Update analysis.py
Browse files- analysis.py +145 -59
analysis.py
CHANGED
@@ -1,13 +1,31 @@
|
|
1 |
from typing import Optional, Tuple, List
|
|
|
2 |
from config import agent, patients_collection, analysis_collection, alerts_collection, logger
|
3 |
from models import RiskLevel
|
4 |
-
from utils import
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
from datetime import datetime
|
6 |
import asyncio
|
7 |
import json
|
8 |
import re
|
9 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
async def create_alert(patient_id: str, risk_data: dict):
|
|
|
11 |
alert_doc = {
|
12 |
"patient_id": patient_id,
|
13 |
"type": "suicide_risk",
|
@@ -16,7 +34,6 @@ async def create_alert(patient_id: str, risk_data: dict):
|
|
16 |
"factors": risk_data["factors"],
|
17 |
"timestamp": datetime.utcnow(),
|
18 |
"acknowledged": False,
|
19 |
-
# Facebook-like notification fields
|
20 |
"notification": {
|
21 |
"type": NotificationType.RISK_ALERT,
|
22 |
"status": NotificationStatus.UNREAD,
|
@@ -27,83 +44,125 @@ async def create_alert(patient_id: str, risk_data: dict):
|
|
27 |
"priority": "high" if risk_data["level"] in ["high", "severe"] else "medium"
|
28 |
}
|
29 |
}
|
30 |
-
await alerts_collection.insert_one(alert_doc)
|
31 |
-
|
32 |
-
# Trigger real-time notification
|
33 |
-
await broadcast_notification(alert_doc["notification"])
|
34 |
|
35 |
-
|
36 |
-
|
37 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
38 |
identifier = patient_id if patient_id else compute_file_content_hash(file_content)
|
39 |
report_data = {"identifier": identifier, "content": report_content, "file_type": file_type}
|
40 |
report_hash = compute_patient_data_hash(report_data)
|
41 |
logger.info(f"🧾 Analyzing report for identifier: {identifier}")
|
42 |
|
43 |
-
|
|
|
|
|
|
|
44 |
if existing_analysis:
|
45 |
logger.info(f"✅ No changes in report data for {identifier}, skipping analysis")
|
46 |
return existing_analysis
|
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 |
async def analyze_patient(patient: dict):
|
|
|
96 |
try:
|
97 |
serialized = serialize_patient(patient)
|
98 |
patient_id = serialized.get("fhir_id")
|
99 |
patient_hash = compute_patient_data_hash(serialized)
|
100 |
logger.info(f"🧾 Analyzing patient: {patient_id}")
|
101 |
|
|
|
102 |
existing_analysis = await analysis_collection.find_one({"patient_id": patient_id})
|
103 |
if existing_analysis and existing_analysis.get("data_hash") == patient_hash:
|
104 |
logger.info(f"✅ No changes in patient data for {patient_id}, skipping analysis")
|
105 |
return
|
106 |
|
|
|
107 |
doc = json.dumps(serialized, indent=2)
|
108 |
message = (
|
109 |
"You are a clinical decision support AI.\n\n"
|
@@ -118,6 +177,7 @@ async def analyze_patient(patient: dict):
|
|
118 |
raw = agent.chat(message=message, history=[], temperature=0.7, max_new_tokens=1024)
|
119 |
structured = structure_medical_response(raw)
|
120 |
|
|
|
121 |
risk_level, risk_score, risk_factors = detect_suicide_risk(raw)
|
122 |
suicide_risk = {
|
123 |
"level": risk_level.value,
|
@@ -125,6 +185,7 @@ async def analyze_patient(patient: dict):
|
|
125 |
"factors": risk_factors
|
126 |
}
|
127 |
|
|
|
128 |
analysis_doc = {
|
129 |
"identifier": patient_id,
|
130 |
"patient_id": patient_id,
|
@@ -141,15 +202,36 @@ async def analyze_patient(patient: dict):
|
|
141 |
upsert=True
|
142 |
)
|
143 |
|
|
|
144 |
if risk_level in [RiskLevel.MODERATE, RiskLevel.HIGH, RiskLevel.SEVERE]:
|
145 |
await create_alert(patient_id, suicide_risk)
|
146 |
|
147 |
logger.info(f"✅ Stored analysis for patient {patient_id}")
|
148 |
|
149 |
except Exception as e:
|
150 |
-
