Spaces:
Sleeping
Sleeping
File size: 42,786 Bytes
6e392b9 be32e06 6e392b9 be32e06 6e392b9 be32e06 a407a02 be32e06 a407a02 be32e06 6e392b9 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 6e392b9 610d2cf be32e06 a407a02 be32e06 a407a02 be32e06 6e392b9 a407a02 6e392b9 a407a02 6e392b9 a407a02 6e392b9 a407a02 6e392b9 a407a02 6e392b9 a407a02 6e392b9 a407a02 6e392b9 a407a02 6e392b9 a407a02 6e392b9 a407a02 6e392b9 a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 be32e06 a407a02 be32e06 a407a02 6e392b9 be32e06 a407a02 6e392b9 be32e06 a407a02 be32e06 610d2cf a407a02 be32e06 a407a02 610d2cf a407a02 6e392b9 be32e06 6e392b9 a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 be32e06 610d2cf a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 610d2cf a407a02 610d2cf be32e06 a407a02 be32e06 6e392b9 a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 610d2cf a407a02 be32e06 a407a02 610d2cf a407a02 6e392b9 a407a02 6e392b9 a407a02 be32e06 a407a02 be32e06 a407a02 610d2cf a407a02 6e392b9 a407a02 6e392b9 a407a02 610d2cf a407a02 610d2cf a407a02 6e392b9 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 6e392b9 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 610d2cf a407a02 be32e06 a407a02 be32e06 a407a02 6e392b9 a407a02 be32e06 a407a02 be32e06 6e392b9 a407a02 6e392b9 a407a02 6e392b9 be32e06 6e392b9 a407a02 610d2cf a407a02 610d2cf a407a02 6e392b9 be32e06 a407a02 6e392b9 a407a02 be32e06 a407a02 610d2cf be32e06 a407a02 610d2cf a407a02 6e392b9 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 a407a02 be32e06 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 |
import streamlit as st
import pandas as pd
import json
import os
from typing import Union, List, Dict, Optional, Tuple
from groq import Groq
from duckduckgo_search import DDGS
from datetime import datetime, timedelta
import time
import numpy as np
import pickle
from dataclasses import dataclass, asdict
import hashlib
from collections import defaultdict
# Set page configuration
st.set_page_config(
page_title="MedAssist - AI Medical Preconsultation",
layout="wide",
initial_sidebar_state="expanded",
page_icon="π₯"
)
# Enhanced CSS for medical theme
st.markdown("""
<style>
/* Medical theme styling */
html, body, .stApp, .main {
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%) !important;
color: #ffffff !important;
}
.medical-header {
background: linear-gradient(45deg, #2c5aa0, #4a90e2) !important;
color: white !important;
padding: 2rem !important;
border-radius: 15px !important;
text-align: center !important;
margin-bottom: 2rem !important;
box-shadow: 0 8px 32px rgba(31, 38, 135, 0.37) !important;
}
.chat-container {
background: rgba(255, 255, 255, 0.1) !important;
border-radius: 15px !important;
padding: 1rem !important;
backdrop-filter: blur(10px) !important;
border: 1px solid rgba(255, 255, 255, 0.2) !important;
margin-bottom: 1rem !important;
max-height: 500px !important;
overflow-y: auto !important;
}
.user-message {
background: linear-gradient(45deg, #4CAF50, #66BB6A) !important;
color: white !important;
padding: 1rem !important;
border-radius: 15px 15px 5px 15px !important;
margin: 0.5rem 0 !important;
margin-left: 2rem !important;
box-shadow: 0 4px 15px rgba(76, 175, 80, 0.4) !important;
}
.assistant-message {
background: rgba(255, 255, 255, 0.15) !important;
color: white !important;
padding: 1rem !important;
border-radius: 15px 15px 15px 5px !important;
margin: 0.5rem 0 !important;
margin-right: 2rem !important;
border-left: 4px solid #2196F3 !important;
backdrop-filter: blur(5px) !important;
}
.agent-status-card {
background: rgba(255, 255, 255, 0.15) !important;
border: 1px solid rgba(255, 255, 255, 0.3) !important;
border-radius: 12px !important;
padding: 1rem !important;
margin: 0.5rem 0 !important;
backdrop-filter: blur(5px) !important;
}
.evolution-metrics {
background: linear-gradient(45deg, #FF6B6B, #FF8E8E) !important;
color: white !important;
padding: 1rem !important;
border-radius: 10px !important;
margin: 0.5rem 0 !important;
}
.warning-box {
background: rgba(255, 152, 0, 0.2) !important;
border: 2px solid #FF9800 !important;
border-radius: 10px !important;
padding: 1.5rem !important;
margin: 1rem 0 !important;
color: white !important;
}
.stButton > button {
