Spaces:
Sleeping
Sleeping
File size: 61,172 Bytes
bf2afe4 9f950d0 bf2afe4 9f950d0 | 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 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 | import streamlit as st
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go
from datetime import datetime
import json
import warnings
from sklearn.ensemble import RandomForestClassifier
from sklearn.preprocessing import LabelEncoder
import numpy as np
import html
import re
warnings.filterwarnings('ignore')
# Import our modules
from config import Config
from database_manager import DatabaseManager
from groq_agent import GroqAgent
from granite_agent import GraniteAgent
from granite_chat import GraniteChatAgent # New import for chat functionality
from reward_system import RewardSystem
from summarisation import create_summarizer
def clean_health_ai_response(raw_response):
"""
Health-specific AI response cleaner with medical disclaimer handling.
"""
if not raw_response or not isinstance(raw_response, str):
return None, {}
try:
response = raw_response.strip()
# Remove HTML tags
response = re.sub(r'<[^>]+>', '', response)
response = html.unescape(response)
# Clean up formatting
response = re.sub(r'\n{3,}', '\n\n', response)
response = re.sub(r' {2,}', ' ', response)
# Remove system prefixes
response = re.sub(r'^(Assistant|AI|Bot|Health Coach):\s*', '', response, flags=re.IGNORECASE)
# Ensure medical disclaimer is properly formatted if present
if "consult" in response.lower() and "doctor" in response.lower():
response = re.sub(
r'(.*consult.*doctor.*)',
r'**Important:** \1',
response,
flags=re.IGNORECASE
)
# Clean up any JSON-like formatting
response = re.sub(r'^\{.*?\}$', '', response, flags=re.DOTALL)
response = re.sub(r'"[^"]*":\s*', '', response)
response = re.sub(r'[{}"\[\],]', '', response)
response = response.strip()
if not response:
return None, {"error": "Response became empty after cleaning"}
metadata = {
"original_length": len(raw_response),
"cleaned_length": len(response),
"has_disclaimer": "consult" in response.lower() and "doctor" in response.lower(),
"response_type": "health_advice"
}
return response, metadata
except Exception as e:
return None, {"error": f"Error cleaning health response: {str(e)}"}
# Page configuration
st.set_page_config(
page_title="π Dynamic Wellness Platform",
page_icon="π",
layout="wide"
)
# Custom CSS with improved styling
st.markdown("""
<style>
.main-title {
font-size: 2.8rem;
color: #1f77b4;
text-align: center;
margin-bottom: 2rem;
font-weight: bold;
text-shadow: 2px 2px 4px rgba(0,0,0,0.1);
}
.task-card {
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
padding: 1.5rem;
border-radius: 15px;
color: white;
margin: 1rem 0;
box-shadow: 0 4px 8px rgba(0,0,0,0.2);
}
.completed-task {
background: linear-gradient(135deg, #56ab2f 0%, #a8e6cf 100%);
padding: 1rem;
border-radius: 10px;
color: white;
margin: 0.5rem 0;
}
.pending-task {
background: linear-gradient(135deg, #ffeaa7 0%, #fab1a0 100%);
padding: 1rem;
border-radius: 10px;
color: #2d3436;
margin: 0.5rem 0;
border-left: 5px solid #e17055;
}
.reward-notification {
background: linear-gradient(135deg, #fd79a8 0%, #fdcb6e 100%);
padding: 1rem;
border-radius: 10px;
color: white;
text-align: center;
font-weight: bold;
margin: 1rem 0;
animation: pulse 2s infinite;
}
@keyframes pulse {
0% { transform: scale(1); }
50% { transform: scale(1.05); }
100% { transform: scale(1); }
}
.ai-analysis {
background: linear-gradient(135deg, #f8f9fa, #e9ecef);
border-left: 5px solid #2196f3;
padding: 1.5rem;
margin: 1rem 0;
border-radius: 10px;
box-shadow: 0 2px 4px rgba(0,0,0,0.1);
color: #000000;
font-size: 16px;
line-height: 1.6;
}
.risk-indicator {
background: white;
border-radius: 10px;
padding: 1rem;
margin: 1rem 0;
border-left: 5px solid;
color: #000000;
font-weight: bold;
}
.risk-low { border-color: #28a745; }
.risk-medium { border-color: #ffc107; }
.risk-high { border-color: #dc3545; }
.wellness-tips {
background: linear-gradient(135deg, #e8f5e8, #f0f8f0);
border-left: 5px solid #28a745;
padding: 1.5rem;
margin: 1rem 0;
border-radius: 10px;
color: #000000;
font-size: 16px;
line-height: 1.8;
}
.progress-summary {
