Add Wikipedia search utility and tool (#1561)

The Python `wikipedia` package gives easy access for searching and
fetching pages from Wikipedia, see https://pypi.org/project/wikipedia/.
It can serve as an additional search and retrieval tool, like the
existing Google and SerpAPI helpers, for both chains and agents.

langchain/__init__.py

@@ -48,6 +48,7 @@ from langchain.utilities.google_search import GoogleSearchAPIWrapper
 from langchain.utilities.google_serper import GoogleSerperAPIWrapper
 from langchain.utilities.searx_search import SearxSearchWrapper
 from langchain.utilities.serpapi import SerpAPIWrapper
+from langchain.utilities.wikipedia import WikipediaAPIWrapper
 from langchain.utilities.wolfram_alpha import WolframAlphaAPIWrapper
 from langchain.vectorstores import FAISS, ElasticVectorSearch
 
@@ -70,6 +71,7 @@ __all__ = [
     "GoogleSearchAPIWrapper",
     "GoogleSerperAPIWrapper",
     "WolframAlphaAPIWrapper",
+    "WikipediaAPIWrapper",
     "Anthropic",
     "Banana",
     "CerebriumAI",

langchain/agents/load_tools.py

@@ -9,12 +9,13 @@ from langchain.chains.api.base import APIChain
 from langchain.chains.llm_math.base import LLMMathChain
 from langchain.chains.pal.base import PALChain
 from langchain.llms.base import BaseLLM
-from langchain.tools.python.tool import PythonREPLTool
 from langchain.requests import RequestsWrapper
 from langchain.tools.base import BaseTool
 from langchain.tools.bing_search.tool import BingSearchRun
 from langchain.tools.google_search.tool import GoogleSearchResults, GoogleSearchRun
+from langchain.tools.python.tool import PythonREPLTool
 from langchain.tools.requests.tool import RequestsGetTool
+from langchain.tools.wikipedia.tool import WikipediaQueryRun
 from langchain.tools.wolfram_alpha.tool import WolframAlphaQueryRun
 from langchain.utilities.bash import BashProcess
 from langchain.utilities.bing_search import BingSearchAPIWrapper
@@ -22,6 +23,7 @@ from langchain.utilities.google_search import GoogleSearchAPIWrapper
 from langchain.utilities.google_serper import GoogleSerperAPIWrapper
 from langchain.utilities.searx_search import SearxSearchWrapper
 from langchain.utilities.serpapi import SerpAPIWrapper
+from langchain.utilities.wikipedia import WikipediaAPIWrapper
 from langchain.utilities.wolfram_alpha import WolframAlphaAPIWrapper
 
 
@@ -124,6 +126,10 @@ def _get_google_search(**kwargs: Any) -> BaseTool:
     return GoogleSearchRun(api_wrapper=GoogleSearchAPIWrapper(**kwargs))
 
 
+def _get_wikipedia(**kwargs: Any) -> BaseTool:
+    return WikipediaQueryRun(api_wrapper=WikipediaAPIWrapper(**kwargs))
+
+
 def _get_google_serper(**kwargs: Any) -> BaseTool:
     return Tool(
         name="Serper Search",
@@ -173,6 +179,7 @@ _EXTRA_OPTIONAL_TOOLS = {
     "google-serper": (_get_google_serper, ["serper_api_key"]),
     "serpapi": (_get_serpapi, ["serpapi_api_key", "aiosession"]),
     "searx-search": (_get_searx_search, ["searx_host"]),
+    "wikipedia": (_get_wikipedia, ["top_k_results"]),
 }
 
 

langchain/tools/wikipedia/__init__.py

@@ -0,0 +1 @@
+"""Wikipedia API toolkit."""

langchain/tools/wikipedia/tool.py

@@ -0,0 +1,25 @@
+"""Tool for the Wolfram Alpha API."""
+
+from langchain.tools.base import BaseTool
+from langchain.utilities.wikipedia import WikipediaAPIWrapper
+
+
+class WikipediaQueryRun(BaseTool):
+    """Tool that adds the capability to search using the Wikipedia API."""
+
+    name = "Wikipedia"
+    description = (
+        "A wrapper around Wikipedia. "
+        "Useful for when you need to answer general questions about "
+        "people, places, companies, historical events, or other subjects. "
+        "Input should be a search query."
+    )
+    api_wrapper: WikipediaAPIWrapper
+
+    def _run(self, query: str) -> str:
+        """Use the Wikipedia tool."""
+        return self.api_wrapper.run(query)
+
+    async def _arun(self, query: str) -> str:
+        """Use the Wikipedia tool asynchronously."""
+        raise NotImplementedError("WikipediaQueryRun does not support async")

langchain/utilities/__init__.py

@@ -7,6 +7,7 @@ from langchain.utilities.google_search import GoogleSearchAPIWrapper
 from langchain.utilities.google_serper import GoogleSerperAPIWrapper
 from langchain.utilities.searx_search import SearxSearchWrapper
 from langchain.utilities.serpapi import SerpAPIWrapper
+from langchain.utilities.wikipedia import WikipediaAPIWrapper
 from langchain.utilities.wolfram_alpha import WolframAlphaAPIWrapper
 
 __all__ = [
@@ -19,4 +20,5 @@ __all__ = [
     "SerpAPIWrapper",
     "SearxSearchWrapper",
     "BingSearchAPIWrapper",
+    "WikipediaAPIWrapper",
 ]

langchain/utilities/wikipedia.py

@@ -0,0 +1,56 @@
+"""Util that calls Wikipedia."""
+from typing import Any, Dict, Optional
+
+from pydantic import BaseModel, Extra, root_validator
+
+
+class WikipediaAPIWrapper(BaseModel):
+    """Wrapper around WikipediaAPI.
+
+    To use, you should have the ``wikipedia`` python package installed.
+    This wrapper will use the Wikipedia API to conduct searches and
+    fetch page summaries. By default, it will return the page summaries
+    of the top-k results of an input search.
+    """
+
+    wiki_client: Any  #: :meta private:
+    top_k_results: int = 3
+
+    class Config:
+        """Configuration for this pydantic object."""
+
+        extra = Extra.forbid
+
+    @root_validator()
+    def validate_environment(cls, values: Dict) -> Dict:
+        """Validate that the python package exists in environment."""
+        try:
+            import wikipedia
+
+            values["wiki_client"] = wikipedia
+        except ImportError:
+            raise ValueError(
+                "Could not import wikipedia python package. "
+                "Please it install it with `pip install wikipedia`."
+            )
+        return values
+
+    def run(self, query: str) -> str:
+        """Run Wikipedia search and get page summaries."""
+        search_results = self.wiki_client.search(query)
+        summaries = []
+        for i in range(min(self.top_k_results, len(search_results))):
+            summary = self.fetch_formatted_page_summary(search_results[i])
+            if summary is not None:
+                summaries.append(summary)
+        return "\n\n".join(summaries)
+
+    def fetch_formatted_page_summary(self, page: str) -> Optional[str]:
+        try:
+            wiki_page = self.wiki_client.page(title=page)
+            return f"Page: {page}\nSummary: {wiki_page.summary}"
+        except (
+            self.wiki_client.exceptions.PageError,
+            self.wiki_client.exceptions.DisambiguationError,
+        ):
+            return None