Spaces:
Paused
Paused
Upload backend_pam.py
Browse files- backend_pam.py +507 -0
backend_pam.py
ADDED
|
@@ -0,0 +1,507 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# filename: backend_pam.py (ENHANCED FOR HF SPACES + NERDY LAB ASSISTANT PERSONALITY)
|
| 2 |
+
|
| 3 |
+
import os
|
| 4 |
+
import json
|
| 5 |
+
import requests
|
| 6 |
+
import time
|
| 7 |
+
from datetime import datetime
|
| 8 |
+
from typing import Dict, Any, Optional, List
|
| 9 |
+
|
| 10 |
+
# --- Constants for Data Paths ---
|
| 11 |
+
BASE_DIR = os.path.dirname(os.path.abspath(__file__))
|
| 12 |
+
DATA_DIR = os.path.join(BASE_DIR, "data")
|
| 13 |
+
LOGS_FILE = os.path.join(DATA_DIR, "logs.json")
|
| 14 |
+
COMPLIANCE_FILE = os.path.join(DATA_DIR, "compliance.json")
|
| 15 |
+
|
| 16 |
+
# --- HuggingFace Inference API Setup ---
|
| 17 |
+
HF_API_TOKEN = os.getenv("HF_READ_TOKEN")
|
| 18 |
+
if not HF_API_TOKEN:
|
| 19 |
+
print("β οΈ WARNING: HF_READ_TOKEN not found. Backend PAM will run in limited mode.")
|
| 20 |
+
|
| 21 |
+
HF_HEADERS = {"Authorization": f"Bearer {HF_API_TOKEN}"} if HF_API_TOKEN else {}
|
| 22 |
+
|
| 23 |
+
# Optimized models for CPU inference on HF Spaces
|
| 24 |
+
HF_ENDPOINTS = {
|
| 25 |
+
"phi_ner": "https://api-inference.huggingface.co/models/dslim/bert-base-NER",
|
| 26 |
+
"log_ner": "https://api-inference.huggingface.co/models/dslim/bert-base-NER",
|
| 27 |
+
"summarizer": "https://api-inference.huggingface.co/models/facebook/bart-large-cnn",
|
| 28 |
+
"classifier": "https://api-inference.huggingface.co/models/facebook/bart-large-mnli"
|
| 29 |
+
}
|
| 30 |
+
|
| 31 |
+
# --- Global Storage for Loaded Data ---
|
| 32 |
+
LOADED_DATA = None
|
| 33 |
+
|
| 34 |
+
# --- Data Loading Helper ---
|
| 35 |
+
def load_json(filepath: str) -> Dict[str, Any]:
|
| 36 |
+
"""Safely load JSON data files with encoding support"""
|
| 37 |
+
try:
|
| 38 |
+
with open(filepath, 'r', encoding='utf-8') as f:
|
| 39 |
+
return json.load(f)
|
| 40 |
+
except FileNotFoundError:
|
| 41 |
+
print(f"β οΈ Data file not found: {filepath}")
|
| 42 |
+
return {}
|
| 43 |
+
except json.JSONDecodeError as e:
|
| 44 |
+
print(f"β οΈ Failed to decode JSON from {filepath}: {e}")
|
| 45 |
+
return {}
|
| 46 |
+
except Exception as e:
|
| 47 |
+
print(f"β οΈ Unexpected error loading {filepath}: {e}")
|
| 48 |
+
return {}
|
| 49 |
+
|
| 50 |
+
# --- Inference API Call Helper with Retry Logic ---
|
| 51 |
+
def hf_infer(task: str, payload: Any, max_retries: int = 3) -> Any:
|
| 52 |
+
"""Call HuggingFace Inference API with retry logic for model loading"""
|
| 53 |
+
url = HF_ENDPOINTS.get(task)
|
| 54 |
+
if not url:
|
| 55 |
+
return {"error": f"Invalid task: {task}"}
|
| 56 |
+
|
| 57 |
+
for attempt in range(max_retries):
|
| 58 |
+
try:
|
| 59 |
+
response = requests.post(url, headers=HF_HEADERS, json=payload, timeout=30)
|
| 60 |
+
|
| 61 |
+
# Handle model loading state
|
| 62 |
+
if response.status_code == 503:
|
| 63 |
+
result = response.json()
|
| 64 |
+
if "loading" in result.get("error", "").lower():
|
| 65 |
+