background: linear-gradient(135deg, #fff3cd, #ffeaa7);
border-left: 5px solid #ffc107;
padding: 1.5rem;
margin: 1rem 0;
border-radius: 10px;
color: #000000;
font-size: 16px;
line-height: 1.6;
}
.stButton > button {
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
color: white;
border: none;
border-radius: 20px;
padding: 0.75rem 2rem;
font-weight: bold;
transition: all 0.3s ease;
}
.stButton > button:hover {
transform: translateY(-2px);
box-shadow: 0 4px 8px rgba(0,0,0,0.2);
}
.bullet-list {
font-size: 16px;
line-height: 1.8;
color: #000000;
margin: 0;
padding-left: 0;
}
.bullet-list li {
margin-bottom: 8px;
list-style-type: none;
color: #000000;
}
/* Chat message styling */
.chat-message {
color: white !important;
background: transparent;
}
.chat-user-message {
color: white !important;
background: rgba(255, 255, 255, 0.1);
padding: 0.5rem;
border-radius: 5px;
margin: 0.2rem 0;
}
.chat-assistant-message {
color: white !important;
background: rgba(255, 255, 255, 0.05);
padding: 0.5rem;
border-radius: 5px;
margin: 0.2rem 0;
}
.stChatMessage {
color: white !important;
}
.stChatMessage p {
color: white !important;
}
.stChatMessage div {
color: white !important;
}
.granite-chat-indicator {
background: linear-gradient(135deg, #4a4a4a 0%, #6c6c6c 100%);
border-left: 5px solid #2196f3;
padding: 0.8rem;
margin: 0.5rem 0;
border-radius: 8px;
color: white;
font-size: 14px;
text-align: center;
}
</style>
""", unsafe_allow_html=True)
# Initialize services
@st.cache_resource
def initialize_services():
config = Config()
db_manager = DatabaseManager(config)
groq_agent = GroqAgent(config)
granite_agent = GraniteAgent(config)
granite_chat_agent = GraniteChatAgent(config) # Add GraniteChatAgent
reward_system = RewardSystem(config, db_manager)
summarizer = create_summarizer(config)
return config, db_manager, groq_agent, granite_agent, granite_chat_agent, reward_system, summarizer
def calculate_health_scores(user_profile):
"""Calculate strict health factor scores based on user profile"""
scores = {}
# Sleep Quality Score (stricter evaluation)
sleep_hours = user_profile.get('Sleep_Hours', 7)
sleep_quality = user_profile.get('Sleep_Quality', 'Fair')
if sleep_hours >= 7 and sleep_hours <= 9 and sleep_quality in ['Excellent', 'Good']:
scores['Sleep Quality'] = 9
elif sleep_hours >= 6 and sleep_hours <= 10 and sleep_quality == 'Good':
scores['Sleep Quality'] = 7
elif sleep_hours >= 6 and sleep_hours <= 10 and sleep_quality == 'Fair':
scores['Sleep Quality'] = 5
else:
scores['Sleep Quality'] = 3
# Stress Level Score (stricter)
stress_level = user_profile.get('Stress_Level', 'Medium')
anxiety_freq = user_profile.get('Anxiety_Frequency', 'Sometimes')
if stress_level == 'Low' and anxiety_freq in ['Never', 'Rarely']:
scores['Stress Management'] = 9
elif stress_level == 'Low' and anxiety_freq == 'Sometimes':
scores['Stress Management'] = 7
elif stress_level == 'Medium' and anxiety_freq in ['Never', 'Rarely']:
scores['Stress Management'] = 6
elif stress_level == 'Medium' and anxiety_freq == 'Sometimes':
scores['Stress Management'] = 4
else:
scores['Stress Management'] = 2
# Work-Life Balance Score (stricter)
work_hours = user_profile.get('Work_Hours', 40)
energy_level = user_profile.get('Energy_Level', 'Medium')
if work_hours <= 40 and energy_level in ['Very High', 'High']:
scores['Work-Life Balance'] = 9
elif work_hours <= 45 and energy_level in ['High', 'Medium']:
scores['Work-Life Balance'] = 7
elif work_hours <= 50 and energy_level == 'Medium':
scores['Work-Life Balance'] = 5
elif work_hours <= 55:
scores['Work-Life Balance'] = 3
else:
scores['Work-Life Balance'] = 1
# Physical Activity Score (stricter)
activity_hours = user_profile.get('Physical_Activity_Hours', 3)
if activity_hours >= 5:
scores['Physical Activity'] = 9
elif activity_hours >= 3:
scores['Physical Activity'] = 7
elif activity_hours >= 1.5:
scores['Physical Activity'] = 5
elif activity_hours >= 0.5:
scores['Physical Activity'] = 3
else:
scores['Physical Activity'] = 1
# Diet & Lifestyle Score (new, stricter)
diet = user_profile.get('Diet', 'Average')
smoking = user_profile.get('Smoking', 'Non-Smoker')
alcohol = user_profile.get('Alcohol_Consumption', 'Rarely')
