KraTUZen's picture
essential changes
fde9314
import uuid
import time
import re
import gradio as gr
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from transformers import StoppingCriteria, StoppingCriteriaList, TextIteratorStreamer
from threading import Thread
import modelscope_studio.components.antd as antd
import modelscope_studio.components.antdx as antdx
import modelscope_studio.components.base as ms
import modelscope_studio.components.pro as pro
from config import DEFAULT_LOCALE, DEFAULT_THEME, get_text, user_config, bot_config, welcome_config
from ui_components.logo import Logo
from ui_components.settings_header import SettingsHeader
# Loading the tokenizer and model from Hugging Face's model hub
tokenizer = AutoTokenizer.from_pretrained("TinyLlama/TinyLlama-1.1B-Chat-v1.0")
model = AutoModelForCausalLM.from_pretrained("TinyLlama/TinyLlama-1.1B-Chat-v1.0")
# Using CUDA for an optimal experience
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model = model.to(device)
# Defining a custom stopping criteria class for the model's text generation
class StopOnTokens(StoppingCriteria):
def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
stop_ids = [2] # IDs of tokens where the generation should stop.
for stop_id in stop_ids:
if input_ids[0][-1] == stop_id:
return True
return False
# Function to generate model predictions with streaming
def generate_response(user_input, history):
stop = StopOnTokens()
messages = "</s>".join([
"</s>".join([
"\n<|user|>:" + item["content"] if item["role"] == "user"
else "\n<|assistant|>:" + item["content"]
for item in history
])
])
messages += f"\n<|user|>:{user_input}\n<|assistant|>:"
model_inputs = tokenizer([messages], return_tensors="pt").to(device)
streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)
generate_kwargs = dict(
**model_inputs,
streamer=streamer,
max_new_tokens=1024,
do_sample=True,
top_p=0.95,
top_k=50,
temperature=0.7,
num_beams=1,
stopping_criteria=StoppingCriteriaList([stop])
)
t = Thread(target=model.generate, kwargs=generate_kwargs)
t.start() # Start generation in a separate thread.
partial_message = ""
for new_token in streamer:
partial_message += new_token
if '</s>' in partial_message:
break
return partial_message
# Define the system prompt for seeding the model's context
SYSTEM_PROMPT = (
"I am LogicLink, Version 5, A state-of-the-art AI chatbot created and engineered by "
"Kratu Gautam"
"I am here to assist you with any queries. How can I help you today?"
)
class Gradio_Events:
_generating = False
@staticmethod
def new_chat(state_value):
# This is CRITICAL - we DO NOT clean up old conversation
# Instead, we leave it in the state to be accessed later
# Create a fresh conversation
new_id = str(uuid.uuid4())
state_value["conversation_id"] = new_id
# Add the new conversation to the list with a default name
state_value["conversations"].append({
"label": "New Chat",
"key": new_id
})
# Seed it with system prompt
state_value["conversation_contexts"][new_id] = {
"history": [{
"role": "system",
"content": SYSTEM_PROMPT,
"key": str(uuid.uuid4()),
"avatar": None
}]
}
# Return updates
return (
gr.update(items=state_value["conversations"]),
gr.update(value=state_value["conversation_contexts"][new_id]["history"]),
gr.update(value=state_value),
gr.update(value="") # empties input
)
@staticmethod
def add_message(input_value, state_value):
input_update = gr.update(value="")
# If input is empty, just return
if not input_value.strip():
conversation = state_value["conversation_contexts"].get(state_value["conversation_id"], {"history": []})
chatbot_update = gr.update(value=conversation["history"])
state_update = gr.update(value=state_value)
return input_update, chatbot_update, state_update
# If there's no active conversation, initialize a new one
if not state_value["conversation_id"]:
random_id = str(uuid.uuid4())
state_value["conversation_id"] = random_id
state_value["conversation_contexts"][random_id] = {"history": [{
"role": "system",
"content": SYSTEM_PROMPT,
"key": str(uuid.uuid4()),
"avatar": None
}]}
# Set the chat name to the first message from user
chat_name = input_value[:20] + ("..." if len(input_value) > 20 else "")
state_value["conversations"].append({
"label": chat_name,
"key": random_id
})
else:
# Get current conversation history
current_id = state_value["conversation_id"]
history = state_value["conversation_contexts"][current_id]["history"]
# If this is the first user message (after system message), update the label
user_messages = [msg for msg in history if msg["role"] == "user"]
if len(user_messages) == 0:
# This is the first user message - update the chat name
chat_name = input_value[:20] + ("..." if len(input_value) > 20 else "")
for i, conv in enumerate(state_value["conversations"]):
if conv["key"] == current_id:
state_value["conversations"][i]["label"] = chat_name
break
# Add the message to history
history = state_value["conversation_contexts"][state_value["conversation_id"]]["history"]
history.append({
"role": "user",
"content": input_value,
"key": str(uuid.uuid4()),
"avatar": None
})
chatbot_update = gr.update(value=history)
return input_update, chatbot_update, gr.update(value=state_value)
@staticmethod
def submit(state_value):
if Gradio_Events._generating:
history = state_value["conversation_contexts"].get(state_value["conversation_id"], {"history": []})["history"]
return (
gr.update(value=history),
gr.update(value=state_value),
gr.update(value="Generation in progress, please wait...")