background: linear-gradient(45deg, #2196F3, #64B5F6) !important;
color: white !important;
border: none !important;
border-radius: 25px !important;
font-weight: bold !important;
padding: 0.75rem 2rem !important;
transition: all 0.3s ease !important;
}
.stButton > button:hover {
transform: translateY(-2px) !important;
box-shadow: 0 8px 25px rgba(33, 150, 243, 0.6) !important;
}
.chat-input {
position: sticky !important;
bottom: 0 !important;
background: rgba(255, 255, 255, 0.1) !important;
padding: 1rem !important;
border-radius: 15px !important;
backdrop-filter: blur(10px) !important;
}
.spinner {
border: 2px solid rgba(255, 255, 255, 0.3);
border-radius: 50%;
border-top: 2px solid #ffffff;
width: 20px;
height: 20px;
animation: spin 1s linear infinite;
display: inline-block;
}
@keyframes spin {
0% { transform: rotate(0deg); }
100% { transform: rotate(360deg); }
}
</style>
""", unsafe_allow_html=True)
@dataclass
class ConversationEntry:
"""Data structure for storing conversation entries"""
timestamp: str
user_input: str
assistant_response: str
symptoms: List[str]
severity_score: float
confidence_score: float
search_queries_used: List[str]
user_feedback: Optional[int] = None # 1-5 rating
was_helpful: Optional[bool] = None
@dataclass
class AgentPerformance:
"""Track agent performance metrics"""
agent_name: str
total_queries: int = 0
successful_responses: int = 0
average_confidence: float = 0.0
user_satisfaction: float = 0.0
learning_rate: float = 0.01
expertise_areas: Dict[str, float] = None
def __post_init__(self):
if self.expertise_areas is None:
self.expertise_areas = defaultdict(float)
class MedicalSearchTool:
"""Enhanced medical search tool with domain-specific optimization"""
def __init__(self):
self.ddgs = DDGS()
self.medical_sources = [
"mayoclinic.org", "webmd.com", "healthline.com", "medlineplus.gov",
"nih.gov", "who.int", "cdc.gov", "ncbi.nlm.nih.gov"
]
def search_medical_info(self, query: str, search_type: str = "symptoms") -> str:
"""Search for medical information with safety considerations"""
try:
# Add medical context to search
medical_queries = {
"symptoms": f"medical symptoms {query} causes diagnosis",
"treatment": f"medical treatment {query} therapy options",
"prevention": f"disease prevention {query} health tips",
"general": f"medical information {query} health facts"
}
enhanced_query = medical_queries.get(search_type, medical_queries["general"])
# Perform search with medical focus
search_results = list(self.ddgs.text(
enhanced_query,
max_results=5,
region='wt-wt',
safesearch='on'
))
if not search_results:
return "No relevant medical information found. Please consult with a healthcare professional."
# Filter and format results with medical authority preference
formatted_results = []
for idx, result in enumerate(search_results, 1):
title = result.get('title', 'No title')
snippet = result.get('body', 'No description')
url = result.get('href', 'No URL')
# Prioritize trusted medical sources
source_trust = "β" if any(source in url for source in self.medical_sources) else ""
formatted_results.append(
f"{idx}. {source_trust} {title}\n"
f" Information: {snippet}\n"
f" Source: {url}\n"
)
return "\n".join(formatted_results)
except Exception as e:
return f"Search temporarily unavailable: {str(e)}"
class GroqLLM:
"""Medical-optimized LLM client"""
def __init__(self, model_name="openai/gpt-oss-20b"):
self.client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
self.model_name = model_name
self.medical_context = """
You are a medical AI assistant for preconsultation guidance.
IMPORTANT: Always remind users that this is not a substitute for professional medical advice.
Provide helpful information while emphasizing the need for proper medical consultation.
"""
def generate_response(self, prompt: str, conversation_history: List[str] = None) -> Tuple[str, float]:
"""Generate response with confidence scoring"""
try:
# Build context with conversation history
context = self.medical_context
if conversation_history:
context += f"\n\nConversation History:\n{chr(10).join(conversation_history[-5:])}"
full_prompt = f"{context}\n\nUser Query: {prompt}\n\nPlease provide helpful medical guidance while emphasizing the importance of professional medical consultation."