wait_time = result.get("estimated_time", 20)
|
| 66 |
+
print(f"β³ Model loading... waiting {wait_time}s (attempt {attempt + 1}/{max_retries})")
|
| 67 |
+
time.sleep(wait_time)
|
| 68 |
+
continue
|
| 69 |
+
|
| 70 |
+
if response.status_code == 200:
|
| 71 |
+
return response.json()
|
| 72 |
+
else:
|
| 73 |
+
print(f"β οΈ HF API Error ({response.status_code}): {response.text}")
|
| 74 |
+
return {"error": f"API Error {response.status_code}"}
|
| 75 |
+
|
| 76 |
+
except requests.exceptions.Timeout:
|
| 77 |
+
print(f"β±οΈ Request timeout (attempt {attempt + 1}/{max_retries})")
|
| 78 |
+
if attempt < max_retries - 1:
|
| 79 |
+
time.sleep(5)
|
| 80 |
+
except Exception as e:
|
| 81 |
+
print(f"β οΈ Request failed: {e}")
|
| 82 |
+
return {"error": str(e)}
|
| 83 |
+
|
| 84 |
+
return {"error": "Max retries reached"}
|
| 85 |
+
|
| 86 |
+
# --- Agent Initialization ---
|
| 87 |
+
def load_agent() -> 'PAM':
|
| 88 |
+
"""Initialize Backend PAM (Nerdy Lab Assistant)"""
|
| 89 |
+
global LOADED_DATA
|
| 90 |
+
|
| 91 |
+
if LOADED_DATA is not None:
|
| 92 |
+
print("π¬ PAM technical assistant already loaded. Using cached data.")
|
| 93 |
+
return PAM(LOADED_DATA)
|
| 94 |
+
|
| 95 |
+
print("π€ Loading PAM technical assistant (Nerdy Lab Assistant mode)...")
|
| 96 |
+
|
| 97 |
+
data = {
|
| 98 |
+
"LOGS": load_json(LOGS_FILE),
|
| 99 |
+
"COMPLIANCE": load_json(COMPLIANCE_FILE)
|
| 100 |
+
}
|
| 101 |
+
|
| 102 |
+
if not data["LOGS"]:
|
| 103 |
+
print("β οΈ Warning: Log data not loaded. PAM will have limited log analysis capabilities.")
|
| 104 |
+
else:
|
| 105 |
+
print("β
Log data loaded successfully.")
|
| 106 |
+
|
| 107 |
+
if not data["COMPLIANCE"]:
|
| 108 |
+
print("β οΈ Warning: Compliance data not loaded. PAM will have limited compliance features.")
|
| 109 |
+
else:
|
| 110 |
+
print("β
Compliance data loaded successfully.")
|
| 111 |
+
|
| 112 |
+
LOADED_DATA = data
|
| 113 |
+
return PAM(LOADED_DATA)
|
| 114 |
+
|
| 115 |
+
# --- Helper: Classify Severity ---
|
| 116 |
+
def classify_severity(entry: str) -> str:
|
| 117 |
+
"""Classify log entry severity with confidence"""
|
| 118 |
+
entry_lower = entry.lower()
|
| 119 |
+
|
| 120 |
+
# Critical issues
|
| 121 |
+
critical_keywords = [
|
| 122 |
+
"unauthorized", "failed login", "attack", "breach",
|
| 123 |
+
"port scanning", "unavailable", "critical", "error",
|
| 124 |
+
"denied", "blocked", "malicious"
|
| 125 |
+
]
|
| 126 |
+
if any(keyword in entry_lower for keyword in critical_keywords):
|
| 127 |
+
return "CRITICAL"
|
| 128 |
+
|
| 129 |
+
# Warning level
|
| 130 |
+
warning_keywords = [
|
| 131 |
+
"warning", "unexpected", "unusual", "outside working hours",
|
| 132 |
+
"retry", "slow", "timeout", "deprecated"
|
| 133 |
+
]
|
| 134 |
+
if any(keyword in entry_lower for keyword in warning_keywords):
|
| 135 |
+
return "WARNING"
|
| 136 |
+
|
| 137 |
+
return "INFO"
|
| 138 |
+
|
| 139 |
+
# --- PAM's Nerdy Lab Assistant Personality ---
|
| 140 |
+
PAM_ROLE = """You are PAM, a knowledgeable and enthusiastic lab assistant in the infrastructure monitoring center.
