app.py

from PIL import Image, ImageDraw, ImageFont
from textwrap import wrap
import requests
import numpy as np
import gradio as gr
import matplotlib.pyplot as plt
import seaborn as sns
from matplotlib.backends.backend_agg import FigureCanvasAgg
from io import BytesIO

import configparser
import tweepy

config = configparser.ConfigParser()
config.read('./config.ini')

api_key = config['twitter']['api_key']
api_key_secret = config['twitter']['api_key_secret']
access_token = config['twitter']['access_token']
access_token_secret = config['twitter']['access_token_secret']

# Authenticate with Twitter
auth = tweepy.OAuthHandler(api_key, api_key_secret)
auth.set_access_token(access_token, access_token_secret)
api = tweepy.API(auth)

tweets = []


def drawTweet(tweet,i):
    
    width, height = 1000, 200
    
    
    image = Image.new('RGBA', (width, height), 'white')

    
    draw = ImageDraw.Draw(image)

    
    font = ImageFont.truetype('SofiaSansCondensed-VariableFont_wght.ttf',  size=36, encoding='utf-16')

    user = tweet.user


    user_tag = user.screen_name
    tweet_text = tweet.full_text


    words = tweet_text.split()
    
    formatted_string = ''
    for i, word in enumerate(words):
        formatted_string += word+' '
        if (i + 1) % 7 == 0:
            formatted_string += '\n'



    draw.multiline_text( (135,50), formatted_string , fill='black' , font=font, embedded_color=True)
    draw.text((135,10), f"@{user_tag}", fill='black',font=font)
    

    
    response = requests.get(user.profile_image_url_https)
    content = response.content

    f = BytesIO(content)

    avatar_size = (100, 100)
    avatar_image = Image.open(f)
    avatar_image = avatar_image.resize(avatar_size)
    image.paste(avatar_image, (10, 10))
    
    
    return image


def collect_tweets(topic):
    

    limit=200
    tweets = tweepy.Cursor(api.search_tweets,q=f"{topic} -filter:retweets", lang="en", tweet_mode='extended', result_type = 'recent').items(limit)

    tweets = [tweet for tweet in tweets]

    images = []
    i = 1
    for tweet in tweets:
        img = drawTweet(tweet,i)
        images.append(img)
    
    sentiment_plot = sentiment_analysis(tweets,topic)

    return images,sentiment_plot

def sentiment_analysis(tweets,topic):
        
        
        tweet_procs = []
        for tweet in tweets:
                tweet_words = []
                for word in tweet.full_text.split(' '):
                        if word.startswith('@') and len(word) > 1:
                            word = '@user'
                        elif word.startswith('https'):
                            word = "http"
                        tweet_words.append(word)
                tweet_proc = " ".join(tweet_words)
                tweet_procs.append(tweet_proc)


        API_URL = "https://api-inference.huggingface.co/models/cardiffnlp/twitter-roberta-base-sentiment"
        headers = {"Authorization": "Bearer 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"}

        print(len(tweet_procs))

        def query(payload):
            response = requests.post(API_URL, headers=headers, json=payload)
            return response.json()

        model_input = {
        "inputs": [tweet_procs[0]]
        }

        for i in range(1,len(tweets)):
            model_input["inputs"].append(tweet_procs[i])
            
        output = query({
        "inputs": model_input["inputs"]})

        negative = 0
        neutral = 0
        positive = 0

        for score in output:
            neg = 0
            neu = 0
            pos = 0
            for labels in score:
                if labels['label'] == 'LABEL_0':
                    neg += labels['score']
                elif labels['label'] == 'LABEL_1':
                    neu += labels['score'] 
                elif labels['label'] == 'LABEL_2':
                    pos += labels['score'] 
            sentiment = max(neg,neu,pos)
            if neg == sentiment:
                negative += 1
            elif neu == sentiment:
                neutral += 1
            elif pos == sentiment:
                positive += 1

        
        sns.barplot(x=["Negative Sentiment", "Neutral Sentiment", "Positive Sentiment"], y = [negative,neutral,positive])
        plt.title(f"Sentiment Analysis on Twitter regarding {topic}")
        canvas = FigureCanvasAgg(plt.gcf())
        canvas.draw()
        plot = np.array(canvas.buffer_rgba())
        return plot





app = gr.Interface(fn=collect_tweets, inputs=gr.Textbox(label="Enter a topic for tweets"),  outputs=[gr.Gallery(label="Generated images", show_label=False, elem_id="gallery").style(grid=[2], height="50"), gr.Image(label="Sentiment Analysis Result")])

app.launch()