parser/msword_parser.py

from typing import Iterator
from langchain_core.documents import Document
from langchain_community.document_loaders.base import BaseBlobParser
from langchain_community.document_loaders.blob_loaders import Blob


class MsWordParser(BaseBlobParser):
    """Parse Microsoft Word documents from a blob."""

    # type: ignore[valid-type]
    def lazy_parse(self, blob: Blob) -> Iterator[Document]:
        """Parse a Microsoft Word document into the Document iterator.

        Args:
            blob: The blob to parse.

        Returns: An iterator of Documents.
        """
        try:
            from docx import Document as DocxDocument
        except ImportError as e:
            raise ImportError(
                "Could not import python-docx, please install with `pip install python-docx`."
            ) from e

        supported_mime_types = [
            "application/msword",
            "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
            "application/vnd.openxmlformats-officedocument.themeManager+xml",
        ]

        # Debugging: Print MIME type
        print(f"Blob MIME type: {blob.mimetype}")

        # type: ignore[attr-defined]
        if blob.mimetype not in supported_mime_types:
            raise ValueError(
                f"This blob type is not supported for this parser. Supported types are: {supported_mime_types}"
            )

        with blob.as_bytes_io() as word_document:  # type: ignore[attr-defined]
            doc = DocxDocument(word_document)
            text = "\n\n".join([para.text for para in doc.paragraphs])
            metadata = {"source": blob.source}  # type: ignore[attr-defined]
            yield Document(page_content=text, metadata=metadata)