Files
Bjorn/bifrost/epoch.py
infinition aac77a3e76 Add Loki and Sentinel utility classes for web API endpoints
- Implemented LokiUtils class with GET and POST endpoints for managing scripts, jobs, and payloads.
- Added SentinelUtils class with GET and POST endpoints for managing events, rules, devices, and notifications.
- Both classes include error handling and JSON response formatting.
2026-03-14 22:33:10 +01:00

293 lines
9.4 KiB
Python

"""
Bifrost — Epoch tracking.
Ported from pwnagotchi/ai/epoch.py + pwnagotchi/ai/reward.py.
"""
import time
import threading
import logging
import os
from logger import Logger
logger = Logger(name="bifrost.epoch", level=logging.DEBUG)
NUM_CHANNELS = 14 # 2.4 GHz channels
# ── Reward function (from pwnagotchi/ai/reward.py) ──────────────
class RewardFunction:
"""Reward signal for RL — higher is better."""
def __call__(self, epoch_n, state):
eps = 1e-20
tot_epochs = epoch_n + eps
tot_interactions = max(
state['num_deauths'] + state['num_associations'],
state['num_handshakes']
) + eps
tot_channels = NUM_CHANNELS
# Positive signals
h = state['num_handshakes'] / tot_interactions
a = 0.2 * (state['active_for_epochs'] / tot_epochs)
c = 0.1 * (state['num_hops'] / tot_channels)
# Negative signals
b = -0.3 * (state['blind_for_epochs'] / tot_epochs)
m = -0.3 * (state['missed_interactions'] / tot_interactions)
i = -0.2 * (state['inactive_for_epochs'] / tot_epochs)
_sad = state['sad_for_epochs'] if state['sad_for_epochs'] >= 5 else 0
_bored = state['bored_for_epochs'] if state['bored_for_epochs'] >= 5 else 0
s = -0.2 * (_sad / tot_epochs)
l_val = -0.1 * (_bored / tot_epochs)
return h + a + c + b + i + m + s + l_val
# ── Epoch state ──────────────────────────────────────────────────
class BifrostEpoch:
"""Tracks per-epoch counters, observations, and reward."""
def __init__(self, config):
self.epoch = 0
self.config = config
# Consecutive epoch counters
self.inactive_for = 0
self.active_for = 0
self.blind_for = 0
self.sad_for = 0
self.bored_for = 0
# Per-epoch action flags & counters
self.did_deauth = False
self.num_deauths = 0
self.did_associate = False
self.num_assocs = 0
self.num_missed = 0
self.did_handshakes = False
self.num_shakes = 0
self.num_hops = 0
self.num_slept = 0
self.num_peers = 0
self.tot_bond_factor = 0.0
self.avg_bond_factor = 0.0
self.any_activity = False
# Timing
self.epoch_started = time.time()
self.epoch_duration = 0
# Channel histograms for AI observation
self.non_overlapping_channels = {1: 0, 6: 0, 11: 0}
self._observation = {
'aps_histogram': [0.0] * NUM_CHANNELS,
'sta_histogram': [0.0] * NUM_CHANNELS,
'peers_histogram': [0.0] * NUM_CHANNELS,
}
self._observation_ready = threading.Event()
self._epoch_data = {}
self._epoch_data_ready = threading.Event()
self._reward = RewardFunction()
def wait_for_epoch_data(self, with_observation=True, timeout=None):
self._epoch_data_ready.wait(timeout)
self._epoch_data_ready.clear()
if with_observation:
return {**self._observation, **self._epoch_data}
return self._epoch_data
def data(self):
return self._epoch_data
def observe(self, aps, peers):
"""Update observation histograms from current AP/peer lists."""
num_aps = len(aps)
if num_aps == 0:
self.blind_for += 1
else:
self.blind_for = 0
bond_unit_scale = self.config.get('bifrost_personality_bond_factor', 20000)
self.num_peers = len(peers)
num_peers = self.num_peers + 1e-10
self.tot_bond_factor = sum(
p.get('encounters', 0) if isinstance(p, dict) else getattr(p, 'encounters', 0)
for p in peers
) / bond_unit_scale
self.avg_bond_factor = self.tot_bond_factor / num_peers
num_aps_f = len(aps) + 1e-10
num_sta = sum(len(ap.get('clients', [])) for ap in aps) + 1e-10
aps_per_chan = [0.0] * NUM_CHANNELS
sta_per_chan = [0.0] * NUM_CHANNELS
peers_per_chan = [0.0] * NUM_CHANNELS
for ap in aps:
ch_idx = ap.get('channel', 1) - 1
if 0 <= ch_idx < NUM_CHANNELS:
aps_per_chan[ch_idx] += 1.0
sta_per_chan[ch_idx] += len(ap.get('clients', []))
for peer in peers:
ch = peer.get('last_channel', 0) if isinstance(peer, dict) else getattr(peer, 'last_channel', 0)
ch_idx = ch - 1
if 0 <= ch_idx < NUM_CHANNELS:
peers_per_chan[ch_idx] += 1.0
# Normalize
aps_per_chan = [e / num_aps_f for e in aps_per_chan]
sta_per_chan = [e / num_sta for e in sta_per_chan]
peers_per_chan = [e / num_peers for e in peers_per_chan]
self._observation = {
'aps_histogram': aps_per_chan,
'sta_histogram': sta_per_chan,
'peers_histogram': peers_per_chan,
}
self._observation_ready.set()
def track(self, deauth=False, assoc=False, handshake=False,
hop=False, sleep=False, miss=False, inc=1):
"""Increment epoch counters."""
