Update app/main.py

app/main.py

@@ -17,6 +17,7 @@ import time
 
 # Set writable paths for cache and data
 cache_dir = '/tmp'
+writable_dir = os.path.join(cache_dir, 'vectors_db')
 nltk_data_path = os.path.join(cache_dir, 'nltk_data')
 
 # Configure NLTK and other library paths
@@ -27,20 +28,12 @@ os.environ['XDG_CACHE_HOME'] = cache_dir
 # Add NLTK data path
 nltk.data.path.append(nltk_data_path)
 
-# Ensure the directory exists
-try:
-    os.makedirs(nltk_data_path, exist_ok=True)
-except OSError as e:
-    print(f"Error creating directory {nltk_data_path}: {e}")
-    raise
+# Ensure the directories exist
+os.makedirs(nltk_data_path, exist_ok=True)
+os.makedirs(writable_dir, exist_ok=True)
 
 # Download required NLTK resources
-try:
-    nltk.download('punkt', download_dir=nltk_data_path)
-    print("NLTK 'punkt' resource downloaded successfully.")
-except Exception as e:
-    print(f"Error downloading NLTK resources: {e}")
-    raise
+nltk.download('punkt', download_dir=nltk_data_path)
 
 def clean_response(response):
     # Remove any leading/trailing whitespace, including newlines
@@ -70,7 +63,7 @@ app.add_middleware(
 openai_api_key = os.environ.get('OPENAI_API_KEY')
 llm = ChatOpenAI(
     api_key=openai_api_key,
-    model_name="gpt-4-turbo-preview",  # or "gpt-3.5-turbo" for a more economical option
+    model_name="gpt-4-turbo-preview",
     temperature=0.7,
     max_tokens=200
 )
@@ -85,10 +78,9 @@ class Query(BaseModel):
 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 in a concise manner. Every response you provide should be relevant to the context of TIET. If a question falls outside of this context, please decline by stating, 'Sorry, I cannot help with that.' If you do not know the answer to a question, do not attempt to fabricate a response; instead, politely decline.
-but avoid sounding boastful or exaggerating. Stay focused on the context provided.
 If the query is not related to TIET or falls outside the context of education, respond with:
         "Sorry, I cannot help with that. I'm specifically designed to answer questions about the Thapar Institute of Engineering and Technology. 
-        For more information, please contact at our toll-free number: 18002024100 or E-mail us at admissions@thapar.edu
+        For more information, please contact our toll-free number: 18002024100 or E-mail us at admissions@thapar.edu
 <context>
 {context}
 </context>
@@ -117,10 +109,11 @@ def vector_embedding():
         encode_kwargs = {'normalize_embeddings': True}
         model_norm = HuggingFaceBgeEmbeddings(model_name=model_name, encode_kwargs=encode_kwargs)
         
+        # Save FAISS vector store to a writable directory
         db = FAISS.from_documents(chunks, model_norm)
-        db.save_local("./vectors_db")
+        db.save_local(writable_dir)
         
-        print("Vector store created and saved successfully.")
+        print(f"Vector store created and saved successfully to {writable_dir}.")
         return {"response": "Vector Store DB Is Ready"}
 
     except Exception as e:
@@ -137,7 +130,7 @@ def get_embeddings():
 def read_item(query: Query):
     try:
         embeddings = get_embeddings()
-        vectors = FAISS.load_local("./vectors_db", embeddings, allow_dangerous_deserialization=True)
+        vectors = FAISS.load_local(writable_dir, 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."}
@@ -154,9 +147,7 @@ def read_item(query: Query):
         # Apply the cleaning function to the response
         cleaned_response = clean_response(response['answer'])
         
-        # For debugging, print the cleaned response
         print("Cleaned response:", repr(cleaned_response))
-        
         return {"response": cleaned_response}
     else:
         return {"response": "No Query Found"}
@@ -167,4 +158,4 @@ def setup():
 
 if __name__ == "__main__":
     import uvicorn
-    uvicorn.run(app, host="0.0.0.0", port=8000)
+    uvicorn.run(app, host="0.0.0.0", port=7860)