setup readme

.env.example

@@ -1,13 +1,27 @@
 RAPIDAPI_KEY=9df2cb5...
 RAPIDAPI_HOST_FLIGHTS=sky-scrapper.p.rapidapi.com #For travel flight information tool
 RAPIDAPI_HOST_PACKAGE=trackingpackage.p.rapidapi.com #For eCommerce order status package tracking tool
-FOOTBALL_DATA_API_KEY=
+FOOTBALL_DATA_API_KEY=....
-# Leave blank to use the built-in mock fixtures generator
 
 STRIPE_API_KEY=sk_test_51J...
 
-LLM_MODEL=openai/gpt-4o # default
+LLM_PROVIDER=openai # default
-LLM_KEY=sk-proj-...
+OPENAI_API_KEY=sk-proj-...
+# or
+#LLM_PROVIDER=grok
+#GROK_API_KEY=xai-your-grok-api-key
+# or
+# LLM_PROVIDER=ollama
+# OLLAMA_MODEL_NAME=qwen2.5:14b
+# or
+# LLM_PROVIDER=google
+# GOOGLE_API_KEY=your-google-api-key
+# or
+# LLM_PROVIDER=anthropic
+# ANTHROPIC_API_KEY=your-anthropic-api-key
+# or
+# LLM_PROVIDER=deepseek
+# DEEPSEEK_API_KEY=your-deepseek-api-key
 
 
 # uncomment and unset these environment variables to connect to the local dev server

Makefile

@@ -1,62 +0,0 @@
-.PHONY: setup install run-worker run-api run-frontend run-train-api run-legacy-worker run-enterprise setup-venv check-python run-dev
-
-# Setup commands
-setup: check-python setup-venv install
-
-check-python:
-	@which python3 >/dev/null 2>&1 || (echo "Python 3 is required. Please install it first." && exit 1)
-
-setup-venv:
-	python3 -m venv venv
-	@echo "Virtual environment created. Don't forget to activate it with 'source venv/bin/activate'"
-
-install:
-	poetry install
-	cd frontend && npm install
-
-# Run commands
-run-worker:
-	poetry run python scripts/run_worker.py
-
-run-api:
-	poetry run uvicorn api.main:app --reload
-
-run-frontend:
-	cd frontend && npx vite
-
-run-train-api:
-	poetry run python thirdparty/train_api.py
-
-run-legacy-worker:
-	poetry run python scripts/run_legacy_worker.py
-
-run-enterprise:
-	cd enterprise && dotnet build && dotnet run
-
-# Development environment setup
-setup-temporal-mac:
-	brew install temporal
-	temporal server start-dev
-
-# Run all development services
-run-dev:
-	@echo "Starting all development services..."
-	@make run-worker & \
-	make run-api & \
-	make run-frontend & \
-	wait
-
-# Help command
-help:
-	@echo "Available commands:"
-	@echo "  make setup              - Create virtual environment and install dependencies"
-	@echo "  make setup-venv         - Create virtual environment only"
-	@echo "  make install            - Install all dependencies"
-	@echo "  make run-worker         - Start the Temporal worker"
-	@echo "  make run-api            - Start the API server"
-	@echo "  make run-frontend       - Start the frontend development server"
-	@echo "  make run-train-api      - Start the train API server"
-	@echo "  make run-legacy-worker  - Start the legacy worker"
-	@echo "  make run-enterprise     - Build and run the enterprise .NET worker"
-	@echo "  make setup-temporal-mac - Install and start Temporal server on Mac"
-	@echo "  make run-dev            - Start all development services (worker, API, frontend) in parallel" 

README.md

@@ -2,13 +2,7 @@
 
 This demo shows a multi-turn conversation with an AI agent running inside a Temporal workflow. The purpose of the agent is to collect information towards a goal, running tools along the way. There's a simple DSL input for collecting information (currently set up to use mock functions to search for public events, search for flights around those events, then create a test Stripe invoice for the trip).
 
-The AI will respond with clarifications and ask for any missing information to that goal. You can configure it to use any LLM supported by [LiteLLM](https://docs.litellm.ai/docs/providers), including:
+The AI will respond with clarifications and ask for any missing information to that goal. You can configure it to use [ChatGPT 4o](https://openai.com/index/hello-gpt-4o/), [Anthropic Claude](https://www.anthropic.com/claude), [Google Gemini](https://gemini.google.com), [Deepseek-V3](https://www.deepseek.com/), [Grok](https://docs.x.ai/docs/overview) or a local LLM of your choice using [Ollama](https://ollama.com).
-- OpenAI models (GPT-4, GPT-3.5)
-- Anthropic Claude models
-- Google Gemini models
-- Deepseek models
-- Ollama models (local)
-- And many more!
 
 It's really helpful to [watch the demo (5 minute YouTube video)](https://www.youtube.com/watch?v=GEXllEH2XiQ) to understand how interaction works.
 
@@ -34,11 +28,7 @@ These are the key elements of an agentic framework:
 For a deeper dive into this, check out the [architecture guide](./architecture.md).
 
 ## Setup and Configuration
-See [the Setup guide](./setup.md) for detailed instructions. The basic configuration requires just two environment variables:
+See [the Setup guide](./setup.md).
-```bash
-LLM_MODEL=openai/gpt-4o  # or any other model supported by LiteLLM
-LLM_KEY=your-api-key-here
-```
 
 ## Customizing Interaction & Tools
 See [the guide to adding goals and tools](./adding-goals-and-tools.md).

activities/tool_activities.py

@@ -1,28 +1,142 @@
 import inspect
 from temporalio import activity
+from ollama import chat, ChatResponse
+from openai import OpenAI
 import json
-from typing import Optional, Sequence
+from typing import Sequence, Optional
 from temporalio.common import RawValue
 import os
 from datetime import datetime
+import google.generativeai as genai
+import anthropic
+import deepseek
 from dotenv import load_dotenv
 from models.data_types import EnvLookupOutput, ValidationInput, ValidationResult, ToolPromptInput, EnvLookupInput
-from litellm import completion
 
 load_dotenv(override=True)
+print(
+    "Using LLM provider: "
+    + os.environ.get("LLM_PROVIDER", "openai")
+    + " (set LLM_PROVIDER in .env to change)"
+)
+
+if os.environ.get("LLM_PROVIDER") == "ollama":
+    print(
+        "Using Ollama (local) model: "
+        + os.environ.get("OLLAMA_MODEL_NAME", "qwen2.5:14b")
+    )
+
 
