Extract all text from this document image with formatting. Group elements by horizontal position into rows.

IMPORTANT - PARAGRAPH MERGING:
Look at the certificate body text that spans multiple visual lines but forms ONE continuous sentence/paragraph. These MUST be merged into a SINGLE element. For example:
"This is to certify that MR TAN... PPT NO:... has no criminal record... certificate... Spain."
This is ONE paragraph spanning 5 visual lines - output as ONE element, ONE row.

Output JSON format:
[
  {
    "row": 1,
    "elements": [
      {
        "bbox_2d": [x1, y1, x2, y2],
        "text_content": "full paragraph text with all lines joined by space",
        "element_type": "title|header|paragraph|label|value|footer",
        "style": {
          "bold": true/false,
          "italic": true/false,
          "font_size": "small|normal|large|xlarge",
          "alignment": "left|center|right"
        }
      }
    ]
  }
]

Rules:
1. MERGE multi-line paragraphs into ONE element (bbox spans all lines)
2. Elements on same horizontal line = same row
3. Keep label:value pairs as separate elements in same row
4. Detect formatting: bold, italic, font_size, alignment