n8n-nodes-notion-advanced 1.2.28-beta → 1.2.30-beta

package/nodes/NotionAdvanced/NotionAITool.node.d.ts

@@ -55,6 +55,13 @@ export declare class NotionAITool implements INodeType {
     static processNestedHtmlInListItem(content: string): string;
     static convertInlineHtmlToMarkdown(content: string): string;
     static processNestedList(listContent: string, listType: 'bulleted_list_item' | 'numbered_list_item', blocks: IDataObject[]): void;
+    static extractListItemsWithBranching(content: string): Array<{
+        text: string;
+        children: Array<{
+            type: string;
+            content: string;
+        }>;
+    }>;
     static extractListItems(content: string): string[];
     static getCalloutEmoji(type: string): string;
     static getCalloutColor(type: string): string;

package/nodes/NotionAdvanced/NotionAITool.node.js

@@ -745,40 +745,77 @@ class NotionAITool {
             return charIndex;
         }
     }
-    // Enhanced hierarchical XML tree structure that catches ALL XML content
+    // Enhanced hierarchical XML tree structure using depth-aware parsing
     static buildXMLTree(content, tagProcessors) {
         var _a;
         const allMatches = [];
-        const processedRanges = [];
-        // Step 1: Collect all XML tags with specific processors
+        // Step 1: Use depth-aware parsing for each tag processor
         tagProcessors.forEach(({ regex, blockCreator, listProcessor }) => {
             var _a;
-            const globalRegex = new RegExp(regex.source, 'gis');
-            let match;
-            while ((match = globalRegex.exec(content)) !== null) {
-                const tagName = ((_a = match[0].match(/<(\w+)/)) === null || _a === void 0 ? void 0 : _a[1]) || 'unknown';
-                const xmlNode = {
-                    id: `${tagName}_${match.index}_${Date.now()}_${Math.random()}`,
-                    tagName,
-                    start: match.index,
-                    end: match.index + match[0].length,
-                    match: match[0],
-                    processor: blockCreator,
-                    groups: match.slice(1),
-                    children: [],
-                    depth: 0,
-                    innerContent: match[0],
-                    replacement: undefined,
-                    listProcessor
-                };
-                allMatches.push(xmlNode);
-                processedRanges.push({ start: xmlNode.start, end: xmlNode.end });
+            const tagPattern = (_a = regex.source.match(/<(\w+)/)) === null || _a === void 0 ? void 0 : _a[1];
+            if (!tagPattern)
+                return;
+            // Find all opening tags of this type
+            let pos = 0;
+            while (pos < content.length) {
+                const openTagStart = content.indexOf(`<${tagPattern}`, pos);
+                if (openTagStart === -1)
+                    break;
+                const openTagEnd = content.indexOf('>', openTagStart);
+                if (openTagEnd === -1)
+                    break;
+                // Find matching closing tag using depth tracking
+                let depth = 1;
+                let searchPos = openTagEnd + 1;
+                let closeTagStart = -1;
+                const openPattern = `<${tagPattern}`;
+                const closePattern = `</${tagPattern}>`;
+                while (searchPos < content.length && depth > 0) {
+                    const nextOpen = content.indexOf(openPattern, searchPos);
+                    const nextClose = content.indexOf(closePattern, searchPos);
+                    if (nextClose === -1)
+                        break;
+                    if (nextOpen !== -1 && nextOpen < nextClose) {
+                        // Found nested opening tag
+                        depth++;
+                        searchPos = nextOpen + openPattern.length;
+                    }
+                    else {
+                        // Found closing tag
+                        depth--;
+                        if (depth === 0) {
+                            closeTagStart = nextClose;
+                            break;
+                        }
+                        searchPos = nextClose + closePattern.length;
+                    }
+                }
+                if (closeTagStart !== -1) {
+                    const fullMatch = content.substring(openTagStart, closeTagStart + closePattern.length);
+                    const innerContent = content.substring(openTagEnd + 1, closeTagStart);
+                    const xmlNode = {
+                        id: `${tagPattern}_${openTagStart}_${Date.now()}_${Math.random()}`,
+                        tagName: tagPattern,
+                        start: openTagStart,
+                        end: closeTagStart + closePattern.length,
+                        match: fullMatch,
+                        processor: blockCreator,
+                        groups: [innerContent], // For list processors, group[0] is the inner content
+                        children: [],
+                        depth: 0,
+                        innerContent,
+                        replacement: undefined,