 class ToolActivities:
     def __init__(self):
-        """Initialize LLM client using LiteLLM."""
+        """Initialize LLM clients based on environment configuration."""
-        self.llm_model = os.environ.get("LLM_MODEL", "openai/gpt-4")
+        self.llm_provider = os.environ.get("LLM_PROVIDER", "openai").lower()
-        self.llm_key = os.environ.get("LLM_KEY")
+        print(f"Initializing ToolActivities with LLM provider: {self.llm_provider}")
-        self.llm_base_url = os.environ.get("LLM_BASE_URL")
+
-        print(f"Initializing ToolActivities with LLM model: {self.llm_model}")
+        # Initialize client variables (all set to None initially)
-        if self.llm_base_url:
+        self.openai_client: Optional[OpenAI] = None
-            print(f"Using custom base URL: {self.llm_base_url}")
+        self.grok_client: Optional[OpenAI] = None
+        self.anthropic_client: Optional[anthropic.Anthropic] = None
+        self.genai_configured: bool = False
+        self.deepseek_client: Optional[deepseek.DeepSeekAPI] = None
+        self.ollama_model_name: Optional[str] = None
+        self.ollama_initialized: bool = False
+
+        # Only initialize the client specified by LLM_PROVIDER
+        if self.llm_provider == "openai":
+            if os.environ.get("OPENAI_API_KEY"):
+                self.openai_client = OpenAI(api_key=os.environ.get("OPENAI_API_KEY"))
+                print("Initialized OpenAI client")
+            else:
+                print("Warning: OPENAI_API_KEY not set but LLM_PROVIDER is 'openai'")
+        
+        elif self.llm_provider == "grok":
+            if os.environ.get("GROK_API_KEY"):
+                self.grok_client = OpenAI(api_key=os.environ.get("GROK_API_KEY"), base_url="https://api.x.ai/v1")
+                print("Initialized grok client")
+            else:
+                print("Warning: GROK_API_KEY not set but LLM_PROVIDER is 'grok'")
+
+        elif self.llm_provider == "anthropic":
+            if os.environ.get("ANTHROPIC_API_KEY"):
+                self.anthropic_client = anthropic.Anthropic(
+                    api_key=os.environ.get("ANTHROPIC_API_KEY")
+                )
+                print("Initialized Anthropic client")
+            else:
+                print(
+                    "Warning: ANTHROPIC_API_KEY not set but LLM_PROVIDER is 'anthropic'"
+                )
+
+        elif self.llm_provider == "google":
+            api_key = os.environ.get("GOOGLE_API_KEY")
+            if api_key:
+                genai.configure(api_key=api_key)
+                self.genai_configured = True
+                print("Configured Google Generative AI")
+            else:
+                print("Warning: GOOGLE_API_KEY not set but LLM_PROVIDER is 'google'")
+
+        elif self.llm_provider == "deepseek":
+            if os.environ.get("DEEPSEEK_API_KEY"):
+                self.deepseek_client = deepseek.DeepSeekAPI(
+                    api_key=os.environ.get("DEEPSEEK_API_KEY")
+                )
+                print("Initialized DeepSeek client")
+            else:
+                print(
+                    "Warning: DEEPSEEK_API_KEY not set but LLM_PROVIDER is 'deepseek'"
+                )
+
+        # For Ollama, we store the model name but actual initialization happens in warm_up_ollama
+        elif self.llm_provider == "ollama":
+            self.ollama_model_name = os.environ.get("OLLAMA_MODEL_NAME", "qwen2.5:14b")
+            print(
+                f"Using Ollama model: {self.ollama_model_name} (will be loaded on worker startup)"
+            )
+        else:
+            print(
+                f"Warning: Unknown LLM_PROVIDER '{self.llm_provider}', defaulting to OpenAI"
+            )
+
+    def warm_up_ollama(self):
+        """Pre-load the Ollama model to avoid cold start latency on first request"""
+        if self.llm_provider != "ollama" or self.ollama_initialized:
+            return False  # No need to warm up if not using Ollama or already warmed up
+
+        try:
+            print(
+                f"Pre-loading Ollama model '{self.ollama_model_name}' - this may take 30+ seconds..."
+            )
+            start_time = datetime.now()
+
+            # Make a simple request to load the model into memory
+            chat(
+                model=self.ollama_model_name,
+                messages=[
+                    {"role": "system", "content": "You are an AI assistant"},
+                    {
+                        "role": "user",
+                        "content": "Hello! This is a warm-up message to load the model.",
+                    },
+                ],
+            )
+
+            elapsed_time = (datetime.now() - start_time).total_seconds()
+            print(f"✅ Ollama model loaded successfully in {elapsed_time:.2f} seconds")
+            self.ollama_initialized = True
+            return True
+        except Exception as e:
+            print(f"❌ Error pre-loading Ollama model: {str(e)}")
+            print(
+                "The worker will continue, but the first actual request may experience a delay."
+            )
+            return False
 
     @activity.defn
-    async def agent_validatePrompt(self, validation_input: ValidationInput) -> ValidationResult:
+    async def agent_validatePrompt(
+        self, validation_input: ValidationInput
+    ) -> ValidationResult:
         """
         Validates the prompt in the context of the conversation history and agent goal.
         Returns a ValidationResult indicating if the prompt makes sense given the context.
@@ -73,7 +187,7 @@ class ToolActivities:
             prompt=validation_prompt, context_instructions=context_instructions
         )
 
-        result = await self.agent_toolPlanner(prompt_input)
+        result = self.agent_toolPlanner(prompt_input)
 
         return ValidationResult(
             validationResult=result.get("validationResult", False),
@@ -81,7 +195,41 @@ class ToolActivities:
         )
 
     @activity.defn
-    async def agent_toolPlanner(self, input: ToolPromptInput) -> dict:
+    def agent_toolPlanner(self, input: ToolPromptInput) -> dict:
+        if self.llm_provider == "ollama":
+            return self.prompt_llm_ollama(input)
+        elif self.llm_provider == "google":
+            return self.prompt_llm_google(input)
+        elif self.llm_provider == "anthropic":
+            return self.prompt_llm_anthropic(input)
+        elif self.llm_provider == "deepseek":
+            return self.prompt_llm_deepseek(input)
+        elif self.llm_provider == "grok":
+            return self.prompt_llm_grok(input)
+        else:
+            return self.prompt_llm_openai(input)
+
+    def parse_json_response(self, response_content: str) -> dict:
+        """
+        Parses the JSON response content and returns it as a dictionary.
+        """
+        try:
+            data = json.loads(response_content)
+            return data
+        except json.JSONDecodeError as e:
+            print(f"Invalid JSON: {e}")
+            raise
+
+    def prompt_llm_openai(self, input: ToolPromptInput) -> dict:
+        if not self.openai_client:
+            api_key = os.environ.get("OPENAI_API_KEY")
+            if not api_key:
+                raise ValueError(
+                    "OPENAI_API_KEY is not set in the environment variables but LLM_PROVIDER is 'openai'"
+                )
+            self.openai_client = OpenAI(api_key=api_key)
+            print("Initialized OpenAI client on demand")
+
         messages = [
             {
                 "role": "system",
@@ -95,53 +243,236 @@ class ToolActivities:
             },
         ]
 
-        try:
+        chat_completion = self.openai_client.chat.completions.create(
-            completion_kwargs = {
+            model="gpt-4o", messages=messages  # was gpt-4-0613
-                "model": self.llm_model,
+        )
-                "messages": messages,
-                "api_key": self.llm_key
-            }
-            
-            # Add base_url if configured
-            if self.llm_base_url:
-                completion_kwargs["base_url"] = self.llm_base_url
 
