refactor(querybot): remove unused imports and update function parameters

- Removed unused imports from `querybot.py` including
  `BaseCallbackManager`, `StreamingStdOutCallbackHandler`, and
  `ChatOpenAI`.
- Updated `make_or_load_vector_index` function to take `service_context`
  as a parameter instead of creating it within the function.
- Added `service_context` as an attribute to the `QueryBot` class.
- Removed the `make_service_context` function as it is no longer needed.

llamabot/bot/querybot.py

@@ -7,9 +7,6 @@
 import tiktoken
 from copy import deepcopy
 from dotenv import load_dotenv
-from langchain.callbacks.base import BaseCallbackManager
-from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
-from langchain.chat_models import ChatOpenAI
 from langchain.schema import AIMessage, HumanMessage, SystemMessage
 from llama_index import (
     Document,
@@ -116,7 +113,7 @@ def __init__(
         # Override vector_index if doc_paths is specified.
         if doc_paths is not None:
             vector_index = make_or_load_vector_index(
-                doc_paths, chunk_sizes=chunk_sizes, use_cache=use_cache
+                service_context, doc_paths, chunk_sizes=chunk_sizes, use_cache=use_cache
             )
 
         # Set object attributes.
@@ -138,6 +135,7 @@ def __init__(
         self.chunk_sizes = chunk_sizes
         self.response_tokens = response_tokens
         self.history_tokens = history_tokens
+        self.service_context = service_context
 
     def __call__(
         self,
@@ -255,23 +253,6 @@ def retrieve(
         return source_nodes
 
 
-def make_service_context():
-    """Make a service context for the QueryBot.
-
-    :returns: A service context.
-    """
-    chat = ChatOpenAI(
-        model_name=default_language_model(),
-        temperature=0.0,
-        streaming=True,
-        verbose=True,
-        callback_manager=BaseCallbackManager([StreamingStdOutCallbackHandler()]),
-    )
-    llm_predictor = LLMPredictor(llm=chat)
-    service_context = ServiceContext.from_defaults(llm_predictor=llm_predictor)
-    return service_context
-
-
 # @validate_call
 def load_index(persist_dir: Path, service_context: ServiceContext):
     """Load an index from disk.
@@ -341,6 +322,7 @@ def make_vector_index(
 
 
 def make_or_load_vector_index(
+    service_context: ServiceContext,
     doc_paths: List[Path] | List[str],
     chunk_sizes: list[int] = [200, 500, 1000, 2000],
     chunk_overlap: int = 0,
@@ -373,7 +355,6 @@ def make_or_load_vector_index(
     # Make persist_dir based on the file hash's hexdigest.
     persist_dir = CACHE_DIR / file_hash_hexdigest
     persist_dir.mkdir(parents=True, exist_ok=True)
-    service_context = make_service_context()
 
     # Step 2: Create the index's split documents.
     split_docs = []