Skip to content

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

test: cover TokenizerFromUpstream empty payload contracts (#18898)

huyua958035bdfe108
diff --git a/test/unit_test/rag/flow/tokenizer/test_schema.py b/test/unit_test/rag/flow/tokenizer/test_schema.py
new file mode 100644
index 0000000000..56315a75d1
--- /dev/null
+++ b/test/unit_test/rag/flow/tokenizer/test_schema.py
@@ -0,0 +1,26 @@
+import pytest
+from pydantic import ValidationError
+
+from rag.flow.tokenizer.schema import TokenizerFromUpstream
+
+
+ ("output_format", "payload_field", "payload"),
+ [
+ ("chunks", "chunks", []),
+ ("json", "json", []),
+ ("markdown", "markdown", ""),
+ ("text", "text", ""),
+ ("html", "html", ""),
+ ],
+)
+def test_empty_upstream_payloads_are_valid(output_format: str, payload_field: str, payload: object) -> None:
+ result = TokenizerFromUpstream(output_format=output_format, **{payload_field: payload})
+
+ assert result.model_dump(by_alias=True)[payload_field] == payload
+
+
[email protected]("output_format", ["chunks", "json", "markdown", "text", "html"])
+def test_missing_upstream_payload_is_rejected(output_format: str) -> None:
+ with pytest.raises(ValidationError):
+ TokenizerFromUpstream(output_format=output_format)