-            response = completion(**completion_kwargs)
+        response_content = chat_completion.choices[0].message.content
-            
+        activity.logger.info(f"ChatGPT response: {response_content}")
-            response_content = response.choices[0].message.content
+
-            activity.logger.info(f"LLM response: {response_content}")
+        # Use the new sanitize function
+        response_content = self.sanitize_json_response(response_content)
+
+        return self.parse_json_response(response_content)
+
+    def prompt_llm_grok(self, input: ToolPromptInput) -> dict:
+        if not self.grok_client:
+            api_key = os.environ.get("GROK_API_KEY")
+            if not api_key:
+                raise ValueError(
+                    "GROK_API_KEY is not set in the environment variables but LLM_PROVIDER is 'grok'"
+                )
+            self.grok_client = OpenAI(api_key=api_key, base_url="https://api.x.ai/v1")
+            print("Initialized grok client on demand")
+
+        messages = [
+            {
+                "role": "system",
+                "content": input.context_instructions
+                + ". The current date is "
+                + datetime.now().strftime("%B %d, %Y"),
+            },
+            {
+                "role": "user",
+                "content": input.prompt,
+            },
+        ]
+
+        chat_completion = self.grok_client.chat.completions.create(
+            model="grok-2-1212", messages=messages  
+        )
+
+        response_content = chat_completion.choices[0].message.content
+        activity.logger.info(f"Grok response: {response_content}")
+
+        # Use the new sanitize function
+        response_content = self.sanitize_json_response(response_content)
+
+        return self.parse_json_response(response_content)
+    def prompt_llm_ollama(self, input: ToolPromptInput) -> dict:
+        # If not yet initialized, try to do so now (this is a backup if warm_up_ollama wasn't called or failed)
+        if not self.ollama_initialized:
+            print(
+                "Ollama model not pre-loaded. Loading now (this may take 30+ seconds)..."
+            )
+            try:
+                self.warm_up_ollama()
+            except Exception:
+                # We already logged the error in warm_up_ollama, continue with the actual request
+                pass
+
+        model_name = self.ollama_model_name or os.environ.get(
+            "OLLAMA_MODEL_NAME", "qwen2.5:14b"
+        )
+        messages = [
+            {
+                "role": "system",
+                "content": input.context_instructions
+                + ". The current date is "
+                + get_current_date_human_readable(),
+            },
+            {
+                "role": "user",
+                "content": input.prompt,
+            },
+        ]
+
+        try:
+            response: ChatResponse = chat(model=model_name, messages=messages)
+            print(f"Chat response: {response.message.content}")
 
             # Use the new sanitize function
-            response_content = self.sanitize_json_response(response_content)
+            response_content = self.sanitize_json_response(response.message.content)
-
             return self.parse_json_response(response_content)
+        except (json.JSONDecodeError, ValueError) as e:
+            # Re-raise JSON-related exceptions to let Temporal retry the activity
+            print(f"JSON parsing error with Ollama response: {str(e)}")
+            raise
         except Exception as e:
-            print(f"Error in LLM completion: {str(e)}")
+            # Log and raise other exceptions that may need retrying
+            print(f"Error in Ollama chat: {str(e)}")
             raise
 
-    def parse_json_response(self, response_content: str) -> dict:
+    def prompt_llm_google(self, input: ToolPromptInput) -> dict:
-        """
+        if not self.genai_configured:
-        Parses the JSON response content and returns it as a dictionary.
+            api_key = os.environ.get("GOOGLE_API_KEY")
-        """
+            if not api_key:
-        try:
+                raise ValueError(
-            data = json.loads(response_content)
+                    "GOOGLE_API_KEY is not set in the environment variables but LLM_PROVIDER is 'google'"
-            return data
+                )
-        except json.JSONDecodeError as e:
+            genai.configure(api_key=api_key)
-            print(f"Invalid JSON: {e}")
+            self.genai_configured = True
-            raise
+            print("Configured Google Generative AI on demand")
+
+        model = genai.GenerativeModel(
+            "models/gemini-1.5-flash",
+            system_instruction=input.context_instructions
+            + ". The current date is "
+            + datetime.now().strftime("%B %d, %Y"),
+        )
+        response = model.generate_content(input.prompt)
+        response_content = response.text
+        print(f"Google Gemini response: {response_content}")
+
+        # Use the new sanitize function
+        response_content = self.sanitize_json_response(response_content)
+
+        return self.parse_json_response(response_content)
+
+    def prompt_llm_anthropic(self, input: ToolPromptInput) -> dict:
+        if not self.anthropic_client:
+            api_key = os.environ.get("ANTHROPIC_API_KEY")
+            if not api_key:
+                raise ValueError(
+                    "ANTHROPIC_API_KEY is not set in the environment variables but LLM_PROVIDER is 'anthropic'"
+                )
+            self.anthropic_client = anthropic.Anthropic(api_key=api_key)
+            print("Initialized Anthropic client on demand")
+
+        response = self.anthropic_client.messages.create(
+            model="claude-3-5-sonnet-20241022",  
+            #model="claude-3-7-sonnet-20250219",  # doesn't do as well
+            max_tokens=1024,
+            system=input.context_instructions
+            + ". The current date is "
+            + get_current_date_human_readable(),
+            messages=[
+                {
+                    "role": "user",
+                    "content": input.prompt,
+                }
+            ],
+        )
+
+        response_content = response.content[0].text
+        print(f"Anthropic response: {response_content}")
+
+        # Use the new sanitize function
+        response_content = self.sanitize_json_response(response_content)
+
+        return self.parse_json_response(response_content)
+
+    def prompt_llm_deepseek(self, input: ToolPromptInput) -> dict:
+        if not self.deepseek_client:
+            api_key = os.environ.get("DEEPSEEK_API_KEY")
+            if not api_key:
+                raise ValueError(
+                    "DEEPSEEK_API_KEY is not set in the environment variables but LLM_PROVIDER is 'deepseek'"
+                )
+            self.deepseek_client = deepseek.DeepSeekAPI(api_key=api_key)
+            print("Initialized DeepSeek client on demand")
+
+        messages = [
+            {
+                "role": "system",
+                "content": input.context_instructions
+                + ". The current date is "
+                + datetime.now().strftime("%B %d, %Y"),
+            },
+            {
+                "role": "user",
+                "content": input.prompt,
+            },
+        ]
+
+        response = self.deepseek_client.chat_completion(prompt=messages)
+        response_content = response
+        print(f"DeepSeek response: {response_content}")
+
+        # Use the new sanitize function
+        response_content = self.sanitize_json_response(response_content)
+
+        return self.parse_json_response(response_content)
 
     def sanitize_json_response(self, response_content: str) -> str:
         """
-        Sanitizes the response content to ensure it's valid JSON.
+        Extracts the JSON block from the response content as a string.
+        Supports:
+        - JSON surrounded by ```json and ```
+        - Raw JSON input
+        - JSON preceded or followed by extra text
+        Rejects invalid input that doesn't contain JSON.
         """
-        # Remove any markdown code block markers
+        try:
-        response_content = response_content.replace("```json", "").replace("```", "")
+            start_marker = "```json"
-        
+            end_marker = "```"
-        # Remove any leading/trailing whitespace
-        response_content = response_content.strip()
-        
-        return response_content
 
