max char check

clients/groq_client.py

@@ -21,7 +21,7 @@ load_dotenv(dotenv_path=".secrets.env", override=True)
 
 # Config
 GROQ_API_KEY = os.getenv("GROQ_API_KEY")
-DATA_PATH = "data.json" 
+DATA_PATH = "data.json"
 CHROMA_PATH = "chroma_db"
 
 TEMPERATURE = float(os.getenv("G_TEMPERATURE", 0.7))
@@ -106,8 +106,10 @@ class GroqClient:
         # This downloads and stores the embedding model locally, eliminating API dependency
         embeddings_model = HuggingFaceEmbeddings(
             model_name="sentence-transformers/all-MiniLM-L6-v2",
-            model_kwargs={'device': 'cpu'},  # Force CPU usage to avoid GPU conflicts
-            encode_kwargs={'normalize_embeddings': True}  # Normalize embeddings for better similarity search
+            model_kwargs={"device": "cpu"},  # Force CPU usage to avoid GPU conflicts
+            encode_kwargs={
+                "normalize_embeddings": True
+            },  # Normalize embeddings for better similarity search
         )
 
         # Clear old data to avoid duplicates
@@ -135,14 +137,19 @@ class GroqClient:
     # ---------------MAIN-----------------
 
     def ask(self, raw_query: str) -> str:
-        q = raw_query
-        if q is None or q == "":
-            return self.handle_unknown_query()
-        if q.lower() in GREETINGS_TRIGGERS:
+        if (
+            not raw_query
+            or raw_query is None
+            or raw_query == ""
+            or len(raw_query) > 1000
+        ):
+            return "Please provide a valid query under 1,000 characters."
+
+        if raw_query.lower() in GREETINGS_TRIGGERS:
             return random.choice(GREETINGS)
 
         try:
-            docs = self.retriever.invoke(q)
+            docs = self.retriever.invoke(raw_query)
         except Exception as e:
             return f"Error retrieving documents: {e}"
 
@@ -152,7 +159,7 @@ class GroqClient:
         context = "\n".join([d.page_content for d in docs])
         fallback = self.handle_unknown_query()
         prompt = self.PROMPT_TEMPLATE.format(
-            context=context, question=q, fallback_response=fallback
+            context=context, question=raw_query, fallback_response=fallback
         )
 
         messages = [
@@ -170,7 +177,6 @@ class GroqClient:
             "compound-beta-mini",
             "llama-3.1-8b-instant",
             "gemma2-9b-it",
-            
         ]
 
         random.shuffle(models_to_try)
@@ -201,4 +207,4 @@ class GroqClient:
                     return f"Error while calling LLM: {e}"
 
         # If all models fail
-        return "I'm temporarily experiencing high demand. Please try again in a few minutes or rephrase your question."
+        return "I'm temporarily experiencing high demand. Please try again in a few minutes or rephrase your question."