)
Gradio_Events._generating = True
# Make sure we have a valid conversation ID
if not state_value["conversation_id"]:
Gradio_Events._generating = False
return (
gr.update(value=[]),
gr.update(value=state_value),
gr.update(value="No active conversation")
)
history = state_value["conversation_contexts"][state_value["conversation_id"]]["history"]
# Assuming the last message is the latest user input
user_input = history[-1]["content"] if (history and history[-1]["role"] == "user") else ""
if not user_input:
Gradio_Events._generating = False
return (
gr.update(value=history),
gr.update(value=state_value),
gr.update(value="No user input provided")
)
# Generate the response from the model
history, response = Gradio_Events.logiclink_chat(user_input, history)
state_value["conversation_contexts"][state_value["conversation_id"]]["history"] = history
Gradio_Events._generating = False
return (
gr.update(value=history),
gr.update(value=state_value),
gr.update(value=response)
)
@staticmethod
def logiclink_chat(user_input, history):
if not user_input:
return history, "No input provided"
try:
start = time.time()
response = generate_response(user_input, history)
elapsed = time.time() - start
# Clean and format the response before appending it
cleaned_response = re.sub(r'\*\(\d+\.\d+s\)\*', '', response).strip()
response_with_time = f"{cleaned_response}\n\n*({elapsed:.2f}s)*"
history.append({
"role": "assistant",
"content": response_with_time,
"key": str(uuid.uuid4()),
"avatar": None
})
return history, response_with_time
except Exception as e:
error_msg = (
f"Generation failed: {str(e)}. "
"Possible causes: insufficient memory, model incompatibility, or input issues."
)
history.append({
"role": "assistant",
"content": error_msg,
"key": str(uuid.uuid4()),
"avatar": None
})
return history, error_msg
@staticmethod
def clear_history(state_value):
if state_value["conversation_id"]:
# Only clear messages after system prompt
current_history = state_value["conversation_contexts"][state_value["conversation_id"]]["history"]
if len(current_history) > 0 and current_history[0]["role"] == "system":
system_message = current_history[0]
state_value["conversation_contexts"][state_value["conversation_id"]]["history"] = [system_message]
else:
state_value["conversation_contexts"][state_value["conversation_id"]]["history"] = []
# Return the cleared history
return (
gr.update(value=state_value["conversation_contexts"][state_value["conversation_id"]]["history"]),
gr.update(value=state_value),
gr.update(value="")
)
return (
gr.update(value=[]),
gr.update(value=state_value),
gr.update(value="")
)
@staticmethod
def delete_conversation(state_value, conversation_key):
# Keep a copy of the conversations before removal
new_conversations = [conv for conv in state_value["conversations"] if conv["key"] != conversation_key]
# Remove the conversation from the list
state_value["conversations"] = new_conversations
# Delete the conversation context
if conversation_key in state_value["conversation_contexts"]:
del state_value["conversation_contexts"][conversation_key]
# If we're deleting the active conversation
if state_value["conversation_id"] == conversation_key:
state_value["conversation_id"] = ""
return gr.update(items=new_conversations), gr.update(value=[]), gr.update(value=state_value)
# If deleting another conversation, keep the current one displayed
return (
gr.update(items=new_conversations),
gr.update(value=state_value["conversation_contexts"].get(
state_value["conversation_id"], {"history": []}
)["history"]),
gr.update(value=state_value)
)
