| # import gradio as gr | |
| # from newspaper import Article | |
| # # Function to scrape the article from a given URL | |
| # def scrape_article(url): | |
| # try: | |
| # # Create an Article object | |
| # article = Article(url) | |
| # # Download and parse the article | |
| # article.download() | |
| # article.parse() | |
| # # Extract the title and text content | |
| # title = article.title | |
| # content = article.text | |
| # if not content: | |
| # return "Failed to scrape content. Please check the URL." | |
| # return f"Title: {title}\n\nContent:\n{content}" | |
| # except Exception as e: | |
| # return f"Error: {str(e)}" | |
| # # Gradio interface | |
| # def main(): | |
| # # Define the Gradio input-output interface | |
| # with gr.Blocks() as interface: | |
| # gr.Markdown("""## Article Scraper using Newspaper3k | |
| # Enter a URL to scrape the article content.""") | |
| # # Input and output widgets | |
| # url_input = gr.Textbox(label="Enter the Article URL") | |
| # output = gr.Textbox(label="Scraped Article", lines=15) | |
| # # Button to trigger scraping | |
| # scrape_button = gr.Button("Scrape Article") | |
| # # Button functionality | |
| # scrape_button.click(scrape_article, inputs=[url_input], outputs=[output]) | |
| # return interface | |
| # if __name__ == "__main__": | |
| # interface = main() | |
| # interface.launch() | |
| import gradio as gr | |
| from newspaper import Article | |
| # Function to scrape the article | |
| def scrape_article(url): | |
| try: | |
| article = Article(url) | |
| article.download() | |
| article.parse() | |
| title = article.title | |
| authors = ", ".join(article.authors) | |
| content = article.text | |
| return { | |
| "Title": title, | |
| "Authors": authors if authors else "Not available", | |
| "Content": content if content else "No content available", | |
| } | |
| except Exception as e: | |
| return {"Error": f"Failed to scrape the article: {str(e)}"} | |
| # Define Gradio interface | |
| def gradio_scraper(url): | |
| result = scrape_article(url) | |
| if "Error" in result: | |
| return result["Error"], "", "" | |
| return result["Title"], result["Authors"], result["Content"] | |
| # Gradio interface | |
| def main(): | |
| interface = gr.Interface( | |
| fn=gradio_scraper, | |
| inputs=gr.Textbox(label="Enter Article URL"), | |
| outputs=[ | |
| gr.Textbox(label="Title"), | |
| gr.Textbox(label="Authors"), | |
| gr.TextArea(label="Content"), | |
| ], | |
| title="Article Scraper", | |
| description="Enter the URL of a news article to scrape its title, authors, and content.", | |
| ) | |
| interface.launch() | |
| if __name__ == "__main__": | |
| main() | |