The input is from the datasets and output should be an image file representing the flow diagram. Ie, it should represent the flow as well as relationships between the datasets. Should use merge, groupby for forming new tables and show the flow for the new tables as well in the diagram.
"the input is a python program that reads the dataset. the dataframe do some transformations and create new [login to view URL] python program (.py) file will be the input to your program and you parse it to create the flow diagram."