logger.error(f"Error analyzing patient: {e}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
151 |
|
152 |
def detect_suicide_risk(text: str) -> Tuple[RiskLevel, float, List[str]]:
|
|
|
153 |
suicide_keywords = [
|
154 |
'suicide', 'suicidal', 'kill myself', 'end my life',
|
155 |
'want to die', 'self-harm', 'self harm', 'hopeless',
|
@@ -159,21 +241,24 @@ def detect_suicide_risk(text: str) -> Tuple[RiskLevel, float, List[str]]:
|
|
159 |
if not explicit_mentions:
|
160 |
return RiskLevel.NONE, 0.0, []
|
161 |
|
162 |
-
assessment_prompt = (
|
163 |
-
"Assess the suicide risk level based on this text. "
|
164 |
-
"Consider frequency, specificity, and severity of statements. "
|
165 |
-
"Respond with JSON format: {\"risk_level\": \"low/moderate/high/severe\", "
|
166 |
-
"\"risk_score\": 0-1, \"factors\": [\"list of risk factors\"]}\n\n"
|
167 |
-
f"Text to assess:\n{text}"
|
168 |
-
)
|
169 |
-
|
170 |
try:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
171 |
response = agent.chat(
|
172 |
message=assessment_prompt,
|
173 |
history=[],
|
174 |
temperature=0.2,
|
175 |
max_new_tokens=256
|
176 |
)
|
|
|
|
|
177 |
json_match = re.search(r'\{.*\}', response, re.DOTALL)
|
178 |
if json_match:
|
179 |
assessment = json.loads(json_match.group())
|
@@ -185,6 +270,7 @@ def detect_suicide_risk(text: str) -> Tuple[RiskLevel, float, List[str]]:
|
|
185 |
except Exception as e:
|
186 |
logger.error(f"Error in suicide risk assessment: {e}")
|
187 |
|
|
|
188 |
risk_score = min(0.1 * len(explicit_mentions), 0.9)
|
189 |
if risk_score > 0.7:
|
190 |
return RiskLevel.HIGH, risk_score, explicit_mentions
|
|
|
1 |
from typing import Optional, Tuple, List
|
2 |
+
from enum import Enum
|
3 |
from config import agent, patients_collection, analysis_collection, alerts_collection, logger
|
4 |
from models import RiskLevel
|
5 |
+
from utils import (
|
6 |
+
structure_medical_response,
|
7 |
+
compute_file_content_hash,
|
8 |
+
compute_patient_data_hash,
|
9 |
+
serialize_patient,
|
10 |
+
broadcast_notification
|
11 |
+
)
|
12 |
from datetime import datetime
|
13 |
import asyncio
|
14 |
import json
|
15 |
import re
|
16 |
|
17 |
+
class NotificationType(str, Enum):
|
18 |
+
RISK_ALERT = "risk_alert"
|
19 |
+
SYSTEM = "system"
|
20 |
+
MESSAGE = "message"
|
21 |
+
|
22 |
+
class NotificationStatus(str, Enum):
|
23 |
+
UNREAD = "unread"
|
24 |
+
READ = "read"
|
25 |
+
ARCHIVED = "archived"
|
26 |
+
|
27 |
async def create_alert(patient_id: str, risk_data: dict):
|
28 |
+
"""Create a new risk alert with notification metadata"""
|
29 |
alert_doc = {
|
30 |
"patient_id": patient_id,
|
31 |
"type": "suicide_risk",
|
|
|
34 |
"factors": risk_data["factors"],
|
35 |
"timestamp": datetime.utcnow(),
|
36 |
"acknowledged": False,
|
|
|
37 |
"notification": {
|
38 |
"type": NotificationType.RISK_ALERT,
|
39 |
"status": NotificationStatus.UNREAD,
|
|
|
44 |
"priority": "high" if risk_data["level"] in ["high", "severe"] else "medium"
|
45 |
}
|
46 |
}
|
|
|
|
|
|
|
|
|
47 |
|
48 |
+
try:
|
49 |
+
await alerts_collection.insert_one(alert_doc)
|
50 |
+
await broadcast_notification(alert_doc["notification"])
|
51 |
+
logger.warning(f"⚠️ Created suicide risk alert for patient {patient_id}")
|
52 |
+
return alert_doc
|
53 |
+
except Exception as e:
|
54 |
+
logger.error(f"Failed to create alert for patient {patient_id}: {str(e)}")
|
55 |
+
raise
|
56 |
+
|
57 |
+
async def analyze_patient_report(
|
58 |
+
patient_id: Optional[str],
|
59 |
+
report_content: str,
|
60 |
+
file_type: str,
|
61 |
+
file_content: bytes
|
62 |
+
):
|
63 |
+
"""Analyze a patient report and create alerts for risks"""
|
64 |
identifier = patient_id if patient_id else compute_file_content_hash(file_content)