+            json_str = None
+
+            # Case 1: JSON surrounded by markers
+            if start_marker in response_content and end_marker in response_content:
+                json_start = response_content.index(start_marker) + len(start_marker)
+                json_end = response_content.index(end_marker, json_start)
+                json_str = response_content[json_start:json_end].strip()
+
+            # Case 2: Text with valid JSON
+            else:
+                # Try to locate the JSON block by scanning for the first `{` and last `}`
+                json_start = response_content.find("{")
+                json_end = response_content.rfind("}")
+
+                if json_start != -1 and json_end != -1 and json_start < json_end:
+                    json_str = response_content[json_start : json_end + 1].strip()
+
+            # Validate and ensure the extracted JSON is valid
+            if json_str:
+                json.loads(json_str)  # This will raise an error if the JSON is invalid
+                return json_str
+
+            # If no valid JSON found, raise an error
+            raise ValueError("Response does not contain valid JSON.")
+
+        except json.JSONDecodeError:
+            # Invalid JSON
+            print(f"Invalid JSON detected in response: {response_content}")
+            raise ValueError("Response does not contain valid JSON.")
+        except Exception as e:
+            # Other errors
+            print(f"Error processing response: {str(e)}")
+            print(f"Full response: {response_content}")
+            raise
+
+    # get env vars for workflow
     @activity.defn
     async def get_wf_env_vars(self, input: EnvLookupInput) -> EnvLookupOutput:
         """ gets env vars for workflow as an activity result so it's deterministic
@@ -167,6 +498,18 @@ class ToolActivities:
 
         return output
 
+
+def get_current_date_human_readable():
+    """
+    Returns the current date in a human-readable format.
+
+    Example: Wednesday, January 1, 2025
+    """
+    from datetime import datetime
+
+    return datetime.now().strftime("%A, %B %d, %Y")
+
+
 @activity.defn(dynamic=True)
 async def dynamic_tool_activity(args: Sequence[RawValue]) -> dict:
     from tools import get_handler

frontend/src/pages/App.jsx

@@ -3,7 +3,7 @@ import NavBar from "../components/NavBar";
 import ChatWindow from "../components/ChatWindow";
 import { apiService } from "../services/api";
 
-const POLL_INTERVAL = 600; // 0.6 seconds
+const POLL_INTERVAL = 500; // 0.5 seconds
 const INITIAL_ERROR_STATE = { visible: false, message: '' };
 const DEBOUNCE_DELAY = 300; // 300ms debounce for user input
 

pyproject.toml

@@ -31,11 +31,16 @@ temporalio = "^1.8.0"
 
 # Standard library modules (e.g. asyncio, collections) don't need to be added
 # since they're built-in for Python 3.8+.
-litellm = "^1.70.0"
+ollama = "^0.4.5"
 pyyaml = "^6.0.2"
 fastapi = "^0.115.6"
 uvicorn = "^0.34.0"
 python-dotenv = "^1.0.1"
+openai = "^1.59.2"
+stripe = "^11.4.1"
+google-generativeai = "^0.8.4"
+anthropic = "0.47.0"
+deepseek = "^1.0.0"
 requests = "^2.32.3"
 pandas = "^2.2.3"
 gtfs-kit = "^10.1.1"

scripts/run_ollama.py

@@ -0,0 +1,23 @@
+from ollama import chat, ChatResponse
+
+
+def main():
+    model_name = "mistral"
+
+    # The messages to pass to the model
+    messages = [
+        {
+            "role": "user",
+            "content": "Why is the sky blue?",
+        }
+    ]
+
+    # Call ollama's chat function
+    response: ChatResponse = chat(model=model_name, messages=messages)
+
+    # Print the full message content
+    print(response.message.content)
+
+
+if __name__ == "__main__":
+    main()

scripts/run_worker.py

@@ -17,18 +17,18 @@ async def main():
     load_dotenv(override=True)
 
     # Print LLM configuration info
-    llm_model = os.environ.get("LLM_MODEL", "openai/gpt-4")
+    llm_provider = os.environ.get("LLM_PROVIDER", "openai").lower()
-    print(f"Worker will use LLM model: {llm_model}")
+    print(f"Worker will use LLM provider: {llm_provider}")
 
     # Create the client
     client = await get_temporal_client()
 
-    # Initialize the activities class
+    # Initialize the activities class once with the specified LLM provider
     activities = ToolActivities()
-    print(f"ToolActivities initialized with LLM model: {llm_model}")
+    print(f"ToolActivities initialized with LLM provider: {llm_provider}")
 
     # If using Ollama, pre-load the model to avoid cold start latency
-    if llm_model.startswith("ollama"):
+    if llm_provider == "ollama":
         print("\n======== OLLAMA MODEL INITIALIZATION ========")
         print("Ollama models need to be loaded into memory on first use.")
         print("This may take 30+ seconds depending on your hardware and model size.")
@@ -51,6 +51,8 @@ async def main():
     print("Worker ready to process tasks!")
     logging.basicConfig(level=logging.WARN)
 
+
+
     # Run the worker
     with concurrent.futures.ThreadPoolExecutor(max_workers=100) as activity_executor:
         worker = Worker(

