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

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

vpc_network_config = NetworkConfig(
    enable_network_isolation=False, security_group_ids=['sg-0363bd1722cdd35da'],
    subnets=['subnet-b649dfef', 'subnet-44c19a21'], encrypt_inter_container_traffic=True
)

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.c5.18xlarge',    # 'ml.t3.medium',  #'ml.c5.4xlarge',  'ml.c5.18xlarge',
    command=['python3'],
    base_job_name="ProcessingJob",
    network_config=vpc_network_config
)

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

run_id = '20241120_PATTERN_DETECTION'

CHUNK_COUNT=1
FOLDER = 'eimpara/pipelines/moments'

script1='/home/ec2-user/SageMaker/ml-collab/eimpara/Moments/ML_pipeline/manual_pattern_detection/regression_script.py'


logging.info(f"Running: {script1}")
for chunk_id in range(CHUNK_COUNT):
    script_processor.run(
            code=script1,
            arguments = ['--dry-run', 'False', 
                         '--run-id', str(run_id), 
                         '--folder', FOLDER,
                         '--chunk-count', str(CHUNK_COUNT), 
                         '--chunk-id', str(chunk_id),
                        ],
            inputs=[
                ProcessingInput(
                    source='s3://dev-cucumbers/eimpara/requirements-OLS.txt',
                    destination = '/opt/ml/processing/input/dependencies'
                )
            ],
            outputs=[
                ProcessingOutput(
                    source='/opt/ml/processing/output', 
                    destination=destination,
                    output_name="Master_output",
                ),
            ]
        )