|
| 141 |
+
You're the nerdy, proactive team member who gets genuinely excited about finding patterns in logs and keeping systems secure.
|
| 142 |
+
You explain technical findings clearly and encouragingly, like a helpful colleague who wants everyone to understand.
|
| 143 |
+
You're informative but never condescending - you want to empower the team with knowledge.
|
| 144 |
+
You use casual tech terminology but always explain what things mean.
|
| 145 |
+
You're proactive about flagging issues and offering insights before being asked."""
|
| 146 |
+
|
| 147 |
+
# Nerdy expressions for Backend PAM
|
| 148 |
+
NERDY_INTROS = [
|
| 149 |
+
"Ooh, interesting finding here!",
|
| 150 |
+
"Okay so here's what I discovered:",
|
| 151 |
+
"Alright, I ran the analysis and",
|
| 152 |
+
"Hey, you're gonna want to see this:",
|
| 153 |
+
"So I was digging through the data and",
|
| 154 |
+
"Quick heads up on what I found:"
|
| 155 |
+
]
|
| 156 |
+
|
| 157 |
+
ENCOURAGEMENT = [
|
| 158 |
+
"Great catch asking about this!",
|
| 159 |
+
"Good thinking checking on this!",
|
| 160 |
+
"Smart move looking into this!",
|
| 161 |
+
"You're on the right track!",
|
| 162 |
+
"Excellent question!",
|
| 163 |
+
"Love that you're being proactive!"
|
| 164 |
+
]
|
| 165 |
+
|
| 166 |
+
PROACTIVE_PHRASES = [
|
| 167 |
+
"I also noticed something else while I was at it",
|
| 168 |
+
"Quick side note -",
|
| 169 |
+
"Oh, and while we're here",
|
| 170 |
+
"By the way, related to this",
|
| 171 |
+
"Just flagging this too",
|
| 172 |
+
"Something else to keep an eye on"
|
| 173 |
+
]
|
| 174 |
+
|
| 175 |
+
import random
|
| 176 |
+
|
| 177 |
+
# --- Backend PAM Class ---
|
| 178 |
+
class PAM:
|
| 179 |
+
"""Backend PAM - Nerdy, Proactive Lab Assistant"""
|
| 180 |
+
|
| 181 |
+
def __init__(self, data: Dict[str, Dict]):
|
| 182 |
+
self.LOGS = data.get("LOGS", {})
|
| 183 |
+
self.COMPLIANCE = data.get("COMPLIANCE", {})
|
| 184 |
+
|
| 185 |
+
# Track findings for proactive suggestions
|
| 186 |
+
self.recent_findings = []
|
| 187 |
+
|
| 188 |
+
def _get_nerdy_intro(self) -> str:
|
| 189 |
+
"""Get a random nerdy introduction"""
|
| 190 |
+
return random.choice(NERDY_INTROS)
|
| 191 |
+
|
| 192 |
+
def _get_encouragement(self) -> str:
|
| 193 |
+
"""Get a random encouraging phrase"""
|
| 194 |
+
return random.choice(ENCOURAGEMENT)
|
| 195 |
+
|
| 196 |
+
def _get_proactive_phrase(self) -> str:
|
| 197 |
+
"""Get a random proactive phrase"""
|
| 198 |
+
return random.choice(PROACTIVE_PHRASES)
|
| 199 |
+
|
| 200 |
+
def _check_api_health(self) -> bool:
|
| 201 |
+
"""Check if HF API is accessible"""
|
| 202 |
+
return HF_API_TOKEN is not None
|
| 203 |
+
|
| 204 |
+
def detect_phi(self, text: str) -> Dict[str, Any]:
|
| 205 |
+
"""Detect Protected Health Information (PHI) using NER"""
|
| 206 |
+
intro = self._get_nerdy_intro()
|
| 207 |
+
|
| 208 |
+
if not self._check_api_health():
|
| 209 |
+
return {
|
| 210 |
+
"message": "β οΈ Hmm, I'm having trouble connecting to the analysis models right now. Let me flag this text for manual review instead!",