setup.md

@@ -14,37 +14,6 @@ If you want to show confirmations/enable the debugging UI that shows tool args,
 SHOW_CONFIRM=True
 ```
 
-### Quick Start with Makefile
-
-We've provided a Makefile to simplify the setup and running of the application. Here are the main commands:
-
-```bash
-# Initial setup
-make setup              # Creates virtual environment and installs dependencies
-make setup-venv         # Creates virtual environment only
-make install            # Installs all dependencies
-
-# Running the application
-make run-worker         # Starts the Temporal worker
-make run-api            # Starts the API server
-make run-frontend       # Starts the frontend development server
-
-# Additional services
-make run-train-api      # Starts the train API server
-make run-legacy-worker  # Starts the legacy worker
-make run-enterprise     # Builds and runs the enterprise .NET worker
-
-# Development environment setup
-make setup-temporal-mac # Installs and starts Temporal server on Mac
-
-# View all available commands
-make help
-```
-
-### Manual Setup (Alternative to Makefile)
-
-If you prefer to run commands manually, follow these steps:
-
 ### Agent Goal Configuration
 
 The agent can be configured to pursue different goals using the `AGENT_GOAL` environment variable in your `.env` file. If unset, default is `goal_choose_agent_type`. 
@@ -56,41 +25,54 @@ GOAL_CATEGORIES=hr,travel-flights,travel-trains,fin
 
 See the section Goal-Specific Tool Configuration below for tool configuration for specific goals.
 
-### LLM Configuration
+### LLM Provider Configuration
 
-Note: We recommend using OpenAI's GPT-4o or Claude 3.5 Sonnet for the best results. There can be significant differences in performance and capabilities between models, especially for complex tasks.
+The agent can use OpenAI's GPT-4o, Google Gemini, Anthropic Claude, or a local LLM via Ollama. Set the `LLM_PROVIDER` environment variable in your `.env` file to choose the desired provider:
 
-The agent uses LiteLLM to interact with various LLM providers. Configure theqfollowing environment variables in your `.env` file:
+- `LLM_PROVIDER=openai` for OpenAI's GPT-4o
+- `LLM_PROVIDER=google` for Google Gemini
+- `LLM_PROVIDER=anthropic` for Anthropic Claude
+- `LLM_PROVIDER=deepseek` for DeepSeek-V3
+- `LLM_PROVIDER=ollama` for running LLMs via [Ollama](https://ollama.ai) (not recommended for this use case)
 
-- `LLM_MODEL`: The model to use (e.g., "openai/gpt-4o", "anthropic/claude-3-sonnet", "google/gemini-pro", etc.)
+### Option 1: OpenAI
-- `LLM_KEY`: Your API key for the selected provider
-- `LLM_BASE_URL`: (Optional) Custom base URL for the LLM provider. Useful for:
-  - Using Ollama with a custom endpoint
-  - Using a proxy or custom API gateway
-  - Testing with different API versions
 
-LiteLLM will automatically detect the provider based on the model name. For example:
+If using OpenAI, ensure you have an OpenAI key for the GPT-4o model. Set this in the `OPENAI_API_KEY` environment variable in `.env`.
-- For OpenAI models: `openai/gpt-4o` or `openai/gpt-3.5-turbo`
-- For Anthropic models: `anthropic/claude-3-sonnet`
-- For Google models: `google/gemini-pro`
-- For Ollama models: `ollama/mistral` (requires `LLM_BASE_URL` set to your Ollama server)
 
-Example configurations:
+### Option 2: Google Gemini
-```bash
-# For OpenAI
-LLM_MODEL=openai/gpt-4o
-LLM_KEY=your-api-key-here
 
-# For Anthropic
+To use Google Gemini:
-LLM_MODEL=anthropic/claude-3-sonnet
-LLM_KEY=your-api-key-here
 
-# For Ollama with custom URL
+1. Obtain a Google API key and set it in the `GOOGLE_API_KEY` environment variable in `.env`.
-LLM_MODEL=ollama/mistral
+2. Set `LLM_PROVIDER=google` in your `.env` file.
-LLM_BASE_URL=http://localhost:11434
-```
 
-For a complete list of supported models and providers, visit the [LiteLLM documentation](https://docs.litellm.ai/docs/providers).
+### Option 3: Anthropic Claude (recommended)
+
+I find that Claude Sonnet 3.5 performs better than the other hosted LLMs for this use case.
+
+To use Anthropic:
+
+1. Obtain an Anthropic API key and set it in the `ANTHROPIC_API_KEY` environment variable in `.env`.
+2. Set `LLM_PROVIDER=anthropic` in your `.env` file.
+
+### Option 4: Deepseek-V3
+
+To use Deepseek-V3:
+
+1. Obtain a Deepseek API key and set it in the `DEEPSEEK_API_KEY` environment variable in `.env`.
+2. Set `LLM_PROVIDER=deepseek` in your `.env` file.
+
+### Option 5: Local LLM via Ollama (not recommended)
+
+To use a local LLM with Ollama:
+
+1. Install [Ollama](https://ollama.com) and the [Qwen2.5 14B](https://ollama.com/library/qwen2.5) model.
+   - Run `ollama run <OLLAMA_MODEL_NAME>` to start the model. Note that this model is about 9GB to download.
+   - Example: `ollama run qwen2.5:14b`
+
+2. Set `LLM_PROVIDER=ollama` in your `.env` file and `OLLAMA_MODEL_NAME` to the name of the model you installed.
+
+Note: I found the other (hosted) LLMs to be MUCH more reliable for this use case. However, you can switch to Ollama if desired, and choose a suitably large model if your computer has the resources.
 
 ## Configuring Temporal Connection
 
@@ -184,7 +166,7 @@ Here is configuration guidance for specific goals. Travel and financial goals ha
 * Requires a Stripe key for the `create_invoice` tool. Set this in the `STRIPE_API_KEY` environment variable in .env
     * It's free to sign up and get a key at [Stripe](https://stripe.com/)
         * Set permissions for read-write on: `Credit Notes, Invoices, Customers and Customer Sessions`
-    * If you don't have a Stripe key, comment out the STRIPE_API_KEY in the .env file, and a dummy invoice will be created rather than a Stripe invoice. The function can be found in `tools/create_invoice.py`
+    * If you're lazy go to `tools/create_invoice.py` and replace the `create_invoice` function with the mock `create_invoice_example` that exists in the same file.
 
 ### Goal: Find a Premier League match, book train tickets to it and invoice the user for the cost (Replay 2025 Keynote)
 - `AGENT_GOAL=goal_match_train_invoice` - Focuses on Premier League match attendance with train booking and invoice generation
@@ -192,7 +174,8 @@ Here is configuration guidance for specific goals. Travel and financial goals ha
     - Note, there is failure built in to this demo (the train booking step) to show how the agent can handle failures and retry. See Tool Configuration below for details.
 #### Configuring Agent Goal: goal_match_train_invoice 
 NOTE: This goal was developed for an on-stage demo and has failure (and its resolution) built in to show how the agent can handle failures and retry.
-* Finding a match requires a key from [Football Data](https://www.football-data.org). Sign up for a free account, then see the 'My Account' page to get your API token. Set `FOOTBALL_DATA_API_KEY` to this value. If the key is omitted, the `SearchFixtures` tool automatically returns mock Premier League fixtures (3 months into the future only).
+* Finding a match requires a key from [Football Data](https://www.football-data.org). Sign up for a free account, then see the 'My Account' page to get your API token. Set `FOOTBALL_DATA_API_KEY` to this value.
+    * If you're lazy go to `tools/search_fixtures.py` and replace the `search_fixtures` function with the mock `search_fixtures_example` that exists in the same file.
 * We use a mock function to search for trains. Start the train API server to use the real API: `python thirdparty/train_api.py`
 * * The train activity is 'enterprise' so it's written in C# and requires a .NET runtime. See the [.NET backend](#net-(enterprise)-backend) section for details on running it.
