main_textblob.py

import polars as pl
from textblob import TextBlob
#from wordcloud import WordCloud
import matplotlib.pyplot as plt


def getTextSubjectivity(txt):
    """
    This Python function uses TextBlob to analyze the subjectivity of a given text.
    
    :param txt: The `getTextSubjectivity` function calculates the subjectivity of a given text using the
    TextBlob library. The `txt` parameter is the text for which you want to determine the subjectivity.
    You can pass any text as an argument to this function, and it will return a value representing the
    subject
    :return: The function `getTextSubjectivity` returns the subjectivity score of the input text using
    TextBlob's sentiment analysis.
    """
    return TextBlob(txt).sentiment.subjectivity

def getTextPolarity(txt):
    """
    The function `getTextPolarity` calculates the polarity of a given text using TextBlob's sentiment
    analysis.
    
    :param txt: The `getTextPolarity` function you provided calculates the polarity of a given text
    using TextBlob library. The `txt` parameter is the text for which you want to determine the
    polarity. You can pass any text as an argument to this function to get its polarity score
    :return: The function `getTextPolarity` returns the polarity of the sentiment of the input text
    `txt`.
    """
    return TextBlob(txt).sentiment.polarity

def getTextAnalysis(n):
    """
    The function `getTextAnalysis` categorizes a number as negative, neutral, or positive based on its
    value.
    
    :param n: The function `getTextAnalysis(n)` takes a numerical input `n` and returns a text analysis
    based on the value of `n`. If `n` is less than 0, it returns "negative". If `n` is equal to 0, it
    returns "neutral". Otherwise, if
    :return: The function `getTextAnalysis` returns a string indicating whether the input number `n` is
    negative, neutral (zero), or positive.
    """
    if n < 0:
        return "negative"
    elif n == 0:
        return "neutral"
    else:
        return "positive"


# Encoding is required to ensure we don't have false "null" values
df = pl.read_csv("sentiment_feedback.csv", ignore_errors=True, encoding='latin-1')
# df = df.drop(columns=['File Name', '_id', 'sentiment', 'sentimentMagnitude', 'url', 'variantId'])

# # Split the 'feedback' column into 'pros' and 'cons'
# df = df.with_columns([
#     pl.col("feedback").str.split("Pros:").list.get(1).str.split("Cons:").list.get(0).str.strip_chars().alias("pros"),
#     pl.col("feedback").str.split("Cons:").list.get(1).str.strip_chars().alias("cons")
#     ])

# Concatenating the feedback to perform a consolidated analysis on the sentiment of the reviewer
df = df.with_columns([pl.concat_str([pl.col("title"), pl.col("feedback").str.replace_all("Pros:", " ").str.replace_all("Cons:", " ")], 
                                    separator=" ").alias("consolidated_feedback")
                                    ])

# Remove the original 'feedback' and 'title' columns, not required
df = df.drop(["feedback", "title"])
print(df.describe())
print(df['consolidated_feedback'][0])

# print(TextBlob(df['consolidated_feedback'][0]).sentiment.polarity)
# Calculate the subjectivity and polarity for each review which is then used to evaluate the sentiment of the reviewer
df = df.with_columns([df['consolidated_feedback'].map_elements(getTextSubjectivity).alias('subjectivity')])
df = df.with_columns([df['consolidated_feedback'].map_elements(getTextPolarity).alias('polarity')])
# print(df.head())


df = df.with_columns(df['polarity'].map_elements(getTextAnalysis).alias('score'))
# print(df.head())

positive = df.filter(pl.col('score') == 'positive')
negative = df.filter(pl.col('score') == 'negative')
neutral = df.filter(pl.col('score') == 'neutral')
print(str(positive.shape[0]/(df.shape[0])*100) + " % of positive reviews")
print(str(negative.shape[0]/(df.shape[0])*100) + " % of negative reviews")
print(str(neutral.shape[0]/(df.shape[0])*100) + " % of neutral reviews")

# print(df.select(pl.col('score'), pl.col('rating')))

# # Creating a word cloud
# positive_words = ' '.join([word for word in df.filter(pl.col('score') == 'positive')['consolidated_feedback']])
# positive_wordCloud = WordCloud(width=600, height=400).generate(positive_words)
# plt.title('Positive Sentiments')
# plt.imshow(positive_wordCloud)
# plt.show()

# negative_words = ' '.join([word for word in df.filter(pl.col('score') == 'negative')['consolidated_feedback']])
# negative_wordCloud = WordCloud(width=600, height=400).generate(negative_words)
# plt.title('Negative Sentiments')
# plt.imshow(positive_wordCloud)
# plt.show()

# neutral_words = ' '.join([word for word in df.filter(pl.col('score') == 'neutral')['consolidated_feedback']])
# neutral_wordCloud = WordCloud(width=600, height=400).generate(neutral_words)
# plt.title('Neutral Sentiments')
# plt.imshow(neutral_wordCloud)
# plt.show()

df.write_csv('results.csv')