From 9cdb176a8e5aeee6bb8bdc0a21b06320b50af72e Mon Sep 17 00:00:00 2001 From: Panos Vagenas <35837085+vagenas@users.noreply.github.com> Date: Fri, 24 Jan 2025 16:58:24 +0100 Subject: [PATCH] update conversion as per review comments, add tests, revert Docling JSON disambiguation, document intricacies Signed-off-by: Panos Vagenas <35837085+vagenas@users.noreply.github.com> --- docling/backend/json/docling_json_backend.py | 36 ++++++++---- docling/datamodel/document.py | 9 +-- tests/test_backend_docling_json.py | 58 ++++++++++++++++++++ tests/test_input_doc.py | 19 +++++++ 4 files changed, 102 insertions(+), 20 deletions(-) create mode 100644 tests/test_backend_docling_json.py diff --git a/docling/backend/json/docling_json_backend.py b/docling/backend/json/docling_json_backend.py index b23f074a..e950f4ce 100644 --- a/docling/backend/json/docling_json_backend.py +++ b/docling/backend/json/docling_json_backend.py @@ -16,11 +16,15 @@ class DoclingJSONBackend(DeclarativeDocumentBackend): self, in_doc: InputDocument, path_or_stream: Union[BytesIO, Path] ) -> None: super().__init__(in_doc, path_or_stream) - self._my_in_doc = in_doc + + # given we need to store any actual conversion exception for raising it from + # convert(), this captures the successful result or the actual error in a + # mutually exclusive way: + self._doc_or_err = self._get_doc_or_err() @override def is_valid(self) -> bool: - return True + return isinstance(self._doc_or_err, DoclingDocument) @classmethod @override @@ -32,15 +36,23 @@ class DoclingJSONBackend(DeclarativeDocumentBackend): def supported_formats(cls) -> set[InputFormat]: return {InputFormat.JSON_DOCLING} + def _get_doc_or_err(self) -> Union[DoclingDocument, Exception]: + try: + json_data: Union[str, bytes] + if isinstance(self.path_or_stream, Path): + with open(self.path_or_stream, encoding="utf-8") as f: + json_data = f.read() + elif isinstance(self.path_or_stream, BytesIO): + json_data = self.path_or_stream.getvalue() + else: + raise RuntimeError(f"Unexpected: {type(self.path_or_stream)=}") + return DoclingDocument.model_validate_json(json_data=json_data) + except Exception as e: + return e + @override - def convert(self) -> DoclingDocument: - json_data: Union[str, bytes] - if isinstance(self.path_or_stream, Path): - with open(self.path_or_stream, encoding="utf-8") as f: - json_data = f.read() - elif isinstance(self.path_or_stream, BytesIO): - json_data = self.path_or_stream.getvalue() + def convert(self): + if isinstance(self._doc_or_err, DoclingDocument): + return self._doc_or_err else: - raise RuntimeError(f"Unexpected: {type(self.path_or_stream)=}") - doc = DoclingDocument.model_validate_json(json_data=json_data) - return doc + raise self._doc_or_err diff --git a/docling/datamodel/document.py b/docling/datamodel/document.py index bdb7d831..a2a93aa3 100644 --- a/docling/datamodel/document.py +++ b/docling/datamodel/document.py @@ -297,7 +297,7 @@ class _DocumentConversionInput(BaseModel): mime = mime or "text/plain" formats = MimeTypeToFormat.get(mime, []) if formats: - if len(formats) == 1 and mime not in ("text/plain", "application/json"): + if len(formats) == 1 and mime not in ("text/plain"): return formats[0] else: # ambiguity in formats return _DocumentConversionInput._guess_from_content( @@ -339,13 +339,6 @@ class _DocumentConversionInput(BaseModel): if InputFormat.XML_USPTO in formats and content_str.startswith("PATN\r\n"): input_format = InputFormat.XML_USPTO - elif mime == "application/json": - if ( - InputFormat.JSON_DOCLING in formats - and '"schema_name": "DoclingDocument"' in content_str - ): - input_format = InputFormat.JSON_DOCLING - return input_format @staticmethod diff --git a/tests/test_backend_docling_json.py b/tests/test_backend_docling_json.py new file mode 100644 index 00000000..a38d9da1 --- /dev/null +++ b/tests/test_backend_docling_json.py @@ -0,0 +1,58 @@ +"""Test methods in module docling.backend.json.docling_json_backend.py.""" + +from io import BytesIO +from pathlib import Path + +import pytest +from pydantic import ValidationError + +from docling.backend.json.docling_json_backend import DoclingJSONBackend +from docling.datamodel.base_models import InputFormat +from docling.datamodel.document import DoclingDocument, InputDocument + +GT_PATH: Path = Path("./tests/data/groundtruth/docling_v2/2206.01062.json") + + +def test_convert_valid_docling_json(): + """Test ingestion of valid Docling JSON.""" + cls = DoclingJSONBackend + path_or_stream = GT_PATH + in_doc = InputDocument( + path_or_stream=path_or_stream, + format=InputFormat.JSON_DOCLING, + backend=cls, + ) + backend = cls( + in_doc=in_doc, + path_or_stream=path_or_stream, + ) + assert backend.is_valid() + + act_doc = backend.convert() + act_data = act_doc.export_to_dict() + + exp_doc = DoclingDocument.load_from_json(GT_PATH) + exp_data = exp_doc.export_to_dict() + + assert act_data == exp_data + + +def test_invalid_docling_json(): + """Test ingestion of invalid Docling JSON.""" + cls = DoclingJSONBackend + path_or_stream = BytesIO(b"{}") + in_doc = InputDocument( + path_or_stream=path_or_stream, + format=InputFormat.JSON_DOCLING, + backend=cls, + filename="foo", + ) + backend = cls( + in_doc=in_doc, + path_or_stream=path_or_stream, + ) + + assert not backend.is_valid() + + with pytest.raises(ValidationError): + backend.convert() diff --git a/tests/test_input_doc.py b/tests/test_input_doc.py index 8b084667..f6c516aa 100644 --- a/tests/test_input_doc.py +++ b/tests/test_input_doc.py @@ -124,6 +124,25 @@ def test_guess_format(tmp_path): doc_path.write_text("xyz", encoding="utf-8") assert dci._guess_format(doc_path) == None + # Valid Docling JSON + test_str = '{"name": ""}' + stream = DocumentStream(name="test.json", stream=BytesIO(f"{test_str}".encode())) + assert dci._guess_format(stream) == InputFormat.JSON_DOCLING + doc_path = temp_dir / "test.json" + doc_path.write_text(test_str, encoding="utf-8") + assert dci._guess_format(doc_path) == InputFormat.JSON_DOCLING + + # Non-Docling JSON + # TODO: Docling JSON is currently the single supported JSON flavor and the pipeline + # will try to validate *any* JSON (based on suffix/MIME) as Docling JSON; proper + # disambiguation seen as part of https://github.com/DS4SD/docling/issues/802 + test_str = "{}" + stream = DocumentStream(name="test.json", stream=BytesIO(f"{test_str}".encode())) + assert dci._guess_format(stream) == InputFormat.JSON_DOCLING + doc_path = temp_dir / "test.json" + doc_path.write_text(test_str, encoding="utf-8") + assert dci._guess_format(doc_path) == InputFormat.JSON_DOCLING + def _make_input_doc(path): in_doc = InputDocument(