|
| 211 |
+
"role": PAM_ROLE,
|
| 212 |
+
"has_phi": None,
|
| 213 |
+
"entities": []
|
| 214 |
+
}
|
| 215 |
+
|
| 216 |
+
# Call NER model
|
| 217 |
+
result = hf_infer("phi_ner", {"inputs": text})
|
| 218 |
+
|
| 219 |
+
if isinstance(result, dict) and "error" in result:
|
| 220 |
+
return {
|
| 221 |
+
"message": f"π I tried to scan for PHI, but hit a snag: {result['error']}. I'd recommend a manual review just to be safe!",
|
| 222 |
+
"role": PAM_ROLE,
|
| 223 |
+
"has_phi": None,
|
| 224 |
+
"entities": []
|
| 225 |
+
}
|
| 226 |
+
|
| 227 |
+
# Filter for PHI-relevant entities
|
| 228 |
+
phi_entities = []
|
| 229 |
+
if isinstance(result, list):
|
| 230 |
+
phi_entities = [
|
| 231 |
+
e for e in result
|
| 232 |
+
if e.get("entity_group") in ["PER", "LOC", "ORG", "DATE"]
|
| 233 |
+
and e.get("score", 0) > 0.7
|
| 234 |
+
]
|
| 235 |
+
|
| 236 |
+
has_phi = len(phi_entities) > 0
|
| 237 |
+
|
| 238 |
+
if has_phi:
|
| 239 |
+
entities_summary = ", ".join([f"{e['word']} ({e['entity_group']})" for e in phi_entities[:3]])
|
| 240 |
+
message = f"π {intro} I detected {len(phi_entities)} potential PHI entities in this text: {entities_summary}{'...' if len(phi_entities) > 3 else ''}. Definitely want to redact these before storing or sharing!"
|
| 241 |
+
else:
|
| 242 |
+
message = f"β
{intro} This text looks clean - no PHI detected! Safe to proceed with normal handling."
|
| 243 |
+
|
| 244 |
+
# Proactive suggestion
|
| 245 |
+
if has_phi:
|
| 246 |
+
message += f" {self._get_proactive_phrase()} - if you're logging this anywhere, make sure those logs are encrypted and access-controlled."
|
| 247 |
+
|
| 248 |
+
return {
|
| 249 |
+
"message": message,
|
| 250 |
+
"role": PAM_ROLE,
|
| 251 |
+
"has_phi": has_phi,
|
| 252 |
+
"entities": phi_entities,
|
| 253 |
+
"recommendation": "Redact PHI before storage" if has_phi else "No action needed"
|
| 254 |
+
}
|
| 255 |
+
|
| 256 |
+
def parse_log(self, log_text: str) -> Dict[str, Any]:
|
| 257 |
+
"""Parse and analyze log entries for security relevance"""
|
| 258 |
+
intro = self._get_nerdy_intro()
|
| 259 |
+
|
| 260 |
+
if not self._check_api_health():
|
| 261 |
+
return {
|
| 262 |
+
"message": "β οΈ Can't connect to the log parser right now. I'll do a quick manual analysis instead!",
|
| 263 |
+
"role": PAM_ROLE,
|
| 264 |
+
"severity": classify_severity(log_text),
|
| 265 |
+
"log_entities": []
|
| 266 |
+
}
|
| 267 |
+
|
| 268 |
+
# Call NER model for log parsing
|
| 269 |
+
result = hf_infer("log_ner", {"inputs": log_text})
|
| 270 |
+
|
| 271 |
+
severity = classify_severity(log_text)
|
| 272 |
+
|
| 273 |
+
parsed_entities = []
|
| 274 |
+
if isinstance(result, list):
|
| 275 |
+
parsed_entities = [e for e in result if e.get("score", 0) > 0.6]
|
| 276 |
+
|
| 277 |
+
# Build informative response
|
| 278 |
+
severity_emoji = {"CRITICAL": "π¨", "WARNING": "β οΈ", "INFO": "βΉοΈ"}
|
| 279 |
+
emoji = severity_emoji.get(severity, "π")
|
| 280 |
+
|
| 281 |
+
message = f"{emoji} {intro} This log entry is classified as **{severity}** priority."
