<?php

namespace AIContentWriter\Campaigns;

use simplehtmldom\HtmlWeb;

defined( 'ABSPATH' ) || exit; // Exit if accessed directly.

/**
 * Class RSSFeed
 *
 * @since 1.7.0
 * @package AIContentWriter/Campaigns
 */
class RSSFeed {

	/**
	 * Generate titles.
	 * This function is responsible for generating titles based on the provided RSS feed.
	 *
	 * @param int $campaign_id The ID of the campaign for which titles are being generated.
	 * @param int $posts_needed The number of posts needed.
	 *
	 * @since 1.7.0
	 * @return bool Whether the titles and source links were generated successfully or not.
	 */
	public static function generate_titles( $campaign_id, $posts_needed ) {
		if ( empty( $campaign_id ) || empty( $posts_needed ) ) {
			return false;
		}

		$feed_link = get_post_meta( $campaign_id, '_aicw_rss_feed_link', true );

		if ( empty( $feed_link ) ) {
			ai_content_writer()->create_log(
				esc_html__( 'No RSS feed link is configured for this campaign. Edit the campaign and add the feed URL.', 'ai-content-writer' ),
				'error',
				$campaign_id
			);

			return false;
		}

		$response = wp_remote_get( $feed_link );

		if ( is_wp_error( $response ) ) {
			ai_content_writer()->create_log(
				sprintf(
				/* translators: 1: RSS feed URL. 2: Error message. */
					esc_html__( 'Could not fetch the RSS feed %1$s: %2$s', 'ai-content-writer' ),
					esc_url( $feed_link ),
					esc_html( $response->get_error_message() )
				),
				'error',
				$campaign_id
			);

			return false;
		}

		$response_code = wp_remote_retrieve_response_code( $response );

		if ( $response_code < 200 || $response_code >= 300 ) {
			ai_content_writer()->create_log(
				sprintf(
				/* translators: 1: RSS feed URL. 2: HTTP status code. */
					esc_html__( 'The RSS feed %1$s returned HTTP status %2$d. Check that the URL is correct and publicly reachable.', 'ai-content-writer' ),
					esc_url( $feed_link ),
					absint( $response_code )
				),
				'error',
				$campaign_id
			);

			return false;
		}

		$body = wp_remote_retrieve_body( $response );

		if ( empty( $body ) ) {
			ai_content_writer()->create_log(
				sprintf(
				/* translators: %s: RSS feed URL. */
					esc_html__( 'The RSS feed %s returned an empty response.', 'ai-content-writer' ),
					esc_url( $feed_link )
				),
				'error',
				$campaign_id
			);

			return false;
		}

		// Suppress warnings and manually handle XML errors.
		$previous_libxml_state = libxml_use_internal_errors( true );
		$rss                   = simplexml_load_string( $body );

		if ( ! $rss ) {
			// Keep the parser error — it is the only clue the user gets about why the feed failed.
			$xml_errors  = libxml_get_errors();
			$xml_message = ! empty( $xml_errors ) ? trim( $xml_errors[0]->message ) : '';
			libxml_clear_errors();
			libxml_use_internal_errors( $previous_libxml_state );

			ai_content_writer()->create_log(
				sprintf(
				/* translators: 1: RSS feed URL. 2: XML parser error message. */
					esc_html__( 'The RSS feed %1$s is not valid XML (%2$s). Make sure the URL points to the feed itself, for example https://example.com/feed, and not to a regular page.', 'ai-content-writer' ),
					esc_url( $feed_link ),
					esc_html( '' !== $xml_message ? $xml_message : __( 'no parser details available', 'ai-content-writer' ) )
				),
				'error',
				$campaign_id
			);

			return false;
		}

		libxml_clear_errors();
		libxml_use_internal_errors( $previous_libxml_state );

		if ( ! isset( $rss->channel->item ) || 0 === count( $rss->channel->item ) ) {
			ai_content_writer()->create_log(
				sprintf(
				/* translators: %s: RSS feed URL. */
					esc_html__( 'The RSS feed %s contains no items. Only RSS feeds with <channel><item> entries are supported; Atom feeds are not.', 'ai-content-writer' ),
					esc_url( $feed_link )
				),
				'error',
				$campaign_id
			);

