WordPress.org
Get WordPress

Plugin Directory

Changeset 3480216


Ignore:
Timestamp:
03/11/2026 01:47:37 PM (4 weeks ago)
Author:
vinsmach
Message:

Version 1.6.3 — fix language and style

Location:
mescio-for-agents/trunk
Files:
4 edited

Legend:

Unmodified
Added
Removed

  • mescio-for-agents/trunk/includes/class-llms-endpoints.php

    r3480063 r3480216  
    238238                $lines[] = '';
    239239                foreach ( $pages as $page ) {
     240                    $page    = self::resolve_translated_post( $page );
    240241                    $desc    = Mescio_For_Agents_Markdown::get_excerpt( $page );
    241242                    $lines[] = '- [' . get_the_title( $page ) . '](' . get_permalink( $page->ID ) . ')'
    … …  
    259260                $lines[] = '';
    260261                foreach ( $posts as $post ) {
     262                    $post    = self::resolve_translated_post( $post );
    261263                    $desc    = Mescio_For_Agents_Markdown::get_excerpt( $post );
    262264                    $date    = get_the_date( 'Y-m-d', $post );
    … …  
    281283                $by_cat = [];
    282284                foreach ( $products as $product ) {
    283                     $cats = get_the_terms( $product->ID, 'product_cat' );
    284                     $cat  = ( $cats && ! is_wp_error( $cats ) ) ? $cats[0]->name : 'Products';
     285                    $product = self::resolve_translated_post( $product );
     286                    $cats    = get_the_terms( $product->ID, 'product_cat' );
     287                    $cat     = ( $cats && ! is_wp_error( $cats ) ) ? $cats[0]->name : 'Products';
    285288                    $by_cat[ $cat ][] = $product;
    286289                }
    … …  
    324327            $lines[] = '';
    325328            foreach ( $cpt_posts as $cpt_post ) {
    326                 $desc    = Mescio_For_Agents_Markdown::get_excerpt( $cpt_post );
    327                 $lines[] = '- [' . get_the_title( $cpt_post ) . '](' . get_permalink( $cpt_post->ID ) . ')'
     329                $cpt_post = self::resolve_translated_post( $cpt_post );
     330                $desc     = Mescio_For_Agents_Markdown::get_excerpt( $cpt_post );
     331                $lines[]  = '- [' . get_the_title( $cpt_post ) . '](' . get_permalink( $cpt_post->ID ) . ')'
    328332                    . ( $desc ? ': ' . $desc : '' );
    329333            }
    … …  
    332336
    333337        return implode( "\n", $lines );
     338    }
     339
     340    /**
     341     * Given a post in the default language, return its translated version
     342     * matching the current language context (WPML or Polylang).
     343     * Falls back to the original post if no translation exists.
     344     */
     345    private static function resolve_translated_post( WP_Post $post ): WP_Post {
     346        // WPML
     347        if ( function_exists( 'icl_object_id' ) ) {
     348            $lang = apply_filters( 'wpml_current_language', null ); // phpcs:ignore WordPress.NamingConventions.PrefixAllGlobals.NonPrefixedHooknameFound
     349            if ( $lang ) {
     350                $translated_id = icl_object_id( $post->ID, $post->post_type, false, $lang );
     351                if ( $translated_id && $translated_id !== $post->ID ) {
     352                    $translated = get_post( $translated_id );
     353                    if ( $translated instanceof WP_Post ) return $translated;
     354                }
     355            }
     356        }
     357        // Polylang
     358        if ( function_exists( 'pll_current_language' ) && function_exists( 'pll_get_post' ) ) {
     359            $lang = pll_current_language( 'slug' );
     360            if ( $lang ) {
     361                $translated_id = pll_get_post( $post->ID, $lang );
     362                if ( $translated_id && $translated_id !== $post->ID ) {
     363                    $translated = get_post( $translated_id );
     364                    if ( $translated instanceof WP_Post ) return $translated;
     365                }
     366            }
     367        }
     368        return $post;
    334369    }
    335370
    … …  
    367402
    368403        foreach ( $all_posts as $post ) {
     404            $post    = self::resolve_translated_post( $post );
    369405            $lines[] = Mescio_For_Agents_Markdown::post_to_markdown( $post );
    370406            $lines[] = '';

