Class XmlTable

docling/backend/xml/uspto_backend.py:1402–1677

Provide a table parser for xml tables in USPTO patent documents. The OASIS Open XML Exchange Table Model can be downloaded from: http://oasis-open.org/specs/soextblx.dtd

class XmlTable:
    """Provide a table parser for xml tables in USPTO patent documents.

    The OASIS Open XML Exchange Table Model can be downloaded from:
    http://oasis-open.org/specs/soextblx.dtd
    """

    class MinColInfoType(TypedDict):
        offset: list[int]
        colwidth: list[int]

    class ColInfoType(MinColInfoType):
        cell_range: list[int]
        cell_offst: list[int]

    def __init__(self, input: str) -> None:
        """Initialize the table parser with the xml content.

        Args:
            input: The xml content.
        """
        self.max_nbr_messages = 2
        self.nbr_messages = 0
        self.empty_text = ""
        self._soup = BeautifulSoup(input, features="xml")

    def _create_tg_range(self, tgs: list[dict[str, Any]]) -> dict[int, ColInfoType]:
        """Create a unified range along the table groups.

        Args:
            tgs: Table group column specifications.

        Returns:
            Unified group column specifications.
        """
        colinfo: dict[int, XmlTable.ColInfoType] = {}

        if len(tgs) == 0:
            return colinfo

        for itg, tg in enumerate(tgs):
            colinfo[itg] = {
                "offset": [],
                "colwidth": [],
                "cell_range": [],
                "cell_offst": [0],
            }
            offst = 0
            for info in tg["colinfo"]:
                cw = info["colwidth"]
                cw = re.sub("pt", "", cw, flags=re.I)
                cw = re.sub("mm", "", cw, flags=re.I)
                try:
                    cw = int(cw)
                except BaseException:
                    cw = float(cw)
                colinfo[itg]["colwidth"].append(cw)
                colinfo[itg]["offset"].append(offst)