			return false;
		}

		$count = 0;
		foreach ( $rss->channel->item as $item ) {
			$title = wp_strip_all_tags( (string) $item->title );
			$link  = esc_url( (string) $item->link );

			if ( empty( $title ) || empty( $link ) ) {
				continue;
			}

			// Check if the title already exists.
			$exists_post_ids = aicw_get_posts_by_title( $title );
			$should_skip     = false;

			foreach ( $exists_post_ids as $post_id ) {
				if ( get_post_meta( $post_id, '_aicw_content_source_link', true ) === $link ) {
					$should_skip = true;
					break;
				}
			}

			if ( $should_skip ) {
				continue;
			}

			// Create a new post.
			$post_id = wp_insert_post(
				array(
					'post_type'    => 'aicw_post',
					'post_title'   => $title,
					'post_content' => '',
					'post_status'  => 'pending',
					'post_author'  => get_post_field( 'post_author', $campaign_id ),
				)
			);

			if ( is_wp_error( $post_id ) ) {
				continue;
			}

			// Add the post meta.
			update_post_meta( $post_id, '_aicw_campaign_id', $campaign_id );
			update_post_meta( $post_id, '_aicw_content_source_link', $link );

			// Update the counter.
			++$count;

			// If the counter is greater than or equal to the target then break the loop.
			if ( $count >= $posts_needed ) {
				break;
			}
		}

		// Update the last run time.
		aicw_update_last_run_time( $campaign_id );

		// Logged after the loop so a partial run (fewer feed items than the target)
		// is reported too, not only a run that hit the target exactly.
		if ( $count > 0 ) {
			ai_content_writer()->create_log(
				sprintf(
				/* translators: %d: Number of posts generated. */
					esc_html__( '%d post titles and source links generated from the RSS feed.', 'ai-content-writer' ),
					absint( $count )
				),
				'success',
				$campaign_id
			);

			return true;
		}

		// The feed parsed fine but produced nothing new — normally every item has
		// already been imported. Logged so the campaign does not look silently broken.
		ai_content_writer()->create_log(
			sprintf(
			/* translators: %s: RSS feed URL. */
				esc_html__( 'No new items found in the RSS feed %s. Every item has already been imported.', 'ai-content-writer' ),
				esc_url( $feed_link )
			),
			'info',
			$campaign_id
		);

		return false;
	}

	/**
	 * Generate content.
	 * This function is responsible for generating content based on the source link.
	 *
	 * @param int    $post_id The ID of the post for which content is being generated.
	 * @param string $title The title of the post.
	 * @param string $keyword The keyword that will be used to generate the content.
	 *
	 * @since 1.7.0
	 * @return bool Whether the content was generated successfully or not.
	 */
	public static function generate_content( $post_id, $title = '', $keyword = '' ) {
		if ( empty( $post_id ) ) {
			return false;
		}

		// Get the source link from the post meta.
		$source_link = get_post_meta( $post_id, '_aicw_content_source_link', true );
		$campaign_id = get_post_meta( $post_id, '_aicw_campaign_id', true );

		if ( empty( $source_link ) ) {
			ai_content_writer()->create_log(
				sprintf(
				/* translators: %d: Post ID. */
					esc_html__( 'Post ID %d has no source link, so its content cannot be fetched from the feed.', 'ai-content-writer' ),
					absint( $post_id )
				),
				'error',
				$campaign_id
			);

			return false;
		}

		// Get the content from the source link.
		$dom  = new HtmlWeb();
		$html = $dom->load( trailingslashit( $source_link ) );

		if ( ! $html ) {
			ai_content_writer()->create_log(
				sprintf(
				/* translators: 1: Source article URL. 2: Post ID. */
					esc_html__( 'Could not fetch the source article %1$s for post ID %2$d. The page may be unreachable or blocking requests.', 'ai-content-writer' ),
					esc_url( $source_link ),
					absint( $post_id )
				),
				'error',
				$campaign_id
			);

			return false;
		}

		// SEO Metadata schema.
		$seo_data = array(
			'description' => '',
			'keywords'    => $keyword,
		);

