多模态导出

输入 [ ]

import datetime
import logging
import time
from pathlib import Path

import datetime import logging import time from pathlib import Path

输入 [ ]

import pandas as pd

import pandas as pd

输入 [ ]

from docling.datamodel.base_models import InputFormat
from docling.datamodel.pipeline_options import PdfPipelineOptions
from docling.document_converter import DocumentConverter, PdfFormatOption
from docling.utils.export import generate_multimodal_pages
from docling.utils.utils import create_hash

from docling.datamodel.base_models import InputFormat from docling.datamodel.pipeline_options import PdfPipelineOptions from docling.document_converter import DocumentConverter, PdfFormatOption from docling.utils.export import generate_multimodal_pages from docling.utils.utils import create_hash

输入 [ ]

_log = logging.getLogger(__name__)

_log = logging.getLogger(__name__)

输入 [ ]

IMAGE_RESOLUTION_SCALE = 2.0

IMAGE_RESOLUTION_SCALE = 2.0

输入 [ ]

def main():
    logging.basicConfig(level=logging.INFO)

    input_doc_path = Path("./tests/data/pdf/2206.01062.pdf")
    output_dir = Path("scratch")

    # Important: For operating with page images, we must keep them, otherwise the DocumentConverter
    # will destroy them for cleaning up memory.
    # This is done by setting AssembleOptions.images_scale, which also defines the scale of images.
    # scale=1 correspond of a standard 72 DPI image
    pipeline_options = PdfPipelineOptions()
    pipeline_options.images_scale = IMAGE_RESOLUTION_SCALE
    pipeline_options.generate_page_images = True

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

    start_time = time.time()

    conv_res = doc_converter.convert(input_doc_path)

    output_dir.mkdir(parents=True, exist_ok=True)

    rows = []
    for (
        content_text,
        content_md,
        content_dt,
        page_cells,
        page_segments,
        page,
    ) in generate_multimodal_pages(conv_res):
        dpi = page._default_image_scale * 72

        rows.append(
            {
                "document": conv_res.input.file.name,
                "hash": conv_res.input.document_hash,
                "page_hash": create_hash(
                    conv_res.input.document_hash + ":" + str(page.page_no - 1)
                ),
                "image": {
                    "width": page.image.width,
                    "height": page.image.height,
                    "bytes": page.image.tobytes(),
                },
                "cells": page_cells,
                "contents": content_text,
                "contents_md": content_md,
                "contents_dt": content_dt,
                "segments": page_segments,
                "extra": {
                    "page_num": page.page_no + 1,
                    "width_in_points": page.size.width,
                    "height_in_points": page.size.height,
                    "dpi": dpi,
                },
            }
        )

    # Generate one parquet from all documents
    df_result = pd.json_normalize(rows)
    now = datetime.datetime.now()
    output_filename = output_dir / f"multimodal_{now:%Y-%m-%d_%H%M%S}.parquet"
    df_result.to_parquet(output_filename)

    end_time = time.time() - start_time

    _log.info(
        f"Document converted and multimodal pages generated in {end_time:.2f} seconds."
    )

    # This block demonstrates how the file can be opened with the HF datasets library
    # from datasets import Dataset
    # from PIL import Image
    # multimodal_df = pd.read_parquet(output_filename)

    # # Convert pandas DataFrame to Hugging Face Dataset and load bytes into image
    # dataset = Dataset.from_pandas(multimodal_df)
    # def transforms(examples):
    #     examples["image"] = Image.frombytes('RGB', (examples["image.width"], examples["image.height"]), examples["image.bytes"], 'raw')
    #     return examples
    # dataset = dataset.map(transforms)

def main(): logging.basicConfig(level=logging.INFO) input_doc_path = Path("./tests/data/pdf/2206.01062.pdf") output_dir = Path("scratch") # Important: For operating with page images, we must keep them, otherwise the DocumentConverter # will destroy them for cleaning up memory. # This is done by setting AssembleOptions.images_scale, which also defines the scale of images. # scale=1 correspond of a standard 72 DPI image pipeline_options = PdfPipelineOptions() pipeline_options.images_scale = IMAGE_RESOLUTION_SCALE pipeline_options.generate_page_images = True doc_converter = DocumentConverter( format_options={ InputFormat.PDF: PdfFormatOption(pipeline_options=pipeline_options) } ) start_time = time.time() conv_res = doc_converter.convert(input_doc_path) output_dir.mkdir(parents=True, exist_ok=True) rows = [] for ( content_text, content_md, content_dt, page_cells, page_segments, page, ) in generate_multimodal_pages(conv_res): dpi = page._default_image_scale * 72 rows.append( { "document": conv_res.input.file.name, "hash": conv_res.input.document_hash, "page_hash": create_hash( conv_res.input.document_hash + ":" + str(page.page_no - 1) ), "image": { "width": page.image.width, "height": page.image.height, "bytes": page.image.tobytes(), }, "cells": page_cells, "contents": content_text, "contents_md": content_md, "contents_dt": content_dt, "segments": page_segments, "extra": { "page_num": page.page_no + 1, "width_in_points": page.size.width, "height_in_points": page.size.height, "dpi": dpi, }, } ) # Generate one parquet from all documents df_result = pd.json_normalize(rows) now = datetime.datetime.now() output_filename = output_dir / f"multimodal_{now:%Y-%m-%d_%H%M%S}.parquet" df_result.to_parquet(output_filename) end_time = time.time() - start_time _log.info( f"Document converted and multimodal pages generated in {end_time:.2f} seconds." ) # This block demonstrates how the file can be opened with the HF datasets library # from datasets import Dataset # from PIL import Image # multimodal_df = pd.read_parquet(output_filename) # # Convert pandas DataFrame to Hugging Face Dataset and load bytes into image # dataset = Dataset.from_pandas(multimodal_df) # def transforms(examples): # examples["image"] = Image.frombytes('RGB', (examples["image.width"], examples["image.height"]), examples["image.bytes"], 'raw') # return examples # dataset = dataset.map(transforms)

输入 [ ]

if __name__ == "__main__":
    main()

if __name__ == "__main__": main()