 * Requires a Stripe key for the `create_invoice` tool. Set this in the `STRIPE_API_KEY` environment variable in .env
@@ -269,4 +252,4 @@ For more details, check out [adding goals and tools guide](./adding-goals-and-to
 [  ] `cd frontend`, `npm install`, `npx vite` <br />
 [ ] Access the UI at `http://localhost:5173` <br />
 
-And that's it! Happy AI Agent Exploring!
+And that's it! Happy AI Agent Exploring!

tools/create_invoice.py

@@ -27,7 +27,7 @@ def ensure_customer_exists(
 def create_invoice(args: dict) -> dict:
     """Create and finalize a Stripe invoice."""
     # If an API key exists in the env file, find or create customer
-    if stripe.api_key is not None and stripe.api_key != "":
+    if stripe.api_key is not None:
         customer_id = ensure_customer_exists(
             args.get("customer_id"), args.get("email", "default@example.com")
         )
@@ -69,3 +69,15 @@ def create_invoice(args: dict) -> dict:
             "invoiceURL": "https://pay.example.com/invoice/12345",
             "reference": "INV-12345",
         }
+
+def create_invoice_example(args: dict) -> dict:
+    """
+    This is an example implementation of the CreateInvoice tool
+    Doesn't call any external services, just returns a dummy response
+    """
+    print("[CreateInvoice] Creating invoice with:", args)
+    return {
+        "invoiceStatus": "generated",
+        "invoiceURL": "https://pay.example.com/invoice/12345",
+        "reference": "INV-12345",
+    }

tools/goal_registry.py

@@ -114,10 +114,10 @@ goal_match_train_invoice = AgentGoal(
     ],
     description="The user wants to book a trip to a city in the UK around the dates of a premier league match. "
     "Help the user find a premier league match to attend, search and book trains for that match and offers to invoice them for the cost of train tickets. "
-    "The user lives in London. Premier league fixtures may be mocked data, so don't worry about valid season dates and teams. "
+    "The user lives in London. "
     "Gather args for these tools in order, ensuring you move the user from one tool to the next: "
     "1. SearchFixtures: Search for fixtures for a team within a specified date range. The user might ask questions about the matches dates and locations to decide on where to go. "
-    "2. SearchTrains: Search for trains to the city of the match. Ensure you list them for the customer to choose from "
+    "2. SearchTrains: Search for trains to the city of the match and list them for the customer to choose from "
     "3. BookTrains: Book the train tickets, used to invoice the user for the cost of the train tickets "
     "4. CreateInvoice: Invoices the user for the cost of train tickets, with total and details inferred from the conversation history ",
     starter_prompt=starter_prompt_generic,
@@ -489,6 +489,6 @@ if multi_goal_mode:
             if tool.name == "ListAgents":
                 list_agents_found = True
                 continue
-        if list_agents_found is False:
+        if list_agents_found == False:
             goal.tools.append(tool_registry.list_agents_tool)
             continue

tools/search_fixtures.py

@@ -1,263 +1,64 @@
 import os
 import requests
-import random
+from datetime import datetime, timedelta
-from datetime import datetime, timedelta, date
 from dotenv import load_dotenv
 
-PREMIER_LEAGUE_CLUBS_DATA = [
-    {"name": "Arsenal FC", "stadium": "Emirates Stadium"},
-    {"name": "Aston Villa FC", "stadium": "Villa Park"},
-    {"name": "AFC Bournemouth", "stadium": "Vitality Stadium"},
-    {"name": "Brentford FC", "stadium": "Gtech Community Stadium"},
-    {"name": "Brighton & Hove Albion FC", "stadium": "American Express Stadium"},
-    {"name": "Chelsea FC", "stadium": "Stamford Bridge"},
-    {"name": "Crystal Palace FC", "stadium": "Selhurst Park"},
-    {"name": "Everton FC", "stadium": "Goodison Park"},
-    {"name": "Fulham FC", "stadium": "Craven Cottage"},
-    {"name": "Ipswich Town FC", "stadium": "Portman Road"},
-    {"name": "Leicester City FC", "stadium": "King Power Stadium"},
-    {"name": "Liverpool FC", "stadium": "Anfield"},
-    {"name": "Manchester City FC", "stadium": "Etihad Stadium"},
-    {"name": "Manchester United FC", "stadium": "Old Trafford"},
-    {"name": "Newcastle United FC", "stadium": "St James' Park"},
-    {"name": "Nottingham Forest FC", "stadium": "City Ground"},
-    {"name": "Southampton FC", "stadium": "St Mary's Stadium"},
-    {"name": "Tottenham Hotspur FC", "stadium": "Tottenham Hotspur Stadium"},
-    {"name": "West Ham United FC", "stadium": "London Stadium"},
-    {"name": "Wolverhampton Wanderers FC", "stadium": "Molineux Stadium"},
-]
-
-
-def get_future_matches(
-    team_name: str,
-    all_clubs_data: list,
-    num_matches: int = 12,
-    date_from: date = None,
-    date_to: date = None,
-) -> list:
-    """Generate a set of future Premier League matches for ``team_name``.
-
-    This is a purely mocked schedule. It returns up to ``num_matches``
-    fixtures, respecting the ``date_from`` and ``date_to`` constraints.
-    Matches are typically on Saturdays or Sundays.
-    """
-    matches = []
-
-    team_details = next((c for c in all_clubs_data if c["name"] == team_name), None)
-    if not team_details:
-        return []
-
-    opponents_pool = [c for c in all_clubs_data if c["name"] != team_name]
-    if not opponents_pool:
-        return []
-
-    # Determine the maximum number of matches we can generate based on opponents
-    # and the requested num_matches
-    num_actual_matches_to_generate = min(num_matches, len(opponents_pool))
-    if num_actual_matches_to_generate == 0:
-        return []
-
-    # Shuffle opponents once and pick them sequentially
-    random.shuffle(opponents_pool)  # Shuffle in place
-
-    # Determine the initial Saturday for match week consideration
-    today_date = date.today()
-    # Default to next Saturday
-    current_match_week_saturday = today_date + timedelta(
-        days=(5 - today_date.weekday() + 7) % 7
-    )
-
-    # If today is Saturday and it's late evening, or if today is Sunday,
-    # advance to the following Saturday.
-    now_time = datetime.now().time()
-    if (
-        today_date.weekday() == 5
-        and now_time > datetime.strptime("20:00", "%H:%M").time()
-    ) or (today_date.weekday() == 6):
-        current_match_week_saturday += timedelta(days=7)
-
-    # If date_from is specified, ensure our starting Saturday is not before it.
-    if date_from:
-        if current_match_week_saturday < date_from:
-            current_match_week_saturday = date_from
-        # Align current_match_week_saturday to be a Saturday on or after the potentially adjusted date
-        current_match_week_saturday += timedelta(
-            days=(5 - current_match_week_saturday.weekday() + 7) % 7
-        )
-
-    opponent_idx = 0
-    while len(matches) < num_actual_matches_to_generate and opponent_idx < len(
-        opponents_pool
-    ):
-        # If the current week's Saturday is already past date_to, stop.
-        if date_to and current_match_week_saturday > date_to:
-            break
-
-        opponent_details = opponents_pool[opponent_idx]
-        is_saturday_game = random.choice([True, True, False])
-        actual_match_date = None
-        kick_off_time = ""
-
-        if is_saturday_game:
-            actual_match_date = current_match_week_saturday
-            kick_off_time = random.choice(["12:30", "15:00", "17:30"])
-        else:  # Sunday game
-            actual_match_date = current_match_week_saturday + timedelta(days=1)
