telegram-md2html 1.0.0 → 1.0.2

package/dist/index.cjs

@@ -293,6 +293,8 @@ class MarkdownConverter {
     }
     /**
      * Wrap token content in HTML tags
+     * FIXED: Removed extra newlines that were being added around code blocks and quotes
+     * Previously added \n before and after, now returns clean tags without extra whitespace
      */
     wrapToken(type, content, language) {
         switch (type) {
@@ -316,7 +318,8 @@ class MarkdownConverter {
                 }
                 const escapedCode = this.options.escapeHtml ? escapeHtml(content) : content;
                 const langAttr = language ? ` class="language-${language}"` : '';
-                return `\n<pre><code${langAttr}>${escapedCode}</code></pre>\n`;
+                // FIXED: Removed \n before and after - now returns just the tag
+                return `<pre><code${langAttr}>${escapedCode}</code></pre>`;
             case 'link':
                 const url = language || '';
                 if (this.hasCustomLinkProcessor) {
@@ -326,9 +329,11 @@ class MarkdownConverter {
                 const escapedText = this.options.escapeHtml ? escapeHtml(content) : content;
                 return `<a href="${escapedUrl}">${escapedText}</a>`;
             case 'quote':
-                return `\n<blockquote>${content.trim()}</blockquote>\n`;
+                // FIXED: Removed \n before and after - now returns just the tag
+                return `<blockquote>${content.trim()}</blockquote>`;
             case 'expandable_quote':
-                return `\n<blockquote expandable>${content.trim()}</blockquote>\n`;
+                // FIXED: Removed \n before and after - now returns just the tag
+                return `<blockquote expandable>${content.trim()}</blockquote>`;
             default:
                 return content;
         }
@@ -360,6 +365,7 @@ class MarkdownConverter {
     }
     /**
      * Process blockquote markers
+     * FIXED: Removed extra newlines from the replacement strings
      */
     processBlockquoteMarkers(text) {
         let result = text;
@@ -367,13 +373,15 @@ class MarkdownConverter {
         const expandableQuoteRegex = /\[EXPANDABLE_QUOTE\](.*?)(?=\n|$)/g;
         result = result.replace(expandableQuoteRegex, (match, content) => {
             const processedContent = this.convertRecursive(content);
-            return `\n<blockquote expandable>${processedContent.trim()}</blockquote>\n`;
+            // FIXED: Removed \n before and after
+            return `<blockquote expandable>${processedContent.trim()}</blockquote>`;
         });
         // Replace regular quote markers (process content recursively)
         const quoteRegex = /\[QUOTE\](.*?)(?=\n|$)/g;
         result = result.replace(quoteRegex, (match, content) => {
             const processedContent = this.convertRecursive(content);
-            return `\n<blockquote>${processedContent.trim()}</blockquote>\n`;
+            // FIXED: Removed \n before and after
+            return `<blockquote>${processedContent.trim()}</blockquote>`;
         });
         return result;
     }
@@ -385,7 +393,8 @@ class MarkdownConverter {
     defaultCodeBlockProcessor(code, language) {
         const escapedCode = this.options.escapeHtml ? escapeHtml(code) : code;
         const langAttr = language ? ` class="language-${language}"` : '';
-        return `\n<pre><code${langAttr}>${escapedCode}</code></pre>\n`;
+        // FIXED: Removed \n before and after in default processor too
+        return `<pre><code${langAttr}>${escapedCode}</code></pre>`;
     }
 }
 

package/dist/index.d.mts

@@ -1,5 +1,5 @@
 import { MarkdownConverter } from './converter.js';
-import { ConvertOptions } from './types.js';
+import type { ConvertOptions } from './types.js';
 /**
  * Convert Telegram-style Markdown to HTML
  * @param text - Markdown text to convert

package/dist/index.d.ts

@@ -1,5 +1,5 @@
 import { MarkdownConverter } from './converter.js';
-import { ConvertOptions } from './types.js';
+import type { ConvertOptions } from './types.js';
 /**
  * Convert Telegram-style Markdown to HTML
  * @param text - Markdown text to convert