diet_score = 9 if diet == 'Healthy' else 5 if diet == 'Average' else 2
smoking_penalty = 0 if smoking == 'Non-Smoker' else -2 if smoking == 'Occasional Smoker' else -4
alcohol_penalty = 0 if alcohol in ['Never', 'Rarely'] else -1 if alcohol == 'Occasionally' else -3
scores['Diet & Lifestyle'] = max(1, diet_score + smoking_penalty + alcohol_penalty)
# Social Media & Digital Wellness (new)
social_media_hours = user_profile.get('Social_Media_Hours', 3)
if social_media_hours <= 1:
scores['Digital Wellness'] = 9
elif social_media_hours <= 2:
scores['Digital Wellness'] = 7
elif social_media_hours <= 4:
scores['Digital Wellness'] = 5
elif social_media_hours <= 6:
scores['Digital Wellness'] = 3
else:
scores['Digital Wellness'] = 1
return scores
def calculate_risk_level(user_profile):
"""Calculate overall risk level based on health scores"""
health_scores = calculate_health_scores(user_profile)
# Calculate weighted average of health scores
# Lower scores indicate higher risk, so we need to invert the scale
total_score = sum(health_scores.values())
max_possible_score = len(health_scores) * 9 # Maximum score per factor is 9
# Convert to risk scale (1-10, where 10 is highest risk)
# If average score is 9, risk should be 1
# If average score is 1, risk should be 10
average_score = total_score / len(health_scores)
risk_level = max(1, min(10, 11 - average_score))
# Apply additional risk factors
additional_risk = 0
# High work hours increase risk
if user_profile.get('Work_Hours', 40) > 60:
additional_risk += 1
# Poor mood significantly increases risk
mood = user_profile.get('Mood', 'Neutral')
if mood in ['Sad', 'Very Sad']:
additional_risk += 2
elif mood == 'Neutral':
additional_risk += 0.5
# High stress with frequent anxiety increases risk
if (user_profile.get('Stress_Level') == 'High' and
user_profile.get('Anxiety_Frequency') in ['Often', 'Always']):
additional_risk += 1.5
# Very low sleep hours increase risk
if user_profile.get('Sleep_Hours', 7) < 5:
additional_risk += 1
# Current medication might indicate existing health issues
if user_profile.get('Medication') == 'Yes':
additional_risk += 0.5
final_risk = min(10, risk_level + additional_risk)
return round(final_risk)
def get_risk_indicator(risk_level):
"""Return risk indicator HTML"""
if risk_level <= 3:
risk_class = "risk-low"
risk_text = "LOW RISK"
risk_description = "Your wellness indicators look good"
elif risk_level <= 6:
risk_class = "risk-medium"
risk_text = "MODERATE RISK"
risk_description = "Some areas need attention"
else:
risk_class = "risk-high"
risk_text = "HIGH RISK"
risk_description = "Important to address wellness concerns"
return f"""
<div class="risk-indicator {risk_class}">
<h4>{risk_text} ({risk_level}/10)</h4>
<p>{risk_description}</p>
</div>
"""
def collect_user_profile():
"""Collect user profile data using the same form as the mental health app"""
st.header("π€ Your Health Profile")
st.markdown("*Please fill out your information to get started with personalized wellness coaching*")
with st.form("user_profile_form"):
st.subheader("π Personal Information")
col1, col2 = st.columns(2)
with col1:
age = st.number_input("Age", min_value=18, max_value=100, value=30)
gender = st.selectbox("Gender", ["Male", "Female", "Non-binary", "Prefer not to say"])
occupation = st.selectbox("Occupation", ["Engineering", "Healthcare", "Education", "IT", "Finance", "Sales", "Other"])
country = st.selectbox("Country", ["USA", "Canada", "UK", "Germany", "Australia", "India", "Other"])
with col2:
consultation = st.selectbox("Previous Mental Health Consultation", ["Yes", "No"])
medication = st.selectbox("Currently Taking Medication", ["Yes", "No"])
diet = st.selectbox("Diet Quality", ["Healthy", "Average", "Unhealthy"])
smoking = st.selectbox("Smoking Habit", ["Non-Smoker", "Occasional Smoker", "Regular Smoker", "Heavy Smoker"])
st.subheader("π Lifestyle Factors")
col3, col4 = st.columns(2)
with col3:
sleep_hours = st.slider("Average Sleep Hours per Night", 4, 12, 7)
work_hours = st.slider("Work Hours per Week", 20, 80, 40)
social_media_hours = st.slider("Social Media Hours per Day", 0, 12, 3)
with col4:
physical_activity = st.slider("Physical Activity Hours per Week", 0, 20, 3)
