mockr

mockr is a Python library for writing MapReduce jobs in an educational setting. It is intended to be used as a conceptual teaching tool.

mockr provides an interface for defining and running MapReduce locally. Simply define your map and reduce functions, input your data and call the run function. Everything is run sequentially and locally.

Note

mockr is only compatible with Python 3.

Installation

Install via pip:

$ pip install mockr

Documentation

Examples and specific information on classes and methods are below.

Job Types

Streaming Jobs

StreamingJob class expects the input to be a string. Newline (“n”) characters delimit “chunks” of data and each line/chunk is sent to a separate map worker.

Pandas Jobs

PandasJob class expects input to be a Pandas DataFrame. The rows of the data frame are equally divided into chunks and each chunk is sent to a separate map worker

Python Jobs

PythonJob class expects input to be a Collections.abc.Sequence type object e.g. Python List. Python Jobs provide two exection methods:

  • the sequence is divided into chunks and each chunk is sent to a separate map worker
  • each item in the list is individually sent to a dedicated map worker

Example Usage

The below example demonstrates counting the number of words in a corpus using MapReduce and the MockStreamingJob class.

More examples here.

examples/streaming_str_ex.py
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
import re
from mockr import run_stream_job

WORD_RE = re.compile(r"[\w']+")

def map_fn(chunk):
    # yield each word in the line
    for word in WORD_RE.findall(chunk):
        yield (word.lower(), 1)

def reduce_fn(key, values):
    yield (key, sum(values))

input_str = "Hello!\nThis is a sample string.\nIt is very simple.\nGoodbye!"

results = run_stream_job(input_str, map_fn, reduce_fn)

print(results)

Output:

[('hello', 1), ('this', 1), ('is', 2), ('a', 1), ('sample', 1), ('string', 1), ('it', 1), ('very', 1), ('simple', 1), ('goodbye', 1)]

Indices and tables