-            kick_off_time = random.choice(["14:00", "16:30"])
-
-        # Check if this specific match date is within the date_to constraint
-        if date_to and actual_match_date > date_to:
-            # If this game is too late, try the next week if possible.
-            # (This mainly affects Sunday games if Saturday was the last valid day)
-            current_match_week_saturday += timedelta(days=7)
-            continue  # Skip adding this match, try next week.
-
-        match_datetime_gmt = (
-            f"{actual_match_date.strftime('%Y-%m-%d')} {kick_off_time} GMT"
-        )
-        is_home_match = random.choice([True, False])
-
-        if is_home_match:
-            team1_name = team_details["name"]
-            team2_name = opponent_details["name"]
-            stadium_name = team_details["stadium"]
-        else:
-            team1_name = opponent_details["name"]
-            team2_name = team_details["name"]
-            stadium_name = opponent_details["stadium"]
-
-        matches.append(
-            {
-                "team1": team1_name,
-                "team2": team2_name,
-                "stadium": stadium_name,
-                "datetime_gmt": match_datetime_gmt,
-            }
-        )
-        opponent_idx += 1
-        current_match_week_saturday += timedelta(
-            days=7
-        )  # Advance to next week's Saturday
-
-    return matches
-
-
 BASE_URL = "https://api.football-data.org/v4"
 
 
 def search_fixtures(args: dict) -> dict:
     load_dotenv(override=True)
-    api_key = os.getenv("FOOTBALL_DATA_API_KEY")
+    api_key = os.getenv("FOOTBALL_DATA_API_KEY", "YOUR_DEFAULT_KEY")
 
     team_name = args.get("team")
     date_from_str = args.get("date_from")
     date_to_str = args.get("date_to")
-
-    if not team_name:
-        return {"error": "Team name is required."}
-
-    parsed_date_from = None
-    if date_from_str:
-        try:
-            parsed_date_from = datetime.strptime(date_from_str, "%Y-%m-%d").date()
-        except ValueError:
-            return {
-                "error": f"Invalid date_from: '{date_from_str}'. Expected format YYYY-MM-DD."
-            }
-
-    parsed_date_to = None
-    if date_to_str:
-        try:
-            parsed_date_to = datetime.strptime(date_to_str, "%Y-%m-%d").date()
-        except ValueError:
-            return {
-                "error": f"Invalid date_to: '{date_to_str}'. Expected format YYYY-MM-DD."
-            }
-
-    if parsed_date_from and parsed_date_to and parsed_date_from > parsed_date_to:
-        return {"error": "date_from cannot be after date_to."}
-
-    # If no API key, fall back to mocked data
-    if not api_key:
-        # Use the parsed date objects (which can be None)
-        fixtures = get_future_matches(
-            team_name,
-            PREMIER_LEAGUE_CLUBS_DATA,
-            date_from=parsed_date_from,
-            date_to=parsed_date_to,
-            # num_matches can be passed explicitly if needed, otherwise defaults to 12
-        )
-        if not fixtures:
-            # Check if the team name itself was invalid, as get_future_matches returns [] for that too
-            team_details_check = next(
-                (c for c in PREMIER_LEAGUE_CLUBS_DATA if c["name"] == team_name), None
-            )
-            if not team_details_check:
-                return {"error": f"Team '{team_name}' not found in mocked data."}
-            # If team is valid, an empty fixtures list means no matches fit the criteria (e.g., date range)
-        return {"fixtures": fixtures}
-
-    # API Key is present, proceed with API logic
-    # The API requires both date_from and date_to
-    if not parsed_date_from or not parsed_date_to:
-        return {
-            "error": "Both date_from and date_to (YYYY-MM-DD) are required for API search."
-        }
-
     headers = {"X-Auth-Token": api_key}
-    # For API calls, team name matching might be case-insensitive or require specific handling
+    team_name = team_name.lower()
-    # The existing logic uses team_name.lower() for the API search path later.
+
+    try:
+        date_from = datetime.strptime(date_from_str, "%Y-%m-%d")
+        date_to = datetime.strptime(date_to_str, "%Y-%m-%d")
+    except ValueError:
+        return {
+            "error": "Invalid date provided. Expected format YYYY-MM-DD for both date_from and date_to."
+        }
 
     # Fetch team ID
     teams_response = requests.get(f"{BASE_URL}/competitions/PL/teams", headers=headers)
     if teams_response.status_code != 200:
-        return {
+        return {"error": "Failed to fetch teams data."}
-            "error": f"Failed to fetch teams data from API (status {teams_response.status_code})."
-        }
 
     teams_data = teams_response.json()
     team_id = None
-    # Using lower() for comparison, assuming API team names might have varied casing
+    for team in teams_data["teams"]:
-    # or the input team_name might not be exact.
+        if team_name in team["name"].lower():
-    # The `ToolDefinition` lists exact names, so direct match might also be an option.
+            team_id = team["id"]
-    for team_api_data in teams_data.get("teams", []):
-        if team_name.lower() in team_api_data.get("name", "").lower():
-            team_id = team_api_data["id"]
             break
 
     if not team_id:
-        return {"error": f"Team '{team_name}' not found via API."}
+        return {"error": "Team not found."}
 
-    date_from_formatted = parsed_date_from.strftime("%Y-%m-%d")
+    date_from_formatted = date_from.strftime("%Y-%m-%d")
-    date_to_formatted = parsed_date_to.strftime("%Y-%m-%d")
+    date_to_formatted = date_to.strftime("%Y-%m-%d")
     fixtures_url = f"{BASE_URL}/teams/{team_id}/matches?dateFrom={date_from_formatted}&dateTo={date_to_formatted}"
-    # print(fixtures_url) # Keep for debugging if necessary
+    print(fixtures_url)
 
     fixtures_response = requests.get(fixtures_url, headers=headers)
     if fixtures_response.status_code != 200:
-        return {
+        return {"error": "Failed to fetch fixtures data."}
-            "error": f"Failed to fetch fixtures data from API (status {fixtures_response.status_code})."
-        }
 
     fixtures_data = fixtures_response.json()
     matching_fixtures = []
 
     for match in fixtures_data.get("matches", []):
-        # Ensure match datetime parsing is robust
+        match_datetime = datetime.strptime(match["utcDate"], "%Y-%m-%dT%H:%M:%SZ")
-        try:
+        if match["competition"]["code"] == "PL":
-            match_datetime_utc = datetime.strptime(
-                match["utcDate"], "%Y-%m-%dT%H:%M:%SZ"
-            )
-        except (ValueError, TypeError):
-            # Skip malformed match entries or log an error
-            continue
-
-        if match.get("competition", {}).get("code") == "PL":
             matching_fixtures.append(
                 {
-                    "date": match_datetime_utc.strftime("%Y-%m-%d"),
+                    "date": match_datetime.strftime("%Y-%m-%d"),
-                    "homeTeam": match.get("homeTeam", {}).get("name", "N/A"),
+                    "homeTeam": match["homeTeam"]["name"],
-                    "awayTeam": match.get("awayTeam", {}).get("name", "N/A"),
+                    "awayTeam": match["awayTeam"]["name"],
                 }
             )
 
@@ -281,69 +82,34 @@ def search_fixtures_example(args: dict) -> dict:
 
     # Validate dates
     try:
-        # Ensure date strings are not None before parsing
+        date_from = datetime.strptime(date_from_str, "%Y-%m-%d")