+                        listProcessor
+                    };
+                    allMatches.push(xmlNode);
+                }
+                pos = openTagEnd + 1;
             }
         });
         // Step 2: Catch ANY remaining XML/HTML tags that weren't processed by specific processors
-        // This prevents ANY XML content from falling through to traditional processing
         const genericXmlRegex = /<[^>]+>[\s\S]*?<\/[^>]+>|<[^>]+\/>/gis;
         let genericMatch;
+        const processedRanges = allMatches.map(node => ({ start: node.start, end: node.end }));
         while ((genericMatch = genericXmlRegex.exec(content)) !== null) {
             const matchStart = genericMatch.index;
             const matchEnd = genericMatch.index + genericMatch[0].length;
@@ -792,7 +829,7 @@ class NotionAITool {
                     start: matchStart,
                     end: matchEnd,
                     match: genericMatch[0],
-                    processor: () => null, // Generic processor that just removes the content
+                    processor: () => null,
                     groups: [],
                     children: [],
                     depth: 0,
@@ -801,12 +838,11 @@ class NotionAITool {
                     listProcessor: undefined
                 };
                 allMatches.push(xmlNode);
-                processedRanges.push({ start: matchStart, end: matchEnd });
             }
         }
         // Sort by start position to maintain document order
         allMatches.sort((a, b) => a.start - b.start);
-        // Build parent-child relationships while preserving ordering
+        // Build parent-child relationships
         const rootNodes = [];
         const nodeStack = [];
         for (const node of allMatches) {
@@ -826,7 +862,7 @@ class NotionAITool {
                 // This is a root node
                 rootNodes.push(node);
             }
-            // Only push self-contained tags to stack (not self-closing)
+            // Push to stack for potential children
             if (!node.match.endsWith('/>') && node.match.includes('</')) {
                 nodeStack.push(node);
             }
@@ -1179,23 +1215,9 @@ class NotionAITool {
                     };
                 }
             },
-            // Standalone list items (only if not already processed in lists): <li>content</li>
-            {
-                regex: /<li\s*[^>]*>(.*?)<\/li>/gis,
-                blockCreator: (content) => {
-                    if (content.trim()) {
-                        // Convert HTML to markdown first, then parse to rich text
-                        const markdownContent = NotionAITool.convertInlineHtmlToMarkdown(content.trim());
-                        return {
-                            type: 'bulleted_list_item',
-                            bulleted_list_item: {
-                                rich_text: NotionAITool.parseBasicMarkdown(markdownContent),
-                            },
-                        };
-                    }
-                    return null;
-                }
-            },
+            // REMOVED: Standalone <li> processor
+            // <li> tags should ONLY be processed within <ul>/<ol> contexts via the list processors above
+            // Having a standalone <li> processor causes XML fragments and double processing
             // Line breaks: <br/> or <br>
             {
                 regex: /<br\s*\/?>/gis,
@@ -1477,61 +1499,45 @@ class NotionAITool {
         processed = processed.replace(/\s+/g, ' ').trim();
         return processed;
     }
-    // Helper function to process nested lists and flatten them for Notion
+    // Helper function to process lists using branch-based approach
+    // Each <ul> and <ol> represents a new branch that contains children
     static processNestedList(listContent, listType, blocks) {
         try {
-            // More robust list item extraction that handles nested <li> tags properly
-            const listItems = NotionAITool.extractListItems(listContent);
-            for (const itemContent of listItems) {
-                if (!itemContent.trim())
+            // Process each <li> element as a potential branch point
+            const listItems = NotionAITool.extractListItemsWithBranching(listContent);
+            for (const item of listItems) {
+                if (!item.text && !item.children.length)
                     continue;
-                // Check if this item contains nested lists
-                const hasNestedList = /<[uo]l\s*[^>]*>/i.test(itemContent);
-                if (hasNestedList) {
-                    // Split content into text parts and nested list parts
-                    const parts = itemContent.split(/(<[uo]l\s*[^>]*>[\s\S]*?<\/[uo]l>)/gi);
-                    for (let i = 0; i < parts.length; i++) {
-                        const part = parts[i].trim();
-                        if (!part)
-                            continue;
-                        // Check if this part is a nested list
-                        const isNestedList = /<[uo]l\s*[^>]*>[\s\S]*?<\/[uo]l>/gi.test(part);