stress_level = st.selectbox("Stress Level", ["Low", "Medium", "High"])
alcohol_consumption = st.selectbox("Alcohol Consumption", ["Never", "Rarely", "Occasionally", "Regularly"])
st.subheader("π Mental Health & Mood")
col5, col6 = st.columns(2)
with col5:
mood = st.selectbox("General Mood", ["Very Happy", "Happy", "Neutral", "Sad", "Very Sad"])
anxiety_frequency = st.selectbox("Anxiety Frequency", ["Never", "Rarely", "Sometimes", "Often", "Always"])
with col6:
sleep_quality = st.selectbox("Sleep Quality", ["Excellent", "Good", "Fair", "Poor"])
energy_level = st.selectbox("Energy Level", ["Very High", "High", "Medium", "Low", "Very Low"])
submitted = st.form_submit_button("π Start My Wellness Journey", type="primary", use_container_width=True)
if submitted:
# Create user profile dictionary
user_profile = {
"user_id": f"user_{datetime.now().strftime('%Y%m%d_%H%M%S')}",
"Age": age,
"Gender": gender,
"Occupation": occupation,
"Country": country,
"Mental_Health_Consultation": consultation,
"Medication": medication,
"Diet": diet,
"Smoking": smoking,
"Sleep_Hours": sleep_hours,
"Work_Hours": work_hours,
"Social_Media_Hours": social_media_hours,
"Physical_Activity_Hours": physical_activity,
"Stress_Level": stress_level,
"Alcohol_Consumption": alcohol_consumption,
"Mood": mood,
"Anxiety_Frequency": anxiety_frequency,
"Sleep_Quality": sleep_quality,
"Energy_Level": energy_level,
"created_at": datetime.now()
}
return user_profile
return None
def display_user_dashboard(user_profile, db_manager, groq_agent, granite_agent, granite_chat_agent, reward_system, summarizer):
"""Display the main user dashboard with all features using the summarizer"""
config = Config()
user_id = user_profile['user_id']
# Header with user info
st.markdown('<h1 class="main-title">π Your Personal Wellness Dashboard</h1>', unsafe_allow_html=True)
st.markdown(f"*Welcome back! Here's your personalized wellness experience.*")
# Top metrics row
col1, col2, col3, col4 = st.columns(4)
reward_summary = reward_system.get_reward_summary(user_id)
with col1:
st.metric("π° Total Coins", reward_summary['total_coins'])
with col2:
st.metric("β‘ Coins Earned", reward_summary['total_earned'])
with col3:
st.metric("β³ Pending Tasks", reward_summary['pending_tasks'])
with col4:
st.metric("π Tasks Done", reward_summary['completed_tasks'])
# Tabs for different sections
tab1, tab2, tab3, tab4 = st.tabs(["π€ AI Health Coach", "π My Tasks", "π¬ Ask Questions", "π My Progress"])
with tab1:
st.header("π€ AI Health Analysis & Coaching")
col1, col2 = st.columns([2, 1])
with col1:
if st.button("π Get AI Health Analysis", type="primary", use_container_width=True):
with st.spinner("π€ Analyzing your health profile..."):
# Calculate risk level based on health scores
risk_level = calculate_risk_level(user_profile)
# Get mental health assessment from Groq
assessment, _ = groq_agent.analyze_mental_health(user_profile)
# Use summarizer for clean display
summarized_assessment = summarizer.summarize_health_analysis(
assessment, risk_level, user_profile
)
# Display risk level indicator
st.markdown(get_risk_indicator(risk_level), unsafe_allow_html=True)
# Display summarized assessment in bullet format
st.markdown(f"""
<div class="ai-analysis">
<h4>π― Your Wellness Assessment</h4>
<div class="bullet-list">
{summarized_assessment}
</div>
<br>
<small><em>AI-Powered Analysis β’ {datetime.now().strftime('%Y-%m-%d %H:%M')}</em></small>
</div>
""", unsafe_allow_html=True)
# Save conversation with both original and summarized versions
db_manager.save_conversation(user_id, {
"type": "health_analysis",
"assessment": assessment,
"summarized_assessment": summarized_assessment,
"risk_level": risk_level
})
# If risk level is concerning, get tasks from Granite
if risk_level >= 4: # Medium to high risk
st.warning("β οΈ Your assessment indicates some areas that need attention. Let me create a personalized wellness plan for you.")
with st.spinner("π― Creating personalized wellness tasks..."):
# Get tasks from Granite
tasks = granite_agent.assign_wellness_tasks(user_profile, assessment, risk_level)
if tasks:
st.success(f"β
Created {len(tasks)} personalized wellness tasks for you!")
