app.py

import gradio as gr
import openai
import os
from anthropic import Anthropic

# Load OpenAI and Anthropic API Keys from environment variables
openai.api_key = os.getenv("openaiapikey")
anthropic_api_key = os.getenv("anthropickey")

# Initialize Anthropic client
client = Anthropic(api_key=anthropic_api_key)

MODEL_NAME = "claude-3-haiku-20240307"

def moderate_text(user_text):
    # OpenAI Moderation
    openai_moderation_result = "Error in OpenAI Moderation"
    try:
        response = openai.Moderation.create(input=user_text)
        moderation_categories = response["results"][0]["categories"]
        moderation_flagged = response["results"][0]["flagged"]

        if moderation_flagged:
            openai_moderation_result = f"OpenAI flags the text for the following categories: {', '.join([category for category, flagged in moderation_categories.items() if flagged])}"
        else:
            openai_moderation_result = "The text is not flagged for any moderation issues by OpenAI."
    except Exception as e:
        openai_moderation_result = f"Error occurred with OpenAI: {e}"

    # Anthropic Moderation (Using Prompt Template)
    anthropic_moderation_result = "Error in Anthropic Moderation"
    try:
        # Create the prompt template for Anthropic's moderation model
        prompt_template = """
        You are a content moderation expert tasked with categorizing user-generated text based on the following guidelines:

        Here is the user-generated text to categorize:
        <user_text>{user_text}</user_text>

        Based on the content, classify this text as either ALLOW or BLOCK. Return nothing else.
        """

        # Format the prompt with the user text
        prompt = prompt_template.format(user_text=user_text)

        # Send the prompt to Claude and get the response
        response = client.messages.create(
            model=MODEL_NAME,
            max_tokens=10,
            messages=[{"role": "user", "content": prompt}]
        ).content[0].text

        # Format the Anthropic moderation result
        anthropic_moderation_result = f"Anthropic's moderation result: {response}"

    except Exception as e:
        anthropic_moderation_result = f"Error occurred with Anthropic: {e}"

    return openai_moderation_result, anthropic_moderation_result


# Create the Gradio interface
iface = gr.Interface(
    fn=moderate_text,
    inputs=gr.Textbox(lines=2, placeholder="Enter text here..."),
    outputs=["text", "text"],
    title="Content Moderation Tool",
    description="Enter some text and get the moderation results from OpenAI and Anthropic."
)

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