Upload folder using huggingface_hub

advanced_rag.py

@@ -34,6 +34,7 @@ import time
 print("Pydantic Version: ")
 print(pydantic.__version__)
 # Add Mistral imports with fallback handling
+
 try:
     from mistralai import Mistral
     MISTRAL_AVAILABLE = True
@@ -45,7 +46,7 @@ except ImportError:
     debug_print("Mistral client library not found. Install with: pip install mistralai")
 
 def debug_print(message: str):
-    print(f"[{datetime.datetime.now().isoformat()}] {message}")
+    print(f"[{datetime.datetime.now().isoformat()}] {message}", flush=True)
 
 def word_count(text: str) -> int:
     return len(text.split())
@@ -447,8 +448,7 @@ class ElevatedRagChain:
                 if not mistral_api_key:
                     raise ValueError("Please set the MISTRAL_API_KEY environment variable to use Mistral API.")
                 try:
-                    from mistralai import Mistral
-                    from mistralai.exceptions import MistralException
+                    from mistralai import Mistral                    
                     debug_print("Mistral library imported successfully")
                 except ImportError:
                     debug_print("Mistral client library not installed. Falling back to Llama pipeline.")
@@ -473,8 +473,7 @@ class ElevatedRagChain:
                                     model="mistral-small-latest",
                                     messages=[{"role": "user", "content": prompt}],
                                     temperature=self.temperature,
-                                    top_p=self.top_p,
-                                    max_tokens=32000
+                                    top_p=self.top_p
                                 )
                                 return response.choices[0].message.content
                             except Exception as e:
@@ -601,16 +600,24 @@ class ElevatedRagChain:
             retrievers=[self.bm25_retriever, self.faiss_retriever],
             weights=[self.bm25_weight, self.faiss_weight]
         )
+        
         base_runnable = RunnableParallel({
             "context": RunnableLambda(self.extract_question) | self.ensemble_retriever,
             "question": RunnableLambda(self.extract_question)
         }) | self.capture_context
-        # Wrap the prompt template in a RunnableLambda
+
+        # Ensure the prompt template is set
         self.rag_prompt = ChatPromptTemplate.from_template(self.prompt_template)
+        if self.rag_prompt is None:
+            raise ValueError("Prompt template could not be created from the given template.")
         prompt_runnable = RunnableLambda(lambda vars: self.rag_prompt.format(**vars))
+        
         self.str_output_parser = StrOutputParser()
         debug_print("Selecting LLM pipeline based on choice: " + self.llm_choice)
         self.llm = self.create_llm_pipeline()
+        if self.llm is None:
+            raise ValueError("LLM pipeline creation failed.")
+        
         def format_response(response: str) -> str:
             input_tokens = count_tokens(self.context + self.prompt_template)
             output_tokens = count_tokens(response)
@@ -620,10 +627,12 @@ class ElevatedRagChain:
             formatted += f"- **Generated using:** {self.llm_choice}\n"
             formatted += f"\n**Conversation History:** {len(self.conversation_history)} conversation(s) considered.\n"
             return formatted
+        
         self.elevated_rag_chain = base_runnable | prompt_runnable | self.llm | format_response
         debug_print("Elevated RAG chain successfully built and ready to use.")
 
 
+
     def get_current_context(self) -> str:
         base_context = "\n".join([str(doc) for doc in self.split_data[:3]]) if self.split_data else "No context available."
         history_summary = "\n\n---\n**Recent Conversations (last 3):**\n"