# Save tasks to database
for task in tasks:
task_id = db_manager.save_task(user_id, task)
if task_id:
st.markdown(f"""
<div class="task-card">
<h5>π― {task['title']}</h5>
<p><strong>Type:</strong> {task['task_type'].replace('_', ' ').title()}</p>
<p><strong>Description:</strong> {task['description']}</p>
<p><strong>Duration:</strong> {task['duration_days']} days</p>
<p><strong>Difficulty:</strong> {task['difficulty'].title()}</p>
<p><strong>Instructions:</strong> {task['instructions']}</p>
<p><strong>Completion Criteria:</strong> {task['completion_criteria']}</p>
<p><strong>Reward:</strong> {reward_system.calculate_task_reward(task['task_type'], task['difficulty'])} coins</p>
</div>
""", unsafe_allow_html=True)
st.info("π Your new tasks have been added to the 'My Tasks' tab. Complete them to earn coins!")
else:
st.error("β Unable to create tasks at this time. Please try again later.")
else:
st.success("π Great news! Your mental health profile looks good. Here are some tips to maintain your wellness:")
# Get wellness tips from Groq and summarize them
with st.spinner("π‘ Getting personalized tips..."):
tips = groq_agent.get_health_tips(user_profile)
summarized_tips = summarizer.summarize_wellness_tips(tips, user_profile)
# Display summarized tips
st.markdown(f"""
<div class="wellness-tips">
<h4>π‘ Personalized Wellness Tips</h4>
<div class="bullet-list">
{summarized_tips}
</div>
</div>
""", unsafe_allow_html=True)
# Save tips conversation
db_manager.save_conversation(user_id, {
"type": "wellness_tips",
"tips": tips,
"summarized_tips": summarized_tips
})
with col2:
st.subheader("π Quick Health Insights")
# Calculate health scores based on user profile
health_factors = calculate_health_scores(user_profile)
fig = go.Figure(go.Bar(
x=list(health_factors.values()),
y=list(health_factors.keys()),
orientation='h',
marker_color=['green' if v >= 7 else 'orange' if v >= 5 else 'red' for v in health_factors.values()],
text=[f"{v}/10" for v in health_factors.values()],
textposition='inside'
))
fig.update_layout(
title="Health Factors Score (1-10)",
xaxis_title="Score",
height=350,
showlegend=False,
xaxis=dict(range=[0, 10]),
font=dict(color="black")
)
st.plotly_chart(fig, use_container_width=True)
# Show calculated risk level
current_risk = calculate_risk_level(user_profile)
st.markdown(get_risk_indicator(current_risk), unsafe_allow_html=True)
# Quick wellness tips button
st.subheader("π‘ Daily Wellness Tips")
if st.button("Get Fresh Tips", use_container_width=True):
with st.spinner("π‘ Generating fresh tips..."):
# Get tips from Groq and summarize
tips = groq_agent.get_health_tips(user_profile)
summarized_tips = summarizer.summarize_wellness_tips(tips, user_profile)
st.markdown(f"""
<div class="wellness-tips">
<div class="bullet-list">
{summarized_tips}
</div>
</div>
""", unsafe_allow_html=True)
with tab2:
st.header("π My Wellness Tasks")
# Get user tasks
pending_tasks = db_manager.get_user_tasks(user_id, "pending")
completed_tasks = db_manager.get_user_tasks(user_id, "completed")
col1, col2 = st.columns(2)
with col1:
st.subheader("β³ Pending Tasks")
if pending_tasks:
for task in pending_tasks:
st.markdown(f"""
<div class="pending-task">
<h5>π― {task['title']}</h5>
<p><strong>Type:</strong> {task['task_type'].replace('_', ' ').title()}</p>
<p><strong>Description:</strong> {task['description']}</p>
<p><strong>Instructions:</strong> {task['instructions']}</p>
<p><strong>Reward:</strong> {reward_system.calculate_task_reward(task['task_type'], task.get('difficulty', 'medium'))} coins</p>
</div>
""", unsafe_allow_html=True)
# Task completion form
with st.expander(f"β
Complete: {task['title']}"):
st.write(f"**Completion Criteria:** {task['completion_criteria']}")
with st.form(f"complete_task_{task['task_id']}"):
st.write("Please provide details about your task completion:")
completion_notes = st.text_area(
"How did you complete this task?",
placeholder="Describe what you did, how it went, any challenges..."
)
quality_rating = st.slider(
"Rate the quality of your completion (1-5)",
min_value=1, max_value=5, value=3
)
exceeded_expectations = st.checkbox(
"I went above and beyond the basic requirements"
)
if st.form_submit_button("π Mark as Completed"):
completion_data = {
"notes": completion_notes,
"quality_rating": quality_rating,
"exceeded_expectations": exceeded_expectations
}
coins_earned = reward_system.award_task_completion(
user_id, task['task_id'], completion_data
)
if coins_earned > 0:
st.markdown(f"""
<div class="reward-notification">
π Congratulations! You earned {coins_earned} coins!