-                        if (isNestedList) {
-                            // Process the nested list
-                            const nestedListMatch = part.match(/<([uo]l)\s*[^>]*>([\s\S]*?)<\/\1>/i);
-                            if (nestedListMatch) {
-                                const [, listTag, innerContent] = nestedListMatch;
-                                const nestedListType = listTag === 'ul' ? 'bulleted_list_item' : 'numbered_list_item';
-                                // Recursively process nested list
-                                NotionAITool.processNestedList(innerContent, nestedListType, blocks);
-                            }
-                        }
-                        else {
-                            // This is text content - clean it and add as a list item
-                            // Only process non-empty text parts as separate list items
-                            const cleanContent = NotionAITool.processNestedHtmlInListItem(part);
-                            if (cleanContent) {
-                                blocks.push({
-                                    type: listType,
-                                    [listType]: {
-                                        rich_text: NotionAITool.parseBasicMarkdown(cleanContent),
-                                    },
-                                });
-                            }
-                        }
+                // Create list item block
+                const listItemBlock = {
+                    type: listType,
+                    [listType]: {
+                        rich_text: [],
+                    },
+                };
+                // Add parent text if present
+                if (item.text && item.text.trim()) {
+                    const cleanText = NotionAITool.processNestedHtmlInListItem(item.text);
+                    if (cleanText) {
+                        listItemBlock[listType].rich_text = NotionAITool.parseBasicMarkdown(cleanText);
                     }
                 }
-                else {
-                    // Simple item without nested lists
-                    const cleanContent = NotionAITool.processNestedHtmlInListItem(itemContent);
-                    if (cleanContent) {
-                        blocks.push({
-                            type: listType,
-                            [listType]: {
-                                rich_text: NotionAITool.parseBasicMarkdown(cleanContent),
-                            },
-                        });
+                // Process child branches and add them as nested children
+                if (item.children.length > 0) {
+                    const childBlocks = [];
+                    for (const child of item.children) {
+                        const childListType = child.type === 'ul' ? 'bulleted_list_item' : 'numbered_list_item';
+                        NotionAITool.processNestedList(child.content, childListType, childBlocks);
                     }
+                    // Add children to the parent block
+                    if (childBlocks.length > 0) {
+                        listItemBlock[listType].children = childBlocks;
+                    }
+                }
+                // Only add the block if it has text or children
+                const listData = listItemBlock[listType];
+                if ((listData.rich_text && listData.rich_text.length > 0) || listData.children) {
+                    blocks.push(listItemBlock);
                 }
             }
         }
@@ -1546,58 +1552,216 @@ class NotionAITool {
             });
         }
     }
-    // Helper function to properly extract list items handling nested <li> tags
-    static extractListItems(content) {
+    // Extract list items with proper branching structure - only process top-level <li> tags
+    static extractListItemsWithBranching(content) {
         const items = [];
-        let currentPos = 0;
-        while (currentPos < content.length) {
-            // Find the next <li> opening tag
-            const liStart = content.indexOf('<li', currentPos);
+        let pos = 0;
+        while (pos < content.length) {
+            // Find next <li> tag at the current level
+            const liStart = content.indexOf('<li', pos);
             if (liStart === -1)
                 break;
-            // Find the end of the opening tag
-            const openTagEnd = content.indexOf('>', liStart);
-            if (openTagEnd === -1)
+            const liOpenEnd = content.indexOf('>', liStart);
+            if (liOpenEnd === -1)
                 break;
-            // Now find the matching closing </li> tag accounting for nesting
-            let depth = 1;
-            let pos = openTagEnd + 1;
-            let itemEnd = -1;
-            while (pos < content.length && depth > 0) {
-                const nextLiOpen = content.indexOf('<li', pos);
-                const nextLiClose = content.indexOf('</li>', pos);
-                // If no more closing tags, we're done
+            // Find the matching </li> using proper depth tracking for nested tags
+            let depth = 0;
