Update app.py

app.py

@@ -1,49 +1,87 @@
 import os
+import threading
+import time
+import subprocess
+
+print("Expanding user path for Ollama")
+OLLAMA = os.path.expanduser("~/ollama")
+
+print("Checking if Ollama exists at the path")
+if not os.path.exists(OLLAMA):
+    print("Ollama not found, downloading it")
+    subprocess.run("curl -L https://ollama.com/download/ollama-linux-amd64 -o ~/ollama", shell=True)
+    os.chmod(OLLAMA, 0o755)
+
+def ollama_service_thread():
+    print("Starting Ollama service thread")
+    subprocess.run("~/ollama serve", shell=True)
+
+print("Creating and starting Ollama service thread")
+OLLAMA_SERVICE_THREAD = threading.Thread(target=ollama_service_thread)
+OLLAMA_SERVICE_THREAD.start()
+
+print("Giving Ollama serve a moment to start")
+time.sleep(10)
+
+print("Setting model to 'gemma2'")
+model = "gemma2"
+
+print(f"Pulling model {model}")
+subprocess.run(f"~/ollama pull {model}", shell=True)
+
+################################################
+################################################
+import copy
 import gradio as gr
-import requests
-
-ollama_url = '0.0.0.0'
-
-console_output = []
-
-def call_ollama_api(prompt):
-    try:
-        response = requests.post(
-            f'{ollama_url}/generate',
-            json={"prompt": prompt}
-        )
-        response.raise_for_status()  # Raise an exception for HTTP errors
-        result = response.json()
-        output = result.get('text', '')
-        console_output.append(f"Prompt: {prompt}\nResponse: {output}")
-        return output
-    except requests.exceptions.RequestException as e:
-        error_message = f"Error: {e}"
-        console_output.append(error_message)
-        return error_message
-
-def chat(prompt):
-    output = call_ollama_api(prompt)
-    return output
-
-def show_console_output():
-    return "\n\n".join(console_output)
+from ollama import Client
+
+print("Initializing Ollama client")
+client = Client(host='http://localhost:11434', timeout=120)
+
+print("Getting Hugging Face token and model ID from environment variables")
+HF_TOKEN = os.environ.get("HF_TOKEN", None)
+MODEL_ID = os.environ.get("MODEL_ID", "google/gemma-2-9b-it")
+MODEL_NAME = MODEL_ID.split("/")[-1]
 
+print("Setting up title and description for Gradio interface")
+TITLE = "<h1><center>ollama-Chat</center></h1>"
+DESCRIPTION = f"""
+<h3>MODEL: <a href="https://hf.co/{MODEL_ID}">{MODEL_NAME}</a></h3>
+<p>Running on Ollama backend.</p>
+"""
+
+
+CSS = """
+.duplicate-button {
+    margin: auto !important;
+    color: white !important;
+    background: black !important;
+    border-radius: 100vh !important;
+}
+h3 {
+    text-align: center;
+}
+"""
+import gradio as gr
+from llama_index.llms.ollama import Ollama
+
+# Initialize the Ollama model
+llm = Ollama(model="llama3", request_timeout=120.0)
+
+# Define the function to get the response from Ollama
+def get_response(question):
+    resp = llm.complete(question)
+    return resp
+
+# Create the Gradio interface
 iface = gr.Interface(
-    fn=chat,
+    fn=get_response,
     inputs="text",
     outputs="text",
-    title="Ollama Chat",
-    description="Chat with Ollama API and see the responses."
-)
-
-console_iface = gr.Interface(
-    fn=show_console_output,
-    inputs=None,
-    outputs="text",
-    title="Console Output",
-    description="See the console outputs here."
+    title="Ask Paul Graham",
+    description="Enter a question to learn more about Paul Graham."
 )
 
-# Launch a tabbed interface
-gr.TabbedInterface([iface, console_iface], ["Chat", "Console"]).launch()
+# Launch the Gradio app
+if __name__ == "__main__":
+    iface.launch()