Update app/main.py

app/main.py

@@ -1,27 +1,43 @@
 import os
 import re
-from openai import OpenAI
-from langchain_openai import ChatOpenAI
+from langchain.llms import OpenAI
+from langchain.chat_models import ChatOpenAI
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain.chains.combine_documents import create_stuff_documents_chain
-from langchain_core.prompts import ChatPromptTemplate
-from langchain.chains import create_retrieval_chain
-from langchain_community.vectorstores import FAISS
-from langchain_community.document_loaders import UnstructuredWordDocumentLoader as DocxLoader
+from langchain.prompts.chat import ChatPromptTemplate
+from langchain.chains import RetrievalQA
+from langchain.vectorstores import FAISS
+from langchain.document_loaders import UnstructuredWordDocumentLoader as DocxLoader
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi import FastAPI
 from pydantic import BaseModel
-from langchain_community.embeddings import HuggingFaceBgeEmbeddings
+from langchain.embeddings.huggingface import HuggingFaceEmbeddings
 import nltk
 import time
 
-os.environ["TRANSFORMERS_CACHE"] = "/path/to/writable/cache"
+# Ensure necessary directories are writable
+cache_dir = "/app/cache"  # Update this to a writable directory path
+os.makedirs(cache_dir, exist_ok=True)
+os.environ["TRANSFORMERS_CACHE"] = cache_dir
 
-# Set up FastAPI app
+# Initialize FastAPI app
 app = FastAPI()
 
-# Get OpenAI API key
-openai_api_key = os.environ.get('OPENAI_API_KEY')
+# Configure CORS (if required)
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# Get OpenAI API key from environment
+openai_api_key = os.environ.get("OPENAI_API_KEY")
+if not openai_api_key:
+    raise ValueError("Please set the OPENAI_API_KEY environment variable.")
+
+# Initialize LLM
 llm = ChatOpenAI(
     api_key=openai_api_key,
     model_name="gpt-4-turbo-preview",
@@ -32,9 +48,11 @@ llm = ChatOpenAI(
 def read_root():
     return {"Hello": "World"}
 
+# Define Pydantic model for query input
 class Query(BaseModel):
     query_text: str
 
+# Utility function to clean responses
 def clean_response(response):
     cleaned = response.strip()
     cleaned = re.sub(r'^["\']+|["\']+$', '', cleaned)
@@ -42,59 +60,40 @@ def clean_response(response):
     cleaned = cleaned.replace('\\n', '')
     return cleaned
 
+# Define the prompt for the chatbot
 prompt = ChatPromptTemplate.from_template(
-"""
-You are a helpful assistant designed specifically for the Thapar Institute of Engineering and Technology (TIET), a renowned technical college. Your task is to answer all queries related to TIET. If a question falls outside of this context, please decline by stating, 'Sorry, I cannot help with that.' For more information, please contact our toll-free number: 18002024100 or email us at [email protected].
-<context>
-{context}
-</context>
-Question: {input}
-"""
+    """
+    You are a helpful assistant designed specifically for the Thapar Institute of Engineering and Technology (TIET), 
+    a renowned technical college. Your task is to answer all queries related to TIET. 
+    If a question falls outside of this context, please decline by stating, 'Sorry, I cannot help with that.' 
+    For more information, please contact our toll-free number: 18002024100 or email us at [email protected].
+    <context>
+    {context}
+    </context>
+    Question: {query}
+    Answer:
+    """
 )
 
+# Load embeddings
+try:
+    embeddings = HuggingFaceEmbeddings(
+        model_name="sentence-transformers/all-MiniLM-L6-v2",  # Ensure this model is valid
+        cache_folder=cache_dir,
+    )
+except Exception as e:
+    raise RuntimeError(f"Failed to initialize embeddings: {e}")
+
+# Example endpoint for handling queries
 @app.post("/chat")
-def read_item(query: Query):
+async def chat(query: Query):
+    context = "Thapar Institute of Engineering and Technology information."
     try:
-        # Load vector store
-        embeddings = HuggingFaceBgeEmbeddings(
-            model_name="BAAI/bge-base-en",
-            encode_kwargs={'normalize_embeddings': True}
+        # Use the LLM to generate a response
+        response = llm.generate(
+            prompt.format(context=context, query=query.query_text)
         )
-        vectors = FAISS.load_local("./vectors_db", embeddings, allow_dangerous_deserialization=True)
-    except Exception as e:
-        print(f"Error loading vector store: {str(e)}")
-        return {"response": "Vector Store Not Found or Error Loading. Please run /setup first."}
-
-    prompt1 = query.query_text
-    if prompt1:
-        start = time.process_time()
-        document_chain = create_stuff_documents_chain(llm, prompt)
-        retriever = vectors.as_retriever()
-        retrieval_chain = create_retrieval_chain(retriever, document_chain)
-        response = retrieval_chain.invoke({'input': prompt1})
-
-        cleaned_response = clean_response(response['answer'])
-        print("Response time:", time.process_time() - start)
+        cleaned_response = clean_response(response)
         return {"response": cleaned_response}
-    else:
-        return {"response": "No Query Found"}
-
-@app.get("/setup")
-def setup():
-    # Example setup function for vector embedding
-    documents = []  # Load your documents here
-    text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=100)
-    chunks = text_splitter.split_documents(documents)
-
-    model_name = "BAAI/bge-base-en"
-    encode_kwargs = {'normalize_embeddings': True}
-    embeddings = HuggingFaceBgeEmbeddings(model_name=model_name, encode_kwargs=encode_kwargs)
-
-    db = FAISS.from_documents(chunks, embeddings)
-    db.save_local("./vectors_db")
-    print("Vector store created and saved successfully.")
-    return {"response": "Vector Store DB Is Ready"}
-
-if __name__ == "__main__":
-    import uvicorn
-    uvicorn.run(app, host="0.0.0.0", port=8000)
+    except Exception as e:
+        return {"error": str(e)}