</div>
""", unsafe_allow_html=True)
st.success(f"β
Task completed successfully! You earned {coins_earned} coins!")
st.balloons()
# Rerun to update the display
st.rerun()
else:
st.error("β There was an issue completing the task. Please try again.")
else:
st.info("π― No pending tasks right now. Get an AI health analysis to receive personalized wellness tasks!")
with col2:
st.subheader("β
Completed Tasks")
if completed_tasks:
for task in completed_tasks[-5:]: # Show last 5 completed tasks
completed_date = task.get('completed_at', datetime.now()).strftime('%Y-%m-%d')
coins_earned = reward_system.calculate_task_reward(
task['task_type'],
task.get('difficulty', 'medium'),
task.get('completion_data', {})
)
st.markdown(f"""
<div class="completed-task">
<h5>β
{task['title']}</h5>
<p><strong>Completed:</strong> {completed_date}</p>
<p><strong>Coins Earned:</strong> {coins_earned}</p>
</div>
""", unsafe_allow_html=True)
else:
st.info("π Completed tasks will appear here as you finish them!")
# Progress summary
if completed_tasks or pending_tasks:
total_tasks = len(completed_tasks) + len(pending_tasks)
completion_rate = len(completed_tasks) / total_tasks * 100 if total_tasks > 0 else 0
fig_progress = go.Figure(go.Indicator(
mode="gauge+number",
value=completion_rate,
title={'text': "Task Completion Rate"},
domain={'x': [0, 1], 'y': [0, 1]},
gauge={
'axis': {'range': [None, 100]},
'bar': {'color': "darkgreen"},
'steps': [
{'range': [0, 50], 'color': "lightgray"},
{'range': [50, 80], 'color': "yellow"},
{'range': [80, 100], 'color': "lightgreen"}
]
}
))
fig_progress.update_layout(height=250, font=dict(color="black"))
st.plotly_chart(fig_progress, use_container_width=True)
# TAB 3 - UPDATED WITH GRANITE CHAT AGENT
with tab3:
st.header("π¬ Ask Your AI Health Coach")
# Add indicator that we're using Granite Chat
st.markdown(f"""
<div class="granite-chat-indicator">
π§ <strong>Powered by IBM Granite Chat AI</strong> - Advanced conversational health coaching with memory
</div>
""", unsafe_allow_html=True)
st.markdown("*Ask any health-related question and get personalized advice based on your profile with conversation memory*")
# Initialize chat history for this user
if f"granite_chat_history_{user_id}" not in st.session_state:
st.session_state[f"granite_chat_history_{user_id}"] = [
{"role": "assistant", "content": "Hello! I'm your personal AI health coach powered by IBM Granite Chat. I know your profile and I maintain conversation memory to provide better, contextual responses. How can I assist you today?"}
]
# Display chat messages with enhanced health-specific styling
for message in st.session_state[f"granite_chat_history_{user_id}"]:
with st.chat_message(message["role"]):
if message["role"] == "assistant":
content = message["content"]
# Check if this message contains medical advice
if "consult" in content.lower() and "doctor" in content.lower():
# Highlight medical disclaimers
st.markdown(f'''
<div class="chat-assistant-message" style="color: white !important;">
<div style="background: rgba(255,193,7,0.2); padding: 10px; border-radius: 5px; border-left: 3px solid #ffc107;">
β οΈ <strong>Medical Advice:</strong><br>
{content}
</div>
</div>
''', unsafe_allow_html=True)
else:
st.markdown(f'<div class="chat-assistant-message" style="color: white !important;">{content}</div>', unsafe_allow_html=True)
else:
st.markdown(f'<div class="chat-user-message" style="color: white !important;">{message["content"]}</div>', unsafe_allow_html=True)
# Chat input
if prompt := st.chat_input("Ask me anything about health and wellness..."):
# Add user message
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "user", "content": prompt})
with st.chat_message("user"):
st.markdown(f'<div class="chat-user-message" style="color: white !important;">{prompt}</div>', unsafe_allow_html=True)
# Generate AI response using Granite Chat agent
with st.chat_message("assistant"):
with st.spinner("π§ Granite Chat AI is thinking..."):
try:
# Use Granite Chat agent for chat response with conversation memory
raw_response = granite_chat_agent.get_chat_response(prompt, user_profile, context=None)
# Use health-specific cleaning function
response, metadata = clean_health_ai_response(raw_response)
if response:
# Add medical disclaimer styling if present
if metadata.get('has_disclaimer', False):
response_html = f'''
<div class="chat-assistant-message" style="color: white !important;">
<div style="background: rgba(255,193,7,0.2); padding: 10px; border-radius: 5px; margin: 5px 0; border-left: 3px solid #ffc107;">
β οΈ <strong>Medical Advice:</strong><br>
{response}
</div>
</div>
'''
else:
response_html = f'<div class="chat-assistant-message" style="color: white !important;">{response}</div>'
st.markdown(response_html, unsafe_allow_html=True)
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "assistant", "content": response})
# Save with enhanced metadata
db_manager.save_conversation(user_id, {
"type": "granite_chat_interaction",
"user_question": prompt,
"ai_response": response,
"agent_used": "granite_chat",
"metadata": metadata,
"has_medical_disclaimer": metadata.get('has_disclaimer', False),
"conversation_length": len(granite_chat_agent.conversation_history)
})
else:
error_msg = "I'm having trouble processing your question right now. Could you please rephrase or try again?"
