Spaces:
Sleeping
Sleeping
File size: 10,782 Bytes
d0ace1e 657a21e d0ace1e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 | """
database/session_store.py — SQLite-based longitudinal session storage.
Stores every triage result per patient and detects deterioration trends
over time using linear regression on confidence scores.
The SessionAgent autonomously triggers alerts when a worsening trend
is detected — no human has to ask for this check.
"""
import os
import sqlite3
import json
from datetime import datetime
import numpy as np
DB_PATH = '/data/sessions.db' if os.path.isdir('/data') else './data/sessions.db'
class SessionStore:
def __init__(self, db_path: str = DB_PATH):
self.db_path = db_path
os.makedirs(os.path.dirname(os.path.abspath(db_path)), exist_ok=True)
self._init_db()
def _init_db(self):
with sqlite3.connect(self.db_path) as conn:
conn.execute("""
CREATE TABLE IF NOT EXISTS sessions (
id INTEGER PRIMARY KEY AUTOINCREMENT,
patient_id TEXT NOT NULL,
timestamp TEXT NOT NULL,
tier INTEGER DEFAULT 1,
copd_confidence REAL DEFAULT 0.0,
pneu_confidence REAL DEFAULT 0.0,
severity TEXT DEFAULT 'LOW',
diagnosis TEXT DEFAULT '',
sound_type TEXT DEFAULT 'Normal',
cough_severity REAL DEFAULT 0.0,
action TEXT DEFAULT '',
symptom_index REAL DEFAULT 0.0,
voice_index REAL DEFAULT 0.0,
drift_score REAL DEFAULT 0.0,
longitudinal_score REAL DEFAULT 0.0
)
""")
# Add new columns to existing DB if upgrading
for col, typedef in [
("symptom_index", "REAL DEFAULT 0.0"),
("voice_index", "REAL DEFAULT 0.0"),
("drift_score", "REAL DEFAULT 0.0"),
("longitudinal_score", "REAL DEFAULT 0.0"),
]:
try:
conn.execute(f"ALTER TABLE sessions ADD COLUMN {col} {typedef}")
except sqlite3.OperationalError:
pass # column already exists
# Patient baselines table
conn.execute("""
CREATE TABLE IF NOT EXISTS patient_baselines (
patient_id TEXT PRIMARY KEY,
voice_features_json TEXT DEFAULT '{}',
cough_embedding BLOB,
created_at TEXT NOT NULL,
updated_at TEXT NOT NULL
)
""")
conn.commit()
# ── Baseline management ───────────────────────────────────────────────────
def get_baseline(self, patient_id: str) -> dict | None:
with sqlite3.connect(self.db_path) as conn:
cursor = conn.execute(
"SELECT * FROM patient_baselines WHERE patient_id = ?",
(patient_id,)
)
row = cursor.fetchone()
if not row:
return None
cols = [d[0] for d in cursor.description]
d = dict(zip(cols, row))
d['voice_features'] = json.loads(d.get('voice_features_json') or '{}')
if d.get('cough_embedding'):
d['cough_embedding'] = np.frombuffer(d['cough_embedding'],
dtype=np.float32)
return d
def save_baseline(self, patient_id: str,
voice_features: dict,
cough_embedding: np.ndarray | None = None):
now = datetime.now().isoformat()
emb_blob = cough_embedding.astype(np.float32).tobytes() \
if cough_embedding is not None else None
with sqlite3.connect(self.db_path) as conn:
if emb_blob is not None:
# Update both voice and cough
conn.execute("""
INSERT INTO patient_baselines
(patient_id, voice_features_json, cough_embedding,
created_at, updated_at)
VALUES (?, ?, ?, ?, ?)
ON CONFLICT(patient_id) DO UPDATE SET
voice_features_json = excluded.voice_features_json,
cough_embedding = excluded.cough_embedding,
updated_at = excluded.updated_at
""", (patient_id, json.dumps(voice_features), emb_blob, now, now))
else:
# Only update voice features — preserve existing cough embedding
conn.execute("""
INSERT INTO patient_baselines
(patient_id, voice_features_json, cough_embedding,
created_at, updated_at)
VALUES (?, ?, NULL, ?, ?)