completion = self.client.chat.completions.create(
model=self.model_name,
messages=[{"role": "user", "content": full_prompt}],
temperature=0.3, # Lower temperature for medical accuracy
max_tokens=1500,
stream=False
)
response = completion.choices[0].message.content if completion.choices else "Unable to generate response"
# Calculate confidence score based on response characteristics
confidence = self._calculate_confidence(response, prompt)
return response, confidence
except Exception as e:
return f"LLM temporarily unavailable: {str(e)}", 0.0
def _calculate_confidence(self, response: str, query: str) -> float:
"""Calculate confidence score based on response quality"""
confidence_factors = 0.0
# Check for medical disclaimers (increases confidence in safety)
if any(phrase in response.lower() for phrase in ["consult", "doctor", "medical professional", "healthcare provider"]):
confidence_factors += 0.3
# Check response length (adequate detail)
if 200 <= len(response) <= 1000:
confidence_factors += 0.2
# Check for structured information
if any(marker in response for marker in ["1.", "β’", "-", "**"]):
confidence_factors += 0.2
# Check for balanced information (not overly certain)
if any(phrase in response.lower() for phrase in ["may", "might", "could", "possible", "typically"]):
confidence_factors += 0.3
return min(confidence_factors, 1.0)
class EvolutionaryMedicalAgent:
"""Evolutionary agent with reinforcement learning capabilities"""
def __init__(self, agent_id: str, specialization: str):
self.agent_id = agent_id
self.specialization = specialization
self.performance = AgentPerformance(agent_name=agent_id)
self.knowledge_base = defaultdict(float)
self.response_patterns = {}
self.learning_memory = []
def process_query(self, query: str, context: str, search_results: str) -> Tuple[str, float]:
"""Process query and adapt based on specialization"""
# Update query count
self.performance.total_queries += 1
# Extract key terms for learning
key_terms = self._extract_medical_terms(query)
# Build specialized response based on agent's expertise
specialized_prompt = f"""
As a {self.specialization} specialist, analyze this medical query:
Query: {query}
Context: {context}
Search Results: {search_results}
Provide specialized insights based on your expertise in {self.specialization}.
Always emphasize the need for professional medical consultation.
"""
# Simulate processing (in real implementation, this would use the LLM)
response = f"Based on my specialization in {self.specialization}, {query.lower()} suggests several considerations. However, please consult with a healthcare professional for proper diagnosis and treatment."
confidence = 0.7 + (self.performance.average_confidence * 0.3)
# Update expertise in relevant areas
for term in key_terms:
self.knowledge_base[term] += 0.1
return response, confidence
def update_from_feedback(self, query: str, response: str, feedback_score: int, was_helpful: bool):
"""Update agent based on user feedback (reinforcement learning)"""
# Calculate reward signal
reward = (feedback_score - 3) / 2 # Convert 1-5 scale to -1 to 1
if was_helpful:
reward += 0.2
# Update performance metrics
if feedback_score >= 3:
self.performance.successful_responses += 1
# Update satisfaction and confidence
self.performance.user_satisfaction = (
(self.performance.user_satisfaction * (self.performance.total_queries - 1) + feedback_score) /
self.performance.total_queries
)
# Store learning memory
self.learning_memory.append({
'query': query,
'response': response,
'reward': reward,
'timestamp': datetime.now().isoformat()
})
# Adapt learning rate based on performance
if self.performance.user_satisfaction > 4.0:
self.performance.learning_rate *= 0.95 # Slow down learning when performing well
elif self.performance.user_satisfaction < 3.0:
self.performance.learning_rate *= 1.1 # Speed up learning when performing poorly
# Update expertise areas based on feedback
terms = self._extract_medical_terms(query)
for term in terms:
self.knowledge_base[term] += reward * self.performance.learning_rate
def _extract_medical_terms(self, text: str) -> List[str]:
"""Extract medical terms from text for learning"""
medical_keywords = [
'pain', 'fever', 'headache', 'nausea', 'fatigue', 'cough', 'cold', 'flu',
'diabetes', 'hypertension', 'infection', 'allergy', 'asthma', 'arthritis',
'anxiety', 'depression', 'insomnia', 'migraine', 'rash', 'swelling'
]
found_terms = []
text_lower = text.lower()
for term in medical_keywords:
if term in text_lower:
found_terms.append(term)
return found_terms
def get_expertise_summary(self) -> Dict:
