Method process_structure

document_qa/grobid_processors.py:159–226

Send a PDF to GROBID and return structured content. Args: input_path: Path to the PDF file. coordinates: If ``True``, include bounding-box coordinate strings in each passage (needed for PDF highlighting). Returns: dict or None: A

(self, input_path, coordinates=False)

        self.grobid_client = grobid_client

    def process_structure(self, input_path, coordinates=False):
        """Send a PDF to GROBID and return structured content.

        Args:
            input_path: Path to the PDF file.
            coordinates: If ``True``, include bounding-box coordinate
                strings in each passage (needed for PDF highlighting).

        Returns:
            dict or None: A dict with keys:

            - ``"biblio"`` — bibliographic metadata (title, authors, DOI, …).
            - ``"passages"`` — list of passage dicts, each containing
              ``text``, ``type``, ``section``, ``subSection``,
              ``passage_id``, and ``coordinates``.
            - ``"filename"`` — stem of the PDF filename.

            Returns ``None`` if GROBID returns a non-200 status.
        """
        try:
            pdf_file, status, text = self.grobid_client.process_pdf(
                "processFulltextDocument",
                input_path,
                consolidate_header=True,
                consolidate_citations=False,
                segment_sentences=False,
                tei_coordinates=coordinates,
                include_raw_citations=False,
                include_raw_affiliations=False,
                generateIDs=True,
            )
        except requests.exceptions.RequestException as exc:
            # Transport-level failure (connection refused, timeout, …).
            # Local/usage errors (bad path, parsing bugs) are intentionally
            # not caught here so they surface with their real traceback.
            raise GrobidServiceError("Grobid service did not respond.") from exc

        if status != 200:
            # Grobid attaches a human-readable reason to error responses
            # (e.g. a 500 body explaining what went wrong). Surface it
            # alongside the status code instead of discarding it.
            reason = text.strip() if text else ""
            message = f"Grobid service returned status {status}."
            if reason:
                message += f" {reason}"
            raise GrobidServiceError(message, status_code=status)

        # Grobid can answer 200 with an empty body (e.g. it gave up on the PDF).
        if not text or not text.strip():
            raise GrobidServiceError("Grobid returned an empty response.", status_code=status)

        # A truncated/corrupted TEI payload makes the XML parser blow up; map
        # that to a clear service error instead of an opaque parsing traceback.
        try:
            document_object = self.parse_grobid_xml(text, coordinates=coordinates)
        except GrobidServiceError:
            raise
        except Exception as exc: