llmsherpa_reader.py

"""
The workaround while https://github.com/nlmatics/llmsherpa/issues/79 is being resolved.
"""

from typing import Any
from io import StringIO
from llmsherpa.readers.layout_reader import Block, Document


class DocTree(Document):
    def __init__(self, blocks: dict[str, Any]):
        super().__init__(blocks)

    def to_html(self) -> str:
        def iter_children(node: Block, level: int, sb: StringIO):
            for child in node.children:
                sb.write(child.to_html())
                print(
                    "-" * level,
                    child.tag,
                    f"({len(child.children)})",
                    child.to_text(),
                )
                iter_children(child, level + 1, sb)

        stream = StringIO()
        iter_children(self.root_node, 0, stream)
        return stream.getvalue()