readme update

README.md

@@ -2,7 +2,7 @@
 
 Work in progress (very early!).
 
-This demo shows a multi-turn conversation with an AI agent running inside a Temporal workflow. The goal is to collect information towards a goal. There's a simple DSL input for collecting information (currently set up to search for flights, see `send_message.py`). The AI will respond with clarifications and ask for any missing information (e.g., origin city, destination, travel dates). It uses a local LLM via Ollama.
+This demo shows a multi-turn conversation with an AI agent running inside a Temporal workflow. The goal is to collect information towards a goal. There's a simple DSL input for collecting information (currently set up to use mock functions to search for events, book flights around those events then generate an invoice for those flights, see `send_message.py`). The AI will respond with clarifications and ask for any missing information to that goal. It uses a local LLM via Ollama.
 
 ## Setup
 * Install [Ollama](https://ollama.com) and the Mistral model (`ollama run qwen2.5:14b`). (note this model is more than 10GB to download).
@@ -16,12 +16,12 @@ From the /scripts directory:
 1. Run the worker: `poetry run python run_worker.py`
 2. In another terminal run the client with a prompt.
 
-    Example: `poetry run python send_message.py 'I want to book a flight'`
+    Example: `poetry run python send_message.py 'Can you find events in march in oceania?'`
 
 3. View the worker's output for the response.
 4. Give followup prompts by signaling the workflow.
 
-    Example: `poetry run python send_message.py 'From San Francisco'`
+    Example: `poetry run python send_message.py 'sure, book flights'`
 5. Get the conversation history summary by querying the workflow.
     
     Example: `poetry run python get_history.py`
@@ -32,10 +32,6 @@ The chat session will end if it has collected enough information to complete the
 Run query get_tool_data to see the data the tool has collected so far.
 
 ## TODO
-- This is currently a good single tool workflow. It could be a child as part of a planning workflow (multiple tools).
+- The LLM prompts move through 3 mock tools (FindEvents, SearchFlights, CreateInvoice) but I should make them contact real APIs.
-- I should integrate another tool. Perhaps something that consumes web sites hunting for destinations to go to in the first place.
+- I need to add a human in the loop confirmation step before it executes any tools.
-- I should make this workflow execute a Search for flights as right now it will finish without doing anything.
-    - Almost there. I have a dynamic activity for the tool's execution. I just need to make it actually do something.
-- I need to add a human in the loop confirmation step before it executes tools.
 - I need to build a chat interface so it's not cli-controlled. Also want to show some 'behind the scenes' of the agents being used as they run.
-- Should I add tools that determine a destination, and also maybe generate a Stripe invoice for the flight?

tools/search_flights.py

@@ -16,8 +16,8 @@ def search_flights(args: dict) -> dict:
     return {
         "tool": "SearchFlights",
         "searchResults": [
-            "Mock flight 1",
+            "QF 123: $1200",
-            "Mock flight 2",
+            "VA 456: $1000",
         ],
         "status": "search-complete",
         "args": args,