Upload 2 files

src/agent.py

@@ -1,6 +1,7 @@
 import json
 
 from langchain_core.messages import SystemMessage, HumanMessage
+from langchain_core.rate_limiters import InMemoryRateLimiter
 from langchain_openai.chat_models import ChatOpenAI
 from langfuse import Langfuse, get_client
 from langfuse.langchain import CallbackHandler
@@ -31,9 +32,19 @@ class Agent:
         :param langfuse_callback_handler: Langfuse callback handler for
             tracking and logging interactions.
         """
+        rate_limiter = InMemoryRateLimiter(
+            # <-- Super slow! We can only make a request once every 10 seconds
+            requests_per_second=0.1,
+            # Wake up every 100 ms to check whether allowed to make a request,
+            check_every_n_seconds=0.1,
+            # Controls the maximum burst size.
+            max_bucket_size=10,
+        )
         self.chat_model = ChatOpenAI(
             model=model,
             api_key=openai_api_key,
+            json_mode=True,
+            rate_limiter=rate_limiter
         )
         with open(system_prompt_path, "r") as file:
             self.system_prompt = file.read()

src/tools.py

@@ -75,13 +75,12 @@ def wiki_search(query: str) -> str:
     Args:
         query: The search query."""
     search_docs = WikipediaLoader(query=query, load_max_docs=2).load()
-    formatted_search_docs = "\n\n---\n\n".join(
-        [
-            {
-                "source": doc.metadata["source"],
-                "page": doc.metadata.get("page", ""),
-                "content": doc.page_content,
-            }
-            for doc in search_docs
-        ])
+    formatted_search_docs = [
+        {
+            "source": doc.metadata["source"],
+            "page": doc.metadata.get("page", ""),
+            "content": doc.page_content,
+        }
+        for doc in search_docs
+    ]
     return json.dumps({"wiki_results": formatted_search_docs})