基于开源组件实现通用文档解析

from decimal import Decimal
from pathlib import Path

from borb.pdf import Document
from borb.pdf import PDF
from borb.pdf import Page
from borb.pdf import Paragraph
from borb.pdf import SingleColumnLayout
from borb.pdf.canvas.font.font import Font
from borb.pdf.canvas.font.simple_font.true_type_font import TrueTypeFont


pdf = Document()
print("add an empty Page")
page = Page()
pdf.add_page(page)
print("use a PageLayout (SingleColumnLayout in this case)")
layout = SingleColumnLayout(page)
font_path: Path = Path('Chinese_font.ttf')
font: Font = TrueTypeFont.true_type_font_from_file(font_path)
print("add a Paragraph object")
layout.add(Paragraph("他们所", font_size=Decimal(20), font=font))
print("store the PDF")
with open(Path("output.pdf"), "wb") as pdf_file_handle:
    PDF.dumps(pdf_file_handle, pdf)
print("load the PDF")
with open(Path("output.pdf"), "rb") as pdf_file_handle:
    doc = PDF.loads(pdf_file_handle)

from pdfminer.high_level import extract_pages
from pdfminer.layout import LTChar, LTTextBoxHorizontal
from document_structure import Page, Character, TextRange, Element, Document


class PdfParse:

    def __init__(self, parse_pdf_path):
        self.parse_pdf_path = parse_pdf_path

    def gen_document_by_pdfminer(self):
        page_list = []
        for page_idx, lt_page in enumerate(extract_pages(self.parse_pdf_path)):
            element_list = []
            for lt_text_box in lt_page._objs:
                if not isinstance(lt_text_box, LTTextBoxHorizontal):
                    continue
                for lt_text_line in lt_text_box._objs:
                    text_range_chars = []
                    for lt_char in lt_text_line._objs:
                        if not isinstance(lt_char, LTChar):
                            continue
                        if lt_char.x0 >= 0 and lt_char.y0 >= 0 and lt_char.width > 0 and lt_char.height > 0:
                            x, y = lt_char.x0, lt_page.height - lt_char.y0 - lt_char.height
                            width, height = lt_char.width, lt_char.height
                            char, font_name = lt_char._text, lt_char.fontname
                            text_range_chars.append(Character(x, y, width, height, page_idx + 1, char, font_name))
                    if text_range_chars:
                        text_range = TextRange(text_range_chars)
                        element_list.append(Element([text_range], element_type='paragraph'))
            page_list.append(Page(lt_page.width, lt_page.height, page_idx + 1, element_list))
        return Document(page_list)

    def gen_document_by_mupdf(self):
        ...


if __name__ == '__main__':
    pdf_parse: PdfParse = PdfParse('test.pdf')
    document_1 = pdf_parse.gen_document_by_pdfminer()