mirror of
https://github.com/zvx-echo6/meshai.git
synced 2026-05-21 23:24:44 +02:00
K7ZVX
3bf5e3dfbc
- Notification pipeline: categories -> rules -> channels - Channels: mesh broadcast, mesh DM, email (SMTP), webhook (generic) - Per-rule severity threshold and category filtering - Quiet hours with emergency override - LLM summarization for mesh delivery over 200 chars only - !subscribe shows available categories, easy mesh subscription - Dashboard notification rules API endpoints - Extensible channel system for future transports (Winlink, JS8Call) - config.yaml notification section with examples Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
64 lines
1.9 KiB
Python
64 lines
1.9 KiB
Python
"""Message summarizer for mesh delivery."""
|
|
|
|
import logging
|
|
from typing import Optional, TYPE_CHECKING
|
|
|
|
if TYPE_CHECKING:
|
|
from ..backends import LLMBackend
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
class MessageSummarizer:
|
|
"""Summarizes long messages for mesh delivery.
|
|
|
|
Only used when:
|
|
- Delivering to mesh channels (broadcast or DM)
|
|
- Message exceeds max_chars (default 200)
|
|
- LLM backend is available
|
|
|
|
Email and webhook channels receive full messages.
|
|
"""
|
|
|
|
def __init__(self, llm_backend: Optional["LLMBackend"] = None):
|
|
self._llm = llm_backend
|
|
|
|
async def summarize(self, message: str, max_chars: int = 195) -> str:
|
|
"""Summarize a message to fit within max_chars.
|
|
|
|
Args:
|
|
message: Original message text
|
|
max_chars: Maximum characters for summary
|
|
|
|
Returns:
|
|
Summarized message, or truncated original if LLM unavailable
|
|
"""
|
|
if len(message) <= max_chars:
|
|
return message
|
|
|
|
if not self._llm:
|
|
return message[:max_chars - 3] + "..."
|
|
|
|
prompt = (
|
|
"Summarize this alert in under %d characters. "
|
|
"Keep severity, location, and key facts. No preamble, just the summary:\n\n%s"
|
|
% (max_chars, message)
|
|
)
|
|
|
|
try:
|
|
# Use the LLM to generate a summary
|
|
response = await self._llm.generate(
|
|
prompt,
|
|
system_prompt="You are a concise alert summarizer. Output only the summary, no explanation.",
|
|
max_tokens=100,
|
|
)
|
|
summary = response.strip()
|
|
|
|
# Ensure it fits
|
|
if len(summary) <= max_chars:
|
|
return summary
|
|
return summary[:max_chars - 3] + "..."
|
|
|
|
except Exception as e:
|
|
logger.debug("LLM summarization failed: %s", e)
|
|
return message[:max_chars - 3] + "..."
|