reviewdataNB.py

import json
import sys
import os.path


# container class for review information
class Review:
    def __init__(self, reviewerID, productID, upVotes, totalVotes, reviewText, overallRating):
        self.reviewerID = reviewerID
        self.productID = productID
        self.upVotes = upVotes
        self.totalVotes = totalVotes
        self.reviewText = reviewText
        self.overallRating = overallRating
        # metrics
        self.favorableRating = float(upVotes) / float(totalVotes)
        self.reviewAge = 0
        # reviewer dict
        self.reviewedAlsoBought = 0
        self.reviewedAlsoViewed = 0
        self.reviewedAlsoBoughtTogether = 0
        self.NumberOfUserReviews = 0
        self.AverageReviewRating = 0
        # product dict
        self.reviewDeviationFromMean = 0
        self.reviewProductMean = 0
        # nlp library
        self.reviewLength = 0
        self.numberStopWords = 0
        self.numberOfPunctuations = 0
        self.averageWordLength = 0
        self.averageSentanceLength = 0
        self.numberExclamationPoints = 0
        self.numberQuestionMarks = 0
        self.readability = 0
        self.namedEntities = 0
        self.numberNouns = 0
        self.numberPassiveVerbs = 0
        self.numberActiveVerbs = 0
        self.numberAdjectives = 0
        self.numberPronous = 0
        self.helpfulLabel = "NA"

    def to_dict(self):
        return {
            'upVotes': self.upVotes,
            'totalVotes': self.totalVotes,
            'overallRating': self.overallRating,
            'favorableRating': self.favorableRating,
            'reviewAge': self.reviewAge,
            'reviewedAlsoBought': self.reviewedAlsoBought,
            'reviewedAlsoViewed': self.reviewedAlsoViewed,
            'reviewedAlsoBoughtTogether': self.reviewedAlsoBoughtTogether,
            'NumberOfUserReviews': self.NumberOfUserReviews,
            'AverageReviewRating': self.AverageReviewRating,
            'reviewDeviationFromMean': self.reviewDeviationFromMean,
            'reviewProductMean': self.reviewProductMean,
            'reviewLength': self.reviewLength,
            'numberStopWords': self.numberStopWords,
            'numberOfPunctuations': self.numberOfPunctuations,
            'averageWordLength': self.averageWordLength,
            'averageSentanceLength': self.averageSentanceLength,
            'numberExclamationPoints': self.numberExclamationPoints,
            'numberQuestionMarks': self.numberExclamationPoints,
            'namedEntities': self.namedEntities,
            'numberNouns': self.numberNouns,
            'numberPassiveVerbs': self.numberPassiveVerbs,
            'numberActiveVerbs': self.numberActiveVerbs,
            'numberAdjectives': self.numberAdjectives,
            'numberPronous': self.numberPronous,
            'readability': self.readability,
            'helpfulLabel': self.helpfulLabel
        }



# Read in review data if there is no pre calculated file
def readInReviewData(fileName, minVoteThreshold, minReviewLength):
    reviewsList = []
    dataStore = []



    for line in open(fileName, 'r'):
        # Get data from json file
        data = json.loads(line)

        # Grab each data component
        reviewerID = data['reviewerID']
        productID = data['asin']
        upVotes = data['helpful'][0]
        totalVotes = data['helpful'][1]
        reviewText = data['reviewText']
        rating = data['overall']

        # Initialize data components into review class
        # Only add to list of reviews if review meets minimum vote threshold and minimum text length threshold
        if (totalVotes >= minVoteThreshold) and (len(reviewText) >= minReviewLength):
            reviewEntry = Review(reviewerID, productID, upVotes, totalVotes, reviewText, rating)
            reviewsList.append(reviewEntry)
            dataStore.append(data)

    return reviewsList