import gradio as gr
import requests
from bs4 import BeautifulSoup

def get_url_content(url):
    response = requests.get(url)
    if response.status_code == 200:
        return response.text
    else:
        return "URL에서 콘텐츠를 가져오는 데 실패했습니다."

def parse_html(html_content):
    soup = BeautifulSoup(html_content, 'html.parser')
    return soup.prettify()

def gradio_fetch_and_parse(url):
    html_content = get_url_content(url)
    parsed_content = parse_html(html_content)
    return parsed_content

def get_main_content(html_content):
    soup = BeautifulSoup(html_content, 'html.parser')

    # 다양한 요소에서 텍스트 추출
    extracted_texts = []

    # <span class="a-list-item">
    extracted_texts.extend([item.get_text(strip=True) for item in soup.find_all('span', class_='a-list-item')])

    # <meta content>
    extracted_texts.extend([meta.get('content', '') for meta in soup.find_all('meta') if meta.get('content')])

    # 추가 요소들
    for tag in ['section', 'article', 'div', 'p', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'header', 'footer', 'aside', 'iframe']:
        extracted_texts.extend([item.get_text(strip=True) for item in soup.find_all(tag)])

    combined_text = ' '.join(extracted_texts)

    if combined_text.strip():
        print("추출된 텍스트:", combined_text)
        return combined_text
    else:
        print("본문 콘텐츠를 찾을 수 없습니다.")
        return ''

def format_script(text):
    sentences = text.split('.')
    script = ""
    for i in range(0, min(len(sentences), 10), 2):
        line = sentences[i].strip() + '. '
        if i+1 < len(sentences):
            line += sentences[i+1].strip() + '\n'
        script += line
        print("현재 스크립트:", script)
    return script

def gradio_fetch_and_format_script(url):
    print("함수 호출됨:", url)
    html_content = get_url_content(url)
    main_content = get_main_content(html_content)
    print("추출된 본문:", main_content)
    script = format_script(main_content)
    print("생성된 스크립트:", script)
    return script

iface_html = gr.Interface(fn=gradio_fetch_and_parse, inputs=gr.Textbox(label="URL을 입력하세요"), outputs=gr.Textbox(label="스크랩된 HTML 콘텐츠"))
iface_script = gr.Interface(fn=gradio_fetch_and_format_script, inputs=gr.Textbox(label="URL을 입력하세요"), outputs=gr.Textbox(label="영상용 스크립트"))
iface_combined = gr.TabbedInterface([iface_html, iface_script], ["HTML 보기", "스크립트 생성"])
iface_combined.launch()