"""Get summary of agent's learned expertise"""
return {
'specialization': self.specialization,
'total_queries': self.performance.total_queries,
'success_rate': (self.performance.successful_responses / max(1, self.performance.total_queries)) * 100,
'user_satisfaction': self.performance.user_satisfaction,
'learning_rate': self.performance.learning_rate,
'top_expertise_areas': dict(sorted(self.knowledge_base.items(), key=lambda x: x[1], reverse=True)[:5])
}
class MedicalConsultationSystem:
"""Main medical consultation system with evolutionary agents"""
def __init__(self):
self.llm = GroqLLM()
self.search_tool = MedicalSearchTool()
self.agents = self._initialize_agents()
self.conversation_history = []
self.conversation_data = []
def _initialize_agents(self) -> Dict[str, EvolutionaryMedicalAgent]:
"""Initialize specialized medical agents"""
return {
"general_practitioner": EvolutionaryMedicalAgent("gp", "General Practice Medicine"),
"symptom_analyzer": EvolutionaryMedicalAgent("symptom", "Symptom Analysis and Triage"),
"wellness_advisor": EvolutionaryMedicalAgent("wellness", "Preventive Care and Wellness"),
"mental_health": EvolutionaryMedicalAgent("mental", "Mental Health and Psychology"),
"emergency_assessor": EvolutionaryMedicalAgent("emergency", "Emergency Assessment and Urgent Care")
}
def process_medical_query(self, user_query: str) -> Dict:
"""Process medical query through evolutionary agent system"""
timestamp = datetime.now().isoformat()
# Determine which agents should handle this query
relevant_agents = self._select_relevant_agents(user_query)
# Search for medical information
search_results = self.search_tool.search_medical_info(user_query, "symptoms")
# Build conversation context
context = "\n".join(self.conversation_history[-3:]) if self.conversation_history else ""
# Get responses from relevant agents
agent_responses = {}
for agent_name in relevant_agents:
agent = self.agents[agent_name]
response, confidence = agent.process_query(user_query, context, search_results)
agent_responses[agent_name] = {
'response': response,
'confidence': confidence,
'specialization': agent.specialization
}
# Generate main LLM response
main_response, main_confidence = self.llm.generate_response(
f"{user_query}\n\nRelevant Information: {search_results}",
self.conversation_history
)
# Combine responses intelligently
final_response = self._combine_responses(main_response, agent_responses)
# Update conversation history
self.conversation_history.extend([
f"User: {user_query}",
f"Assistant: {final_response}"
])
# Extract symptoms for analysis
symptoms = self._extract_symptoms(user_query)
severity_score = self._assess_severity(user_query, symptoms)
# Store conversation data
conversation_entry = ConversationEntry(
timestamp=timestamp,
user_input=user_query,
assistant_response=final_response,
symptoms=symptoms,
severity_score=severity_score,
confidence_score=main_confidence,
search_queries_used=[user_query]
)
self.conversation_data.append(conversation_entry)
return {
'response': final_response,
'confidence': main_confidence,
'severity_score': severity_score,
'symptoms_detected': symptoms,
'agents_consulted': relevant_agents,
'agent_responses': agent_responses,
'search_performed': True
}
def _select_relevant_agents(self, query: str) -> List[str]:
"""Select most relevant agents for the query"""
query_lower = query.lower()
relevant_agents = ["general_practitioner"] # Always include GP
# Mental health keywords
mental_health_keywords = ["stress", "anxiety", "depression", "sleep", "mood", "worry", "panic", "sad"]
if any(keyword in query_lower for keyword in mental_health_keywords):
relevant_agents.append("mental_health")
# Emergency keywords
emergency_keywords = ["severe", "intense", "emergency", "urgent", "chest pain", "difficulty breathing", "blood"]
if any(keyword in query_lower for keyword in emergency_keywords):
relevant_agents.append("emergency_assessor")
# Wellness keywords
wellness_keywords = ["prevention", "healthy", "nutrition", "exercise", "lifestyle", "diet"]
if any(keyword in query_lower for keyword in wellness_keywords):
relevant_agents.append("wellness_advisor")
# Always include symptom analyzer for health queries
if any(keyword in query_lower for keyword in ["pain", "ache", "hurt", "symptom", "feel"]):
relevant_agents.append("symptom_analyzer")
return list(set(relevant_agents))
def _combine_responses(self, main_response: str, agent_responses: Dict) -> str:
"""Intelligently combine responses from multiple agents"""
if not agent_responses:
return main_response
combined = main_response + "\n\n**Specialist Insights:**\n"
