Update app.py

app.py

@@ -1,61 +1,78 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+from rpc import load_index, generate
 
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+def setup():
+    """ Downloads the memory mapped vector index (~10GB), installs NGT and loads the index"""
+    import os
+    import requests
+    import subprocess
+    from stream_unzip import stream_unzip
+    
+    DATASET_URL = os.environ.get("DATASET_URL")
+    if not DATASET_URL:
+        raise ValueError("DATASET_URL must be set in the environment")
+    extract_dir = "/dev/shm/rpc-vecdb"
+    os.makedirs(extract_dir, exist_ok=True)
+    response = requests.get(DATASET_URL, stream=True)
+    response.raise_for_status()
+    
+    print("Starting streaming extraction to /dev/shm...")
+    for filename, filesize, file_iter in stream_unzip(response.iter_content(chunk_size=8192)):
+        if isinstance(filename, bytes):
+            filename = filename.decode('utf-8')
+        file_path = os.path.join(extract_dir, filename)
+        os.makedirs(os.path.dirname(file_path), exist_ok=True)
+        with open(file_path, 'wb') as f_out:
+            for chunk in file_iter:
+                f_out.write(chunk)
+        print(f"Extracted: {filename} -> {file_path}")
+    
+    files = os.listdir(extract_dir)
+    files = [f for f in files if os.path.isfile(os.path.join(extract_dir, f))]
+    for f in files: print(f)
+    print("Index extracted")
+
+    print("Installing NGT...")
+    subprocess.check_call(["bash", "install_ngt.sh"])
+    print("NGT installed")
+
+    print("Loading index...")
+    load_index(extract_dir)
+    print("Index loaded")
+    
 
 
 def respond(
     message,
     history: list[tuple[str, str]],
-    system_message,
+    user_name,
+    ai_name,
+    use_rpc,
     max_tokens,
     temperature,
-    top_p,
 ):
-    messages = [{"role": "system", "content": system_message}]
-
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
+    prompt = "<s>"
+    
+    for m in history:
+        prompt += f"{user_name}: {m[0]}\n{ai_name}: {m[1]}\n"
+    
+    for response in generate(prompt, use_rpc=use_rpc, max_tokens=max_tokens)
         yield response
 
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
+
 demo = gr.ChatInterface(
     respond,
     additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
+        gr.Textbox(value="Peter", label="User name"),
+        gr.Textbox(value="Sarah", label="Assistant name"),
+        gr.Checkbox(
+            label="Use RPC",
+            info="Compare Normal vs. RPC-Enhanced Model",
+            value=True
         ),
+        gr.Slider(minimum=1, maximum=320, value=128, step=1, label="Max new tokens"),
+        gr.Slider(minimum=0.1, maximum=3.0, value=0.2, step=0.1, label="Temperature (only used without RPC)"),
     ],
 )