sylvain
a8727654af
- agent1.py : bot XMPP connecté à Ollama avec boucle agentique - skills/web_search.py : recherche DuckDuckGo (ddgs) - skills/web_read.py : lecture et extraction de pages web - skills/memory.py : mémoire persistante SQLite (REMEMBER/RECALL) - skills/loader.py : chargement dynamique des skills
39 lines
1.1 KiB
Python
39 lines
1.1 KiB
Python
"""
|
|
Skill : READ
|
|
Télécharge une page web et la convertit en texte lisible.
|
|
"""
|
|
import urllib.request
|
|
from bs4 import BeautifulSoup
|
|
|
|
SKILL_NAME = "read"
|
|
TRIGGER = "READ:"
|
|
|
|
MAX_CHARS = 4000
|
|
|
|
def execute(args: str) -> str:
|
|
url = args.strip()
|
|
if not url:
|
|
return "Erreur : URL vide."
|
|
try:
|
|
req = urllib.request.Request(url, headers={"User-Agent": "Mozilla/5.0"})
|
|
with urllib.request.urlopen(req, timeout=15) as r:
|
|
html = r.read()
|
|
|
|
soup = BeautifulSoup(html, "html.parser")
|
|
|
|
# Supprimer scripts, styles, nav
|
|
for tag in soup(["script", "style", "nav", "footer", "header"]):
|
|
tag.decompose()
|
|
|
|
text = soup.get_text(separator="\n")
|
|
lines = [l.strip() for l in text.splitlines() if l.strip()]
|
|
content = "\n".join(lines)
|
|
|
|
if len(content) > MAX_CHARS:
|
|
content = content[:MAX_CHARS] + "\n...[tronqué]"
|
|
|
|
return "Contenu de {} :\n{}".format(url, content)
|
|
|
|
except Exception as e:
|
|
return "Erreur lors de la lecture de {} : {}".format(url, e)
|