package/dist/index.esm.js

@@ -0,0 +1,419 @@
+class MarkdownTokenizer {
+    constructor(text) {
+        this.text = text;
+    }
+    /**
+     * Tokenize the markdown text
+     */
+    tokenize() {
+        const tokens = [];
+        let pos = 0;
+        const text = this.text;
+        while (pos < text.length) {
+            // Skip if inside code block
+            if (this.isInsideCodeBlock(text, pos)) {
+                pos++;
+                continue;
+            }
+            // Try to match each token type (from outermost to innermost)
+            const token = this.matchToken(pos);
+            if (token) {
+                tokens.push(token);
+                pos = token.end;
+            }
+            else {
+                pos++;
+            }
+        }
+        return tokens.sort((a, b) => a.start - b.start);
+    }
+    matchToken(start) {
+        const text = this.text;
+        const remaining = text.slice(start);
+        // Skip if we're inside a quote marker
+        if (remaining.startsWith('[QUOTE]') || remaining.startsWith('[EXPANDABLE_QUOTE]')) {
+            return null;
+        }
+        // Match code block (triple backticks) - highest priority
+        const codeBlockMatch = remaining.match(/^```(\w+)?\n([\s\S]*?)```/);
+        if (codeBlockMatch) {
+            return {
+                type: 'code_block',
+                content: codeBlockMatch[2],
+                language: codeBlockMatch[1],
+                start: start,
+                end: start + codeBlockMatch[0].length
+            };
+        }
+        // Match inline code
+        const inlineCodeMatch = remaining.match(/^`([^`\n]+)`/);
+        if (inlineCodeMatch && !this.isInsideInlineCode(text, start)) {
+            return {
+                type: 'inline_code',
+                content: inlineCodeMatch[1],
+                start: start,
+                end: start + inlineCodeMatch[0].length
+            };
+        }
+        // Match spoiler
+        const spoilerMatch = remaining.match(/^\|\|([^|\n]+?)\|\|/);
+        if (spoilerMatch) {
+            return {
+                type: 'spoiler',
+                content: spoilerMatch[1],
+                start: start,
+                end: start + spoilerMatch[0].length
+            };
+        }
+        // Match strikethrough
+        const strikethroughMatch = remaining.match(/^~~([^~\n]+?)~~/);
+        if (strikethroughMatch) {
+            return {
+                type: 'strikethrough',
+                content: strikethroughMatch[1],
+                start: start,
+                end: start + strikethroughMatch[0].length
+            };
+        }
+        // Match bold
+        const boldMatch = remaining.match(/^\*\*([^*\n]+?)\*\*/);
+        if (boldMatch) {
+            return {
+                type: 'bold',
+                content: boldMatch[1],
+                start: start,
+                end: start + boldMatch[0].length
+            };
+        }
+        // Match underline
+        const underlineMatch = remaining.match(/^__([^_\n]+?)__/);
+        if (underlineMatch) {
+            return {
+                type: 'underline',
+                content: underlineMatch[1],
+                start: start,
+                end: start + underlineMatch[0].length
+            };
+        }
+        // Match italic with asterisk
+        const italicAsteriskMatch = remaining.match(/^\*([^*\n][^*]*?)\*/);
+        if (italicAsteriskMatch && italicAsteriskMatch[1].trim().length > 0) {
+            // Don't match if it's part of bold (**)
+            if (start > 0 && text[start - 1] === '*' && start < text.length - 1 && text[start + 1] === '*') {
+                return null;
+            }
+            return {
+                type: 'italic',
+                content: italicAsteriskMatch[1],
+                start: start,
+                end: start + italicAsteriskMatch[0].length
+            };
+        }
+        // Match italic with underscore
+        const italicUnderscoreMatch = remaining.match(/^_([^_\n]+?)_/);
+        if (italicUnderscoreMatch && italicUnderscoreMatch[1].trim().length > 0) {
+            // Don't match if it's part of underline (__)
+            if (start > 0 && text[start - 1] === '_' && start < text.length - 1 && text[start + 1] === '_') {
+                return null;
+            }
+            return {
+                type: 'italic',
+                content: italicUnderscoreMatch[1],
+                start: start,
+                end: start + italicUnderscoreMatch[0].length
+            };
+        }
+        // Match link
+        const linkMatch = remaining.match(/^\[([^\]]+?)\]\(([^)]+?)\)/);
+        if (linkMatch) {
+            return {
+                type: 'link',
+                content: linkMatch[1],
+                start: start,
+                end: start + linkMatch[0].length,
+                language: linkMatch[2]
+            };
+        }
+        return null;
+    }
+    isInsideCodeBlock(text, position) {
+        // Check for code blocks
+        const codeBlockRegex = /```[\s\S]*?```/g;
+        let match;
+        while ((match = codeBlockRegex.exec(text)) !== null) {
+            if (position > match.index && position < match.index + match[0].length) {
+                // But allow matching the closing ``` itself
+                if (position >= match.index + match[0].length - 3) {
+                    return false;
+                }
+                return true;
+            }
+        }