for agent_name, data in agent_responses.items():
if data['confidence'] > 0.6: # Only include confident responses
combined += f"\nβ’ **{data['specialization']}**: {data['response'][:200]}...\n"
return combined
def _extract_symptoms(self, query: str) -> List[str]:
"""Extract symptoms from user query"""
common_symptoms = [
'fever', 'headache', 'nausea', 'pain', 'cough', 'fatigue', 'dizziness',
'rash', 'swelling', 'shortness of breath', 'chest pain', 'abdominal pain'
]
query_lower = query.lower()
detected_symptoms = [symptom for symptom in common_symptoms if symptom in query_lower]
return detected_symptoms
def _assess_severity(self, query: str, symptoms: List[str]) -> float:
"""Assess severity of reported symptoms (0-10 scale)"""
severity_score = 0.0
query_lower = query.lower()
# High severity indicators
high_severity = ["severe", "intense", "unbearable", "emergency", "chest pain", "difficulty breathing"]
medium_severity = ["moderate", "persistent", "recurring", "worse", "concerning"]
if any(indicator in query_lower for indicator in high_severity):
severity_score += 7.0
elif any(indicator in query_lower for indicator in medium_severity):
severity_score += 4.0
else:
severity_score += 2.0
# Add points for multiple symptoms
severity_score += min(len(symptoms) * 0.5, 2.0)
return min(severity_score, 10.0)
def update_agent_performance(self, query_index: int, feedback_score: int, was_helpful: bool):
"""Update agent performance based on user feedback"""
if query_index < len(self.conversation_data):
entry = self.conversation_data[query_index]
entry.user_feedback = feedback_score
entry.was_helpful = was_helpful
# Update all agents that were involved in this query
for agent in self.agents.values():
agent.update_from_feedback(entry.user_input, entry.assistant_response, feedback_score, was_helpful)
def get_system_metrics(self) -> Dict:
"""Get comprehensive system performance metrics"""
total_conversations = len(self.conversation_data)
if total_conversations == 0:
return {"status": "No conversations yet"}
avg_confidence = np.mean([entry.confidence_score for entry in self.conversation_data])
avg_severity = np.mean([entry.severity_score for entry in self.conversation_data])
feedback_entries = [entry for entry in self.conversation_data if entry.user_feedback is not None]
avg_feedback = np.mean([entry.user_feedback for entry in feedback_entries]) if feedback_entries else 0
return {
"total_conversations": total_conversations,
"average_confidence": avg_confidence,
"average_severity": avg_severity,
"average_user_feedback": avg_feedback,
"agent_performance": {name: agent.get_expertise_summary() for name, agent in self.agents.items()}
}
# Initialize session state
if 'medical_system' not in st.session_state:
st.session_state.medical_system = MedicalConsultationSystem()
if 'chat_messages' not in st.session_state:
st.session_state.chat_messages = []
medical_system = st.session_state.medical_system
# Main interface
st.markdown("""
<div class="medical-header">
<h1>π₯ MedAssist - AI Medical Preconsultation</h1>
<p>Advanced AI-powered medical guidance with evolutionary learning agents</p>
</div>
""", unsafe_allow_html=True)
# Medical disclaimer
st.markdown("""
<div class="warning-box">
<h3>β οΈ Important Medical Disclaimer</h3>
<p>This AI system provides general health information and is NOT a substitute for professional medical advice, diagnosis, or treatment. Always consult with qualified healthcare professionals for medical concerns. In case of emergency, contact emergency services immediately.</p>
</div>
""", unsafe_allow_html=True)
# Main layout
col1, col2 = st.columns([3, 1])
with col1:
st.markdown("### π¬ Medical Consultation Chat")
# Chat display area
chat_container = st.container()
with chat_container:
st.markdown('<div class="chat-container">', unsafe_allow_html=True)
for i, message in enumerate(st.session_state.chat_messages):
if message["role"] == "user":
st.markdown(f'<div class="user-message">π€ <strong>You:</strong> {message["content"]}</div>', unsafe_allow_html=True)
else:
st.markdown(f'<div class="assistant-message">π€ <strong>MedAssist:</strong> {message["content"]}</div>', unsafe_allow_html=True)
# Add feedback buttons for assistant messages
col_a, col_b, col_c = st.columns([1, 1, 8])
with col_a:
if st.button("π", key=f"helpful_{i}"):
medical_system.update_agent_performance(i//2, 5, True)
st.success("Feedback recorded!")
with col_b:
if st.button("π", key=f"not_helpful_{i}"):
medical_system.update_agent_performance(i//2, 2, False)
st.info("Feedback recorded. We'll improve!")