		$meta_desc = $html->find( 'meta[name=description]', 0 );
		if ( ! empty( $meta_desc ) ) {
			$seo_data['description'] = trim( $meta_desc->content );
		}

		$meta_keywords = $html->find( 'meta[name=keywords]', 0 );
		if ( ! empty( $meta_keywords ) ) {
			$seo_data['keywords'] = trim( $meta_keywords->content );
		}

		// Get the post title, content and error.
		$result = array(
			'title'   => '',
			'content' => '',
		);

		// Extract title - try different common title selectors.
		$possible_title_selectors = array(
			'h1.entry-title',
			'h1.post-title',
			'h1.article-title',
			'article h1',
			'.article-header h1',
			'.post-header h1',
			'header h1',
			'h1',
			'.headline',
			'.title',
			'#title',
		);

		// Try to find the title using the possible selectors.
		foreach ( $possible_title_selectors as $selector ) {
			$title_element = $html->find( $selector, 0 );
			if ( $title_element ) {
				$result['title'] = trim( $title_element->plaintext );
				break;
			}
		}

		// If title still not found, use the page title.
		if ( empty( $result['title'] ) ) {
			$title_element = $html->find( 'title', 0 );
			if ( $title_element ) {
				$result['title'] = trim( $title_element->plaintext );
			}
		}

		// Extract content - try different common content containers.
		$content_elements           = null;
		$possible_content_selectors = array(
			'article',
			'.post-content',
			'.entry-content',
			'.article-content',
			'.content',
			'#content',
			'.post-body',
			'.article-body',
			'.story-content',
			'.main-content',
			'main',
			'.post',
			'#main-content',
			'.story',
			'.entry',
			'.blog-post',
			'.post-entry',
		);

		// Try to identify the main content container.
		foreach ( $possible_content_selectors as $selector ) {
			$element = $html->find( $selector, 0 );
			if ( $element ) {
				$content_elements = $element;
				break;
			}
		}

		// If no content found with selector approach, use paragraph fallback.
		if ( empty( (array) $content_elements ) ) {
			// Find all paragraphs and analyze which ones likely belong to the main content.
			$all_paragraphs = $html->find( 'p' );

			// Filter out small paragraphs that likely aren't article content and paragraphs in common non-content areas like headers, footers, sidebars.
			$content_paragraphs = array();
			foreach ( $all_paragraphs as $p ) {
				$text   = trim( $p->plaintext );
				$length = strlen( $text );

				// Skip empty paragraphs.
				if ( $length < 10 ) { continue;
				}

				// Skip paragraphs in common non-content areas.
				$parent       = $p->parent();
				$parent_tag   = $parent ? strtolower( $parent->tag ) : '';
				$parent_class = $parent ? strtolower( $parent->class ) : '';
				$parent_id    = $parent ? strtolower( $parent->id ) : '';

				$skip_keywords = array( 'footer', 'header', 'sidebar', 'comment', 'menu', 'nav', 'widget' );
				$should_skip   = false;
				foreach ( $skip_keywords as $skip_keyword ) {
					if ( strpos( $parent_class, $skip_keyword ) !== false || strpos( $parent_id, $skip_keyword ) !== false || strpos( $parent_tag, $skip_keyword ) !== false ) {
						$should_skip = true;
						break;
					}
				}

				if ( ! $should_skip ) {
					$content_paragraphs[] = $text;
				}
			}

			// Join all valid paragraphs.
			$result['content'] = implode( "\n\n", $content_paragraphs );
		} else {
			// Clean up the content - Remove common non-content elements.
			foreach ( $content_elements->find( 'aside, .sidebar, .widget, nav, .navigation, .menu, .comment, .comments, .post-meta, .related-posts, .social-share, .share-buttons, .sharing, .author-bio, script, style, iframe, form, .advertisement, .ads, .ad-container' ) as $remove ) {
				$remove->outertext = '';
				$remove->clear();
			}

