import gradio as gr
import spacy
import re
import os
import requests

nlp = None

def load_spacy_model():
    global nlp
    if not nlp:
        try:
            nlp = spacy.load("en_core_web_sm")
        except OSError:
            print("Downloading Spacy model...")
            os.system("python -m spacy download en_core_web_sm")
            nlp = spacy.load("en_core_web_sm")

def detect_ai_content(text, download_file=False):
    # Load Spacy model if not already loaded
    load_spacy_model()

    # Count the number of words in the text
    word_count = len(text.split())

    # Analyze the text using Spacy
    doc = nlp(text)

    # Count the number of tokens that are not in Spacy's default stop word list
    non_stopword_tokens = [token for token in doc if not token.is_stop]
    non_stopword_count = len(non_stopword_tokens)

    # Calculate the percentage of non-stopword tokens
    percentage_ai = (1 - non_stopword_count / word_count) * 100

    # Clean the text by removing extra spaces, line breaks and special characters
    cleaned_text = re.sub(r'\s+', ' ', text).strip()
    cleaned_text = re.sub(r'[^\w\s]', '', cleaned_text)

    # Return a dictionary with the percentage of AI-generated content and the cleaned text
    result = {
        "text": cleaned_text,
        "percentage": f"{percentage_ai:.2f}% AI-generated content"
    }

    # Download cleaned text as file if option is enabled
    if download_file:
        with open('cleaned_text.txt', 'w') as f:
            f.write(cleaned_text)
        result['download_link'] = 'cleaned_text.txt'

    return result

def upload_text_file(file):
    with open(file.name, 'r') as f:
        text = f.read()
    return text

def get_example_text():
    # Get example text from external API
    try:
        response = requests.get('https://baconipsum.com/api/?type=all-meat&sentences=1')
        if response.ok:
            return response.json()[0]
    except:
        pass

    # Use a fallback example text if API request fails
    return "This is a sample text with no AI-generated content."

# Create Gradio interface
inputs = [
    gr.inputs.Textbox(lines=5, label="Enter text to analyze:"),
    gr.inputs.Checkbox(label="Download cleaned text as file", default=False)
]

outputs = gr.JSON(
    label="Results",
    allow_download=True,
    download_filename="ai_content_analysis.json"
)

file_input = gr.inputs.File(label="Upload a text file:")
file_output = gr.outputs.Textbox(label="File contents:")

examples = [
    {"text": get_example_text(), "percentage": "0.00% AI-generated content"},
    {"text": "This is a text generated by a chatbot.", "percentage": "100.00% AI-generated content"}
]

iface = gr.Interface(
    fn=detect_ai_content,
    inputs=inputs + [file_input],
    outputs=[outputs,file_output],
    title="AI-Generated Content Detector",
    description="This app uses a simple stylometry analysis technique to estimate the percentage of AI-generated content in a given text.",
    show_input=True,
    show_output=True,)
iface.launch()