JumpStart Uncompressed Format Causing AttributeError for JumpStart ID Local Container Mode

[gwang111]

Describe the bug
When building a JumpStart ID model with ModelBuilder, an AttributeError occurs when fetching the JumpStart resources from S3

To reproduce

model_builder = ModelBuilder(
    model="huggingface-llm-falcon-7b-bf16",
    schema_builder=SchemaBuilder(sample_input, sample_output),
    model_path="/path",
    mode=Mode.LOCAL_CONTAINER
)

builder = model_builder.build()

Expected behavior
The uncompressed model resources should be fetched from S3

Screenshots or logs

INFO:botocore.credentials:Found credentials from IAM Role: BaseNotebookInstanceEc2InstanceRole
sagemaker.config INFO - Not applying SDK defaults from location: /etc/xdg/sagemaker/config.yaml
sagemaker.config INFO - Not applying SDK defaults from location: /home/ec2-user/.config/sagemaker/config.yaml
INFO:botocore.credentials:Found credentials from IAM Role: BaseNotebookInstanceEc2InstanceRole
WARNING:sagemaker.jumpstart:Using model 'huggingface-llm-falcon-7b-bf16' with wildcard version identifier '*'. You can pin to version '2.0.0' for more stable results. Note that models may have different input/output signatures after a major version upgrade.
ModelBuilder: INFO:     JumpStart Model ID detected.
INFO:sagemaker.jumpstart:No instance type selected for inference hosting endpoint. Defaulting to ml.g5.2xlarge.
ModelBuilder: INFO:     JumpStart ID huggingface-llm-falcon-7b-bf16 is packaged with Image URI: 763104351884.dkr.ecr.us-east-1.amazonaws.com/huggingface-pytorch-tgi-inference:2.0.1-tgi1.1.0-gpu-py39-cu118-ubuntu20.04
ModelBuilder: INFO:     Building for TGI JumpStart Model ID...
ModelBuilder: WARNING:     75.98389809091147 percent of docker disk space at /var/lib/docker is used. Please consider freeing up disk space or increasing the EBS volume if you are on a SageMaker Notebook.
ModelBuilder: INFO:     Downloading JumpStart artifacts from S3...
{'S3DataSource': {'S3Uri': 's3://jumpstart-cache-prod-us-east-1/huggingface-llm/huggingface-llm-falcon-7b-bf16/artifacts/inference-prepack/v1.0.1/', 'S3DataType': 'S3Prefix', 'CompressionType': 'None'}}
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
File <timed exec>:1

File ~/anaconda3/envs/python3/lib/python3.10/site-packages/sagemaker/serve/builder/model_builder.py:579, in ModelBuilder.build(self, mode, role_arn, sagemaker_session)
    577 if isinstance(self.model, str):
    578     if self._is_jumpstart_model_id():
--> 579         return self._build_for_jumpstart()
    580     if self._is_djl():
    581         return self._build_for_djl()

File ~/anaconda3/envs/python3/lib/python3.10/site-packages/sagemaker/serve/builder/jumpstart_builder.py:261, in JumpStart._build_for_jumpstart(self)
    258     self.pysdk_model = pysdk_model
    259     self.image_uri = self.pysdk_model.image_uri
--> 261     self._build_for_tgi_jumpstart()
    262 else:
    263     raise ValueError(
    264         "JumpStart Model ID was not packaged with djl-inference or tgi-inference container."
    265     )

File ~/anaconda3/envs/python3/lib/python3.10/site-packages/sagemaker/serve/builder/jumpstart_builder.py:222, in JumpStart._build_for_tgi_jumpstart(self)
    220 if self.mode == Mode.LOCAL_CONTAINER:
    221     if not hasattr(self, "prepared_for_tgi"):
--> 222         self.prepared_for_tgi = prepare_tgi_js_resources(
    223             model_path=self.model_path,
    224             js_id=self.model,
    225             dependencies=self.dependencies,
    226             model_data=self.pysdk_model.model_data,
    227         )
    228     self._prepare_for_mode()
    229 elif self.mode == Mode.SAGEMAKER_ENDPOINT and hasattr(self, "prepared_for_tgi"):

File ~/anaconda3/envs/python3/lib/python3.10/site-packages/sagemaker/serve/model_server/tgi/prepare.py:74, in prepare_tgi_js_resources(model_path, js_id, shared_libs, dependencies, model_data)
     58 """Prepare serving when a JumpStart model id is given
     59 
     60 Args:
   (...)
     70 
     71 """
     72 model_path, code_dir = _create_dir_structure(model_path)
---> 74 return _copy_jumpstart_artifacts(model_data, js_id, code_dir)

File ~/anaconda3/envs/python3/lib/python3.10/site-packages/sagemaker/serve/model_server/tgi/prepare.py:22, in _copy_jumpstart_artifacts(model_data, js_id, code_dir)
     19 with _tmpdir(directory=str(code_dir)) as js_model_dir:
     20     # TODO: remove if block after 10/30 once everything is shifted to uncompressed
     21     print(model_data)
---> 22     if model_data.endswith("tar.gz"):
     23         subprocess.run(["aws", "s3", "cp", model_data, js_model_dir])
     25         logger.info("Uncompressing JumpStart artifacts for faster loading...")

AttributeError: 'dict' object has no attribute 'endswith'

System information
A description of your system. Please provide:

• SageMaker Python SDK version: 2.199
• Framework name (eg. PyTorch) or algorithm (eg. KMeans): TGI
• Framework version: Latest
• Python version: 3.10
• CPU or GPU: GPU
• Custom Docker image (Y/N): N

Additional context
Add any other context about the problem here.

[gwang111]

Fix Here: #4281