-        if date_from_str is None or date_to_str is None:
+        date_to = datetime.strptime(date_to_str, "%Y-%m-%d")
-            raise ValueError("Date strings cannot be None")
-        date_from_obj = datetime.strptime(date_from_str, "%Y-%m-%d")
-        date_to_obj = datetime.strptime(date_to_str, "%Y-%m-%d")
     except ValueError:
         return {
             "error": "Invalid date provided. Expected format YYYY-MM-DD for both date_from and date_to."
         }
 
     # Calculate 3 reasonable fixture dates within the given range
-    date_range = (date_to_obj - date_from_obj).days
+    date_range = (date_to - date_from).days
-    if date_range < 0:  # date_from is after date_to
-        return {"fixtures": []}  # No fixtures possible
-
-    fixture_dates_timestamps = []
     if date_range < 21:
-        # If range is less than 3 weeks, use evenly spaced fixtures if possible
+        # If range is less than 3 weeks, use evenly spaced fixtures
-        if date_range >= 2:  # Need at least some gap for 3 fixtures
+        fixture_dates = [
-            fixture_dates_timestamps = [
+            date_from + timedelta(days=max(1, date_range // 3)),
-                date_from_obj
+            date_from + timedelta(days=max(2, date_range * 2 // 3)),
-                + timedelta(days=max(0, date_range // 4)),  # Closer to start
+            date_to - timedelta(days=min(2, date_range // 4)),
-                date_from_obj + timedelta(days=max(1, date_range // 2)),  # Middle
+        ]
-                date_to_obj - timedelta(days=max(0, date_range // 4)),  # Closer to end
-            ]
-        elif date_range == 1:  # Only two days
-            fixture_dates_timestamps = [date_from_obj, date_to_obj]
-        elif date_range == 0:  # Only one day
-            fixture_dates_timestamps = [date_from_obj]
-        else:  # date_range is negative, handled above, or 0 (single day)
-            fixture_dates_timestamps = [date_from_obj] if date_range == 0 else []
-
     else:
-        # Otherwise space them out by weeks, ensuring they are within the bounds
+        # Otherwise space them out by weeks
-        d1 = date_from_obj + timedelta(days=7)
+        fixture_dates = [
-        d2 = date_from_obj + timedelta(days=14)
+            date_from + timedelta(days=7),
-        d3 = date_to_obj - timedelta(days=7)  # Potential third game from the end
+            date_from + timedelta(days=14),
+            date_to - timedelta(days=7),
+        ]
 
-        fixture_dates_timestamps.append(d1)
+    # Ensure we only have 3 dates
-        if d2 <= date_to_obj and d2 > d1:  # ensure d2 is valid and distinct
+    fixture_dates = fixture_dates[:3]
-            fixture_dates_timestamps.append(d2)
-        if (
-            d3 >= date_from_obj and d3 > d2 and d3 <= date_to_obj
-        ):  # ensure d3 is valid and distinct
-            fixture_dates_timestamps.append(d3)
-        elif (
-            d3 < date_from_obj and len(fixture_dates_timestamps) < 3
-        ):  # if d3 is too early, try using date_to_obj itself if distinct
-            if date_to_obj not in fixture_dates_timestamps:
-                fixture_dates_timestamps.append(date_to_obj)
-
-    # Ensure unique dates and sort, then take up to 3.
-    fixture_dates_timestamps = sorted(
-        list(
-            set(
-                f_date
-                for f_date in fixture_dates_timestamps
-                if date_from_obj <= f_date <= date_to_obj
-            )
-        )
-    )
-    fixture_dates_final = fixture_dates_timestamps[:3]
 
+    # Expanded pool of opponent teams to avoid team playing against itself
     all_opponents = [
         "Manchester United FC",
         "Leicester City FC",
@@ -354,35 +120,35 @@ def search_fixtures_example(args: dict) -> dict:
         "Tottenham Hotspur FC",
         "West Ham United FC",
         "Everton FC",
-        "Generic Opponent A",
-        "Generic Opponent B",
-        "Generic Opponent C",  # Fallbacks
     ]
 
+    # Select opponents that aren't the same as the requested team
     available_opponents = [
         team for team in all_opponents if team.lower() != team_name.lower()
     ]
 
-    # Ensure we have enough opponents for the number of fixtures we'll generate
+    # Ensure we have at least 3 opponents
-    if len(available_opponents) < len(fixture_dates_final):
+    if len(available_opponents) < 3:
-        needed = len(fixture_dates_final) - len(available_opponents)
+        # Add generic opponents if needed
-        for i in range(needed):
+        additional_teams = [f"Opponent {i} FC" for i in range(1, 4)]
-            available_opponents.append(f"Placeholder Opponent {i+1}")
+        available_opponents.extend(additional_teams)
 
-    opponents = available_opponents[: len(fixture_dates_final)]
+    # Take only the first 3 opponents
+    opponents = available_opponents[:3]
 
+    # Generate fixtures - always exactly 3
     fixtures = []
-    for i, fixture_date_obj in enumerate(fixture_dates_final):
+    for i, fixture_date in enumerate(fixture_dates):
-        if i >= len(opponents):  # Should not happen with the logic above
+        date_str = fixture_date.strftime("%Y-%m-%d")
-            break
+
-        date_str = fixture_date_obj.strftime("%Y-%m-%d")
+        # Alternate between home and away games
-        if i % 2 == 0:  # Home game
+        if i % 2 == 0:
-            fixtures.append(
-                {"date": date_str, "homeTeam": team_name, "awayTeam": opponents[i]}
-            )
-        else:  # Away game
             fixtures.append(
                 {"date": date_str, "homeTeam": opponents[i], "awayTeam": team_name}
             )
+        else:
+            fixtures.append(
+                {"date": date_str, "homeTeam": team_name, "awayTeam": opponents[i]}
+            )
 
     return {"fixtures": fixtures}

tools/tool_registry.py

@@ -90,7 +90,7 @@ search_flights_tool = ToolDefinition(
 
 search_trains_tool = ToolDefinition(
     name="SearchTrains",
-    description="Search for trains between two English cities. Returns a list of train information for the user to choose from. Present the list to the user.",
+    description="Search for trains between two English cities. Returns a list of train information for the user to choose from.",
     arguments=[
         ToolArgument(
             name="origin",
@@ -156,7 +156,7 @@ create_invoice_tool = ToolDefinition(
 
 search_fixtures_tool = ToolDefinition(
     name="SearchFixtures",
-    description="Search for upcoming fixtures for a given team within a date range inferred from the user's description. Ignore valid premier league dates. Valid teams this season are Arsenal FC, Aston Villa FC, AFC Bournemouth, Brentford FC, Brighton & Hove Albion FC, Chelsea FC, Crystal Palace FC, Everton FC, Fulham FC, Ipswich Town FC, Leicester City FC, Liverpool FC, Manchester City FC, Manchester United FC, Newcastle United FC, Nottingham Forest FC, Southampton FC, Tottenham Hotspur FC, West Ham United FC, Wolverhampton Wanderers FC",
+    description="Search for upcoming fixtures for a given team within a date range inferred from the user's description. Valid teams this 24/25 season are Arsenal FC, Aston Villa FC, AFC Bournemouth, Brentford FC, Brighton & Hove Albion FC, Chelsea FC, Crystal Palace FC, Everton FC, Fulham FC, Ipswich Town FC, Leicester City FC, Liverpool FC, Manchester City FC, Manchester United FC, Newcastle United FC, Nottingham Forest FC, Southampton FC, Tottenham Hotspur FC, West Ham United FC, Wolverhampton Wanderers FC",
     arguments=[
         ToolArgument(
             name="team",