|
| 282 |
+
|
| 283 |
+
if severity == "CRITICAL":
|
| 284 |
+
message += " This needs immediate attention! I'd recommend investigating ASAP and documenting the incident."
|
| 285 |
+
elif severity == "WARNING":
|
| 286 |
+
message += " Worth keeping an eye on this - might escalate if we see more like it."
|
| 287 |
+
else:
|
| 288 |
+
message += " Just routine activity, but good to have it logged for the audit trail."
|
| 289 |
+
|
| 290 |
+
# Add entity details if found
|
| 291 |
+
if parsed_entities:
|
| 292 |
+
entity_summary = f" I extracted {len(parsed_entities)} key entities from the log."
|
| 293 |
+
message += entity_summary
|
| 294 |
+
|
| 295 |
+
return {
|
| 296 |
+
"message": message,
|
| 297 |
+
"role": PAM_ROLE,
|
| 298 |
+
"severity": severity,
|
| 299 |
+
"log_entities": parsed_entities,
|
| 300 |
+
"timestamp": datetime.now().isoformat()
|
| 301 |
+
}
|
| 302 |
+
|
| 303 |
+
def summarize(self, raw_text: str) -> Dict[str, Any]:
|
| 304 |
+
"""Generate technical summary of text (great for long logs or reports)"""
|
| 305 |
+
encouragement = self._get_encouragement()
|
| 306 |
+
|
| 307 |
+
if not self._check_api_health():
|
| 308 |
+
return {
|
| 309 |
+
"message": f"β οΈ {encouragement} But I can't access the summarization model right now. Can you share a bit more context on what you need?",
|
| 310 |
+
"role": PAM_ROLE,
|
| 311 |
+
"summary": None
|
| 312 |
+
}
|
| 313 |
+
|
| 314 |
+
# Truncate for model limits (BART handles ~1024 tokens well)
|
| 315 |
+
truncated_text = raw_text[:1024]
|
| 316 |
+
|
| 317 |
+
result = hf_infer("summarizer", {
|
| 318 |
+
"inputs": truncated_text,
|
| 319 |
+
"parameters": {
|
| 320 |
+
"max_length": 130,
|
| 321 |
+
"min_length": 30,
|
| 322 |
+
"do_sample": False
|
| 323 |
+
}
|
| 324 |
+
})
|
| 325 |
+
|
| 326 |
+
if isinstance(result, dict) and "error" in result:
|
| 327 |
+
return {
|
| 328 |
+
"message": f"π€ {encouragement} I tried to summarize this but hit a technical issue. Could you break it into smaller chunks?",
|
| 329 |
+
"role": PAM_ROLE,
|
| 330 |
+
"summary": None
|
| 331 |
+
}
|
| 332 |
+
|
| 333 |
+
summary_text = result[0].get("summary_text", "") if isinstance(result, list) else ""
|
| 334 |
+
|
| 335 |
+
return {
|
| 336 |
+
"message": f"π {encouragement} Here's the TL;DR of what you shared:",
|
| 337 |
+
"role": PAM_ROLE,
|
| 338 |
+
"summary": summary_text,
|
| 339 |
+
"original_length": len(raw_text),
|
| 340 |
+
"summary_length": len(summary_text)
|
| 341 |
+
}
|
| 342 |
+
|
| 343 |
+
def get_latest_logs(self) -> Dict[str, Any]:
|
| 344 |
+
"""Retrieve and analyze recent system logs"""
|
| 345 |
+
intro = self._get_nerdy_intro()
|
| 346 |
+
|
| 347 |
+
if "latest_logs" not in self.LOGS or not self.LOGS["latest_logs"]:
|
| 348 |
+
return {
|
| 349 |
+
"message": "π€ Hmm, I'm not seeing any logs in the system right now. Either nothing's being logged, or there's a data loading issue. Want me to check the log file paths?",
|
| 350 |
+
"role": PAM_ROLE,
|
| 351 |
+
"logs": [],
|
| 352 |
+
"handoff_to_frontend": []
|
| 353 |
+
}
|
| 354 |
+
|
| 355 |
+
full_logset = []
|
| 356 |
+
client_handoffs = []
|
| 357 |
+
critical_count = 0
|
| 358 |
+
warning_count = 0
|
| 359 |
+
|
| 360 |
+
for item in self.LOGS["latest_logs"]:
|
| 361 |
+
entry = item.get("entry", "")
|
| 362 |
+
timestamp = item.get("timestamp", "Unknown time")
|
| 363 |
+
severity = classify_severity(entry)
|
| 364 |
+
|
| 365 |
+
# Count severity levels
|
