{
    "q1": {
        "type": "blank", 
        "question": "\n<p>\nThe concepts beind the functional programming paradigm requires functions\nto be ____, meaning functions only rely on their given inputs to produce\nan output.  A benefit of using such ____ functions is the reduction of\n____, which occur when there are changes performed within a function's\noperation that are outside its scope.\n</p>\n\n<p>\nTo support this programming paradigm, Python supports ____ functions,\nwhich means you can pass functions as arguments to other functions.\nThree common higher-order functions in Python are:\n</p>\n\n\n<ul>\n\n  <li>____: This function takes in a function and applies it to each item\n  in an iterable to create a new iterable object.</li>\n  \n  <li>____: This function takes in a function and applies it to each item\n  in an iterable to create a new iterable object that contains only\n  elements that returned <tt>True</tt>.</li>\n  \n  <li>____: This function takes in a function and applies it to each item\n  in an iterable to accumulate or aggregate into a single value.</li>\n\n</ul>\n\n<p>\nTo create anonymous functions that we can pass to the higher-order\nfunctions above, we can use ____ expressions.\n</p>\n\n<p>\nBecause the first two higher-order functions above return iterables, we\ncan re-write them using ____, which are a more Pythonic way of making\nlists.\n</p>\n\n<p>\nRather than materializing or populating a whole list or iterable, we can\nproduce items one-by-one using a ____ function that contains the\n<b><tt>yield</tt></b> keyword.\n</p>\n"
    }, 
    "q2": {
        "type": "order", 
        "question": "\n\n<p>\nMapReduce is a programming model created by Google and is used for\nprocessing large data sets across a large cluster of commodity machines.\nInspired by the functional programming paradigm, a typical MapReduce\napplication goes through the following three phases:\n</p>\n\n<ol>\n\n  <li>Worker nodes apply the first function to each item in the local\n  data set and writes the intermediate output to a temporary\n  storage.</li>\n\n  <li>Worker nodes redistribute the intermediate output such that all the\n  data belonging to one key is located on the same worker node.</li>\n\n  <li>Worker nodes process each group of intermediate output data with\n  second function to produce final result.</li>\n\n</ol>\n\n<p>\nOrder the following items to match the descriptions of the phases above.\n</p>\n", 
        "responses": {
            "map": "Map", 
            "reduce": "Reduce", 
            "shuffle": "Shuffle"
        }
    }
}