File size: 7,386 Bytes
4f3daba
86de97d
 
 
 
4f3daba
86de97d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b5cd9e9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
86de97d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4f3daba
 
 
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
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
import gradio as gr
from openai import AzureOpenAI
import os
from dotenv import load_dotenv
import time

def load_environment():
    """Load environment variables."""
    load_dotenv(override=True)

def initialize_openai_client():
    """Initialize the Azure OpenAI client."""
    return AzureOpenAI(
        azure_endpoint=os.getenv("AZURE_OPENAI_ENDPOINT"),
        api_key=os.getenv("AZURE_OPENAI_API_KEY"),
        api_version="2024-10-01-preview"
    )

def create_assistant(client, vector_store_id):
    """Create an assistant with specified configuration."""
    return client.beta.assistants.create(
        model="gpt-4o",
        instructions="ζŒ‡η€ΊγŒγͺγ„ι™γ‚Šγ€ζ—₯本θͺžγ§ε›žη­”してください。",
        tools=[{
            "type": "file_search",
            "file_search": {"ranking_options": {"ranker": "default_2024_08_21", "score_threshold": 0}}
        }],
        tool_resources={"file_search": {"vector_store_ids": [vector_store_id]}},
        temperature=0
    )

def create_thread(client):
    """Create a new thread."""
    return client.beta.threads.create()

def clear_thread(_):
    """Clear the chat history and reset the thread."""
    global thread
    thread = create_thread(client)
    return [], ""

def get_annotations(msg):
    annotations = msg.content[0].text.annotations
    file_ids = []
    if annotations:
        for annotation in annotations:
            file_id = annotation.file_citation.file_id
            if file_id in file_ids:
                continue

            print("file_id", file_id)

            cited_file = client.files.retrieve(file_id)

            print("filename", cited_file.filename)

            try:
                content = client.files.content(file_id)
            except Exception as e:
                print(e)
                pass

            file_ids.append(file_id)

    return file_ids

def get_chatbot_response(client, thread_id, assistant_id, message):
    """Get chatbot response for a given message."""
    client.beta.threads.messages.create(
        thread_id=thread_id,
        role="user",
        content=message  # Ensure the content is an object with a `text` key
    )

    run = client.beta.threads.runs.create(
        thread_id=thread_id,
        assistant_id=assistant_id
    )

    while run.status in ["queued", "in_progress", "cancelling"]:
        time.sleep(1)
        run = client.beta.threads.runs.retrieve(
            thread_id=thread_id,
            run_id=run.id
        )

    if run.status == "completed":
        messages = client.beta.threads.messages.list(thread_id=thread_id)

        for msg in messages:
            # file_ids = get_annotations(msg)

            main_text = msg.content[0].text.value
            # main_text += "\n> aaa"
            return main_text

    elif run.status == "requires_action":
        # Handle cases where the assistant requires further action
        pass

    return "Unable to retrieve a response."  # Fallback response

def chatbot_response(history, message):

    """Wrapper function to generate chatbot response."""
    global thread
    # Get response from the API
    assistant_response = get_chatbot_response(client, thread.id, assistant.id, message)

    # Update chat history
    history.append({"role": "user", "content": message})
    history.append({"role": "assistant", "content": assistant_response})

    return history, ""

# Load environment variables
load_environment()

# Initialize OpenAI client
client = initialize_openai_client()

# Define vector store ID
vector_store_id = os.getenv("AZURE_OPENAI_VECTOR_STORE_ID")

# Create assistant and thread
assistant = create_assistant(client, vector_store_id)
thread = create_thread(client)

with gr.Blocks() as demo:
    import gradio as gr
from openai import AzureOpenAI
import os
from dotenv import load_dotenv
import time

def load_environment():
    """Load environment variables."""
    load_dotenv(override=True)

def initialize_openai_client():
    """Initialize the Azure OpenAI client."""
    return AzureOpenAI(
        azure_endpoint=os.getenv("AZURE_OPENAI_ENDPOINT"),
        api_key=os.getenv("AZURE_OPENAI_API_KEY"),
        api_version="2024-10-01-preview"
    )

def create_assistant(client, vector_store_id):
    """Create an assistant with specified configuration."""
    return client.beta.assistants.create(
        model="gpt-4o",
        instructions="",
        tools=[{
            "type": "file_search",
            "file_search": {"ranking_options": {"ranker": "default_2024_08_21", "score_threshold": 0}}
        }],
        tool_resources={"file_search": {"vector_store_ids": [vector_store_id]}},
        temperature=0
    )

def create_thread(client):
    """Create a new thread."""
    return client.beta.threads.create()

def clear_thread(_):
    """Clear the chat history and reset the thread."""
    global thread
    thread = create_thread(client)
    return [], ""

def get_chatbot_response(client, thread_id, assistant_id, message):
    """Get chatbot response for a given message."""
    client.beta.threads.messages.create(
        thread_id=thread_id,
        role="user",
        content=message  # Ensure the content is an object with a `text` key
    )

    run = client.beta.threads.runs.create(
        thread_id=thread_id,
        assistant_id=assistant_id
    )

    while run.status in ["queued", "in_progress", "cancelling"]:
        time.sleep(1)
        run = client.beta.threads.runs.retrieve(
            thread_id=thread_id,
            run_id=run.id
        )

    if run.status == "completed":
        messages = client.beta.threads.messages.list(thread_id=thread_id)

        for msg in messages:
            main_text = msg.content[0].text.value
            return main_text

    return "Unable to retrieve a response."  # Fallback response

def chatbot_response(history, message):
    """Wrapper function to generate chatbot response."""
    global thread
    # Get response from the API
    assistant_response = get_chatbot_response(client, thread.id, assistant.id, message)

    # Update chat history
    history.append({"role": "user", "content": message})
    history.append({"role": "assistant", "content": assistant_response})

    return history, ""

# Load environment variables
load_environment()

# Initialize OpenAI client
client = initialize_openai_client()

# Define vector store ID
vector_store_id = os.getenv("AZURE_OPENAI_VECTOR_STORE_ID")

# Create assistant and thread
assistant = create_assistant(client, vector_store_id)
thread = create_thread(client)

with gr.Blocks() as demo:
    gr.Markdown("""
        # Azure OpenAI Assistants API x Gradio x Zenn
        This is a Gradio demo of Retrieval-Augmented Generation (RAG) using the Azure OpenAI Assistants API, applied to [Zenn articles](https://zenn.dev/nakamura196).
    """)

    chatbot = gr.Chatbot(type="messages")
    msg = gr.Textbox()
    clear = gr.ClearButton([msg, chatbot])

    def respond(message, chat_history):

        bot_message = get_chatbot_response(client, thread.id, assistant.id, message)

        chat_history.append({"role": "user", "content": message})
        chat_history.append({"role": "assistant", "content": bot_message})
        return "", chat_history

    msg.submit(respond, [msg, chatbot], [msg, chatbot])

    clear.click(clear_thread, [chatbot])

if __name__ == "__main__":
    demo.launch()