Spaces:
Sleeping
Sleeping
Update tools/web_search.py
Browse files- tools/web_search.py +56 -20
tools/web_search.py
CHANGED
|
@@ -1,27 +1,63 @@
|
|
| 1 |
from typing import Any, Optional
|
| 2 |
from smolagents.tools import Tool
|
| 3 |
-
import duckduckgo_search
|
|
|
|
|
|
|
| 4 |
|
| 5 |
-
class
|
| 6 |
-
name = "
|
| 7 |
-
description = "Performs a
|
| 8 |
inputs = {'query': {'type': 'string', 'description': 'The search query to perform.'}}
|
| 9 |
output_type = "string"
|
| 10 |
|
| 11 |
-
def
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
from typing import Any, Optional
|
| 2 |
from smolagents.tools import Tool
|
| 3 |
+
# import duckduckgo_search
|
| 4 |
+
import requests
|
| 5 |
+
from xml.etree import ElementTree
|
| 6 |
|
| 7 |
+
class PubmedSearchTool(Tool):
|
| 8 |
+
name = "pubmed_search"
|
| 9 |
+
description = "Performs a pubmed search based on your query then returns the top search results."
|
| 10 |
inputs = {'query': {'type': 'string', 'description': 'The search query to perform.'}}
|
| 11 |
output_type = "string"
|
| 12 |
|
| 13 |
+
def pubmed_search_tool(query: str):
|
| 14 |
+
"""Outil pour rechercher des articles médicaux sur PubMed en fonction d'une requête."""
|
| 15 |
+
|
| 16 |
+
base_url = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi"
|
| 17 |
+
params = {
|
| 18 |
+
"db": "pubmed",
|
| 19 |
+
"term": query, # Terme de recherche (ex: "diabetes treatment")
|
| 20 |
+
"retmax": 5, # Nombre maximum de résultats
|
| 21 |
+
"retmode": "xml" # Format XML pour récupérer les résultats
|
| 22 |
+
}
|
| 23 |
+
|
| 24 |
+
response = requests.get(base_url, params=params)
|
| 25 |
+
|
| 26 |
+
if response.status_code != 200:
|
| 27 |
+
return "Erreur lors de la recherche sur PubMed."
|
| 28 |
+
|
| 29 |
+
# Parser la réponse XML pour extraire les IDs des articles
|
| 30 |
+
root = ElementTree.fromstring(response.content)
|
| 31 |
+
article_ids = [id_elem.text for id_elem in root.findall(".//Id")]
|
| 32 |
+
|
| 33 |
+
if not article_ids:
|
| 34 |
+
return "Aucun article trouvé pour cette requête."
|
| 35 |
+
|
| 36 |
+
# Construire les liens vers PubMed
|
| 37 |
+
pubmed_links = [f"https://pubmed.ncbi.nlm.nih.gov/{article_id}/" for article_id in article_ids]
|
| 38 |
+
|
| 39 |
+
return "\n".join(pubmed_links)
|
| 40 |
|
| 41 |
+
# class DuckDuckGoSearchTool(Tool):
|
| 42 |
+
# name = "web_search"
|
| 43 |
+
# description = "Performs a duckduckgo web search based on your query (think a Google search) then returns the top search results."
|
| 44 |
+
# inputs = {'query': {'type': 'string', 'description': 'The search query to perform.'}}
|
| 45 |
+
# output_type = "string"
|
| 46 |
+
|
| 47 |
+
# def __init__(self, max_results=10, **kwargs):
|
| 48 |
+
# super().__init__()
|
| 49 |
+
# self.max_results = max_results
|
| 50 |
+
# try:
|
| 51 |
+
# from duckduckgo_search import DDGS
|
| 52 |
+
# except ImportError as e:
|
| 53 |
+
# raise ImportError(
|
| 54 |
+
# "You must install package `duckduckgo_search` to run this tool: for instance run `pip install duckduckgo-search`."
|
| 55 |
+
# ) from e
|
| 56 |
+
# self.ddgs = DDGS(**kwargs)
|
| 57 |
+
|
| 58 |
+
# def forward(self, query: str) -> str:
|
| 59 |
+
# results = self.ddgs.text(query, max_results=self.max_results)
|
| 60 |
+
# if len(results) == 0:
|
| 61 |
+
# raise Exception("No results found! Try a less restrictive/shorter query.")
|
| 62 |
+
# postprocessed_results = [f"[{result['title']}]({result['href']})\n{result['body']}" for result in results]
|
| 63 |
+
# return "## Search Results\n\n" + "\n\n".join(postprocessed_results)
|