  • mescio-for-agents/trunk/includes/class-markdown-generator.php

    r3480063 r3480216  
    316316    /**
    317317     * Return the AI description or fall back to WP excerpt.
     318     *
     319     * On multilingual sites (WPML, Polylang) the $post object may be the
     320     * default-language post even when the index is being built for a translated
     321     * URL. We resolve the correct translated post ID so the excerpt is always
     322     * in the right language.
     323     *
     324     * Shortcodes (WPBakery, WoodMart, etc.) are stripped before returning so
     325     * that builder markup never leaks into the llms.txt index.
    318326     */
    319327    public static function get_excerpt( WP_Post $post ): string {
     328        // Resolve the translated post if a language plugin is active.
     329        $post = self::maybe_get_translated_post( $post );
     330
    320331        $ai = get_post_meta( $post->ID, self::META_DESCRIPTION, true );
    321         if ( $ai ) return wp_strip_all_tags( $ai );
    322         if ( $post->post_excerpt ) return wp_strip_all_tags( $post->post_excerpt );
    323         // Auto-trim from content
    324         $plain = wp_strip_all_tags( $post->post_content );
     332        if ( $ai ) return self::strip_shortcode_noise( wp_strip_all_tags( $ai ) );
     333
     334        if ( $post->post_excerpt ) {
     335            return self::strip_shortcode_noise( wp_strip_all_tags( $post->post_excerpt ) );
     336        }
     337
     338        // Auto-trim from content — strip shortcode tags first, then HTML.
     339        $plain = self::strip_shortcode_noise( $post->post_content );
     340        $plain = wp_strip_all_tags( $plain );
    325341        $plain = preg_replace( '/\s+/', ' ', $plain );
    326342        $words = explode( ' ', trim( $plain ) );
    327343        return implode( ' ', array_slice( $words, 0, 40 ) );
     344    }
     345
     346    /**
     347     * Strip page-builder shortcode tags (and their content) from a raw string.
     348     * Used by get_excerpt() so that llms.txt index lines are clean.
     349     *
     350     * This mirrors the logic in html_to_markdown() but operates on raw
     351     * post_content / post_excerpt before do_shortcode() is ever called.
     352     */
     353    private static function strip_shortcode_noise( string $text ): string {
     354        $builders = [ 'vc_', 'et_pb_', 'fl_', 'fusion_', 'cs_', 'x_', 'woodmart_' ];
     355        foreach ( $builders as $prefix ) {
     356            for ( $i = 0; $i < 10; $i++ ) {
     357                $before = $text;
     358                // Self-closing
     359                $text = preg_replace( '/\[' . $prefix . '[^\]]*\/\]/s', '', $text );
     360                // Open+content+close pairs
     361                $text = preg_replace( '/\[' . $prefix . '[^\]]*\].*?\[\/' . $prefix . '[^\]]*\]/s', '', $text );
     362                // Orphaned open/close tags
     363                $text = preg_replace( '/\[\/?(' . $prefix . ')[^\]]*\]/', '', $text );
     364                if ( $text === $before ) break;
     365            }
     366        }
     367        // SiteOrigin
     368        $text = preg_replace( '/\[\/?so[w]?_[^\]]*\]/s', '', $text );
     369        // Generic: any shortcode tag longer than 200 chars (base64 blobs)
     370        $text = preg_replace( '/\[[^\]]{200,}\]/', '', $text );
     371        // Any remaining shortcode bracket pairs
     372        $text = preg_replace( '/\[[^\]]+\]/', '', $text );
     373
     374        return trim( $text );
     375    }
     376
     377    /**
     378     * If WPML or Polylang is active, return the translated version of $post
     379     * matching the current language context. Falls back to $post unchanged.
     380     */
     381    private static function maybe_get_translated_post( WP_Post $post ): WP_Post {
     382        // WPML
     383        if ( function_exists( 'icl_object_id' ) ) {
     384            $lang        = apply_filters( 'wpml_current_language', null ); // phpcs:ignore WordPress.NamingConventions.PrefixAllGlobals.NonPrefixedHooknameFound
     385            $translated  = $lang ? icl_object_id( $post->ID, $post->post_type, false, $lang ) : null;
     386            if ( $translated && $translated !== $post->ID ) {
     387                $p = get_post( $translated );
     388                if ( $p instanceof WP_Post ) return $p;
     389            }
     390        }
     391        // Polylang
     392        if ( function_exists( 'pll_current_language' ) && function_exists( 'pll_get_post' ) ) {
     393            $lang       = pll_current_language( 'slug' );
     394            $translated = $lang ? pll_get_post( $post->ID, $lang ) : null;
     395            if ( $translated && $translated !== $post->ID ) {
     396                $p = get_post( $translated );
     397                if ( $p instanceof WP_Post ) return $p;
     398            }
     399        }
     400        return $post;
    328401    }
    329402

  • mescio-for-agents/trunk/mescio-for-agents.php

    r3480063 r3480216  
    44 * Plugin URI:        https://wordpress.org/plugins/mescio-for-agents/
    55 * Description:       Mescio for Agents serves your posts, pages and WooCommerce products as clean Markdown to AI agents and GPT crawlers — using HTTP content negotiation (Accept: text/markdown). Human visitors never notice a thing.
    6  * Version:           1.6.2
     6 * Version:           1.6.3
    77 * Requires at least: 6.0
    88 * Requires PHP:      8.0
    … …  
    5252
    5353    /** Plugin version — must match the Version header above. */
    54     const VERSION = '1.6.2';
     54    const VERSION = '1.6.3';
    5555
    5656    /** Post types served by default (filterable via mescio_enabled_post_types). */

  • mescio-for-agents/trunk/readme.txt

    r3480063 r3480216  
    55Tested up to: 6.9
    66Requires PHP: 8.0
    7 Stable tag: 1.6.2
     7Stable tag: 1.6.3
    88License: GPLv2 or later
    99License URI: https://www.gnu.org/licenses/gpl-2.0.html
    … …  
    133133== Changelog ==
    134134
    135 = 1.6.2 =
    136 * Fix: page builder shortcodes (WPBakery, WoodMart, Divi, ecc.) ora rimossi correttamente prima dell'esecuzione — elimina CSS/attributi blob dall'output Markdown
    137 * Fix: content negotiation (Accept: text/markdown) ora funziona correttamente su siti multilingua con WPML/Polylang
     135= 1.6.3 =
     136* Fix: shortcode builder tags (WPBakery, WoodMart, ecc.) rimossi anche dagli excerpt nel llms.txt index
     137* Fix: llms.txt ora serve titoli, excerpt e permalink nella lingua corretta su siti multilingua (WPML/Polylang)
     138* Fix: llms-full.txt ora serve il contenuto completo dei post nella lingua corretta (WPML/Polylang)
     139
    138140
    139141= 1.6.1 =
Note: See TracChangeset for help on using the changeset viewer.