+        return false;
+    }
+    isInsideInlineCode(text, position) {
+        // Check for inline code
+        const inlineCodeRegex = /`[^`\n]*`/g;
+        let match;
+        while ((match = inlineCodeRegex.exec(text)) !== null) {
+            if (position > match.index && position < match.index + match[0].length) {
+                // But allow matching the closing ` itself
+                if (position === match.index + match[0].length - 1) {
+                    return false;
+                }
+                return true;
+            }
+        }
+        return false;
+    }
+}
+
+/**
+ * Escapes HTML special characters (but not double-escape)
+ */
+function escapeHtml(text) {
+    if (!text)
+        return text;
+    // Replace & first (but not if it's already an entity)
+    let result = text.replace(/&(?!#?\w+;)/g, '&amp;');
+    result = result.replace(/</g, '&lt;');
+    result = result.replace(/>/g, '&gt;');
+    result = result.replace(/"/g, '&quot;');
+    result = result.replace(/'/g, '&#39;');
+    return result;
+}
+/**
+ * Escapes Telegram HTML special characters
+ */
+function escapeTelegramHtml(text) {
+    if (!text)
+        return text;
+    // For Telegram, we only need to escape &, <, >, and "
+    let result = text.replace(/&(?!#?\w+;)/g, '&amp;');
+    result = result.replace(/</g, '&lt;');
+    result = result.replace(/>/g, '&gt;');
+    result = result.replace(/"/g, '&quot;');
+    return result;
+}
+/**
+ * Appends missing code block delimiters
+ */
+function autoCloseCodeBlocks(text) {
+    // Count triple backticks
+    const tripleBacktickCount = (text.match(/```/g) || []).length;
+    // If odd number, add closing backticks
+    if (tripleBacktickCount % 2 === 1) {
+        return text + '\n```';
+    }
+    return text;
+}
+
+class MarkdownConverter {
+    constructor(options = {}) {
+        this.hasCustomLinkProcessor = !!options.linkProcessor;
+        this.hasCustomCodeBlockProcessor = !!options.codeBlockProcessor;
+        this.options = {
+            escapeHtml: options.escapeHtml ?? true,
+            autoCloseCodeBlocks: options.autoCloseCodeBlocks ?? true,
+            linkProcessor: options.linkProcessor || this.defaultLinkProcessor.bind(this),
+            codeBlockProcessor: options.codeBlockProcessor || this.defaultCodeBlockProcessor.bind(this)
+        };
+    }
+    /**
+     * Convert markdown text to Telegram HTML
+     */
+    convert(text) {
+        // Auto-close code blocks if enabled
+        let processedText = this.options.autoCloseCodeBlocks
+            ? autoCloseCodeBlocks(text)
+            : text;
+        // First pass: convert blockquotes (they should be at line starts)
+        processedText = this.preprocessBlockquotes(processedText);
+        // Convert the text recursively
+        let result = this.convertRecursive(processedText);
+        // Process blockquote markers
+        result = this.processBlockquoteMarkers(result);
+        // Only trim if there's actual content (not just whitespace)
+        if (result.trim() === '') {
+            return text; // Return original text (spaces) if result is empty
+        }
+        return result.trim();
+    }
+    /**
+     * Recursively convert markdown, handling nested styles
+     */
+    convertRecursive(text, depth = 0) {
+        if (depth > 10)
+            return text; // Prevent infinite recursion
+        // Tokenize the text
+        const tokenizer = new MarkdownTokenizer(text);
+        const tokens = tokenizer.tokenize();
+        // If no tokens found, return the text as-is (with HTML escaping)
+        if (tokens.length === 0) {
+            return this.options.escapeHtml ? escapeTelegramHtml(text) : text;
+        }
+        let result = '';
+        let lastPos = 0;
+        for (const token of tokens) {
+            // Add text before token
+            if (token.start > lastPos) {
+                const textBefore = text.slice(lastPos, token.start);
+                result += this.options.escapeHtml ? escapeTelegramHtml(textBefore) : textBefore;
+            }
+            // Handle code blocks specially (no recursive parsing inside)
+            if (token.type === 'code_block') {
+                const codeContent = this.options.escapeHtml ? escapeHtml(token.content) : token.content;
+                result += this.wrapToken(token.type, codeContent, token.language);
+                lastPos = token.end;
+                continue;
+            }
+            // Handle inline code specially (no recursive parsing inside)
+            if (token.type === 'inline_code') {
+                const codeContent = this.options.escapeHtml ? escapeHtml(token.content) : token.content;
+                result += `<code>${codeContent}</code>`;
+                lastPos = token.end;
+                continue;
+            }
+            // Process other token content recursively
+            const tokenContent = this.convertRecursive(token.content, depth + 1);
+            // Wrap the content in appropriate HTML tags
+            result += this.wrapToken(token.type, tokenContent, token.language);
+            lastPos = token.end;
+        }
+        // Add remaining text
+        if (lastPos < text.length) {
+            const remainingText = text.slice(lastPos);
+            result += this.options.escapeHtml ? escapeTelegramHtml(remainingText) : remainingText;
+        }
+        return result;
+    }
+    /**
+     * Wrap token content in HTML tags
+     * FIXED: Removed extra newlines that were being added around code blocks and quotes
