From 3a2a140d815bc9af88747ae4c24f325d0a1d67d8 Mon Sep 17 00:00:00 2001
From: Eli Waltuch <eliwaltuch@gmail.com>
Date: Wed, 29 Oct 2025 15:08:30 +0200
Subject: [PATCH] week2 assignment: Return of the JedAI

The second generation of the JedAI Master from 8759d43c2f7ebe4ea83834a8072395f9681e0cfc
(week1/community-contributions/week1-jedi-master.py) is back with:
* a Gradio UI
* Text Streaming
* Model Selection using examples from openai, anthropic, ollama local, and ollama cloud

How to run:
$ python3 week2/community-contributions/week2-jedi-master.py

Signed-off-by: Eli Waltuch <eliwaltuch@gmail.com>
---
 .../week2-jedi-master.py                      | 81 +++++++++++++++++++
 1 file changed, 81 insertions(+)
 create mode 100644 week2/community-contributions/week2-jedi-master.py

diff --git a/week2/community-contributions/week2-jedi-master.py b/week2/community-contributions/week2-jedi-master.py
new file mode 100644
index 0000000..4718834
--- /dev/null
+++ b/week2/community-contributions/week2-jedi-master.py
@@ -0,0 +1,81 @@
+#!/usr/bin/python3
+
+import os
+from dotenv import load_dotenv
+from openai import OpenAI
+import gradio as gr
+
+MODEL_ENDPOINTS = {
+        "gpt-4.1-mini": {"type": "openai", "base_url": "https://api.openai.com/v1", "api_key": ""},
+        "claude-haiku-4-5": {"type": "anthropic", "base_url": "https://api.anthropic.com/v1/", "api_key": ""},
+        "gemma3n:e2b": {"type": "ollama", "base_url": "http://localhost:11434/v1", "api_key": ""}, # small ollama model that runs on-device
+        "qwen3-vl:235b-cloud": {"type": "ollama", "base_url": "http://localhost:11434/v1", "api_key": ""}, # large ollama model that runs in the cloud
+}
+
+def load_api_keys():
+    # Load environment variables in a file called .env
+    load_dotenv(override=True)
+    openai_key = os.getenv('OPENAI_API_KEY')
+    anthropic_key = os.getenv('ANTHROPIC_API_KEY')
+    KEYS = {"openai": openai_key, "anthropic": anthropic_key}
+
+    # Check the keys
+    if not openai_key:
+        raise RuntimeError("Error: No OpenAI API key was found!")
+    elif not openai_key.startswith("sk-proj-"):
+        raise RuntimeError("Error: An OpenAI API key was found, but it doesn't start sk-proj-; please check you're using the right key")
+    elif openai_key.strip() != openai_key:
+        raise RuntimeError("Error: An OpenAI API key was found, but it looks like it might have space or tab characters at the start or end - please remove them!")
+    if not anthropic_key:
+        raise RuntimeError("Error: No Anthropic API key was found!")
+    elif not anthropic_key.startswith("sk-ant-"):
+        raise RuntimeError("Error: An Antrhopic API key was found, but it doesn't start sk-ant-; please check you're using the right key")
+    elif anthropic_key.strip() != anthropic_key:
+        raise RuntimeError("Error: An Anthropic API key was found, but it looks like it might have space or tab characters at the start or end - please remove them!")
+    else:
+        # add the verified keys to global MODEL_ENDPOINTS struct
+        for model, cfg in MODEL_ENDPOINTS.items():
+            cfg["api_key"] = KEYS.get(cfg["type"], "")
+        return "API keys found and look good so far!"
+
+def ask_llm(user_prompt, history, model):
+    system_prompt = """
+    You are a wise Jedi Master and an excellent teacher.
+    You will answer any question you are given by breaking it down into small steps
+    that even a complete beginner will understand.
+    When answering, speak as if you are Yoda from the Star Wars universe.
+    Also, refer to the user as "My young Padawan"
+    End every answer with "May the force be with you, always."
+    """
+    base_url = MODEL_ENDPOINTS.get(model, {}).get("base_url", "https://api.openai.com/v1")
+    api_key = MODEL_ENDPOINTS.get(model, {}).get("api_key", "")
+    client = OpenAI(base_url=base_url, api_key=api_key)
+    history = [{"role":h["role"], "content":h["content"]} for h in history]
+    messages = [{"role": "system", "content": system_prompt}] + history + [{"role": "user", "content": user_prompt}]
+    stream = client.chat.completions.create(model=model, messages=messages, stream=True)
+    response = ""
+    for chunk in stream:
+        response += chunk.choices[0].delta.content or ''
+        yield response
+    #return response.choices[0].message.content
+
+def main():
+    load_api_keys()
+    with gr.Blocks() as demo:
+        gr.Markdown("### Return of the JedAI")
+        model_dropdown = gr.Dropdown(
+            label="Select Model",
+            choices=[
+                "gpt-4.1-mini",
+                "claude-haiku-4-5",
+                "gemma3n:e2b",
+                "qwen3-vl:235b-cloud"
+            ],
+            value="gpt-4.1-mini",
+            interactive=True
+        )
+        chat = gr.ChatInterface(fn=ask_llm, type="messages", additional_inputs=[model_dropdown])
+        demo.launch()
+
+if __name__ == "__main__":
+    main()