File size: 8,265 Bytes
2e77168
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03d90b7
2e77168
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f5c4c58
2e77168
 
 
 
 
 
 
 
 
 
 
 
f5c4c58
2e77168
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
import base64
import os
import time

import gradio as gr
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 modelscope_studio.components.pro.chatbot import (ChatbotBotConfig,
                                                      ChatbotPromptsConfig,
                                                      ChatbotWelcomeConfig)
from modelscope_studio.components.pro.multimodal_input import \
  MultimodalInputUploadConfig
from openai import OpenAI

client = OpenAI(
    base_url='https://api-inference.modelscope.cn/v1/',
    api_key=os.getenv("MODELSCOPE_ACCESS_TOKEN"),  # ModelScope Token
)

model = "Qwen/Qwen2.5-VL-72B-Instruct"


def prompt_select(input_value, e: gr.EventData):
    input_value["text"] = e._data["payload"][0]["value"]["description"]
    return gr.update(value=input_value)


def clear():
    return gr.update(value=None)


def cancel(chatbot_value):
    chatbot_value[-1]["loading"] = False
    chatbot_value[-1]["status"] = "done"
    chatbot_value[-1]["footer"] = "Chat completion paused"

    return gr.update(value=chatbot_value), gr.update(loading=False), gr.update(
        disabled=False)


def retry(chatbot_value, e: gr.EventData):
    index = e._data["payload"][0]["index"]
    chatbot_value = chatbot_value[:index]

    yield gr.update(loading=True), gr.update(value=chatbot_value), gr.update(
        disabled=True)
    for chunk in submit(None, chatbot_value):
        yield chunk


def image_to_base64(image_path):
    with open(image_path, "rb") as image_file:
        encoded_string = base64.b64encode(image_file.read()).decode('utf-8')
    return f"data:image/jpeg;base64,{encoded_string}"


def format_history(history):
    messages = [{"role": "system", "content": "You are a helpful assistant."}]
    for item in history:
        if item["role"] == "user":
            messages.append({
                "role":
                "user",
                "content": [{
                    "type": "image_url",
                    "image_url": image_to_base64(file)
                } for file in item["content"][0]["content"]
                            if os.path.exists(file)] +
                [{
                    "type": "text",
                    "text": item["content"][1]["content"]
                }]
            })
        elif item["role"] == "assistant":
            # ignore thought message
            messages.append({"role": "assistant", "content": item["content"]})
    return messages


def submit(input_value, chatbot_value):
    if input_value is not None:
        chatbot_value.append({
            "role":
            "user",
            "content": [{
                "type": "file",
                "content": [f for f in input_value["files"]]
            }, {
                "type": "text",
                "content": input_value["text"]
            }],
        })
    history_messages = format_history(chatbot_value)
    chatbot_value.append({
        "role": "assistant",
        "content": "",
        "loading": True,
        "status": "pending"
    })
    yield {
        input: gr.update(value=None, loading=True),
        clear_btn: gr.update(disabled=True),
        chatbot: gr.update(value=chatbot_value)
    }

    try:
        response = client.chat.completions.create(model=model,
                                                  messages=history_messages,
                                                  stream=True)
        start_time = time.time()

        for chunk in response:
            chatbot_value[-1]["content"] += chunk.choices[0].delta.content
            chatbot_value[-1]["loading"] = False

            yield {chatbot: gr.update(value=chatbot_value)}

        chatbot_value[-1]["footer"] = "{:.2f}".format(time.time() -
                                                      start_time) + 's'
        chatbot_value[-1]["status"] = "done"
        yield {
            clear_btn: gr.update(disabled=False),
            input: gr.update(loading=False),
            chatbot: gr.update(value=chatbot_value),
        }
    except Exception as e:
        chatbot_value[-1]["loading"] = False
        chatbot_value[-1]["status"] = "done"
        chatbot_value[-1]["content"] = "Failed to respond, please try again."
        yield {
            clear_btn: gr.update(disabled=False),
            input: gr.update(loading=False),
            chatbot: gr.update(value=chatbot_value),
        }
        raise e


with gr.Blocks() as demo, ms.Application(), antdx.XProvider(), ms.AutoLoading(
):
    with antd.Flex(vertical=True, gap="middle", elem_style=dict(height=800)):
        chatbot = pro.Chatbot(
            # for flex=1 to fill the remaining space
            height=0,
            elem_style=dict(flex=1),
            welcome_config=ChatbotWelcomeConfig(
                variant="borderless",
                icon=
                "https://assets.alicdn.com/g/qwenweb/qwen-webui-fe/0.0.44/static/favicon.png",
                title=f"Hello, I'm {model}",
                description="You can upload images and text to get started.",
                prompts=ChatbotPromptsConfig(
                    title="How can I help you today?",
                    styles={
                        "list": {
                            "width": '100%',
                        },
                        "item": {
                            "flex": 1,
                        },
                    },
                    items=[{
                        "label":
                        "πŸ“… Make a plan",
                        "children": [{
                            "description":
                            "Help me with a plan to start a business"
                        }, {
                            "description":
                            "Help me with a plan to achieve my goals"
                        }, {
                            "description":
                            "Help me with a plan for a successful interview"
                        }]
                    }, {
                        "label":
                        "πŸ–‹ Help me write",
                        "children": [{
                            "description":
                            "Help me write a story with a twist ending"
                        }, {
                            "description":
                            "Help me write a blog post on mental health"
                        }, {
                            "description":
                            "Help me write a letter to my future self"
                        }]
                    }])),
            user_config=dict(
                avatar="https://api.dicebear.com/7.x/miniavs/svg?seed=3"),
            bot_config=ChatbotBotConfig(
                header=model,
                actions=["copy", "retry"],
                avatar=
                "https://assets.alicdn.com/g/qwenweb/qwen-webui-fe/0.0.44/static/favicon.png"
            ),
        )

        with pro.MultimodalInput(upload_config=MultimodalInputUploadConfig(
                max_count=4, multiple=True, accept="image/*")) as input:
            with ms.Slot("prefix"):
                with antd.Button(value=None, color="default",
                                 variant="text") as clear_btn:
                    with ms.Slot("icon"):
                        antd.Icon("ClearOutlined")
        clear_btn.click(fn=clear, outputs=[chatbot])
        submit_event = input.submit(fn=submit,
                                    inputs=[input, chatbot],
                                    outputs=[input, chatbot, clear_btn])
        input.cancel(fn=cancel,
                     inputs=[chatbot],
                     outputs=[chatbot, input, clear_btn],
                     cancels=[submit_event],
                     queue=False)
        chatbot.retry(fn=retry,
                      inputs=[chatbot],
                      outputs=[input, chatbot, clear_btn])
        chatbot.welcome_prompt_select(fn=prompt_select,
                                      inputs=[input],
                                      outputs=[input])
if __name__ == "__main__":
    demo.queue().launch()