Upload 2 files

app.py

@@ -0,0 +1,36 @@
+import gradio as gr
+import subprocess
+import os
+
+def run_command(command, check=True):
+    """Run a shell command and handle output/errors."""
+    try:
+        subprocess.run(command, shell=True, check=check)
+        print("Done!")
+    except subprocess.CalledProcessError as e:
+        print(f"Error running command: {command}\n{e}")
+        exit(1)
+
+run_command("python setup_llama.py")
+
+def check_function():
+    if os.path.exists("Llama-3.1-8B-Lexi-Uncensored_V2_Q4.gguf"):
+        return True
+    else:
+        print("Llama-3.1-8B-Lexi-Uncensored_V2_Q4.gguf not found. Please run setup_llama.py first.")
+        return False
+
+def button_click(input_text):
+    if check_function():
+        return f"Function succeeded! Your input: {input_text}"
+    else:
+        return "Function failed. Please try again."
+
+with gr.Blocks() as demo:
+    text_input = gr.Textbox(label="Enter some text")
+    button = gr.Button("Run Function")
+    output = gr.Textbox(label="Result")
+
+    button.click(fn=button_click, inputs=text_input, outputs=output)
+
+demo.launch()

setup_llama.py

@@ -0,0 +1,67 @@
+import os
+import subprocess
+import time
+import requests
+os.environ["OLLAMA_HOST"] = "0.0.0.0"
+os.environ["NGROK_AUTHTOKEN"] = None
+
+def run_command(command, check=True):
+    """Run a shell command and handle output/errors."""
+    try:
+        subprocess.run(command, shell=True, check=check)
+    except subprocess.CalledProcessError as e:
+        print(f"Error running command: {command}\n{e}")
+        exit(1)
+
+def download_model():
+    """Download the Llama model from HuggingFace."""
+    model_url = "https://huggingface.co/Orenguteng/Llama-3.1-8B-Lexi-Uncensored-V2-GGUF/resolve/main/Llama-3.1-8B-Lexi-Uncensored_V2_Q4.gguf"
+    print(f"Downloading model from {model_url}...")
+    run_command(f"wget {model_url} -O llama.gguf")
+
+def start_ollama_service():
+    """Start Ollama serve in the background."""
+    print("Starting Ollama service...")
+    run_command("ollama serve &")
+    while True:
+        try:
+            response = requests.get("http://localhost:11434")
+            if response.status_code == 200:
+                break
+        except requests.ConnectionError:
+            time.sleep(2)
+
+def create_model():
+    """Create the model using Ollama."""
+    print("Creating the model in Ollama...")
+    run_command("ollama create llama -f Modelfile")
+
+def setup_ngrok():
+    run_command("pip install pyngrok")
+    ngrok_authtoken = os.getenv("NGROK_AUTHTOKEN")
+    if not ngrok_authtoken:
+        print("Error: NGROK_AUTHTOKEN secret not found.")
+        exit(1)
+
+    run_command(f"./ngrok authtoken {ngrok_authtoken}")
+    run_command("ngrok http 11434 &")
+
+    # Wait for the public URL from ngrok
+    public_url = None
+    while not public_url:
+        try:
+            tunnels_response = requests.get("http://127.0.0.1:4040/api/tunnels")
+            public_url = tunnels_response.json()['tunnels'][0]['public_url']
+        except (requests.ConnectionError, KeyError, IndexError):
+            time.sleep(2)
+
+    print(f"Ngrok public URL: {public_url}")
+    return public_url
+
+if __name__ == "__main__":
+    run_command("pip install ollama")
+    download_model()
+    create_model()
+    start_ollama_service()
+    setup_ngrok()
+    print("Ollama service is running and accessible through ngrok.")