app.py

import os
import praw
import gradio as gr
from transformers import TextClassificationPipeline, AutoModelForSequenceClassification, AutoTokenizer



client_id = os.environ["client_id"]
client_secret = os.environ["client_secret"]
user_agent = os.environ["user_agent"]

reddit = praw.Reddit(client_id =client_id,
                     client_secret =client_secret, user_agent =user_agent)


model_name = "ProsusAI/finbert"
tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=True)
model = AutoModelForSequenceClassification.from_pretrained(model_name, num_labels = 3)
pipe = TextClassificationPipeline(model=model, tokenizer=tokenizer, max_length=64, truncation=True, padding = 'max_length')


def reddit_analysis(subreddit_name, num_posts):
    
    local_score = 0
    local_titles = []
    subreddit = reddit.subreddit(subreddit_name)
    if int(num_posts) > 16:
        return "Number of posts should be less than 15"
    else:
        for post in subreddit.new(limit=int(num_posts)):
            
            prediction = pipe(post.title)
            local_titles.append(post.title)
            
            if prediction[0]["label"] == "negative":
                local_score-= prediction[0]["score"]
            elif prediction[0]["label"] == "positive":
                local_score+= prediction[0]["score"]
                
        titles_string = "\n".join(local_titles)
            
        return local_score, titles_string
        
        #print(post.title)
        #print(post.selftext)

        
        
total_score = 0
text_list = []
def manual_analysis(text):
    
    global total_score
    prediction = pipe(text)
    
    text_list.append(text)
    if prediction[0]["label"] == "negative":
        total_score-= prediction[0]["score"]
    elif prediction[0]["label"] == "positive":
        total_score+= prediction[0]["score"]
        
    return prediction, total_score


with gr.Blocks() as demo:
    with gr.Tab("Seperate Analysis"): 
        first_title = """<p><h1 align="center" style="font-size: 24px;">Analyse texts manually</h1></p>"""
        gr.HTML(first_title)
        with gr.Row():
            with gr.Column():
                text = gr.Textbox(label="text")
                analyse = gr.Button("Analyse")


            with gr.Column():
                label_score = gr.Textbox(label="Label/Score")
                average_score = gr.Textbox(label="Average Score")

        analyse.click(fn=manual_analysis, inputs=text, outputs=[label_score, average_score], api_name="Calc1")
        
    with gr.Tab("Mass Analysis"):
        second_title = """<p><h1 align="center" style="font-size: 24px;">Analyse latest posts from Reddit</h1></p>"""
        gr.HTML(second_title)
        with gr.Row():
            with gr.Column():
                subreddit_name = gr.Textbox(label="Subreddit Name")
            
                num_post = gr.Textbox(label="Number of Posts")
                analyse = gr.Button("Analyse")
            with gr.Column():
                average_score = gr.Textbox(label="Average Score")
                tifu_titles = gr.Textbox(label="Tifu Titles")

        analyse.click(fn=reddit_analysis, inputs=[subreddit_name, num_post], outputs=[average_score, tifu_titles], api_name="Calc2")
        


demo.launch()