+     * Previously added \n before and after, now returns clean tags without extra whitespace
+     */
+    wrapToken(type, content, language) {
+        switch (type) {
+            case 'bold':
+                return `<b>${content}</b>`;
+            case 'italic':
+                return `<i>${content}</i>`;
+            case 'underline':
+                return `<u>${content}</u>`;
+            case 'strikethrough':
+                return `<s>${content}</s>`;
+            case 'spoiler':
+                return `<span class="tg-spoiler">${content}</span>`;
+            case 'inline_code':
+                // Already handled above
+                return `<code>${content}</code>`;
+            case 'code_block':
+                // Already handled above, but handle custom processor
+                if (this.hasCustomCodeBlockProcessor) {
+                    return this.options.codeBlockProcessor(content, language);
+                }
+                const escapedCode = this.options.escapeHtml ? escapeHtml(content) : content;
+                const langAttr = language ? ` class="language-${language}"` : '';
+                // FIXED: Removed \n before and after - now returns just the tag
+                return `<pre><code${langAttr}>${escapedCode}</code></pre>`;
+            case 'link':
+                const url = language || '';
+                if (this.hasCustomLinkProcessor) {
+                    return this.options.linkProcessor(url, content);
+                }
+                const escapedUrl = this.options.escapeHtml ? escapeHtml(url) : url;
+                const escapedText = this.options.escapeHtml ? escapeHtml(content) : content;
+                return `<a href="${escapedUrl}">${escapedText}</a>`;
+            case 'quote':
+                // FIXED: Removed \n before and after - now returns just the tag
+                return `<blockquote>${content.trim()}</blockquote>`;
+            case 'expandable_quote':
+                // FIXED: Removed \n before and after - now returns just the tag
+                return `<blockquote expandable>${content.trim()}</blockquote>`;
+            default:
+                return content;
+        }
+    }
+    /**
+     * Preprocess blockquotes to mark them before other parsing
+     */
+    preprocessBlockquotes(text) {
+        const lines = text.split('\n');
+        const processedLines = [];
+        for (const line of lines) {
+            const trimmedLine = line.trim();
+            // Only treat lines starting with > at the beginning of line as blockquotes
+            if (trimmedLine.startsWith('**>')) {
+                // Expandable blockquote
+                const content = trimmedLine.substring(3).trim();
+                processedLines.push(`[EXPANDABLE_QUOTE]${content}`);
+            }
+            else if (trimmedLine.startsWith('>')) {
+                // Regular blockquote
+                const content = trimmedLine.substring(1).trim();
+                processedLines.push(`[QUOTE]${content}`);
+            }
+            else {
+                processedLines.push(line);
+            }
+        }
+        return processedLines.join('\n');
+    }
+    /**
+     * Process blockquote markers
+     * FIXED: Removed extra newlines from the replacement strings
+     */
+    processBlockquoteMarkers(text) {
+        let result = text;
+        // Replace expandable quote markers (process content recursively)
+        const expandableQuoteRegex = /\[EXPANDABLE_QUOTE\](.*?)(?=\n|$)/g;
+        result = result.replace(expandableQuoteRegex, (match, content) => {
+            const processedContent = this.convertRecursive(content);
+            // FIXED: Removed \n before and after
+            return `<blockquote expandable>${processedContent.trim()}</blockquote>`;
+        });
+        // Replace regular quote markers (process content recursively)
+        const quoteRegex = /\[QUOTE\](.*?)(?=\n|$)/g;
+        result = result.replace(quoteRegex, (match, content) => {
+            const processedContent = this.convertRecursive(content);
+            // FIXED: Removed \n before and after
+            return `<blockquote>${processedContent.trim()}</blockquote>`;
+        });
+        return result;
+    }
+    defaultLinkProcessor(url, text) {
+        const escapedUrl = this.options.escapeHtml ? escapeHtml(url) : url;
+        const escapedText = this.options.escapeHtml ? escapeHtml(text) : text;
+        return `<a href="${escapedUrl}">${escapedText}</a>`;
+    }
+    defaultCodeBlockProcessor(code, language) {
+        const escapedCode = this.options.escapeHtml ? escapeHtml(code) : code;
+        const langAttr = language ? ` class="language-${language}"` : '';
+        // FIXED: Removed \n before and after in default processor too
+        return `<pre><code${langAttr}>${escapedCode}</code></pre>`;
+    }
+}
+
+/**
+ * Convert Telegram-style Markdown to HTML
+ * @param text - Markdown text to convert
+ * @param options - Conversion options
+ * @returns Telegram-compatible HTML
+ */
+function markdownToHtml(text, options) {
+    const converter = new MarkdownConverter(options);
+    return converter.convert(text);
+}
+/**
+ * Create a converter instance with custom options
+ */
+function createConverter(options) {
+    return new MarkdownConverter(options);
+}
+var index = {
+    markdownToHtml,
+    createConverter,
+    MarkdownConverter
+};
+
+export { MarkdownConverter, createConverter, index as default, markdownToHtml };

