Skip to content

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

fix(agent/tools): run a DuckDuckGo news search when the channel is news (#19176)

MarMar Labs293de5aad78a
diff --git a/agent/tools/duckduckgo.py b/agent/tools/duckduckgo.py
index bd072bd242..8056f75569 100644
--- a/agent/tools/duckduckgo.py
+++ b/agent/tools/duckduckgo.py
@@ -77,7 +77,10 @@ class DuckDuckGo(ToolBase, ABC):
return
try:
- if kwargs.get("topic", "general") == "general":
+ # Test for "news" rather than for a web-search value: callers send either
+ # "text" or "general" for the web search, so "news" is the only stable
+ # discriminator between the two branches.
+ if (kwargs.get("channel") or self._param.channel) != "news":
with DDGS() as ddgs:
if self.check_if_canceled("DuckDuckGo processing"):
return
diff --git a/test/unit_test/agent/tools/test_duckduckgo_channel.py b/test/unit_test/agent/tools/test_duckduckgo_channel.py
new file mode 100644
index 0000000000..f28225d9e2
--- /dev/null
+++ b/test/unit_test/agent/tools/test_duckduckgo_channel.py
@@ -0,0 +1,94 @@
+#
+# Copyright 2025 The InfiniFlow Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+import pytest
+
+pytest.importorskip("duckduckgo_search")
+
+import agent.tools.duckduckgo as ddg_module # noqa: E402
+from agent.tools.duckduckgo import DuckDuckGo, DuckDuckGoParam # noqa: E402
+
+
+class _FakeDDGS:
+ """Records which search method the tool chose."""
+
+ calls: list[str] = []
+
+ def __enter__(self):
+ return self
+
+ def __exit__(self, *_):
+ return False
+
+ def text(self, *_a, **_kw):
+ _FakeDDGS.calls.append("text")
+ return []
+
+ def news(self, *_a, **_kw):
+ _FakeDDGS.calls.append("news")
+ return []
+
+
+def _make_tool(channel):
+ # Bypass the canvas-bound __init__ the way test_google_unit.py does, and stub
+ # the canvas-touching helpers so _invoke's branch is the only thing exercised.
+ tool = DuckDuckGo.__new__(DuckDuckGo)
+ param = DuckDuckGoParam()
+ param.max_retries = 0
+ param.delay_after_error = 0
+ param.channel = channel
+ tool._param = param
+ tool.check_if_canceled = lambda *a, **k: False
+ tool._retrieve_chunks = lambda *a, **k: None
+ tool.set_output = lambda *a, **k: None
+ return tool
+
+
[email protected](autouse=True)
+def _fake_ddgs(monkeypatch):
+ _FakeDDGS.calls = []
+ monkeypatch.setattr(ddg_module, "DDGS", _FakeDDGS)
+
+
[email protected]("channel", ["news"])
+def test_news_channel_runs_a_news_search(channel):
+ # The node's own parameter is what the canvas passes; selecting News must not
+ # fall through to the general web search.
+ _make_tool(channel)._invoke(query="q")
+ assert _FakeDDGS.calls == ["news"]
+
+
+def test_news_channel_from_kwargs_runs_a_news_search():
+ # An LLM tool call supplies `channel` in kwargs rather than on the param.
+ _make_tool("text")._invoke(query="q", channel="news")
+ assert _FakeDDGS.calls == ["news"]
+
+
[email protected]("channel", ["text", "general"])
+def test_non_news_channels_run_a_text_search(channel):
+ # Callers send either value for a web search; both must keep it.
+ _make_tool(channel)._invoke(query="q")
+ assert _FakeDDGS.calls == ["text"]
+
+
+def test_blank_channel_falls_back_to_the_param():
+ # A blank kwarg must not be read as a channel; the node's setting wins.
+ _make_tool("news")._invoke(query="q", channel="")
+ assert _FakeDDGS.calls == ["news"]