File size: 7,667 Bytes
b527097
4754c75
3568413
b527097
4754c75
 
 
 
26aec96
 
82e5cca
 
b527097
3568413
 
809f87e
 
283e426
3568413
 
809f87e
283e426
 
809f87e
 
 
 
 
283e426
3568413
 
809f87e
283e426
 
809f87e
 
 
 
 
283e426
3568413
 
283e426
 
 
 
 
 
 
 
 
3568413
 
283e426
 
 
 
 
 
 
 
 
 
 
3568413
 
283e426
 
 
 
 
 
809f87e
283e426
 
3568413
 
809f87e
283e426
 
809f87e
283e426
3568413
 
 
 
 
26aec96
3568413
 
 
26aec96
 
 
3568413
 
b527097
 
 
3568413
b527097
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
26aec96
 
 
 
4754c75
b527097
26aec96
b527097
3568413
 
b527097
3568413
b527097
 
3568413
b527097
 
 
 
 
 
 
 
3568413
 
 
 
b527097
3568413
b527097
3568413
 
 
 
b527097
3568413
b527097
 
 
 
3568413
b527097
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3568413
 
4754c75
 
 
 
 
 
 
 
 
 
 
26aec96
 
 
 
 
 
 
 
4754c75
26aec96
 
 
 
 
 
 
 
4754c75
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
82e5cca
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
import requests
from pydantic import BaseModel, Field
from huggingface_hub import InferenceClient
from openai import OpenAI
from bs4 import BeautifulSoup
from markdownify import markdownify as md
from langchain_core.tools import tool, Tool
from langchain_experimental.utilities import PythonREPL
from pypdf import PdfReader
from io import BytesIO
from youtube_transcript_api import YouTubeTranscriptApi
from pytube import extract


# --- Basic operations --- #

@tool
def multiply(a: float, b: float) -> float:
    """Multiplies two numbers.

    Args:
        a (float): the first number
        b (float): the second number
    """
    return a * b


@tool
def add(a: float, b: float) -> float:
    """Adds two numbers.

    Args:
        a (float): the first number
        b (float): the second number
    """
    return a + b


@tool
def subtract(a: float, b: float) -> int:
    """Subtracts two numbers.

    Args:
        a (float): the first number
        b (float): the second number
    """
    return a - b


@tool
def divide(a: float, b: float) -> float:
    """Divides two numbers.

    Args:
        a (float): the first float number
        b (float): the second float number
    """
    if b == 0:
        raise ValueError("Cannot divided by zero.")
    return a / b


@tool
def modulus(a: int, b: int) -> int:
    """Get the modulus of two numbers.

    Args:
        a (int): the first number
        b (int): the second number
    """
    return a % b


@tool
def power(a: float, b: float) -> float:
    """Get the power of two numbers.

    Args:
        a (float): the first number
        b (float): the second number
    """
    return a**b


# --- Functions --- #

@tool
def query_image(query: str, image_url: str, need_reasoning: bool = False) -> str:
    """Ask anything about an image using a Vision Language Model

    Args:
        query (str): The query about the image, e.g. how many persons are on the image?
        image_url (str): The URL to the image
        need_reasoning (bool): Set to True for complex query that require a reasoning model to answer properly. Set to False otherwise.
    """

    # PROVIDER = 'huggingface'
    PROVIDER = 'openai'

    try:
        if PROVIDER == 'huggingface':
            client = InferenceClient(provider="nebius")
            completion = client.chat.completions.create(
                # model="google/gemma-3-27b-it",
                model="Qwen/Qwen2.5-VL-72B-Instruct",
                messages=[
                    {
                        "role": "user",
                        "content": [
                            {
                                "type": "text",
                                "text": query
                            },
                            {
                                "type": "image_url",
                                "image_url": {
                                    "url": image_url
                                }
                            }
                        ]
                    }
                ],
                max_tokens=512,
            )
            return completion.choices[0].message