st.markdown(f'<div class="chat-message" style="color: white !important;">{error_msg}</div>', unsafe_allow_html=True)
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "assistant", "content": error_msg})
except Exception as e:
error_msg = f"I encountered an error while processing your question. Please try again or rephrase your question."
st.markdown(f'<div class="chat-message" style="color: white !important;">{error_msg}</div>', unsafe_allow_html=True)
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "assistant", "content": error_msg})
# Quick question buttons - now using Granite Chat
st.subheader("π Quick Questions")
col1, col2, col3 = st.columns(3)
quick_questions = [
"How can I reduce my stress levels?",
"What's the best sleep routine for me?",
"How much exercise should I be doing?",
"How can I improve my work-life balance?",
"What foods should I eat for better mood?",
"How can I manage my social media usage?"
]
for i, question in enumerate(quick_questions[:6]):
col = [col1, col2, col3][i % 3]
with col:
if st.button(question, key=f"granite_quick_q_{i}"):
# Add question to chat and trigger response
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "user", "content": question})
# Generate AI response using Granite Chat
with st.spinner("π§ Granite Chat AI is thinking..."):
try:
# Use Granite Chat agent for chat response
raw_response = granite_chat_agent.get_chat_response(question, user_profile, context=None)
# Use health-specific cleaning function
response, metadata = clean_health_ai_response(raw_response)
if response:
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "assistant", "content": response})
# Save with enhanced metadata
db_manager.save_conversation(user_id, {
"type": "granite_chat_interaction",
"user_question": question,
"ai_response": response,
"agent_used": "granite_chat",
"metadata": metadata,
"has_medical_disclaimer": metadata.get('has_disclaimer', False),
"conversation_length": len(granite_chat_agent.conversation_history)
})
else:
error_msg = "I'm having trouble processing your question right now. Please try again."
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "assistant", "content": error_msg})
except Exception as e:
error_msg = f"I encountered an error while processing your question. Please try again."
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "assistant", "content": error_msg})
# Rerun to show the new messages
st.rerun()
# Add Granite Chat-specific features
st.markdown("---")
st.subheader("π§ Granite Chat AI Features")
col1, col2, col3 = st.columns(3)
with col1:
if st.button("π‘ Get Wellness Advice", use_container_width=True, key="granite_wellness_advice"):
with st.spinner("π§ Granite Chat AI generating wellness advice..."):
advice = granite_chat_agent.get_wellness_advice("general wellness based on my profile", user_profile)
cleaned_advice, _ = clean_health_ai_response(advice)
if cleaned_advice:
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "assistant", "content": cleaned_advice})
st.rerun()
with col2:
if st.button("β Ask Health Question", use_container_width=True, key="granite_health_question"):
question = "What should I focus on most for better health based on my profile?"
with st.spinner("π§ Granite Chat AI answering..."):
answer = granite_chat_agent.answer_question(question, user_profile)
cleaned_answer, _ = clean_health_ai_response(answer)
if cleaned_answer:
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "user", "content": question})
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "assistant", "content": cleaned_answer})
st.rerun()
with col3:
if st.button("π€ Get Support", use_container_width=True, key="granite_support"):
concern = "feeling overwhelmed with my wellness goals"
with st.spinner("π§ Granite Chat AI providing support..."):
support = granite_chat_agent.provide_support(concern, user_profile)
cleaned_support, _ = clean_health_ai_response(support)
if cleaned_support:
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "user", "content": f"I'm {concern}"})
st.session_state[f"granite_chat_history_{user_id}"].append({"role": "assistant", "content": cleaned_support})
st.rerun()
# Granite Chat conversation controls
st.markdown("---")
st.subheader("ποΈ Chat Controls")
col1, col2, col3 = st.columns(3)
with col1:
if st.button("ποΈ Clear Chat History", use_container_width=True, key="granite_clear_chat"):
st.session_state[f"granite_chat_history_{user_id}"] = [
{"role": "assistant", "content": "Hello! I'm your personal AI health coach powered by IBM Granite Chat. How can I assist you today?"}
]
granite_chat_agent.clear_conversation_history()
st.success("Chat history cleared!")