			/**
			 * Allow cleanup HTML elements from the content.
			 * This will allow users to clean up specific elements using HTML tags, classes, or IDs.
			 *
			 * @param object $content_elements The content elements.
			 * @param int    $post_id The post ID.
			 *
			 * @since 2.0.2
			 * @return object The modified content elements.
			 */
			$content_elements = apply_filters( 'aicw_cleanup_content_elements', $content_elements, $post_id );

			// Define allowed HTML tags.
			$allowed_html = array( 'p', 'h2', 'h3', 'h4', 'h5', 'h6', 'ol', 'ul', 'blockquote' );

			// Extract content with HTML structure preserved.
			$paragraphs = array();
			foreach ( $content_elements->find( 'p, h2, h3, h4, h5, h6, ol, ul, blockquote' ) as $node ) {
				// Get the tag name.
				$tag_name = strtolower( $node->tag );

				// Check if this tag is in our allowed tags list.
				if ( in_array( $tag_name, $allowed_html, true ) ) {
					// For list elements, need to preserve structure.
					if ( 'ol' === $tag_name || 'ul' === $tag_name ) {
						$list_items = array();
						foreach ( $node->find( 'li' ) as $li ) {
							$li_text = trim( $li->plaintext );
							if ( ! empty( $li_text ) ) {
								$list_items[] = "<li>{$li_text}</li>";
							}
						}
						if ( ! empty( $list_items ) ) {
							$list_html    = "<{$tag_name}>" . implode( '', $list_items ) . "</{$tag_name}>";
							$paragraphs[] = $list_html;
						}
					} else {
						// For other tags like p, h2, etc., just get inner text.
						$text = trim( $node->plaintext );
						if ( ! empty( $text ) ) {
							$paragraphs[] = "<{$tag_name}>{$text}</{$tag_name}>";
						}
					}
				}
			}

			// Join all elements with proper spacing.
			$result['content'] = implode( "\n", $paragraphs );

			// If no paragraphs found, use the plain text of the content element.
			if ( empty( $result['content'] ) ) {
				$result['content'] = trim( $content_elements->plaintext );
			}
		}

		// Clean the content.
		$result['content'] = preg_replace( '/\s+/', ' ', $result['content'] ); // Normalize whitespace.
		$result['content'] = preg_replace( '/\n\s*\n/', "\n\n", $result['content'] ); // Remove extra line breaks.
		$result['content'] = trim( $result['content'] );

		// Clean the title.
		$result['title'] = preg_replace( '/\s+/', ' ', $result['title'] ); // Normalize whitespace.
		$result['title'] = trim( $result['title'] );

		// Release memory.
		$html->clear();
		unset( $html );

		// Check if the content is empty.
		if ( isset( $result['content'] ) && empty( $result['content'] ) ) {
			ai_content_writer()->create_log(
				sprintf(
				/* translators: 1: Source article URL. 2: Post ID. */
					esc_html__( 'No readable article content could be extracted from %1$s for post ID %2$d. The page layout may not be supported.', 'ai-content-writer' ),
					esc_url( $source_link ),
					absint( $post_id )
				),
				'error',
				$campaign_id
			);

			return false;
		}

		// Update the post content.
		$updated = wp_update_post(
			array(
				'ID'           => $post_id,
				'post_title'   => wp_strip_all_tags( isset( $result['title'] ) ? $result['title'] : $title ),
				'post_content' => wp_kses_post( $result['content'] ),
			)
		);

		if ( is_wp_error( $updated ) ) {
			ai_content_writer()->create_log(
				sprintf(
				/* translators: 1: Post ID. 2: Error message. */
					esc_html__( 'Could not save the generated content for post ID %1$d: %2$s', 'ai-content-writer' ),
					absint( $post_id ),
					esc_html( $updated->get_error_message() )
				),
				'error',
				$campaign_id
			);

			return false;
		}

		return true;
	}
}