st.markdown('</div>', unsafe_allow_html=True)
# Chat input
with st.container():
st.markdown('<div class="chat-input">', unsafe_allow_html=True)
user_input = st.text_input("Describe your symptoms or health concerns:",
placeholder="e.g., I've been having headaches for 3 days...",
key="medical_input")
col_send, col_clear = st.columns([1, 4])
with col_send:
send_message = st.button("Send π€", type="primary")
with col_clear:
if st.button("Clear Chat ποΈ"):
st.session_state.chat_messages = []
st.rerun()
st.markdown('</div>', unsafe_allow_html=True)
with col2:
st.markdown("### π€ AI Agent Status")
# Agent status display
for agent_name, agent in medical_system.agents.items():
expertise = agent.get_expertise_summary()
st.markdown(f"""
<div class="agent-status-card">
<h4>{agent.specialization}</h4>
<p><strong>Queries:</strong> {expertise['total_queries']}</p>
<p><strong>Success Rate:</strong> {expertise['success_rate']:.1f}%</p>
<p><strong>Satisfaction:</strong> {expertise['user_satisfaction']:.1f}/5</p>
<p><strong>Learning Rate:</strong> {expertise['learning_rate']:.3f}</p>
</div>
""", unsafe_allow_html=True)
st.markdown("### π System Metrics")
metrics = medical_system.get_system_metrics()
if "total_conversations" in metrics:
st.markdown(f"""
<div class="evolution-metrics">
<p><strong>Total Chats:</strong> {metrics['total_conversations']}</p>
<p><strong>Avg Confidence:</strong> {metrics['average_confidence']:.2f}</p>
<p><strong>Avg Severity:</strong> {metrics['average_severity']:.1f}/10</p>
<p><strong>User Rating:</strong> {metrics['average_user_feedback']:.1f}/5</p>
</div>
""", unsafe_allow_html=True)
# Process user input
if send_message and user_input:
# Add user message
st.session_state.chat_messages.append({"role": "user", "content": user_input})
# Show thinking indicator
with st.spinner("π§ AI agents are analyzing your query..."):
# Process the query
result = medical_system.process_medical_query(user_input)
# Add assistant response
response_content = result['response']
# Add severity and confidence info
if result['severity_score'] > 7:
response_content += f"\n\nβ οΈ **High severity detected ({result['severity_score']:.1f}/10). Please seek immediate medical attention if symptoms are severe.**"
elif result['severity_score'] > 4:
response_content += f"\n\nβ‘ **Moderate severity detected ({result['severity_score']:.1f}/10). Consider scheduling a medical appointment.**"
if result['symptoms_detected']:
response_content += f"\n\nπ **Detected symptoms:** {', '.join(result['symptoms_detected'])}"
response_content += f"\n\nπ€ **Confidence Score:** {result['confidence']:.2f} | **Agents Consulted:** {', '.join(result['agents_consulted'])}"
st.session_state.chat_messages.append({"role": "assistant", "content": response_content})
st.rerun()
# Sidebar with additional features
with st.sidebar:
st.markdown("### π οΈ System Controls")
if st.button("π Reset System"):
st.session_state.medical_system = MedicalConsultationSystem()
st.session_state.chat_messages = []
st.rerun()
st.markdown("### π Learning Analytics")
if st.button("π View Detailed Analytics"):
st.session_state.show_analytics = True
if st.button("πΎ Export Chat History"):
if st.session_state.chat_messages:
chat_data = {
'timestamp': datetime.now().isoformat(),
'messages': st.session_state.chat_messages,
'system_metrics': medical_system.get_system_metrics()
}
st.download_button(
label="Download Chat Data",
data=json.dumps(chat_data, indent=2),
file_name=f"medical_chat_{datetime.now().strftime('%Y%m%d_%H%M%S')}.json",
mime="application/json"
)
else:
st.warning("No chat history to export")
st.markdown("### π― Quick Health Topics")
quick_topics = [
"Common cold symptoms",
"Headache causes",
"Stress management",
"Sleep problems",
"Healthy diet tips",
"Exercise recommendations"
]
for topic in quick_topics:
if st.button(f"π‘ {topic}", key=f"topic_{topic.replace(' ', '_')}"):
st.session_state.chat_messages.append({"role": "user", "content": f"Tell me about {topic.lower()}"})
with st.spinner("π§ Processing..."):
result = medical_system.process_medical_query(f"Tell me about {topic.lower()}")
response_content = result['response']
if result['symptoms_detected']:
response_content += f"\n\nπ **Related symptoms:** {', '.join(result['symptoms_detected'])}"
