Update main.py and start_server.sh scripts

main.py

@@ -1,7 +1,7 @@
 from langchain.schema import AIMessage, HumanMessage
 import gradio as gr
 from langchain_community.llms import Ollama
-
+import time
 def parse_model_names(path):
     """Parses the model file to extract value-label pairs for the dropdown."""
     choices = []
@@ -19,39 +19,66 @@ models = parse_model_names("models.txt")
 
 
 def predict(message, history, model):
-    print("Predicting", message, history, models[model][1]), 
+    print("Predicting", message, history, model), 
     llm = Ollama(model=models[model][1], timeout=1000)  # Instantiate Ollama with the selected model
     history_langchain_format = []
-    for human, ai in history:
-        history_langchain_format.append(HumanMessage(content=human))
-        history_langchain_format.append(AIMessage(content=ai))
-    history_langchain_format.append(HumanMessage(content=message))
+    for m in message:
+        history_langchain_format.append(HumanMessage(content=m[0]))
+        if m[1] is not None:
+            history_langchain_format.append(AIMessage(content=m[1]))
     try:
         chat_response = llm.invoke(history_langchain_format)
     except Exception as e:  # Use a general exception handler here
         chat_response = "Error: " + str(e)
-        
-    return chat_response
+    
+    return [(chat_response, )]
 
 
 
-with gr.Blocks(fill_height=True) as demo:
-    with gr.Row():
-        model_dropdown = gr.Dropdown(label="Select LLM Model", choices=models, info="Select the model you want to chat with", type="index")
+# with gr.Blocks(fill_height=True) as demo:
+#     with gr.Row():
 
-        # We use a state variable to track the current model
-        model_state = gr.State(value=model_dropdown.value)
 
-        def update_model(selected_model):
-            print("Model selected", selected_model)
-            model_state.value = selected_model
-            return selected_model
+#         def update_model(selected_model):
+#             print("Model selected", selected_model)
+#             model_state.value = selected_model
+#             return selected_model
         
         
-    chat = gr.ChatInterface(predict,
-        additional_inputs=[ model_dropdown ],
+#     chat = gr.ChatInterface(predict,
+#         additional_inputs=[ model_dropdown ],
         
+#     )
+      
+
+def print_like_dislike(x: gr.LikeData):
+    print(x.index, x.value, x.liked)
+
+def add_message(history, message):
+    for x in message["files"]:
+        history.append(((x,), None))
+    if message["text"] is not None:
+        history.append((message["text"], None))
+    return history, gr.MultimodalTextbox(value=None, interactive=False)
+
+with gr.Blocks() as demo:
+    model_dropdown = gr.Dropdown(label="Select LLM Model", choices=models, info="Select the model you want to chat with", type="index")
+    model_state = gr.State(value=model_dropdown.value)
+    chatbot = gr.Chatbot(
+        [],
+        elem_id="chatbot",
+        bubble_full_width=False
     )
+
+    chat_input = gr.MultimodalTextbox(interactive=True, file_types=["image"], placeholder="Enter message or upload file...", show_label=False)
+
+    chat_msg = chat_input.submit(add_message, [chatbot, chat_input], [chatbot, chat_input])
+    bot_msg = chat_msg.then(predict, [chatbot, chat_input, model_dropdown], chatbot, api_name="bot_response")
+    bot_msg.then(lambda: gr.MultimodalTextbox(interactive=True), None, [chat_input])
+
+    chatbot.like(print_like_dislike, None, None)
+
+demo.queue()
         
 
 if __name__ == "__main__":

start_server.sh

@@ -6,7 +6,7 @@ ollama pull llama3:8b   > /dev/null 2>&1
 ollama pull gemma:2b    > /dev/null 2>&1
 ollama pull gemma:7b    > /dev/null 2>&1
 
-ollama create mistral4k:7b --file .\mistral7b.Modelfile > /dev/null 2>&1
-ollama create llama38k:8b  --file .\llama38b.Modelfile  > /dev/null 2>&1
+ollama create mistral4k:7b --file /home/user/app/mistral7b.Modelfile
+ollama create llama38k:8b --file /home/user/app/llama38b.Modelfile
 
 python main.py