Tables recognized as images

[philipplelidis]

Bug

When trying to parse the documentation attached, some of the tables are wrongfully recognized as images. Even though the following option is set: pipeline_options.generate_table_images = False

Steps to reproduce

import pathlib
from docling.datamodel.base_models import InputFormat
from docling.datamodel.pipeline_options import (
    PdfPipelineOptions,
    TesseractOcrOptions,
)
from docling.document_converter import DocumentConverter, PdfFormatOption
import os
from docling_core.types.doc import ImageRefMode, PictureItem, TableItem


pdf_path = "Operating_Manual_PlanTeam-SPACE_first20.pdf"
model_name = os.path.expanduser("~/models/models--Hiveurban--multilingual-e5-large-pooled/snapshots/72881863bb80d223adce6fdb057972bdbdaddf9f")
output_dir = pathlib.Path("images_docling")

pipeline_options = PdfPipelineOptions()
pipeline_options.do_ocr = False
pipeline_options.ocr_options = TesseractOcrOptions(force_full_page_ocr=True, lang=["deu"])
pipeline_options.do_table_structure = True
pipeline_options.table_structure_options.do_cell_matching = True
pipeline_options.images_scale = 2.0
pipeline_options.generate_picture_images = True
pipeline_options.generate_table_images = False

converter = DocumentConverter(format_options={InputFormat.PDF: PdfFormatOption(pipeline_options=pipeline_options)},)

conv_res = converter.convert(pdf_path)
print("Dokument name:", conv_res.document.name)
print("Number of groups:", len(conv_res.document.groups))
print("Number of texts:", len(conv_res.document.texts))
print("Number of tables:", len(conv_res.document.tables))
print("Number of images:", len(conv_res.document.pictures))

output_dir.mkdir(parents=True, exist_ok=True)
doc_filename = conv_res.input.file.stem

table_counter = 0
picture_counter = 0
for element, _level in conv_res.document.iterate_items():
    if isinstance(element, PictureItem):
        picture_counter += 1
        element_image_filename = (
            output_dir / f"{doc_filename}-picture-{picture_counter}.png"
        )
        with element_image_filename.open("wb") as fp:
            element.get_image(conv_res.document).save(fp, "PNG")


# Save markdown with embedded pictures
md_filename = output_dir / f"{doc_filename}-with-images.md"
conv_res.document.save_as_markdown(md_filename, image_mode=ImageRefMode.EMBEDDED)

Docling version

2.34.0

Python version

3.10.12

Operating_Manual_PlanTeam-SPACE_first20.pdf

[BelaidCH]

@PeterStaar-IBM Can you please assign this to me. I am really interested.

[PeterStaar-IBM]

@PeterStaar-IBM Can you please assign this to me. I am really interested.

of course!