NeoAnd/Docling
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: allow computing page images on-demand with scale and cache them (#36)

Browse Source 
* feat: allow computing page images on-demand and cache them

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* feat: expose scale for export of page images and document elements

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* fix comment

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

---------

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
This commit is contained in:
Michele Dolfi
2024-08-20 13:27:19 +02:00
committed by GitHub
parent c253dd743a
commit 78347bf679
9 changed files with 104 additions and 77 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
docling/datamodel/document.py
View File

@@ -1,7 +1,7 @@
import logging
from io import BytesIO
from pathlib import Path, PurePath
from typing import ClassVar, Dict, Iterable, List, Optional, Type, Union
from typing import ClassVar, Dict, Iterable, List, Optional, Tuple, Type, Union
from docling_core.types import BaseCell, BaseText
from docling_core.types import BoundingBox as DsBoundingBox
@@ -21,6 +21,7 @@ from docling.datamodel.base_models import (
DocumentStream,
FigureElement,
Page,
PageElement,
TableElement,
TextElement,
)
@@ -302,6 +303,20 @@ class ConvertedDocument(BaseModel):
else:
return ""
def render_element_images(
self, element_types: Tuple[PageElement] = (FigureElement,)
):
for element in self.assembled.elements:
if isinstance(element, element_types):
page_ix = element.page_no
scale = self.pages[page_ix]._default_image_scale
crop_bbox = element.cluster.bbox.scaled(scale=scale).to_top_left_origin(
page_height=self.pages[page_ix].size.height * scale
)
cropped_im = self.pages[page_ix].image.crop(crop_bbox.as_tuple())
yield element, cropped_im
class DocumentConversionInput(BaseModel):