Update agent.py

agent.py

@@ -11,7 +11,7 @@ from serpapi import GoogleSearch
 from langgraph.graph import StateGraph
 from langchain_core.language_models.llms import LLM
 from langchain_core.messages import SystemMessage, HumanMessage
-from langchain.tools import tool
+from langchain_core.tools import tool
 from langchain_community.document_loaders import WikipediaLoader, ArxivLoader
 
 # ────────────────
@@ -32,39 +32,91 @@ with open("system_prompt.txt","r",encoding="utf-8") as f:
     SYSTEM_PROMPT = f.read().strip()
 
 # ────────────────
-# 4️⃣ Define your tools (unchanged semantics)
+# 4️⃣ Define your tools (with docstrings)
 # ────────────────
+
 @tool
 def calculator(expr: str) -> str:
+    """
+    Evaluate the given Python expression and return its result as a string.
+    Returns "Error" if evaluation fails.
+    """
     try:
         return str(eval(expr))
-    except:
+    except Exception:
         return "Error"
 
+
 @tool
 def retrieve_docs(query: str, k: int = 3) -> str:
+    """
+    Perform vector similarity search over the FAISS index.
+    
+    Args:
+        query: the user’s query string to embed and search for.
+        k: the number of nearest documents to return (default 3).
+    
+    Returns:
+        The top-k document contents concatenated into one string.
+    """
     q_emb = EMBEDDER.encode([query]).astype("float32")
     D, I = INDEX.search(q_emb, k)
     return "\n\n---\n\n".join(DOCS[i] for i in I[0])
 
+
 SERPAPI_KEY = os.getenv("SERPAPI_KEY")
+
 @tool
 def web_search(query: str, num_results: int = 5) -> str:
-    params = {"engine":"google","q":query,"num":num_results,"api_key":SERPAPI_KEY}
+    """
+    Run a Google search via SerpAPI and return the top snippets.
+    
+    Args:
+        query: the search query.
+        num_results: how many results to fetch (default 5).
+    
+    Returns:
+        A newline-separated list of snippet strings.
+    """
+    params = {
+        "engine":       "google",
+        "q":            query,
+        "num":          num_results,
+        "api_key":      SERPAPI_KEY,
+    }
     res = GoogleSearch(params).get_dict().get("organic_results", [])
     return "\n".join(f"- {r.get('snippet','')}" for r in res)
 
+
 @tool
 def wiki_search(query: str) -> str:
+    """
+    Search Wikipedia for up to 2 pages matching `query`.
+    
+    Args:
+        query: the topic to look up on Wikipedia.
+    
+    Returns:
+        The combined page contents of the top-2 Wikipedia results.
+    """
     pages = WikipediaLoader(query=query, load_max_docs=2).load()
     return "\n\n---\n\n".join(d.page_content for d in pages)
 
+
 @tool
 def arxiv_search(query: str) -> str:
+    """
+    Search ArXiv for up to 3 papers matching `query` and return abstracts.
+    
+    Args:
+        query: the search query for ArXiv.
+    
+    Returns:
+        The first 1000 characters of each of the top-3 ArXiv abstracts.
+    """
     papers = ArxivLoader(query=query, load_max_docs=3).load()
     return "\n\n---\n\n".join(d.page_content[:1000] for d in papers)
 
-
 # ────────────────
 # 5️⃣ Define your State schema
 # ────────────────