# (The remainder of your Gradio UI code remains largely unchanged.)
css = """
:root {
--color-red: #ff4444;
--color-blue: #1e88e5;
--color-black: #000000;
--color-dark-gray: #121212;
}
.gradio-container { background: var(--color-black) !important; color: white !important; }
.gr-textbox textarea, .ms-gr-ant-input-textarea { background: var(--color-dark-gray) !important; border: 2px solid var(--color-blue) !important; color: white !important; }
.gr-chatbot { background: var(--color-dark-gray) !important; border: 2px solid var(--color-red) !important; }
.gr-textbox.output-textbox { background: var(--color-dark-gray) !important; border: 2px solid var(--color-red) !important; color: white !important; margin-bottom: 10px; }
.gr-chatbot .user { background: var(--color-blue) !important; border-color: var(--color-blue) !important; }
.gr-chatbot .bot { background: var(--color-dark-gray) !important; border: 1px solid var(--color-red) !important; }
.gr-button { background: var(--color-blue) !important; border-color: var(--color-blue) !important; }
.gr-chatbot .tool { background: var(--color-dark-gray) !important; border: 1px solid var(--color-red) !important; }
"""
with gr.Blocks(css=css, fill_width=True, title="LogicLinkV5") as demo:
state = gr.State({
"conversation_contexts": {},
"conversations": [],
"conversation_id": "",
})
with ms.Application(), antdx.XProvider(theme=DEFAULT_THEME, locale=DEFAULT_LOCALE), ms.AutoLoading():
with antd.Row(gutter=[20, 20], wrap=False, elem_id="chatbot"):
# Left Column
with antd.Col(md=dict(flex="0 0 260px", span=24, order=0), span=0, order=1):
with ms.Div(elem_classes="chatbot-conversations"):
with antd.Flex(vertical=True, gap="small", elem_style=dict(height="100%")):
Logo()
with antd.Button(color="primary", variant="filled", block=True, elem_classes="new-chat-btn") as new_chat_btn:
ms.Text(get_text("New Chat", "新建对话"))
with ms.Slot("icon"):
antd.Icon("PlusOutlined")
with antdx.Conversations(elem_classes="chatbot-conversations-list") as conversations:
with ms.Slot('menu.items'):
with antd.Menu.Item(label="Delete", key="delete", danger=True) as conversation_delete_menu_item:
with ms.Slot("icon"):
antd.Icon("DeleteOutlined")
# Right Column
with antd.Col(flex=1, elem_style=dict(height="100%")):
with antd.Flex(vertical=True, gap="small", elem_classes="chatbot-chat"):
chatbot = pro.Chatbot(elem_classes="chatbot-chat-messages", height=600,
welcome_config=welcome_config(), user_config=user_config(),
bot_config=bot_config())
output_textbox = gr.Textbox(label="LatestOutputTextbox", lines=1,
elem_classes="output-textbox", interactive=True)
with antdx.Suggestion(items=[]):
with ms.Slot("children"):
with antdx.Sender(placeholder="Type your message...", elem_classes="chat-input") as input:
with ms.Slot("prefix"):
with antd.Flex(gap=4):
with antd.Button(type="text", elem_classes="clear-btn") as clear_btn:
with ms.Slot("icon"):
antd.Icon("ClearOutlined")
# Event Handlers
input.submit(fn=Gradio_Events.add_message, inputs=[input, state],
outputs=[input, chatbot, state]).then(
fn=Gradio_Events.submit, inputs=[state],
outputs=[chatbot, state, output_textbox]
)
new_chat_btn.click(fn=Gradio_Events.new_chat,
inputs=[state],
outputs=[conversations, chatbot, state, input],
queue=False)
clear_btn.click(fn=Gradio_Events.clear_history, inputs=[state],
outputs=[chatbot, state, output_textbox])
conversations.menu_click(
fn=lambda state_value, e: (
# If there's no payload, skip
gr.skip() if (e is None or not isinstance(e, dict) or 'key' not in e._data['payload'][0] or 'menu_key' not in e._data['payload'][1])
else (
# Extract keys
(lambda conv_key, action_key: (
# If "delete", remove that convo
Gradio_Events.delete_conversation(state_value, conv_key)
if action_key == "delete"
# If other action, do nothing
else (
gr.update(items=state_value["conversations"]),
gr.update(value=state_value["conversation_contexts"]
.get(state_value["conversation_id"], {"history": []})
["history"]),
gr.update(value=state_value)
)
))(
e._data['payload'][0]['key'],
e._data['payload'][1]['key']
)
)
),
inputs=[state],
outputs=[conversations, chatbot, state],
queue=False
)
demo.queue().launch(share=True, debug=True)