        elif PROVIDER == 'openai':
            if need_reasoning:
                model_name = "o4-mini"
            else:
                model_name = "gpt-4.1-mini"
            client = OpenAI()
            response = client.responses.create(
                model=model_name,
                input=[{
                    "role": "user",
                    "content": [
                        {"type": "input_text", "text": query},
                        {
                            "type": "input_image",
                            "image_url": image_url,
                        },
                    ],
                }],
            )

            return response.output_text

        else:
            raise AttributeError(f'PROVIDER must be "openai" or "huggingface", received "{PROVIDER}"')

    except Exception as e:
        return f"query_image failed: {e}"


@tool
def automatic_speech_recognition(file_url: str, file_extension: str) -> str:
    """Transcribe an audio file to text

    Args:
        file_url (str): the URL to the audio file
        file_extension (str): the file extension, e.g. mp3
    """

    # PROVIDER = 'huggingface'
    PROVIDER = 'openai'

    try:
        if PROVIDER == 'huggingface':
            client = InferenceClient(provider="fal-ai")
            return client.automatic_speech_recognition(file_url, model="openai/whisper-large-v3")

        elif PROVIDER == 'openai':
            # download the audio file
            response = requests.get(file_url)
            response.raise_for_status()
            # write to disk
            file_extension = file_extension.replace('.','')
            with open(f'tmp.{file_extension}', 'wb') as file:
                file.write(response.content)

            audio_file = open(f'tmp.{file_extension}', "rb")
            client = OpenAI()
            transcription = client.audio.transcriptions.create(
                model="whisper-1",
                file=audio_file
            )
            return transcription.text

        else:
            raise AttributeError(f'PROVIDER must be "openai" or "huggingface", received "{PROVIDER}"')

    except Exception as e:
        return f"automatic_speech_recognition failed: {e}"


@tool
def get_webpage_content(page_url: str) -> str:
    """Load a web page and return it to markdown if possible

    Args:
        page_url (str): the URL of web page to get
    """
    try:
        r = requests.get(page_url)
        r.raise_for_status()
        text = ""
        # special case if page is a PDF file
        if r.headers.get('Content-Type', '') == 'application/pdf':
            pdf_file = BytesIO(r.content)
            reader = PdfReader(pdf_file)
            for page in reader.pages:
                text += page.extract_text()
        else:
            soup = BeautifulSoup((r.text), 'html.parser')
            if soup.body:
                # convert to markdown
                text = md(str(soup.body))
            else:
                # return the raw content
                text = r.text
        return text
    except Exception as e:
        return f"get_webpage_content failed: {e}"


# ======= Python code interpreter =======
# WARNING: Python REPL can execute arbitrary code on the host machine (e.g., delete files, make network requests). Use with caution.

class PythonREPLInput(BaseModel):
    code: str = Field(description="The Python code string to execute.")

python_repl = PythonREPL()

python_repl_tool = Tool(
    name="python_repl",
    description="""A Python REPL shell (Read-Eval-Print Loop).
Use this to execute single or multi-line python commands.
Input should be syntactically valid Python code.
Always end your code with `print(...)` to see the output.
Do NOT execute code that could be harmful to the host system.
You are allowed to download files from URLs.
Do NOT send commands that block indefinitely (e.g., `input()`).""",
    func=python_repl.run,
    args_schema=PythonREPLInput
)

@tool
def get_youtube_transcript(page_url: str) -> str:
    """Get the transcript of a YouTube video

    Args:
        page_url (str): YouTube URL of the video
    """
    try:
        # get video ID from URL
        video_id = extract.video_id(page_url)

        # get transcript
        ytt_api = YouTubeTranscriptApi()
        transcript = ytt_api.fetch(video_id)

        # keep only text
        txt = '\n'.join([s.text for s in transcript.snippets])
        return txt
    except Exception as e:
        return f"get_youtube_transcript failed: {e}"