IDAPICourseworkLibrary.py

#! /usr/bin/env python
# -*- coding: utf-8 -*-
import Image
import glob, os
import numpy
# Function to read in data from a file in the format defined by Duncan
def ReadFile(filename):
    f = open(filename);
    # the first line contains the number of variables. the function int converts a
    # string to an integer
    noVariables = int(f.readline())
    # the second line contains the number of root nodes
    noRoots = int(f.readline());
    # the third line contains the number of states of each variable
    # this command extracts a list of integers. The split method breaks the line into a list of substrings
    # The map function applies a function (int) to a list (the substrings) to produce a list of integers.
    noStates = map(int,((f.readline()).split()))
    # the fourth line contains a single integer the number of data points
    noDataPoints = int(f.readline())
    # all the subsequent lines of the file are data points. Each line is extracted as a list of integers which is 
    # appended to the list datain.
    datain = []
    for x in range(noDataPoints):
        datain.append(map(int,((f.readline()).split())))
    f.close()
    return [noVariables, noRoots, noStates, noDataPoints, datain]
# Function to write an array to a results file
# the array is assumed to be either of proababilities of dependencies
def AppendArray(filename, anArray):
    f = open(filename, 'a')
    for row in range(anArray.shape[0]):
        for col in range(anArray.shape[1]):
            f.write( '%6.3f ' % (anArray[row,col]))
        f.write('\n')
    f.write('\n\n')
    f.close()
#Function to write a list to a results file
def AppendList(filename, aList):
    f = open(filename, 'a')
    for row in range(aList.shape[0]):
        f.write( '%6.3f ' % (aList[row]))
    f.write('\n\n')
    f.close()
#Function to write a string to a results file
def AppendString(filename, aString):
    f = open(filename, 'a')
    f.write('%s\n' % (aString))
    f.close()
#
# Image handline functions
# 
# These functions turn images into data sets and vice versa
#
# Function to turn a principal component into an image and save it. The assumed resolution os 92 by 112 pixels.
# The component is a one dimensional representation of an image with each row concatinated
def SaveEigenface(component,filename):
    theMax = max(component)
    theMin = min(component)
    scale = 255.0/(theMax-theMin)
    eigenfaceImage = map(int,(component - theMin) * scale)
    im = Image.new('L',(92,112))
    for y in range(im.size[1]):
        for x in range(im.size[0]):
            im.putpixel((x,y),eigenfaceImage[x+92*y])
    im.save(filename)
#
# Function to convert images into a data format equivalent to the above format where each row of an array is 
# one image with rows concatinated into a single vector.
# The images for this project are assumed to be all of resolution 92 by 112 pixels and are taken from
# the current directory in .pgm format
def ReadImages():
    datain = []
    for infile in glob.glob("*.pgm"):
        #filename, ext = os.path.splitext(infile)
        #print filename
        im = Image.open(infile)
        pixels = []
        for y in range(im.size[1]):
            for x in range(im.size[0]):
                pixels.append(im.getpixel((x,y)))
        datain.append(pixels)
    return datain

def ReadOneImage(filename):
    datain = []
    im = Image.open(filename)
    for y in range(im.size[1]):
        for x in range(im.size[0]):
            datain.append(im.getpixel((x,y)))
    return datain

#Functions to save and read an eigenface basis to a file
#Needed for testing tasks 4.4 to 4.6 in the event that task 4.3 is unattempted or unsucessful
def WriteEigenfaceBasis(pcBasis):
    f = open("EigenfaceBasis.txt", "w")
    for row in range(pcBasis.shape[0]):
        for col in range(pcBasis.shape[1]):
            f.write( '%12.10f ' % (pcBasis[row,col]))
        f.write('\n')
    f.write('\n\n')
    f.close()
def ReadEigenfaceBasis():
    f = open("PrincipalComponents.txt");
    datain = []
    for line in range(10):
        datain.append(map(float,(f.readline().split())))
    f.close()
    return numpy.array(datain)