René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Model serialization drops keys","articleBody":"There may be an issue affecting the `serialize_model` methods of the Pydantic models in this library.\n\nTaking the [`DocumentContent`](https://github.com/gleanwork/api-client-python/blob/56b3b86fba975cd22b00cc29c592f0e01143c284/src/glean/api_client/models/documentcontent.py#L16) model as an example, we see:\n\n```py\nclass DocumentContent(BaseModel):\n    full_text_list: Annotated[\n        Optional[List[str]], pydantic.Field(alias=\"fullTextList\")\n    ] = None\n    r\"\"\"The plaintext content of the document.\"\"\"\n\n    @model_serializer(mode=\"wrap\")\n    def serialize_model(self, handler):\n        optional_fields = set([\"fullTextList\"])\n        serialized = handler(self)\n        m = {}\n\n        for n, f in type(self).model_fields.items():\n            k = f.alias or n\n            val = serialized.get(k)\n\n            if val != UNSET_SENTINEL:\n                if val is not None or k not in optional_fields:\n                    m[k] = val\n\n        return m\n```\n\nThis model uses a field alias that, when constructing the Pydantic object from an API response, will map the `fullTextList` field of the JSON object to the `full_text_list` field of the Pydantic object.\n\nHowever, the model serializer uses:\n\n```py\n...\nk = f.alias or n\nval = serialized.get(k)\n...\n```\n\nwhich means that the field alias (`fullTextList`) will be used to extract the value rather than the Pydantic field name. This results in `value` being `None` and in missing keys in the returned dictionary `m` when the field name and its alias are different.\n\nTo support this claim, please find attached a [`documents.json`](https://github.com/user-attachments/files/24693667/documents.json) file that contains an anonymized response collected from the Glean API (`/rest/api/v1/getdocuments` [endpoint](https://developers.glean.com/api/client-api/documents/getdocuments)).\n\nAnd below is a simple `debug.py` script to run alongside it:\n\n```py\nimport pathlib\n\nfrom glean.api_client import models\nfrom glean.api_client.utils.unmarshal_json_response import unmarshal_json_response\n\n\nclass DummyHttpResponse:\n    def __init__(self, text):\n        self.status_code = 200\n        self.text = text\n\n\nwith pathlib.Path(\"documents.json\").open(\"r\") as f:\n    http_res = DummyHttpResponse(\n        text=f.read(),\n    )\n\n\ndocuments_response = unmarshal_json_response(models.GetDocumentsResponse, http_res)\n\n\nassert isinstance(documents_response, models.GetDocumentsResponse)\nassert documents_response.documents is not None\nassert isinstance(\n    documents_response.documents[\"https://company.com/Test\"].content,\n    models.DocumentContent,\n)\nassert (\n    documents_response.documents[\"https://company.com/Test\"].content.full_text_list[0]\n    == \"This is a test document.\"\n)\n\nserialized_document_response = documents_response.model_dump()\n\nassert isinstance(serialized_document_response, dict)\nassert serialized_document_response[\"documents\"] is not None\n\n# Here's the problem: no `full_text_list` or `fullTextList` in the serialized response!\nassert (\n    len(serialized_document_response[\"documents\"][\"https://company.com/Test\"][\"content\"])\n    \u003e 0\n)\n```\n\nRunning it yields:\n\n```shell\n$ ls    \ndebug.py  documents.json\n\n$ python debug.py\nTraceback (most recent call last):\n  File \"/workspace/app/debug/debug.py\", line 40, in \u003cmodule\u003e\n    len(serialized_document_response[\"documents\"][\"https://company.com/Test\"][\"content\"])\n    \u003e 0\nAssertionError\n```\n\nNote that I'm using:\n\n```\npydantic_core==2.41.5\npydantic==2.12.5\nglean-api-client==0.11.27\n```","author":{"url":"https://github.com/lucas-bremond","@type":"Person","name":"lucas-bremond"},"datePublished":"2026-01-18T05:36:56.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":1},"url":"https://github.com/105/api-client-python/issues/105"}