144,7 @@ def get_pipeline(
         inputs=[
             ProcessingInput(
                 source=input_data,
-                destination="/opt/ml/processing/input/data.csv",
+                destination="/opt/ml/processing/input",
             ),
         ],
         outputs=[
@@ -153,7 +153,7 @@ def get_pipeline(
         code="source_scripts/preprocessing.py",
         job_arguments=[
             "--input-data",
-            "/opt/ml/processing/input/data.csv",
+            input_data,
             "--do-train-test-split",
             "False",
         ],

modules/sagemaker/sagemaker-templates/templates/batch_inference/seed_code/build_app/source_scripts/preprocessing.py

@@ -62,23 +62,13 @@ def merge_two_dicts(x: Dict[str, Any], y: Dict[str, Any]) -> Dict[str, Any]:
     base_dir = "/opt/ml/processing"
     pathlib.Path(f"{base_dir}/data").mkdir(parents=True, exist_ok=True)
     input_data = args.input_data
-    logger.info("Input data path: %s", input_data)
+    logger.info("Input data S3 URL: %s", input_data)
 
-    # Check if input_data is an S3 URL or a local path
-    if input_data.startswith("s3://"):
-        # Download from S3
-        bucket = input_data.split("/")[2]
-        key = "/".join(input_data.split("/")[3:])
-        logger.info("Downloading data from bucket: %s, key: %s", bucket, key)
-        fn = f"{base_dir}/data/abalone-dataset.csv"
-        s3 = boto3.resource("s3")
-        s3.Bucket(bucket).download_file(key, fn)
-    else:
-        # Use local path (file already downloaded by ProcessingInput)
-        logger.info("Using local input data path")
-        fn = input_data
+    # Extract filename from S3 URL - file is already downloaded by ProcessingInput
+    filename = input_data.split('/')[-1]
+    fn = f"/opt/ml/processing/input/{filename}"
+    logger.info("Reading file: %s", fn)
 
-    logger.debug("Reading downloaded data.")
     df = pd.read_csv(
         fn,
         header=None,