import sagemaker
import boto3
from sagemaker.processing import ScriptProcessor, ProcessingInput, ProcessingOutput, FrameworkProcessor
from sagemaker import get_execution_role

role = get_execution_role()
bucket = 'dev-cucumbers'
sagemaker_session = sagemaker.session.Session(default_bucket = bucket)

script_processor = ScriptProcessor(
    image_uri='683313688378.dkr.ecr.us-east-1.amazonaws.com/sagemaker-scikit-learn:0.20.0-cpu-py3',
    sagemaker_session=sagemaker_session,
    role=role,
    instance_count=1,
    instance_type='ml.c5.4xlarge',  #'ml.c5.9xlarge',
    command=['python3'],
    base_job_name="ProcessingJob"
)

destination = '/opt/ml/processing/input'

code='/home/ec2-user/SageMaker/ml-collab/eimpara/Moments/TikTok/LAGS_tiktok_metal.py'


script_processor.run(
    code=code,
    inputs=[
        ProcessingInput(
            source='s3://dev-cucumbers/eimpara/TikTok_analysis/', # also tried with empty string as the data source for Fourier_and_ARIMA.ipynb is SnowFlake and called in the jupyter notebook
            destination=destination
        ),
        ProcessingInput(
            source='s3://dev-cucumbers/eimpara/requirements.txt',
            destination = '/opt/ml/processing/input/dependencies'
        )
    ],
    outputs=[
        ProcessingOutput(
            source='/opt/ml/processing/output',  # also tried with empty string
            destination=destination,
            output_name="tiktok_genre_analysis",
        ),
    ]
)