refactor: let excel use lazy image loader (#13558)

### What problem does this PR solve?

let excel use lazy image loader

### Type of change

- [x] Refactoring

---------

Co-authored-by: Yingfeng <yingfeng.zhang@gmail.com>

deepdoc/parser/docx_parser.py

@@ -26,7 +26,7 @@ from docx.image.exceptions import (
     UnexpectedEndOfFileError,
     UnrecognizedImageError,
 )
-from rag.utils.lazy_image import LazyDocxImage
+from rag.utils.lazy_image import LazyImage
 
 class RAGFlowDocxParser:
     def get_picture(self, document, paragraph):
@@ -66,7 +66,7 @@ class RAGFlowDocxParser:
                 image_blobs.append(image_blob)
         if not image_blobs:
             return None
-        return LazyDocxImage(image_blobs)
+        return LazyImage(image_blobs)
 
 
     def __extract_table_content(self, tb):

deepdoc/parser/excel_parser.py

@@ -18,9 +18,9 @@ from io import BytesIO
 
 import pandas as pd
 from openpyxl import Workbook, load_workbook
-from PIL import Image
 
 from rag.nlp import find_codec
+from rag.utils.lazy_image import LazyImage
 
 # copied from `/openpyxl/cell/cell.py`
 ILLEGAL_CHARACTERS_RE = re.compile(r"[\000-\010]|[\013-\014]|[\016-\037]")
@@ -122,7 +122,7 @@ class RAGFlowExcelParser:
         for img in images:
             try:
                 img_bytes = img._data()
-                pil_img = Image.open(BytesIO(img_bytes)).convert("RGB")
+                lazy_img = LazyImage([img_bytes])
 
                 anchor = img.anchor
                 if hasattr(anchor, "_from") and hasattr(anchor, "_to"):
@@ -139,7 +139,7 @@ class RAGFlowExcelParser:
 
                 item = {
                     "sheet": sheetname or ws.title,
-                    "image": pil_img,
+                    "image": lazy_img,
                     "image_description": "",
                     "row_from": r1,
                     "col_from": c1,

deepdoc/parser/figure_parser.py

@@ -75,7 +75,7 @@ def vision_figure_parser_figure_xlsx_wrapper(images,callback=None, **kwargs):
         vision_model = None
     if vision_model:
         figures_data = [((
-                        img["image"],   # Image.Image
+                        img["image"],   # Image.Image or LazyImage (converted by ensure_pil_image)
                         [img["image_description"]]     # description list (must be list)
                     ),
                     [