import json
import os
from sys import exit as sysexit
import pandas

import sfn

# Map for event fields - {task type: detail field name}
task_detail_map = {
    'ExecutionFailed': 'executionFailedEventDetails',
    'MapStateEntered': 'stateEnteredEventDetails',
    'ChoiceStateEntered': 'stateEnteredEventDetails',
    'ChoiceStateExited': 'stateExitedEventDetails',
    'MapIterationSucceeded': 'mapIterationSucceededEventDetails',
    'TaskStateExited': 'stateExitedEventDetails',
    'LambdaFunctionStarted': None,
    'MapIterationStarted': 'mapIterationStartedEventDetails',
    'LambdaFunctionScheduled': 'lambdaFunctionScheduledEventDetails',
    'ExecutionStarted': 'executionStartedEventDetails',
    'LambdaFunctionFailed': 'lambdaFunctionFailedEventDetails',
    'LambdaFunctionSucceeded': 'lambdaFunctionSucceededEventDetails',
    'MapStateStarted': 'mapStateStartedEventDetails',
    'TaskStateEntered': 'stateEnteredEventDetails'
}

SFN_NAME = os.environ.get('SFN_NAME')
SFN_ARN = os.environ.get('SFN_ARN')
SFN_EXEC_NAME = os.environ.get('SFN_EXEC_NAME')
SFN_EXEC_ARN = \
    SFN_ARN.replace('stateMachine', 'execution') + ':' + SFN_EXEC_NAME

if not SFN_ARN:
    sysexit('You must export SFN_ARN in your environment with the SFN ARN.')

if not SFN_NAME:
    sysexit('You must export SFN_NAME in your environment with the SFN name.')

if not SFN_EXEC_NAME:
    sysexit('You must export SFN_EXEC_NAME in your environment with the SFN '
            'execution name.')


def process_exec_list(exec_list):
    """Prune a list of executions to the important information."""

    # Smooth out datatypes
    exec_dict = json.loads(
        json.dumps(exec_list, indent=4, sort_keys=True, default=str))

    # Creat csv
    csv = list()

    print('Processing rows.')
    for event in exec_dict:
        row = get_row_from_event(event)

        csv.append(row)

    return csv


def get_row_from_event(event):
    """Generate a CSV-shaped row from a single SFN event."""
    detail_key = task_detail_map.get(event['type'])

    error_detail = None
    cause_detail = None
    input_row = dict()  # Default no payload data
    name = None

    # Process input/output payload data if the event has a payload
    if detail_key:
        details = event[detail_key]

        # Load vars
        input_detail = json.loads(details.get('input', '{}'))
        output_detail = json.loads(details.get('output', '{}'))
        cause_detail = details.get('cause')
        error_detail = details.get('error')
        name = details.get('name')

        if input_detail:
            input_row = parse_input_field(input_detail)
        elif output_detail:
            input_row = parse_input_field(output_detail)

    main_row = {
        'exec_order_id': event['id'],
        'prev_exec_order_id': event['previousEventId'],
        'timestamp': event['timestamp'],
        'task_name': name,
        'type': event['type'],
        'error': error_detail,
        'cause': cause_detail
    }

    # Combine fields
    row = {
        **main_row,
        **input_row
    }

    return row


# Override this function to process different payload shapes
def parse_input_field(detail):
    """Parse input/output field and return a formatted dict of values."""

    input_block = dict(asset_field=None)

    # Assign vals
    input_block['product_detail'] = detail.get('product', {})
    input_block['correlation_id'] = detail.get('correlation_id')
    input_block['asset_type'] = detail.get('asset_type')
    input_block['item_index'] = detail.get('item_index')

    # Get correct asset field
    if input_block['asset_type'] == 'cover':
        input_block['asset_field'] = 'artwork'
    if input_block['asset_type'] == 'audio':
        input_block['asset_field'] = 'track'

    input_row = {
        'item_index': input_block['item_index'],
        'asset_type': input_block['asset_type'],
        'correlation_id': input_block['correlation_id'],
        'product_id': input_block['product_detail'].get('product_id'),
        'upc': input_block['product_detail'].get('upc'),
        'vendor_id': input_block['product_detail'].get('vendor_id'),
        'subaccount_id': input_block['product_detail'].get('subaccount_id'),
        'release_name': input_block['product_detail'].get('release_name'),
        'project_code': input_block['product_detail'].get('project_code'),
        'project_name': input_block['product_detail'].get('project_name'),
        'project_id': input_block['product_detail'].get('project_id'),
        'bucket': input_block['product_detail'].get(input_block['asset_field'],
                                                    {}).get('bucket') if
        input_block['asset_field'] else None,
        'key': input_block['product_detail'].get(input_block['asset_field'],
                                                 {}).get('key') if
        input_block['asset_field'] else None,
        'filename': input_block['product_detail'].get(
            input_block['asset_field'], {}).get('filename') if input_block[
            'asset_field'] else None,
        'ows_assets_filename': input_block['product_detail'].get(
            input_block['asset_field'], {}).get(
            'ows_assets_filename') if input_block['asset_field'] else None,
    }

    return input_row


if __name__ == '__main__':
    print(f'SFN Execution ARN: {SFN_EXEC_ARN}')
    print('Getting Execution History.... This will take some time.')
    exec_list = sfn.get_execution_log_by_arn(SFN_EXEC_ARN, max_res=200)

    print('Converting history to dict')
    output = process_exec_list(exec_list)

    # Convert to dataframe
    print('Converting to Pandas DataFrame')
    df = pandas.DataFrame(output)

    # Write to Excel
    print(f'Writing XLSX out to: {SFN_EXEC_NAME}.xlsx')
    df.to_excel(SFN_EXEC_ARN+'.xlsx')