[BUG]: 按照逻辑，目前的markdown中如果包含了图片，可能会触发报错的问题

[pengjunfeng11]

🐛 Bug description [Please make everyone to understand it]

针对 opencontext/context_processing/processor/document_processor.py下的_process_document_page_by_page支持了.md文档，但是在具体的_extract_vlm_pages方法中又不支持md文档中的图片做处理的逻辑进行修复。 具体代码如下：

    def _process_document_page_by_page(self, raw_context: RawContextProperties, file_path: str, file_ext: str) -> List[ProcessedContext]:
		.......


        # 1. Analyze pages
        if file_ext == ".pdf":
            page_infos = self._document_converter.analyze_pdf_pages(file_path, self._text_threshold)
        elif file_ext in [".docx", ".doc"]:
            page_infos = self._document_converter.analyze_docx_pages(file_path)
        elif file_ext == ".md":
            page_infos = self._document_converter.analyze_markdown_pages(file_path)
        elif file_ext == ".txt":
            return self._process_txt_file(raw_context, file_path)
        else:
            raise ValueError(f"Unsupported file type for page-by-page: {file_ext}")

        # 2. Classify pages
        text_pages = [p for p in page_infos if not p.has_visual_elements]
        vlm_pages = [p for p in page_infos if p.has_visual_elements]

        logger.info(f"Document analysis: {len(text_pages)} text pages, {len(vlm_pages)} visual pages")

        # 3. Process visual pages (extract text)
        vlm_texts = {}  # dict: page_number -> extracted_text
        if vlm_pages:
            vlm_text_list = self._extract_vlm_pages(file_path, vlm_pages)
            # Associate extracted text with page numbers
            for page_info, text in zip(vlm_pages, vlm_text_list):
                vlm_texts[page_info.page_number] = text

但是在_extract_vlm_pages方法中：

    def _extract_vlm_pages(self, file_path: str, page_infos: List[PageInfo]) -> List[str]:
	        """Extract text from visual pages using VLM, returns extracted text list (in page order)"""
	        file_ext = Path(file_path).suffix.lower()
	
	        if file_ext in [".docx", ".doc"]:
	            return self._process_vlm_pages_with_doc_images(page_infos)
	
	        # For PDF and other formats, convert pages to images
	        # Convert document to images
	        all_images = self._document_converter.convert_to_images(file_path)

也就是会走到 self._document_converter.convert_to_images(file_path)中 同时，在 self._document_converter.convert_to_images(file_path)中，逻辑如下：

    def convert_to_images(self, file_path: str) -> List[Image.Image]:
        """Convert document to image list"""
        if not os.path.exists(file_path):
            raise FileNotFoundError(f"File not found: {file_path}")

        file_ext = Path(file_path).suffix.lower()
        logger.info(f"Converting document to images: {file_path} (type: {file_ext})")

        if file_ext == ".pdf":
            return self._convert_pdf_to_images(file_path)
        elif file_ext in [".png", ".jpg", ".jpeg", ".gif", ".bmp", ".webp"]:
            return self._load_image(file_path)
        elif file_ext in [".pptx", ".ppt"]:
            return self._convert_pptx_to_images(file_path)
        else:
            raise ValueError(f"Unsupported file format: {file_ext}")

也就是说，对于.md文档， self._document_converter.convert_to_images(file_path)将会raise error：Unsupported file format: .md

对于该bug，修改如下： 修改_extract_vlm_pages方法为：

    def _extract_vlm_pages(self, file_path: str, page_infos: List[PageInfo]) -> List[str]:
        """Extract text from visual pages using VLM, returns extracted text list (in page order)"""
        file_ext = Path(file_path).suffix.lower()

        if file_ext in [".docx", ".doc", ".md"]:
            return self._process_vlm_pages_with_doc_images(page_infos)

使其将.md文档交由self._process_vlm_pages_with_doc_images(page_infos)方法处理

🧑‍💻 Step to reproduce

理论上只要本地的markdown文档中包含了图片都会有error

👾 Expected result

不报错，正常处理vlm逻辑

🚑 Any additional information

No response

🛠️ MineContext Version

最新版本

💻 Platform Details

如上