app.py

import gradio as gr
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
import os

# Get Hugging Face token from environment variable
HF_TOKEN = os.getenv("HF_TOKEN")
if not HF_TOKEN:
    raise ValueError("Please set HF_TOKEN environment variable with your Hugging Face access token")

# Load model and tokenizer
model_name = "iimran/AnalyserV1"
tokenizer = AutoTokenizer.from_pretrained(model_name, token=HF_TOKEN)
model = AutoModelForSequenceClassification.from_pretrained(model_name, token=HF_TOKEN)

def classify_complaint(text):
    inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True, max_length=256)
    with torch.no_grad():
        outputs = model(**inputs)
    return model.config.id2label[torch.argmax(outputs.logits).item()]

# Create Gradio interface
demo = gr.Interface(
    fn=classify_complaint,
    inputs=gr.Textbox(lines=3, placeholder="Enter your complaint here...", label="Complaint Text"),
    outputs=gr.Label(label="Predicted Category"),
    title="Complaint Category Classifier",
    description="Automatically classify community complaints into specific categories",
    examples=[
        ["I wanted to bring to your attention that a huge big truck has been parked on Main Street"],
        ["There are overgrown bushes on Oak Road that pose a fire risk"],
        ["I am writing to report an ongoing issue of excessive noise from the construction site located at [specific location, if known]. The noise persists during night hours, significantly disrupting the peace and causing inconvenience to residents in the area.  This continuous disturbance violates acceptable noise regulations and affects the well-being of those who need restful sleep during these hours. I kindly request immediate action to address this matter by enforcing noise restrictions during night hours or implementing measures to minimize the disturbance. Thank you for your attention to this urgent concern. I look forward to your prompt response."]
    ]
)

if __name__ == "__main__":
    demo.launch()