+            let searchPos = liOpenEnd + 1; // Start after the opening <li> tag
+            let liEnd = -1;
+            while (searchPos < content.length) {
+                const nextLiOpen = content.indexOf('<li', searchPos);
+                const nextLiClose = content.indexOf('</li>', searchPos);
+                // Handle case where no more closing tags
                 if (nextLiClose === -1)
                     break;
-                // If there's an opening tag before the next closing tag, increase depth
+                // If there's an opening tag before the next closing tag
                 if (nextLiOpen !== -1 && nextLiOpen < nextLiClose) {
                     depth++;
-                    pos = nextLiOpen + 3; // Move past '<li'
+                    searchPos = nextLiOpen + 3; // Move past '<li'
                 }
                 else {
                     // Found a closing tag
-                    depth--;
                     if (depth === 0) {
-                        itemEnd = nextLiClose + 5; // Include the '</li>'
+                        // This is our matching closing tag
+                        liEnd = nextLiClose;
+                        break;
+                    }
+                    else {
+                        // This closing tag belongs to a nested li
+                        depth--;
+                        searchPos = nextLiClose + 5; // Move past '</li>'
+                    }
+                }
+            }
+            if (liEnd === -1) {
+                // No matching closing tag found
+                pos = liOpenEnd + 1;
+                continue;
+            }
+            // Extract the content between <li> and </li>
+            const fullItemContent = content.substring(liOpenEnd + 1, liEnd);
+            if (!fullItemContent.trim()) {
+                pos = liEnd + 5;
+                continue;
+            }
+            const item = { text: '', children: [] };
+            // Process the content to separate text from nested lists
+            let contentPos = 0;
+            let textParts = [];
+            while (contentPos < fullItemContent.length) {
+                // Look for the next nested list (ul or ol)
+                const nextUlStart = fullItemContent.indexOf('<ul', contentPos);
+                const nextOlStart = fullItemContent.indexOf('<ol', contentPos);
+                let nextListStart = -1;
+                let listType = '';
+                if (nextUlStart !== -1 && (nextOlStart === -1 || nextUlStart < nextOlStart)) {
+                    nextListStart = nextUlStart;
+                    listType = 'ul';
+                }
+                else if (nextOlStart !== -1) {
+                    nextListStart = nextOlStart;
+                    listType = 'ol';
+                }
+                if (nextListStart === -1) {
+                    // No more nested lists - add remaining text
+                    const remainingText = fullItemContent.substring(contentPos);
+                    if (remainingText.trim()) {
+                        textParts.push(remainingText);
+                    }
+                    break;
+                }
+                // Add text before the nested list
+                const textBefore = fullItemContent.substring(contentPos, nextListStart);
+                if (textBefore.trim()) {
+                    textParts.push(textBefore);
+                }
+                // Find the end of this nested list
+                const listOpenEnd = fullItemContent.indexOf('>', nextListStart);
+                if (listOpenEnd === -1) {
+                    // Malformed list tag
+                    textParts.push(fullItemContent.substring(contentPos));
+                    break;
+                }
+                // Track depth to find the matching closing tag
+                let listDepth = 1;
+                let listSearchPos = listOpenEnd + 1;
+                let listEnd = -1;
+                const openTag = `<${listType}`;
+                const closeTag = `</${listType}>`;
+                while (listSearchPos < fullItemContent.length && listDepth > 0) {
+                    const nextListOpen = fullItemContent.indexOf(openTag, listSearchPos);
+                    const nextListClose = fullItemContent.indexOf(closeTag, listSearchPos);
+                    if (nextListClose === -1)
                         break;
+                    if (nextListOpen !== -1 && nextListOpen < nextListClose) {
+                        listDepth++;
+                        listSearchPos = nextListOpen + openTag.length;
                     }
                     else {
-                        pos = nextLiClose + 5; // Move past '</li>'
+                        listDepth--;
+                        if (listDepth === 0) {
+                            listEnd = nextListClose + closeTag.length;
+                            break;
+                        }
+                        listSearchPos = nextListClose + closeTag.length;
                     }
                 }
+                if (listEnd !== -1) {
+                    // Extract the content between <ul>/<ol> and </ul>/<ol>