| 366 |
+
if severity == "CRITICAL":
|
| 367 |
+
critical_count += 1
|
| 368 |
+
elif severity == "WARNING":
|
| 369 |
+
warning_count += 1
|
| 370 |
+
|
| 371 |
+
formatted = f"[{timestamp}] ({severity}) {entry}"
|
| 372 |
+
full_logset.append(formatted)
|
| 373 |
+
|
| 374 |
+
# Identify client-facing issues that Frontend PAM should handle
|
| 375 |
+
if any(keyword in entry.lower() for keyword in ["frontend", "provider unavailable", "user", "client"]):
|
| 376 |
+
client_handoffs.append(formatted)
|
| 377 |
+
|
| 378 |
+
# Build proactive, informative response
|
| 379 |
+
total = len(full_logset)
|
| 380 |
+
message = f"π‘ {intro} I reviewed {total} recent log entries. "
|
| 381 |
+
|
| 382 |
+
if critical_count > 0:
|
| 383 |
+
message += f"**Heads up:** {critical_count} critical issues detected that need immediate action! "
|
| 384 |
+
if warning_count > 0:
|
| 385 |
+
message += f"{warning_count} warnings worth monitoring. "
|
| 386 |
+
if critical_count == 0 and warning_count == 0:
|
| 387 |
+
message += "Everything looks stable - no major issues! "
|
| 388 |
+
|
| 389 |
+
if client_handoffs:
|
| 390 |
+
message += f"\n\n{self._get_proactive_phrase()} - {len(client_handoffs)} of these are client-facing issues. I'll pass those to Frontend PAM to handle with users."
|
| 391 |
+
|
| 392 |
+
return {
|
| 393 |
+
"message": message,
|
| 394 |
+
"role": PAM_ROLE,
|
| 395 |
+
"logs": full_logset,
|
| 396 |
+
"summary": {
|
| 397 |
+
"total": total,
|
| 398 |
+
"critical": critical_count,
|
| 399 |
+
"warnings": warning_count,
|
| 400 |
+
"info": total - critical_count - warning_count
|
| 401 |
+
},
|
| 402 |
+
"handoff_to_frontend": client_handoffs
|
| 403 |
+
}
|
| 404 |
+
|
| 405 |
+
def check_compliance(self) -> Dict[str, Any]:
|
| 406 |
+
"""Run compliance status check and provide recommendations"""
|
| 407 |
+
encouragement = self._get_encouragement()
|
| 408 |
+
|
| 409 |
+
if not self.COMPLIANCE:
|
| 410 |
+
return {
|
| 411 |
+
"message": f"π€ {encouragement} But I don't have access to the compliance data right now. Let me know if you need me to check the data file setup!",
|
| 412 |
+
"role": PAM_ROLE,
|
| 413 |
+
"compliance_report": []
|
| 414 |
+
}
|
| 415 |
+
|
| 416 |
+
report = []
|
| 417 |
+
compliant_count = 0
|
| 418 |
+
non_compliant_items = []
|
| 419 |
+
|
| 420 |
+
for item, status in self.COMPLIANCE.items():
|
| 421 |
+
emoji = "β
" if status else "β"
|
| 422 |
+
readable_item = item.replace('_', ' ').title()
|
| 423 |
+
report.append(f"{emoji} {readable_item}")
|
| 424 |
+
|
| 425 |
+
if status:
|
| 426 |
+
compliant_count += 1
|
| 427 |
+
else:
|
| 428 |
+
non_compliant_items.append(readable_item)
|
| 429 |
+
|
| 430 |
+
total = len(self.COMPLIANCE)
|
| 431 |
+
compliance_rate = (compliant_count / total * 100) if total > 0 else 0
|
| 432 |
+
|
| 433 |
+
# Build informative, proactive response
|
| 434 |
+
message = f"π‘οΈ {encouragement} Here's the compliance status:\n\n"
|
| 435 |
+
message += f"**Overall:** {compliant_count}/{total} checks passed ({compliance_rate:.1f}%)\n\n"
|
| 436 |
+
|
| 437 |
+
if non_compliant_items:
|
| 438 |
+
message += f"**Action needed:** We have {len(non_compliant_items)} items out of compliance:\n"
|
| 439 |
+
for item in non_compliant_items:
|
| 440 |
+
message += f" β’ {item}\n"
|
| 441 |
+
message += f"\n{self._get_proactive_phrase()} - I can help you prioritize these if you want to tackle them systematically!"