package/dist/index.mjs

@@ -289,6 +289,8 @@ class MarkdownConverter {
     }
     /**
      * Wrap token content in HTML tags
+     * FIXED: Removed extra newlines that were being added around code blocks and quotes
+     * Previously added \n before and after, now returns clean tags without extra whitespace
      */
     wrapToken(type, content, language) {
         switch (type) {
@@ -312,7 +314,8 @@ class MarkdownConverter {
                 }
                 const escapedCode = this.options.escapeHtml ? escapeHtml(content) : content;
                 const langAttr = language ? ` class="language-${language}"` : '';
-                return `\n<pre><code${langAttr}>${escapedCode}</code></pre>\n`;
+                // FIXED: Removed \n before and after - now returns just the tag
+                return `<pre><code${langAttr}>${escapedCode}</code></pre>`;
             case 'link':
                 const url = language || '';
                 if (this.hasCustomLinkProcessor) {
@@ -322,9 +325,11 @@ class MarkdownConverter {
                 const escapedText = this.options.escapeHtml ? escapeHtml(content) : content;
                 return `<a href="${escapedUrl}">${escapedText}</a>`;
             case 'quote':
-                return `\n<blockquote>${content.trim()}</blockquote>\n`;
+                // FIXED: Removed \n before and after - now returns just the tag
+                return `<blockquote>${content.trim()}</blockquote>`;
             case 'expandable_quote':
-                return `\n<blockquote expandable>${content.trim()}</blockquote>\n`;
+                // FIXED: Removed \n before and after - now returns just the tag
+                return `<blockquote expandable>${content.trim()}</blockquote>`;
             default:
                 return content;
         }
@@ -356,6 +361,7 @@ class MarkdownConverter {
     }
     /**
      * Process blockquote markers
+     * FIXED: Removed extra newlines from the replacement strings
      */
     processBlockquoteMarkers(text) {
         let result = text;
@@ -363,13 +369,15 @@ class MarkdownConverter {
         const expandableQuoteRegex = /\[EXPANDABLE_QUOTE\](.*?)(?=\n|$)/g;
         result = result.replace(expandableQuoteRegex, (match, content) => {
             const processedContent = this.convertRecursive(content);
-            return `\n<blockquote expandable>${processedContent.trim()}</blockquote>\n`;
+            // FIXED: Removed \n before and after
+            return `<blockquote expandable>${processedContent.trim()}</blockquote>`;
         });
         // Replace regular quote markers (process content recursively)
         const quoteRegex = /\[QUOTE\](.*?)(?=\n|$)/g;
         result = result.replace(quoteRegex, (match, content) => {
             const processedContent = this.convertRecursive(content);
-            return `\n<blockquote>${processedContent.trim()}</blockquote>\n`;
+            // FIXED: Removed \n before and after
+            return `<blockquote>${processedContent.trim()}</blockquote>`;
         });
         return result;
     }
@@ -381,7 +389,8 @@ class MarkdownConverter {
     defaultCodeBlockProcessor(code, language) {
         const escapedCode = this.options.escapeHtml ? escapeHtml(code) : code;
         const langAttr = language ? ` class="language-${language}"` : '';
-        return `\n<pre><code${langAttr}>${escapedCode}</code></pre>\n`;
+        // FIXED: Removed \n before and after in default processor too
+        return `<pre><code${langAttr}>${escapedCode}</code></pre>`;
     }
 }