Function parse_markdown_with_images

examples/run_leetcode_agent.py:59–131

Parse markdown text and convert it to a message content format that supports images. Supports standard markdown image syntax: -  - image without alt text -  - image with alt text Example: Input: "**Example:**\n\n![](https://example.com/i

(markdown_text: str)

    result: str = Field(description="The generated code")

def parse_markdown_with_images(markdown_text: str) -> Union[str, list]:
    """
    Parse markdown text and convert it to a message content format that supports images.
    
    Supports standard markdown image syntax:
    - ![](url) - image without alt text
    - ![alt text](url) - image with alt text
    
    Example:
        Input: "**Example:**\n\n![](https://example.com/image.jpg)\n\nText after"
        Output: [
            ContentPartText(text="**Example:**\n\n"),
            ContentPartImage(image_url=ImageURL(url="https://example.com/image.jpg", ...)),
            ContentPartText(text="\n\nText after")
        ]
    
    Args:
        markdown_text: The markdown text that may contain image references like ![](url)
    
    Returns:
        If no images found: returns the original string
        If images found: returns a list of ContentPartText and ContentPartImage objects
    """
    # Pattern to match markdown image syntax: ![](url) or ![alt](url)
    # Matches: ![optional alt text](image_url)
    image_pattern = r'!\[([^\]]*)\]\(([^)]+)\)'
    
    # Find all image matches with their positions
    matches = list(re.finditer(image_pattern, markdown_text))
    
    if not matches:
        # No images found, return as plain string
        return markdown_text
    
    # Build content list with text and image parts
    content_parts = []
    last_end = 0
    
    for match in matches:
        # Add text before the image (including whitespace)
        text_before = markdown_text[last_end:match.start()]
        # Only add non-empty text parts (preserve whitespace if it's meaningful)
        if text_before:
            content_parts.append(ContentPartText(text=text_before))
        
        # Extract image URL (group 2 is the URL in parentheses)
        image_url = match.group(2)
        
        # Determine media type from URL extension (handle URLs with query parameters)
        media_type = 'image/png'  # default fallback
        # Extract path before query parameters (e.g., "image.jpg?v=1" -> "image.jpg")
        url_path = image_url.split('?')[0].lower()
        if url_path.endswith('.jpg') or url_path.endswith('.jpeg'):
            media_type = 'image/jpeg'
        elif url_path.endswith('.png'):
            media_type = 'image/png'
        elif url_path.endswith('.gif'):
            media_type = 'image/gif'