Skip to content

RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.

fix(sdk): preserve JSON document downloads (#19367)

Wenchao And5ebaf6f07db
diff --git a/sdk/python/ragflow_sdk/modules/document.py b/sdk/python/ragflow_sdk/modules/document.py
index 4df6d7d526..c5141f5b49 100644
--- a/sdk/python/ragflow_sdk/modules/document.py
+++ b/sdk/python/ragflow_sdk/modules/document.py
@@ -14,11 +14,13 @@
# limitations under the License.
#
-import json
+import logging
from .base import Base
from .chunk import Chunk
+logger = logging.getLogger(__name__)
+
class Document(Base):
class ParserConfig(Base):
@@ -64,16 +66,19 @@ class Document(Base):
def download(self):
res = self.get(f"/datasets/{self.dataset_id}/documents/{self.id}")
- error_keys = set(["code", "message"])
+ content_disposition = res.headers.get("Content-Disposition", "")
+ if content_disposition.lstrip().lower().startswith("attachment"):
+ logger.debug("Document.download returning attachment content dataset_id=%s document_id=%s", self.dataset_id, self.id)
+ return res.content
+
try:
response = res.json()
- actual_keys = set(response.keys())
- if actual_keys == error_keys:
+ if isinstance(response, dict) and "code" in response and response["code"] != 0:
raise Exception(response.get("message"))
- else:
- return res.content
- except json.JSONDecodeError:
- return res.content
+ except ValueError:
+ pass
+ logger.debug("Document.download returning non-attachment content dataset_id=%s document_id=%s", self.dataset_id, self.id)
+ return res.content
def list_chunks(self, page=1, page_size=30, keywords="", id=""):
data = {"keywords": keywords, "page": page, "page_size": page_size, "id": id}
diff --git a/test/testcases/test_sdk_api/test_file_management_within_dataset/test_download_document.py b/test/testcases/test_sdk_api/test_file_management_within_dataset/test_download_document.py
index c6fad07ce1..da5a806743 100644
--- a/test/testcases/test_sdk_api/test_file_management_within_dataset/test_download_document.py
+++ b/test/testcases/test_sdk_api/test_file_management_within_dataset/test_download_document.py
@@ -17,9 +17,93 @@ from concurrent.futures import ThreadPoolExecutor, as_completed
import pytest
from common import bulk_upload_documents
+from ragflow_sdk.modules.document import Document
from utils import compare_by_hash
+class _DownloadResponse:
+ def __init__(self, content, payload, headers=None):
+ self.content = content
+ self._payload = payload
+ self.headers = headers or {}
+
+ def json(self):
+ if isinstance(self._payload, Exception):
+ raise self._payload
+ return self._payload
+
+
+ "content,payload",
+ [
+ (b"[]", []),
+ (b"[1, 2]", [1, 2]),
+ (b"null", None),
+ (b'"hello"', "hello"),
+ (b"42", 42),
+ (
+ b'{"code": 0, "message": "file content"}',
+ {"code": 0, "message": "file content"},
+ ),
+ ],
+)
+def test_download_preserves_attached_json_bytes(monkeypatch, content, payload):
+ response = _DownloadResponse(
+ content,
+ payload,
+ {"Content-Disposition": 'attachment; filename="sample.json"'},
+ )
+ document = Document(None, {"id": "doc", "dataset_id": "dataset"})
+ monkeypatch.setattr(document, "get", lambda *_args, **_kwargs: response)
+
+ assert document.download() == content
+
+
+def test_download_still_raises_api_error_without_attachment(monkeypatch):
+ response = _DownloadResponse(
+ b'{"code": 102, "message": "document not found"}',
+ {"code": 102, "message": "document not found"},
+ {"Content-Type": "application/json"},
+ )
+ document = Document(None, {"id": "doc", "dataset_id": "dataset"})
+ monkeypatch.setattr(document, "get", lambda *_args, **_kwargs: response)
+
+ with pytest.raises(Exception, match="document not found"):
+ document.download()
+
+
+def test_download_raises_api_error_with_extra_response_fields(monkeypatch):
+ response = _DownloadResponse(
+ b'{"code": 102, "message": "document not found", "data": null}',
+ {"code": 102, "message": "document not found", "data": None},
+ {"Content-Type": "application/json"},
+ )
+ document = Document(None, {"id": "doc", "dataset_id": "dataset"})
+ monkeypatch.setattr(document, "get", lambda *_args, **_kwargs: response)
+
+ with pytest.raises(Exception, match="document not found"):
+ document.download()
+
+
+ "content,payload",
+ [
+ (b'{"code": 0, "message": "success"}', {"code": 0, "message": "success"}),
+ (b"plain text", ValueError("not JSON")),
+ ],
+)
+def test_download_preserves_non_error_content_without_attachment(monkeypatch, content, payload):
+ response = _DownloadResponse(content, payload, {"Content-Type": "application/json"})
+ document = Document(None, {"id": "doc", "dataset_id": "dataset"})
+ monkeypatch.setattr(document, "get", lambda *_args, **_kwargs: response)
+
+ assert document.download() == content
+
+
@pytest.mark.p1
@pytest.mark.parametrize(
"generate_test_files",