In this assignment you will implement and test three classes for a Spreadsheet application (no GUI at this stage). You will also have to make a design decision, based on performance, and draw the class diagram. The classes are:
Your design/tests/implementation must use exactly the following externally visible, public API (API.py). This implies that all other attributes and methods used in the design must be private (internal to the class). Note: private attributes/methods names start with __ in Python. API documentation in HTML can be automatically generated with pydoc -w. For the assignment, this leads to api.html.
'''
### ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ###
CellData, CellCoordinate, SpreadsheetData classes API.
Hans Vangheluwe & Jean-Sébastien Bolduc
Modelling, Simulation and Design Lab
School of Computer Science
McGill University
last modified: 01/21/02
### ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ###
'''
# Types of objects that can be stored in a CellData object
dataTypes = (types.IntType,)
### ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ###
### class CellData
class CellData:
"""
Encapsulated Integer data
"""
def __init__(self, value = 0):
''' value:Integer -> -- CellData constructor.
A TypeError is raised on bad argument.
'''
def getValue(self):
''' -> :Integer'''
def setValue(self, value = 0):
''' value:Integer ->
A TypeError is raised on bad argument.
'''
def __str__(self):
''' -> :String
Return the string representation of the object value (Integer).
'''
### ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ###
### class CellCoordinate
class CellCoordinate:
"""
Encapsulated coordinates of cells in a spreadsheet.
Note: might in the future want to implement "coordinate arithmetic"
by means of __add__ etc.
"""
def __init__(self, row = 1, column = 1):
''' row:PositiveInteger, column:PositiveInteger ->
CellCoordinate constructor.
A KeyError is raised on bad arguments.
'''
def getRow(self):
''' -> :Integer'''
def getColumn(self):
''' -> :Integer'''
def setRow(self, row = 1):
'''row:PositiveInteger ->
A KeyError is raised on bad argument.
'''
def setColumn(self, column = 1):
'''column:PositiveInteger ->
A KeyError is raised on bad argument.
'''
def __str__(self):
''' -> :String
Return the string representation of the cell coordinate.
Example: with row=2, column=88, __str__ will return <CellCoordinate:2,88>
'''
### ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ###
### class SpreadsheetData
class SpreadsheetData:
"""
Encapsulates a dynamically sized spreadsheet structure
containing CellData data and indexed by CellCoordinate coordinates.
"""
def __init__(self):
''' __init__() -> -- SpreadsheetData constructor.'''
def __setitem__(self, coord, data):
'''coord:CellCoordinate, data:CellData ->
Update the content of cell indexed by ``coord'' with ``data''.
A KeyError is raised on bad coordinate,
A TypeError is raised on bad value.
Example use: sd[CellCoordinate(3,4)] = CellData(33)
'''
def __getitem__(self, coord):
'''coord:CellCoordinate -> :CellData | None
Return the content of a cell indexed by ``coord''
(return None if the cell is empty).
A KeyError is raised on bad coordinate.
Example use: sd[CellCoordinate(3,4)]
'''
def __delitem__(self, coord):
'''coord:CellCoordinate ->
Empty the cell indexed by ``coord''.
A KeyError is raised on bad coordinate.
Example use: del sd[CellCoordinate(3,4)]
'''
def getLU(self):
''' -> :CellCoordinate | None
Return a CellCoordinate containing the
Left-most non-empty column, and Upper-most non-emtpy row.
Return None in case of an empty spreadsheet
'''
def getRB(self):
''' -> :CellCoordinate | None
Return a CellCoordinate containing the Right-most non-empty column,
and Bottom-most non-emtpy row.
Return None in case of an empty spreadsheet
'''
def __str__(self):
''' -> :String
Return the string representation of the SpreadSheet.
This looks like a table of values with spaces for empty cells.
The row and column indexes are also shown.
'''
Use ``dia'' to draw the class diagram. Show all attributes and methods of the classes. produce a GIF or JPEG image. Both dia file and image file must be submitted.
For each class, write a unittest test script. Each script shall test the class for
Use PyUnit (aka unittest.py) to write your scripts.
To test the test scripts, make a ``dummy'' implementation of all classes (just implement all class methods with pass) and run all tests. They should all fail.
Implement the classes CellData and CellCoordinate. Both should pass their respective tests.
Implement the SpreadsheetData class, using Python lists (i.e., an array is a list of lists) as an internal data structure. Empty cells are represented by None list entries. Internally, cells will be indexed like
self.__data[row][column]
This prototype should pass all tests.
Implement the SpreadsheetData class, this time using a Python dictionary as an internal data structure. The dictionary keys should be tuples containing the row and column. Internally, cells will be indexed like
self.__data[ (row, column) ]
This prototype should pass all tests.
Write a script (you do not have to use PyUnit) to evaluate the performance of both prototypes 1 and 2. For each prototype, we expect the following experiments:
Timing must include the time to instantiate the SpreadsheetData class. Perform experiments with different spreadsheet sizes n, and plot the results (time as a function of n). This should result in 4 plots (Full and Sparse combined with Prototypes 1 and 2). You must submit 4 GIF or JPEG files. For each plot, try a sufficient number of n values to obtain a curve. nmax for each plot should be determined by the additional requirement that response time should not be more than 15 seconds. nmax will obviously depend on the power and load of the machine you run this test on. Thus, to make meaningful conclusions, all performance tests must be run on the same machine.
Comment on the results: