Spaces:
Sleeping
Sleeping
Added VisitWebpageTool
Browse files
app.py
CHANGED
|
@@ -20,6 +20,47 @@ huggingface_hub.login(os.getenv('HF_TOKEN'))
|
|
| 20 |
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
|
| 21 |
|
| 22 |
# Initialize the search tool
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
search_tool = DuckDuckGoSearchTool()
|
| 24 |
|
| 25 |
@tool
|
|
@@ -82,7 +123,7 @@ with open("prompts.yaml", 'r') as stream:
|
|
| 82 |
|
| 83 |
agent = CodeAgent(
|
| 84 |
model=model,
|
| 85 |
-
tools=[final_answer, say_hello, image_generation_tool, search_tool], ## add your tools here (don't remove final answer)
|
| 86 |
max_steps=6,
|
| 87 |
verbosity_level=1,
|
| 88 |
grammar=None,
|
|
|
|
| 20 |
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
|
| 21 |
|
| 22 |
# Initialize the search tool
|
| 23 |
+
|
| 24 |
+
class VisitWebpageTool(Tool):
|
| 25 |
+
name = "visit_webpage"
|
| 26 |
+
description = (
|
| 27 |
+
"Visits a webpage at the given url and reads its content as a markdown string. Use this to browse webpages."
|
| 28 |
+
)
|
| 29 |
+
inputs = {
|
| 30 |
+
"url": {
|
| 31 |
+
"type": "string",
|
| 32 |
+
"description": "The url of the webpage to visit.",
|
| 33 |
+
}
|
| 34 |
+
}
|
| 35 |
+
output_type = "string"
|
| 36 |
+
|
| 37 |
+
def forward(self, url: str) -> str:
|
| 38 |
+
try:
|
| 39 |
+
import re
|
| 40 |
+
|
| 41 |
+
import requests
|
| 42 |
+
from markdownify import markdownify
|
| 43 |
+
from requests.exceptions import RequestException
|
| 44 |
+
|
| 45 |
+
from smolagents.utils import truncate_content
|
| 46 |
+
except ImportError as e:
|
| 47 |
+
raise ImportError(
|
| 48 |
+
"You must install packages `markdownify` and `requests` to run this tool: for instance run `pip install markdownify requests`."
|
| 49 |
+
) from e
|
| 50 |
+
try:
|
| 51 |
+
response = requests.get(url, timeout=20)
|
| 52 |
+
response.raise_for_status() # Raise an exception for bad status codes
|
| 53 |
+
markdown_content = markdownify(response.text).strip()
|
| 54 |
+
markdown_content = re.sub(r"\n{3,}", "\n\n", markdown_content)
|
| 55 |
+
return truncate_content(markdown_content, 40000)
|
| 56 |
+
|
| 57 |
+
except requests.exceptions.Timeout:
|
| 58 |
+
return "The request timed out. Please try again later or check the URL."
|
| 59 |
+
except RequestException as e:
|
| 60 |
+
return f"Error fetching the webpage: {str(e)}"
|
| 61 |
+
except Exception as e:
|
| 62 |
+
return f"An unexpected error occurred: {str(e)}"
|
| 63 |
+
|
| 64 |
search_tool = DuckDuckGoSearchTool()
|
| 65 |
|
| 66 |
@tool
|
|
|
|
| 123 |
|
| 124 |
agent = CodeAgent(
|
| 125 |
model=model,
|
| 126 |
+
tools=[final_answer, say_hello, image_generation_tool, search_tool, VisitWebpageTool()], ## add your tools here (don't remove final answer)
|
| 127 |
max_steps=6,
|
| 128 |
verbosity_level=1,
|
| 129 |
grammar=None,
|