|
65 |
report_data = {"identifier": identifier, "content": report_content, "file_type": file_type}
|
66 |
report_hash = compute_patient_data_hash(report_data)
|
67 |
logger.info(f"🧾 Analyzing report for identifier: {identifier}")
|
68 |
|
69 |
+
# Check for existing analysis
|
70 |
+
existing_analysis = await analysis_collection.find_one(
|
71 |
+
{"identifier": identifier, "report_hash": report_hash}
|
72 |
+
)
|
73 |
if existing_analysis:
|
74 |
logger.info(f"✅ No changes in report data for {identifier}, skipping analysis")
|
75 |
return existing_analysis
|
76 |
|
77 |
+
try:
|
78 |
+
# Generate analysis
|
79 |
+
prompt = (
|
80 |
+
"You are a clinical decision support AI. Analyze the following patient report:\n"
|
81 |
+
"1. Summarize the patient's medical history.\n"
|
82 |
+
"2. Identify risks or red flags (including mental health and suicide risk).\n"
|
83 |
+
"3. Highlight missed diagnoses or treatments.\n"
|
84 |
+
"4. Suggest next clinical steps.\n"
|
85 |
+
f"\nPatient Report ({file_type}):\n{'-'*40}\n{report_content[:10000]}"
|
86 |
+
)
|
87 |
|
88 |
+
raw_response = agent.chat(
|
89 |
+
message=prompt,
|
90 |
+
history=[],
|
91 |
+
temperature=0.7,
|
92 |
+
max_new_tokens=1024
|
93 |
+
)
|
94 |
+
structured_response = structure_medical_response(raw_response)
|
95 |
|
96 |
+
# Detect suicide risk
|
97 |
+
risk_level, risk_score, risk_factors = detect_suicide_risk(raw_response)
|
98 |
+
suicide_risk = {
|
99 |
+
"level": risk_level.value,
|
100 |
+
"score": risk_score,
|
101 |
+
"factors": risk_factors
|
102 |
+
}
|
103 |
|
104 |
+
# Store analysis
|
105 |
+
analysis_doc = {
|
106 |
+
"identifier": identifier,
|
107 |
+
"patient_id": patient_id,
|
108 |
+
"timestamp": datetime.utcnow(),
|
109 |
+
"summary": structured_response,
|
110 |
+
"suicide_risk": suicide_risk,
|
111 |
+
"raw": raw_response,
|
112 |
+
"report_hash": report_hash,
|
113 |
+
"file_type": file_type
|
114 |
+
}
|
115 |
|
116 |
+
await analysis_collection.update_one(
|
117 |
+
{"identifier": identifier, "report_hash": report_hash},
|
118 |
+
{"$set": analysis_doc},
|
119 |
+
upsert=True
|
120 |
+
)
|
121 |
|
122 |
+
# Create alert if risk detected
|
123 |
+
if patient_id and risk_level in [RiskLevel.MODERATE, RiskLevel.HIGH, RiskLevel.SEVERE]:
|
124 |
+
await create_alert(patient_id, suicide_risk)
|
125 |
+
|
126 |
+
logger.info(f"✅ Stored analysis for identifier {identifier}")
|
127 |
+
return analysis_doc
|
128 |
|
129 |
+
except Exception as e:
|
130 |
+
logger.error(f"Error analyzing report for {identifier}: {str(e)}")
|
131 |
+
error_alert = {
|
132 |
+
"identifier": identifier,
|
133 |
+
"type": "system_error",
|
134 |
+
"level": "high",
|
135 |
+
"message": f"Report analysis failed: {str(e)}",
|
136 |
+
"timestamp": datetime.utcnow(),
|
137 |
+
"acknowledged": False,
|
138 |
+
"notification": {
|
139 |
+
"type": NotificationType.SYSTEM,
|
140 |
+
"status": NotificationStatus.UNREAD,
|
141 |
+
"title": "Report Analysis Error",
|
142 |
+
"message": f"Failed to analyze report for {'patient ' + patient_id if patient_id else 'unknown identifier'}",
|
143 |
+
"icon": "❌",
|
144 |
+
"action_url": "/system/errors",
|
145 |
+
"priority": "high"
|
146 |
+
}
|
147 |
+
}
|
148 |
+
await alerts_collection.insert_one(error_alert)
|
149 |
+
raise
|
150 |
|
151 |
async def analyze_patient(patient: dict):
|
152 |
+
"""Analyze complete patient record and create alerts for risks"""
|
153 |
try:
|
154 |
serialized = serialize_patient(patient)
|
155 |
patient_id = serialized.get("fhir_id")
|
156 |
patient_hash = compute_patient_data_hash(serialized)
|
157 |
logger.info(f"🧾 Analyzing patient: {patient_id}")
