#!/usr/bin/env python3

import re
import matplotlib.pyplot as plt
import numpy as np

class Article:
    def __init__(self, dictionary):
        for i, j in dictionary.items():
            setattr(self, i, j)

class Node:
	def __init__(self, article=None, number=None, next=None):
		self.article = article
		self.number = number
		self.next = next
	def __str__(self):
		if self.next != None:
			return "[{}, {}] -> {}".format(self.article, self.number, self.next)
		else:
			return "[{}, {}]".format(self.article, self.number)
	def count(self):
		if self.next != None:
			return self.number + self.next.count()
		else:
			return self.number

class LinkedList:
	def __init__(self, article, word):
		self.head = Node(article, 0)
		self.currentNode = self.head
		self.word = word
	def newNode(self, article):
		node = Node(article, 0)
		if self.head.article == None:
			self.head = node
			self.currentNode = node
		else:
			self.currentNode.next = node
			self.currentNode = node
	def show(self):
		print("[{}] -> {}".format(self.word, self.head))

# initialize variables
articles = []
keys = []
hashTable = {}
histogramNumbers = []

# read dataset
collection = open('collection.txt', 'r')
articles = collection.read()
collection.close()

# seperate articles into list
articles = re.sub('<doc>', ' ', articles)
articles = articles.split('</doc>')[:-1]

for i in range(len(articles)):
	# Seperate article items based on tags
	articles[i] = re.sub('<.?p>|</.*?>', ' ', articles[i])
	articles[i] = re.sub('\n', '', articles[i])
	articles[i] = re.sub('> ', ' <', articles[i])
	articles[i] = articles[i][2:]
	articles[i] = articles[i].split(' <')

	# Create Article object of article items
	tempDict = {}
	for j in range(0, len(articles[i]), 2):
		tempDict[articles[i][j]] = articles[i][j+1]
	articles[i] = Article(tempDict)

	# define what items of article are used
	tempWords = ""
	for j in vars(articles[i]).values():
		tempWords = tempWords + str(j)

	# lowercase and modify words to split into list
	keys.append(tempWords.lower())
	keys[i] = keys[i].replace("'s", "")
	keys[i] = keys[i].replace("' ", "")
	keys[i] = keys[i].replace(" '", "")
	keys[i] = re.sub(r'[;|-|_|:|<|>|(|)|+|/|*|"|?|!|.|,|(|)]', r'', keys[i])
	keys[i] = keys[i].split()

# create hashtable with the words as keys
for i in range(len(keys)):
	for word in keys[i]:
		articleNumber = i + 1
		if word not in hashTable.keys():
			hashTable[word] = LinkedList(articleNumber, word)
			hashTable[word].head.number += 1
		else:
			cursor = hashTable[word]
			if cursor.head.article == articleNumber:
				cursor.head.number += 1
			else:
				cursor = cursor.head
				while cursor.article != articleNumber:
					if cursor.next != None:
						cursor = cursor.next
					else:
						hashTable[word].newNode(articleNumber)
				cursor.number += 1

# print the hash table for each key and keep track of count for plot
for key in hashTable:
	hashTable[key].show()
	histogramNumbers.append(hashTable[key].head.count())

# plot histogram
bins = np.arange(max(histogramNumbers)) - 0.5
plt.hist(histogramNumbers, bins, alpha=0.75, color='g', linewidth=0.75, edgecolor='black')
plt.title("(b)")
plt.xlabel("Total count of words")
plt.ylabel("Frequency of total count")
x1,x2,y1,y2 = plt.axis()
plt.axis((1,21,y1,y2))
plt.xticks(np.arange(1, 20, 2))
plt.xlim([0.5, 20.5])
plt.show()