RK-A2

OpenAI assistants API can take files as inputs. change set_prompt_mods to take an additional input:

"file_urls" : ["<url1>", "<url2>", ...]"

<urlk> will be Google Cloud Storage URL.

Ensure that these files are associated with the assistant when it is created. Example code below.

import openai
from openai import OpenAI

1. 1. OpenAI Setup
   client = OpenAI(api_key="YOUR_API_KEY")

1. 2. File Preparation and Upload (Do this once for initial setup)
   def upload_files(file_paths):
   file_ids = []
   for path in file_paths:
   with open(path, "rb") as file:
   file_response = client.files.create(file=file, purpose="assistants")
   file_ids.append(file_response.id)
   return file_ids

1. Replace with your actual file paths
   file_paths = ["your_document1.txt", "your_document2.pdf", ...]
   uploaded_file_ids = upload_files(file_paths)

1. 3. Create Assistant
   assistant = client.beta.assistants.create(
   model="gpt-3.5-turbo-1106",
   instructions="You are a helpful assistant. Answer questions using the provided files. If the answer is not in the files, say you couldn't find it.",
   tools=[{"type": "file_search"}],
   file_ids=uploaded_file_ids # Attach files
   )

1. 4. Create Thread and Get User's Question
   thread = client.beta.threads.create()

while True:
user_input = input("Ask your question: ")

1. 5. Add Message to Thread
   client.beta.threads.messages.create(
   thread_id=thread.id,
   role="user",
   content=user_input
   )

1. 6. Run the Assistant
   run = client.beta.threads.runs.create(
   thread_id=thread.id,
   assistant_id=assistant.id
   )

1. 7. Retrieve Assistant's Response
   while True:
   run_status = client.beta.threads.runs.retrieve(thread_id=thread.id, run_id=run.id)
   if run_status.status == 'completed':
   messages = client.beta.threads.messages.list(thread_id=thread.id)
   assistant_message = messages.data⁰.content⁰.text # Latest message
   print("Answer:", assistant_message.value)
   break