+                    const listContent = fullItemContent.substring(listOpenEnd + 1, listEnd - closeTag.length);
+                    item.children.push({
+                        type: listType,
+                        content: listContent
+                    });
+                    contentPos = listEnd;
+                }
+                else {
+                    // Malformed nested list - treat remaining as text
+                    textParts.push(fullItemContent.substring(contentPos));
+                    break;
+                }
             }
-            if (itemEnd !== -1) {
-                // Extract the content between <li...> and </li>
-                const fullMatch = content.substring(liStart, itemEnd);
-                const innerMatch = fullMatch.match(/<li[^>]*>([\s\S]*)<\/li>$/);
-                if (innerMatch) {
-                    items.push(innerMatch[1]);
+            // Combine all text parts and clean them
+            if (textParts.length > 0) {
+                const combinedText = textParts.join(' ').trim();
+                const cleanText = NotionAITool.processNestedHtmlInListItem(combinedText);
+                if (cleanText) {
+                    item.text = cleanText;
                 }
-                currentPos = itemEnd;
             }
-            else {
-                // Malformed HTML, skip this tag
-                currentPos = openTagEnd + 1;
+            // Only add items that have either text or children
+            if (item.text.trim() || item.children.length > 0) {
+                items.push(item);
+            }
+            pos = liEnd + 5; // Move past </li>
+        }
+        return items;
+    }
+    // Helper function to properly extract list items handling nested <li> tags
+    static extractListItems(content) {
+        const items = [];
+        // Use a more robust regex approach that respects nesting
+        // This regex captures the complete <li>...</li> blocks including nested content
+        const liRegex = /<li[^>]*>((?:[^<]|<(?!\/li>))*?(?:<[uo]l[^>]*>[\s\S]*?<\/[uo]l>(?:[^<]|<(?!\/li>))*?)*?)<\/li>/gi;
+        let match;
+        while ((match = liRegex.exec(content)) !== null) {
+            const itemContent = match[1];
+            if (itemContent && itemContent.trim()) {
+                items.push(itemContent.trim());
+            }
+        }
+        // Fallback to the old depth-tracking method if regex fails
+        if (items.length === 0) {
+            let currentPos = 0;
+            while (currentPos < content.length) {
+                // Find the next <li> opening tag
+                const liStart = content.indexOf('<li', currentPos);
+                if (liStart === -1)
+                    break;
+                // Find the end of the opening tag
+                const openTagEnd = content.indexOf('>', liStart);
+                if (openTagEnd === -1)
+                    break;
+                // Now find the matching closing </li> tag accounting for nesting
+                let depth = 1;
+                let pos = openTagEnd + 1;
+                let itemEnd = -1;
+                while (pos < content.length && depth > 0) {
+                    const nextLiOpen = content.indexOf('<li', pos);
+                    const nextLiClose = content.indexOf('</li>', pos);
+                    // If no more closing tags, we're done
+                    if (nextLiClose === -1)
+                        break;
+                    // If there's an opening tag before the next closing tag, increase depth
+                    if (nextLiOpen !== -1 && nextLiOpen < nextLiClose) {
+                        depth++;
+                        pos = nextLiOpen + 3; // Move past '<li'
+                    }
+                    else {
+                        // Found a closing tag
+                        depth--;
+                        if (depth === 0) {
+                            itemEnd = nextLiClose + 5; // Include the '</li>'
+                            break;
+                        }
+                        else {
+                            pos = nextLiClose + 5; // Move past '</li>'
+                        }
+                    }
+                }
+                if (itemEnd !== -1) {
+                    // Extract the content between <li...> and </li>
+                    const fullMatch = content.substring(liStart, itemEnd);
+                    const innerMatch = fullMatch.match(/<li[^>]*>([\s\S]*)<\/li>$/);
+                    if (innerMatch) {
+                        items.push(innerMatch[1]);
+                    }
+                    currentPos = itemEnd;
+                }
+                else {
+                    // Malformed HTML, skip this tag
+                    currentPos = openTagEnd + 1;
+                }
             }
         }
         return items;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "n8n-nodes-notion-advanced",
-  "version": "1.2.28-beta",
+  "version": "1.2.30-beta",
   "description": "Advanced n8n Notion nodes: Full-featured workflow node + AI Agent Tool for intelligent Notion automation with 25+ block types (BETA)",
   "scripts": {},
   "files": [