response_content += f"\n\nπ€ **Confidence:** {result['confidence']:.2f}"
st.session_state.chat_messages.append({"role": "assistant", "content": response_content})
st.rerun()
# Analytics Dashboard (if requested)
if st.session_state.get('show_analytics', False):
st.markdown("---")
st.markdown("## π Detailed System Analytics")
metrics = medical_system.get_system_metrics()
if "agent_performance" in metrics:
# Agent Performance Comparison
st.markdown("### π€ Agent Performance Analysis")
agent_data = []
for agent_name, performance in metrics["agent_performance"].items():
agent_data.append({
'Agent': performance['specialization'],
'Success Rate (%)': performance['success_rate'],
'User Satisfaction': performance['user_satisfaction'],
'Learning Rate': performance['learning_rate'],
'Total Queries': performance['total_queries']
})
if agent_data:
df_agents = pd.DataFrame(agent_data)
st.dataframe(df_agents, use_container_width=True)
# Performance charts
col1, col2 = st.columns(2)
with col1:
st.markdown("#### Success Rate by Agent")
if not df_agents.empty:
st.bar_chart(df_agents.set_index('Agent')['Success Rate (%)'])
with col2:
st.markdown("#### User Satisfaction by Agent")
if not df_agents.empty:
st.bar_chart(df_agents.set_index('Agent')['User Satisfaction'])
# Conversation Analysis
st.markdown("### π¬ Conversation Analysis")
if medical_system.conversation_data:
conversation_df = pd.DataFrame([asdict(entry) for entry in medical_system.conversation_data])
col1, col2, col3 = st.columns(3)
with col1:
st.metric("Total Conversations", len(conversation_df))
avg_confidence = conversation_df['confidence_score'].mean()
st.metric("Average Confidence", f"{avg_confidence:.2f}")
with col2:
avg_severity = conversation_df['severity_score'].mean()
st.metric("Average Severity", f"{avg_severity:.1f}/10")
feedback_data = conversation_df[conversation_df['user_feedback'].notna()]
if not feedback_data.empty:
avg_feedback = feedback_data['user_feedback'].mean()
st.metric("Average User Rating", f"{avg_feedback:.1f}/5")
with col3:
symptoms_detected = sum(len(symptoms) for symptoms in conversation_df['symptoms'])
st.metric("Total Symptoms Detected", symptoms_detected)
helpful_responses = conversation_df['was_helpful'].sum() if 'was_helpful' in conversation_df else 0
st.metric("Helpful Responses", helpful_responses)
# Severity distribution
st.markdown("#### Severity Score Distribution")
severity_counts = conversation_df['severity_score'].value_counts().sort_index()
st.bar_chart(severity_counts)
# Most common symptoms
st.markdown("#### Most Common Symptoms")
all_symptoms = []
for symptoms_list in conversation_df['symptoms']:
all_symptoms.extend(symptoms_list)
if all_symptoms:
symptom_counts = pd.Series(all_symptoms).value_counts().head(10)
st.bar_chart(symptom_counts)
else:
st.info("No symptoms data available yet")
# Timeline analysis
st.markdown("#### Usage Timeline")
conversation_df['timestamp'] = pd.to_datetime(conversation_df['timestamp'])
daily_usage = conversation_df.groupby(conversation_df['timestamp'].dt.date).size()
st.line_chart(daily_usage)
else:
st.info("No conversation data available for analysis yet")
# Learning Progress
st.markdown("### π§ AI Learning Progress")
for agent_name, agent in medical_system.agents.items():
with st.expander(f"π {agent.specialization} Learning Details"):
expertise = agent.get_expertise_summary()
st.write(f"**Total Experience:** {expertise['total_queries']} queries processed")
st.write(f"**Current Learning Rate:** {expertise['learning_rate']:.4f}")
st.write(f"**Performance Trend:** {'Improving' if expertise['user_satisfaction'] > 3.5 else 'Learning'}")
if expertise['top_expertise_areas']:
st.write("**Top Expertise Areas:**")
for area, score in expertise['top_expertise_areas'].items():
st.write(f" β’ {area.title()}: {score:.2f}")
# Learning memory (last few interactions)
if hasattr(agent, 'learning_memory') and agent.learning_memory:
st.write("**Recent Learning Events:**")
for memory in agent.learning_memory[-3:]:
reward_emoji = "β
" if memory['reward'] > 0 else "β" if memory['reward'] < 0 else "β‘οΈ"
st.write(f" {reward_emoji} Reward: {memory['reward']:.2f} | Query: {memory['query'][:50]}...")