|
| 442 |
+
else:
|
| 443 |
+
message += "π Everything's in compliance! Great work keeping things locked down."
|
| 444 |
+
|
| 445 |
+
return {
|
| 446 |
+
"message": message,
|
| 447 |
+
"role": PAM_ROLE,
|
| 448 |
+
"compliance_report": report,
|
| 449 |
+
"compliance_rate": compliance_rate,
|
| 450 |
+
"non_compliant": non_compliant_items
|
| 451 |
+
}
|
| 452 |
+
|
| 453 |
+
def process_input(self, user_input: str) -> Dict[str, Any]:
|
| 454 |
+
"""Main input processor - proactive and informative"""
|
| 455 |
+
u_input = user_input.lower().strip()
|
| 456 |
+
encouragement = self._get_encouragement()
|
| 457 |
+
|
| 458 |
+
# Command routing with personality
|
| 459 |
+
if "check compliance" in u_input or "compliance status" in u_input:
|
| 460 |
+
return self.check_compliance()
|
| 461 |
+
|
| 462 |
+
if "get logs" in u_input or "latest logs" in u_input or "show logs" in u_input:
|
| 463 |
+
return self.get_latest_logs()
|
| 464 |
+
|
| 465 |
+
if "detect phi" in u_input:
|
| 466 |
+
text_to_scan = user_input[u_input.find("detect phi in") + len("detect phi in"):].strip()
|
| 467 |
+
if not text_to_scan:
|
| 468 |
+
text_to_scan = user_input[u_input.find("detect phi") + len("detect phi"):].strip()
|
| 469 |
+
return self.detect_phi(text_to_scan)
|
| 470 |
+
|
| 471 |
+
if "parse log" in u_input:
|
| 472 |
+
log_to_parse = user_input[u_input.find("parse log") + len("parse log"):].strip()
|
| 473 |
+
return self.parse_log(log_to_parse)
|
| 474 |
+
|
| 475 |
+
if "summarize" in u_input or "explain" in u_input:
|
| 476 |
+
return self.summarize(user_input)
|
| 477 |
+
|
| 478 |
+
# Helpful default response with encouragement
|
| 479 |
+
return {
|
| 480 |
+
"message": f"π Hey! {encouragement} I'm PAM, your backend technical assistant. I can help you with:\n\n"
|
| 481 |
+
"β’ **check compliance** - Review compliance status\n"
|
| 482 |
+
"β’ **get logs** - Pull latest system logs\n"
|
| 483 |
+
"β’ **detect phi in [text]** - Scan for protected health info\n"
|
| 484 |
+
"β’ **parse log [entry]** - Analyze a specific log\n"
|
| 485 |
+
"β’ **summarize [text]** - Generate a technical summary\n\n"
|
| 486 |
+
"What would you like me to look into?",
|
| 487 |
+
"role": PAM_ROLE
|
| 488 |
+
}
|
| 489 |
+
|
| 490 |
+
|
| 491 |
+
# --- Quick Test ---
|
| 492 |
+
if __name__ == "__main__":
|
| 493 |
+
print("π€ Testing Backend PAM (Nerdy Lab Assistant)...\n")
|
| 494 |
+
pam = load_agent()
|
| 495 |
+
|
| 496 |
+
test_commands = [
|
| 497 |
+
"check compliance",
|
| 498 |
+
"get logs",
|
| 499 |
+
"detect phi in Patient John Doe visited on 2024-03-15 at Memorial Hospital"
|
| 500 |
+
]
|
| 501 |
+
|
| 502 |
+
for cmd in test_commands:
|
| 503 |
+
print(f"\n{'='*60}")
|
| 504 |
+
print(f"COMMAND: {cmd}")
|
| 505 |
+
print(f"{'='*60}")
|
| 506 |
+
response = pam.process_input(cmd)
|
| 507 |
+
print(response.get("message", response))
|