In SageMaker, how to get the value from Http Header?

I have a machine learning classification model that was trained outside of SageMaker. The model is in Scikit-learn format. To run this model, the preprocessing step requires the binary content of a file and its metadata, such as the filename, author, and creation date to generate the feature vector. Since the payload is used to send the binary content, I utilize the Header to send additional information to SageMaker.

Referring to this post, I created a function called input_handler(data, context) in the inference.py file to retrieve the custom header value. However, this function was not invoked during the process. So I am not sure if the function input_handler works for SKLearnModel or not. Do we have other alternatives to get values from the header?

I used the class sagemaker.sklearn.SKLearnModel for deploying the model."

This is how I deployed the model

from sagemaker.sklearn import SKLearnModel
from sagemaker import get_execution_role


role = get_execution_role()
model_path = f's3://{bucket_name}/{model_name}.tar.gz'
entry_point = 'inference_ransomware.py'
dependencies = ...

model = SKLearnModel(
    model_data=model_path,
    role=role,
    entry_point=entry_point,
    dependencies=dependencies,
    framework_version='1.2-1'
)

predictor = model.deploy(
    initial_instance_count=1,
    instance_type='ml.m4.xlarge',
)

filename = ...

with open(filename, mode='rb') as file:
    byte_data = file.read()
    predictor.predict(byte_data, initial_args={'CustomAttributes': json.dumps({"filename": filename})})