Upload 4 files

.env

@@ -3,4 +3,5 @@ SUPABASE_URL=https://qgggbukkivjakulcmprh.supabase.co
 SUPABASE_SERVICE_KEY=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJzdXBhYmFzZSIsInJlZiI6InFnZ2didWtraXZqYWt1bGNtcHJoIiwicm9sZSI6InNlcnZpY2Vfcm9sZSIsImlhdCI6MTc1MjIxOTE4NSwiZXhwIjoyMDY3Nzk1MTg1fQ.XxiYK6LzPGWKPiCBbF_moqfmpwCBsFLzw0aBNgxpN9M
 GROQ_API_KEY=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
 TAVILY_API_KEY=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
+OPENAI_API_KEY=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
 

agent.py

@@ -113,6 +113,16 @@ def arvix_search(query: str) -> str:
     return {"arvix_results": formatted_search_docs}
 
 
+@tool
+def wolfram_alpha_query(query: str) -> str:
+    import wolframalpha
+    client = wolframalpha.Client(os.environ['WOLFRAM_APP_ID'])
+    res = client.query(query)
+    try:
+        return next(res.results).text
+    except StopIteration:
+        return "No result found."
+
 
 # load the system prompt from the file
 with open("system_prompt.txt", "r", encoding="utf-8") as f:
@@ -150,12 +160,19 @@ tools = [
     wiki_search,
     web_search,
     arvix_search,
+    wolfram_alpha_query
 ]
 
 # Build graph function
-def build_graph(provider: str = "groq"):
+def build_graph(provider: str = "openai"):
     """Build the graph"""
     # Load environment variables from .env file
+    if provider == "openai":
+        from langchain.chat_models import ChatOpenAI
+    llm = ChatOpenAI(model_name="gpt-4", temperature=0)
+        elif provider == "anthropic":
+    from langchain.chat_models import ChatAnthropic
+    llm = ChatAnthropic(model="claude-v1", temperature=0)
     if provider == "google":
         # Google Gemini
         llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash", temperature=0)

app.py

@@ -24,7 +24,7 @@ class BasicAgent:
         self.graph = build_graph()
 
     def __call__(self, question: str) -> str:
-        print(f"Agent received question : {question}...")
+        print(f"Agent received question : {question}")
         # Wrap the question in a HumanMessage from langchain_core
         messages = [HumanMessage(content=question)]
         messages = self.graph.invoke({"messages": messages})

requirements.txt

@@ -17,3 +17,6 @@ wikipedia
 pgvector
 python-dotenv
 sentence-transformers
+langchain 
+openai 
+anthropic