if st.button("π Close Analytics"):
st.session_state.show_analytics = False
st.rerun()
# Health Tips Section
st.markdown("---")
st.markdown("### π Daily Health Tips")
health_tips = [
"π§ Stay hydrated: Aim for 8-10 glasses of water daily",
"πΆ Take regular walks: Even 10 minutes can boost your mood",
"π΄ Maintain sleep hygiene: 7-9 hours of quality sleep is essential",
"π₯ Eat colorful foods: Variety ensures you get different nutrients",
"π§ Practice mindfulness: Just 5 minutes of meditation can reduce stress",
"π± Take breaks from screens: Follow the 20-20-20 rule",
"π€ Stay connected: Social connections are vital for mental health",
"βοΈ Get sunlight: 15 minutes of sunlight helps with Vitamin D"
]
# Display a random tip
import random
daily_tip = random.choice(health_tips)
st.info(f"**π‘ Today's Health Tip:** {daily_tip}")
# Emergency Resources Section
st.markdown("### π¨ Emergency Resources")
emergency_col1, emergency_col2 = st.columns(2)
with emergency_col1:
st.markdown("""
**π When to Seek Immediate Help:**
- Chest pain or difficulty breathing
- Severe allergic reactions
- Loss of consciousness
- Severe bleeding
- Signs of stroke (FAST test)
- Severe burns
""")
with emergency_col2:
st.markdown("""
**π Emergency Contacts:**
- Emergency Services: 911 (US), 112 (EU)
- Poison Control: 1-800-222-1222 (US)
- Mental Health Crisis: 988 (US)
- Text HOME to 741741 (Crisis Text Line)
**π₯ Find Nearest Hospital:**
Use your maps app or call emergency services
""")
# Data Persistence and Learning Enhancement
class DataPersistence:
"""Handle data persistence for learning and analytics"""
def __init__(self, data_dir: str = "medical_ai_data"):
self.data_dir = data_dir
os.makedirs(data_dir, exist_ok=True)
def save_conversation_data(self, system: MedicalConsultationSystem):
"""Save conversation data for future learning"""
try:
data_file = os.path.join(self.data_dir, f"conversations_{datetime.now().strftime('%Y%m%d')}.json")
conversations = []
for entry in system.conversation_data:
conversations.append(asdict(entry))
with open(data_file, 'w') as f:
json.dump(conversations, f, indent=2)
return True
except Exception as e:
st.error(f"Failed to save data: {str(e)}")
return False
def save_agent_knowledge(self, system: MedicalConsultationSystem):
"""Save agent learning data"""
try:
for agent_name, agent in system.agents.items():
agent_file = os.path.join(self.data_dir, f"agent_{agent_name}_knowledge.pkl")
agent_data = {
'knowledge_base': dict(agent.knowledge_base),
'performance': asdict(agent.performance),
'learning_memory': agent.learning_memory[-100:] # Keep last 100 entries
}
with open(agent_file, 'wb') as f:
pickle.dump(agent_data, f)
return True
except Exception as e:
st.error(f"Failed to save agent knowledge: {str(e)}")
return False
def load_agent_knowledge(self, system: MedicalConsultationSystem):
"""Load previously saved agent knowledge"""
try:
for agent_name, agent in system.agents.items():
agent_file = os.path.join(self.data_dir, f"agent_{agent_name}_knowledge.pkl")
if os.path.exists(agent_file):
with open(agent_file, 'rb') as f:
agent_data = pickle.load(f)
# Restore knowledge base
agent.knowledge_base = defaultdict(float, agent_data.get('knowledge_base', {}))
# Restore learning memory
agent.learning_memory = agent_data.get('learning_memory', [])
# Restore performance metrics
if 'performance' in agent_data:
perf_data = agent_data['performance']
agent.performance.total_queries = perf_data.get('total_queries', 0)
agent.performance.successful_responses = perf_data.get('successful_responses', 0)
agent.performance.average_confidence = perf_data.get('average_confidence', 0.0)
agent.performance.user_satisfaction = perf_data.get('user_satisfaction', 0.0)
agent.performance.learning_rate = perf_data.get('learning_rate', 0.01)
return True
except Exception as e:
st.error(f"Failed to load agent knowledge: {str(e)}")
return False
# Initialize data persistence
if 'data_persistence' not in st.session_state:
st.session_state.data_persistence = DataPersistence()
# Load previous learning data when system starts
if 'knowledge_loaded' not in st.session_state:
st.session_state.data_persistence.load_agent_knowledge(medical_system)
st.session_state.knowledge_loaded = True
# Auto-save functionality
if len(st.session_state.chat_messages) > 0 and len(st.session_state.chat_messages) % 10 == 0:
# Save data every 10 messages
st.session_state.data_persistence.save_conversation_data(medical_system)
st.session_state.data_persistence.save_agent_knowledge(medical_system)
# Footer with system information
st.markdown("---")
st.markdown("""
<div style="text-align: center; padding: 2rem; opacity: 0.8;">
<p><strong>MedAssist v1.0</strong> | AI-Powered Medical Preconsultation System</p>
<p>π€ Evolutionary Learning Agents β’ π Real-time Medical Search β’ π¬ Intelligent Chat Interface</p>
<p><small>β οΈ This system is for informational purposes only and is not a substitute for professional medical advice</small></p>
</div>
""", unsafe_allow_html=True) |