ON CONFLICT(patient_id) DO UPDATE SET
voice_features_json = excluded.voice_features_json,
updated_at = excluded.updated_at
""", (patient_id, json.dumps(voice_features), now, now))
conn.commit()
# ── Session management ────────────────────────────────────────────────────
def save_session(self,
patient_id: str,
triage_result: dict,
copd_conf: float,
pneu_conf: float,
tier: int = 1,
sound_type: str = 'Normal',
cough_severity: float = 0.0,
symptom_index: float = 0.0,
voice_index: float = 0.0,
drift_score: float = 0.0,
longitudinal_score: float = 0.0):
with sqlite3.connect(self.db_path) as conn:
conn.execute("""
INSERT INTO sessions
(patient_id, timestamp, tier, copd_confidence, pneu_confidence,
severity, diagnosis, sound_type, cough_severity, action,
symptom_index, voice_index, drift_score, longitudinal_score)
VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
""", (
patient_id,
datetime.now().isoformat(),
tier,
round(copd_conf, 4),
round(pneu_conf, 4),
triage_result.get('severity', 'LOW'),
triage_result.get('diagnosis', ''),
sound_type,
round(float(cough_severity), 4),
triage_result.get('recommended_action', ''),
round(float(symptom_index), 4),
round(float(voice_index), 4),
round(float(drift_score), 4),
round(float(longitudinal_score), 4),
))
conn.commit()
def get_sessions(self, patient_id: str, n: int = 10) -> list:
with sqlite3.connect(self.db_path) as conn:
cursor = conn.execute("""
SELECT * FROM sessions
WHERE patient_id = ?
ORDER BY timestamp DESC
LIMIT ?
""", (patient_id, n))
rows = cursor.fetchall()
cols = [d[0] for d in cursor.description]
return [dict(zip(cols, row)) for row in rows]
def get_latest_session(self, patient_id: str) -> dict | None:
sessions = self.get_sessions(patient_id, n=1)
return sessions[0] if sessions else None
def get_all_patient_ids(self) -> list:
with sqlite3.connect(self.db_path) as conn:
cursor = conn.execute(
"SELECT DISTINCT patient_id FROM sessions ORDER BY patient_id"
)
return [row[0] for row in cursor.fetchall()]
def check_deterioration(self,
patient_id: str,
window: int = 5,
slope_threshold: float = 0.05,
conf_threshold: float = 0.45) -> list | None:
sessions = self.get_sessions(patient_id, n=window)
if len(sessions) < 3:
return None
sessions = list(reversed(sessions))
alerts = []
# Check COPD / Pneumonia audio confidence (Tier 2 sessions)
for disease, col in [('COPD', 'copd_confidence'),
('Pneumonia', 'pneu_confidence')]:
confidences = [s[col] for s in sessions]
current_conf = confidences[-1]
x = np.arange(len(confidences), dtype=float)
slope = float(np.polyfit(x, confidences, 1)[0])
if slope > slope_threshold and current_conf > conf_threshold:
alerts.append({
'disease': disease,
'current_confidence': round(current_conf, 4),
'trend_slope': round(slope, 4),
'sessions_analysed': len(sessions),
'message': (
f"DETERIORATION ALERT: {disease} risk increasing. "
f"Current: {current_conf:.0%}. "
f"Rising +{slope:.3f}/session over {len(sessions)} sessions. "
"Urgent clinical review recommended."
),
})
# Check longitudinal score trend (Tier 1 sessions)
long_scores = [s.get('longitudinal_score', 0.0) for s in sessions]
if any(s > 0 for s in long_scores):
x = np.arange(len(long_scores), dtype=float)
slope = float(np.polyfit(x, long_scores, 1)[0])
current = long_scores[-1]
if slope > slope_threshold and current > conf_threshold:
alerts.append({
'disease': 'Overall Health',
'current_confidence': round(current, 4),
'trend_slope': round(slope, 4),
'sessions_analysed': len(sessions),
'message': (
f"DETERIORATION ALERT: Overall respiratory risk increasing. "
f"Longitudinal score: {current:.0%}. "
f"Rising +{slope:.3f}/session. Doctor review recommended."
),
})
return alerts if alerts else None
|