if deauth:
self.num_deauths += inc
self.did_deauth = True
self.any_activity = True
if assoc:
self.num_assocs += inc
self.did_associate = True
self.any_activity = True
if miss:
self.num_missed += inc
if hop:
self.num_hops += inc
# Reset per-channel flags on hop
self.did_deauth = False
self.did_associate = False
if handshake:
self.num_shakes += inc
self.did_handshakes = True
if sleep:
self.num_slept += inc
def next(self):
"""Transition to next epoch — compute reward, update streaks, reset counters."""
# Update activity streaks
if not self.any_activity and not self.did_handshakes:
self.inactive_for += 1
self.active_for = 0
else:
self.active_for += 1
self.inactive_for = 0
self.sad_for = 0
self.bored_for = 0
sad_threshold = self.config.get('bifrost_personality_sad_epochs', 25)
bored_threshold = self.config.get('bifrost_personality_bored_epochs', 15)
if self.inactive_for >= sad_threshold:
self.bored_for = 0
self.sad_for += 1
elif self.inactive_for >= bored_threshold:
self.sad_for = 0
self.bored_for += 1
else:
self.sad_for = 0
self.bored_for = 0
now = time.time()
self.epoch_duration = now - self.epoch_started
# System metrics
cpu = _cpu_load()
mem = _mem_usage()
temp = _temperature()
# Cache epoch data for other threads
self._epoch_data = {
'duration_secs': self.epoch_duration,
'slept_for_secs': self.num_slept,
'blind_for_epochs': self.blind_for,
'inactive_for_epochs': self.inactive_for,
'active_for_epochs': self.active_for,
'sad_for_epochs': self.sad_for,
'bored_for_epochs': self.bored_for,
'missed_interactions': self.num_missed,
'num_hops': self.num_hops,
'num_peers': self.num_peers,
'tot_bond': self.tot_bond_factor,
'avg_bond': self.avg_bond_factor,
'num_deauths': self.num_deauths,
'num_associations': self.num_assocs,
'num_handshakes': self.num_shakes,
'cpu_load': cpu,
'mem_usage': mem,
'temperature': temp,
}
self._epoch_data['reward'] = self._reward(self.epoch + 1, self._epoch_data)
self._epoch_data_ready.set()
logger.info(
"[epoch %d] dur=%ds blind=%d sad=%d bored=%d inactive=%d active=%d "
"hops=%d missed=%d deauths=%d assocs=%d shakes=%d reward=%.3f",
self.epoch, int(self.epoch_duration), self.blind_for,
self.sad_for, self.bored_for, self.inactive_for, self.active_for,
self.num_hops, self.num_missed, self.num_deauths, self.num_assocs,
self.num_shakes, self._epoch_data['reward'],
)
# Reset for next epoch
self.epoch += 1
self.epoch_started = now
self.did_deauth = False
self.num_deauths = 0
self.num_peers = 0
self.tot_bond_factor = 0.0
self.avg_bond_factor = 0.0
self.did_associate = False
self.num_assocs = 0
self.num_missed = 0
self.did_handshakes = False
self.num_shakes = 0
self.num_hops = 0
self.num_slept = 0
self.any_activity = False
# ── System metric helpers ────────────────────────────────────────
def _cpu_load():
try:
return os.getloadavg()[0]
except (OSError, AttributeError):
return 0.0
def _mem_usage():
try:
with open('/proc/meminfo', 'r') as f:
lines = f.readlines()
total = int(lines[0].split()[1])
available = int(lines[2].split()[1])
return (total - available) / total if total else 0.0
except Exception:
return 0.0
def _temperature():
try:
with open('/sys/class/thermal/thermal_zone0/temp', 'r') as f:
return int(f.read().strip()) / 1000.0
except Exception:
return 0.0