|
158 |
|
159 |
+
# Check for existing analysis
|
160 |
existing_analysis = await analysis_collection.find_one({"patient_id": patient_id})
|
161 |
if existing_analysis and existing_analysis.get("data_hash") == patient_hash:
|
162 |
logger.info(f"✅ No changes in patient data for {patient_id}, skipping analysis")
|
163 |
return
|
164 |
|
165 |
+
# Generate analysis
|
166 |
doc = json.dumps(serialized, indent=2)
|
167 |
message = (
|
168 |
"You are a clinical decision support AI.\n\n"
|
|
|
177 |
raw = agent.chat(message=message, history=[], temperature=0.7, max_new_tokens=1024)
|
178 |
structured = structure_medical_response(raw)
|
179 |
|
180 |
+
# Detect suicide risk
|
181 |
risk_level, risk_score, risk_factors = detect_suicide_risk(raw)
|
182 |
suicide_risk = {
|
183 |
"level": risk_level.value,
|
|
|
185 |
"factors": risk_factors
|
186 |
}
|
187 |
|
188 |
+
# Store analysis
|
189 |
analysis_doc = {
|
190 |
"identifier": patient_id,
|
191 |
"patient_id": patient_id,
|
|
|
202 |
upsert=True
|
203 |
)
|
204 |
|
205 |
+
# Create alert if risk detected
|
206 |
if risk_level in [RiskLevel.MODERATE, RiskLevel.HIGH, RiskLevel.SEVERE]:
|
207 |
await create_alert(patient_id, suicide_risk)
|
208 |
|
209 |
logger.info(f"✅ Stored analysis for patient {patient_id}")
|
210 |
|
211 |
except Exception as e:
|
212 |
+
logger.error(f"Error analyzing patient: {str(e)}")
|
213 |
+
error_alert = {
|
214 |
+
"patient_id": patient_id if 'patient_id' in locals() else "unknown",
|
215 |
+
"type": "system_error",
|
216 |
+
"level": "high",
|
217 |
+
"message": f"Patient analysis failed: {str(e)}",
|
218 |
+
"timestamp": datetime.utcnow(),
|
219 |
+
"acknowledged": False,
|
220 |
+
"notification": {
|
221 |
+
"type": NotificationType.SYSTEM,
|
222 |
+
"status": NotificationStatus.UNREAD,
|
223 |
+
"title": "Analysis Error",
|
224 |
+
"message": f"Failed to analyze patient {patient_id if 'patient_id' in locals() else 'unknown'}",
|
225 |
+
"icon": "❌",
|
226 |
+
"action_url": "/system/errors",
|
227 |
+
"priority": "high"
|
228 |
+
}
|
229 |
+
}
|
230 |
+
await alerts_collection.insert_one(error_alert)
|
231 |
+
raise
|
232 |
|
233 |
def detect_suicide_risk(text: str) -> Tuple[RiskLevel, float, List[str]]:
|
234 |
+
"""Detect suicide risk level from text analysis"""
|
235 |
suicide_keywords = [
|
236 |
'suicide', 'suicidal', 'kill myself', 'end my life',
|
237 |
'want to die', 'self-harm', 'self harm', 'hopeless',
|
|
|
241 |
if not explicit_mentions:
|
242 |
return RiskLevel.NONE, 0.0, []
|
243 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
244 |
try:
|
245 |
+
# Get AI assessment
|
246 |
+
assessment_prompt = (
|
247 |
+
"Assess the suicide risk level based on this text. "
|
248 |
+
"Consider frequency, specificity, and severity of statements. "
|
249 |
+
"Respond with JSON format: {\"risk_level\": \"low/moderate/high/severe\", "
|
250 |
+
"\"risk_score\": 0-1, \"factors\": [\"list of risk factors\"]}\n\n"
|
251 |
+
f"Text to assess:\n{text}"
|
252 |
+
)
|
253 |
+
|
254 |
response = agent.chat(
|
255 |
message=assessment_prompt,
|
256 |
history=[],
|
257 |
temperature=0.2,
|
258 |
max_new_tokens=256
|
259 |
)
|
260 |
+
|
261 |
+
# Parse response
|
262 |
json_match = re.search(r'\{.*\}', response, re.DOTALL)
|
263 |
if json_match:
|
264 |
assessment = json.loads(json_match.group())
|
|
|
270 |
except Exception as e:
|
271 |
logger.error(f"Error in suicide risk assessment: {e}")
|
272 |
|
273 |
+
# Fallback heuristic if AI assessment fails
|
274 |
risk_score = min(0.1 * len(explicit_mentions), 0.9)
|
275 |
if risk_score > 0.7:
|
276 |
return RiskLevel.HIGH, risk_score, explicit_mentions
|