Drop langchain-community (closes the last item in #510)

src/pyproject.toml

@@ -22,7 +22,6 @@ optional = true
 [tool.poetry.dependencies]
 python = "<3.14,>=3.10"
 langchain-core = ">=1.5.0,<2.0"
-langchain-community = ">=0.4.2,<0.5"
 langchain-text-splitters = ">=1.1.2,<2.0"
 python-dotenv = "^1.0.0"
 langchain-openai = ">=1.3.5,<2.0"
@@ -53,6 +52,11 @@ nltk = "^3.9.1"
 # optional even though nothing in sherpa_ai/ imports it directly. It was
 # previously pulled in transitively by `unstructured`.
 transformers = "^5.14.1"
+# database/user_usage_tracker.py imports sqlalchemy directly for the usage-
+# tracking DB. It was previously only a transitive dependency of
+# langchain-community (via langchain-classic); dropping that package
+# exposed this as a missing direct dependency.
+sqlalchemy = "^2.0.51"
 
 [tool.poetry.group.test.dependencies]
 pytest = "^9.0.3"

src/sherpa_ai/tools.py

@@ -3,10 +3,10 @@
 from typing import Any, List, Tuple, Union
 
 import requests
-from langchain_community.utilities import GoogleSerperAPIWrapper
 from langchain_core.tools import BaseTool
 from langchain_core.vectorstores import VectorStoreRetriever
 from loguru import logger
+from pydantic import BaseModel, Field
 from typing_extensions import Literal
 
 import sherpa_ai.config as cfg
@@ -19,6 +19,34 @@
 HTTP_GET_TIMEOUT = 20.0
 
 
+class GoogleSerperAPIWrapper(BaseModel):
+    """Wrapper around the Serper.dev Google Search API.
+
+    Replaces langchain_community's GoogleSerperAPIWrapper, which was the
+    package's only remaining use of langchain-community (a project that
+    now prints its own "being sunset, no longer actively maintained"
+    deprecation warning on import).
+    """
+
+    url: str = "https://google.serper.dev/search"
+    api_key: str = Field(default_factory=lambda: cfg.SERPER_API_KEY or "")
+
+    def search(self, query: str) -> dict:
+        """Query the Serper.dev Google Search API directly."""
+        headers = {
+            "X-API-KEY": self.api_key,
+            "Content-Type": "application/json",
+        }
+        response = requests.post(
+            self.url,
+            headers=headers,
+            params={"q": query},
+            timeout=HTTP_GET_TIMEOUT,
+        )
+        response.raise_for_status()
+        return response.json()
+
+
 def get_tools(memory, config):
     """Factory function to create and configure a set of tools for the agent.
 
@@ -328,8 +356,7 @@ def _run_single_query(
             Link: https://example.com/python
         """
         logger.debug(f"Search query: {query}")
-        google_serper = GoogleSerperAPIWrapper()
-        search_results = google_serper._google_serper_api_results(query)
+        search_results = GoogleSerperAPIWrapper().search(query)
         logger.debug(f"Google Search Result: {search_results}")
 
         # case 1: answerBox in the result dictionary

src/tests/conftest.py

@@ -46,7 +46,7 @@ def mock_env(external_api):
     os.environ["OPENAI_API_KEY"] = "dummy"
 
     with mock.patch(
-        "langchain_community.utilities.GoogleSerperAPIWrapper._google_serper_api_results"
+        "sherpa_ai.tools.GoogleSerperAPIWrapper.search"
     ) as mock_search, mock.patch("sherpa_ai.utils.scrape_with_url") as mock_scrape:
         mock_search.return_value = GOOGLE_SEARCH_MOCK
         # mock_socket.side_effect = guard

src/tests/unit_tests/output_parsers/test_citation_validation.py

@@ -77,7 +77,7 @@ def test_task_agent_succeeds(get_llm, external_api):  # noqa: F811
         }
 
         with mock.patch(
-            "langchain_community.utilities.GoogleSerperAPIWrapper._google_serper_api_results"
+            "sherpa_ai.tools.GoogleSerperAPIWrapper.search"
         ) as mock_search:
             mock_search.return_value = GOOGLE_SEARCH_MOCK
             result = task_agent.run()

src/tests/unit_tests/tools/test_search_tool.py

@@ -1,11 +1,16 @@
 import re
-from unittest.mock import patch
+from unittest.mock import MagicMock, patch
 
 import pytest
 from loguru import logger
 
+import sherpa_ai.config as cfg
 from sherpa_ai.config import AgentConfig
-from sherpa_ai.tools import SearchTool
+from sherpa_ai.tools import GoogleSerperAPIWrapper, SearchTool
+
+# Captured at collection time, before the autouse `mock_env` fixture (conftest.py)
+# patches GoogleSerperAPIWrapper.search for every test.
+_real_search = GoogleSerperAPIWrapper.search
 
 
 def _extract_links(search_result: str) -> list:
@@ -16,6 +21,27 @@ def _extract_links(search_result: str) -> list:
     return re.findall(r"Link:(\S+)", search_result)
 
 
+def test_google_serper_search_calls_serper_api_directly():
+    """Regression test for the langchain-community removal: this hits
+    Serper's API directly (no GoogleSerperAPIWrapper in between), so pin
+    down the request shape it relies on."""
+    mock_response = MagicMock()
+    mock_response.json.return_value = {"organic": []}
+
+    with patch.object(cfg, "SERPER_API_KEY", "test-key"), \
+         patch("sherpa_ai.tools.requests.post", return_value=mock_response) as mock_post:
+        result = _real_search(GoogleSerperAPIWrapper(), "what is the weather today?")
+
+    mock_post.assert_called_once_with(
+        "https://google.serper.dev/search",
+        headers={"X-API-KEY": "test-key", "Content-Type": "application/json"},
+        params={"q": "what is the weather today?"},
+        timeout=pytest.approx(20.0),
+    )
+    mock_response.raise_for_status.assert_called_once()
+    assert result == {"organic": []}
+
+
 def test_formulate_search_query():
     config = AgentConfig(verbose=True)
     search_tool = SearchTool(config=config)