st.rerun()
with col2:
if st.button("π Conversation Summary", use_container_width=True, key="granite_conv_summary"):
summary = granite_chat_agent.get_conversation_summary()
st.info(summary)
with col3:
# Chat personality selector
personality_type = st.selectbox("π Chat Personality",
["supportive", "professional", "casual", "direct"],
key="granite_personality_select",
help="Choose how the AI should respond to you")
if st.button("Set Personality", use_container_width=True, key="granite_set_personality"):
granite_chat_agent.set_chat_personality(personality_type)
st.success(f"Chat personality set to: {personality_type}")
# Display conversation statistics
st.markdown("---")
st.subheader("π Chat Statistics")
col1, col2, col3 = st.columns(3)
total_messages = len(st.session_state.get(f"granite_chat_history_{user_id}", []))
user_messages = len([msg for msg in st.session_state.get(f"granite_chat_history_{user_id}", []) if msg["role"] == "user"])
assistant_messages = len([msg for msg in st.session_state.get(f"granite_chat_history_{user_id}", []) if msg["role"] == "assistant"])
with col1:
st.metric("Total Messages", total_messages)
with col2:
st.metric("Your Questions", user_messages)
with col3:
st.metric("AI Responses", assistant_messages)
# Show Granite Chat agent internal conversation history length
granite_internal_history = len(granite_chat_agent.conversation_history)
st.info(f"π§ Granite Chat AI is maintaining {granite_internal_history} conversation turns in memory for better context.")
with tab4:
st.header("π My Wellness Progress")
# Get user's progress data
all_tasks = db_manager.get_user_tasks(user_id)
completed_tasks = [task for task in all_tasks if task.get('status') == 'completed']
conversations = list(db_manager.db[config.CONVERSATIONS_COLLECTION].find(
{"user_id": user_id}
).sort("timestamp", -1).limit(10))
col1, col2 = st.columns(2)
with col1:
st.subheader("π Achievement Summary")
# Use summarizer for progress summary
progress_summary = summarizer.create_progress_summary(completed_tasks, user_profile)
# Display progress summary
st.markdown(f"""
<div class="progress-summary">
<h4>π Your Progress Highlights</h4>
<div class="bullet-list">
{progress_summary}
</div>
</div>
""", unsafe_allow_html=True)
# Calculate achievements
reward_summary = reward_system.get_reward_summary(user_id)
st.metric("Total Coins Earned", reward_summary['total_earned'])
st.metric("Current Coin Balance", reward_summary['total_coins'])
st.metric("Tasks Completed", reward_summary['completed_tasks'])
st.metric("Pending Tasks", reward_summary['pending_tasks'])
# Current risk level display
current_risk = calculate_risk_level(user_profile)
st.markdown(f"""
<div class="risk-indicator risk-{'low' if current_risk <= 3 else 'medium' if current_risk <= 6 else 'high'}">
<h4>Current Risk Level: {current_risk}/10</h4>
<p>Based on your health profile analysis</p>
</div>
""", unsafe_allow_html=True)
# Task completion by type
if all_tasks:
completed_by_type = {}
for task in all_tasks:
if task.get('status') == 'completed':
task_type = task['task_type'].replace('_', ' ').title()
completed_by_type[task_type] = completed_by_type.get(task_type, 0) + 1
if completed_by_type:
fig_tasks = px.pie(
values=list(completed_by_type.values()),
names=list(completed_by_type.keys()),
title="Completed Tasks by Type"
)
fig_tasks.update_layout(font=dict(color="black"))
st.plotly_chart(fig_tasks, use_container_width=True)
with col2:
st.subheader("π Progress Timeline")
if all_tasks:
# Create timeline of task completions
task_timeline = []
for task in all_tasks:
if task.get('status') == 'completed' and task.get('completed_at'):
task_timeline.append({
'date': task['completed_at'].strftime('%Y-%m-%d'),
'task': task['title'],
'coins': reward_system.calculate_task_reward(
task['task_type'],
task.get('difficulty', 'medium')
)
})
if task_timeline:
timeline_df = pd.DataFrame(task_timeline)
# Group by date and sum coins
daily_coins = timeline_df.groupby('date')['coins'].sum().reset_index()
fig_timeline = px.line(
daily_coins,
x='date',
y='coins',
title="Daily Coins Earned",
markers=True
)
fig_timeline.update_layout(font=dict(color="black"))
st.plotly_chart(fig_timeline, use_container_width=True)
else:
st.info("Complete some tasks to see your progress timeline!")
else:
st.info("Your progress will appear here as you complete tasks and interact with the AI coach!")
# Health scores visualization
st.subheader("π Health Factor Trends")
health_scores = calculate_health_scores(user_profile)
fig_health = go.Figure(data=go.Scatterpolar(
r=list(health_scores.values()),
theta=list(health_scores.keys()),
fill='toself',
name='Health Scores'
))
fig_health.update_layout(
polar=dict(
radialaxis=dict(
visible=True,
range=[0, 10]
)),
showlegend=False,
title="Health Factors Radar Chart",
font=dict(color="black")
)
st.plotly_chart(fig_health, use_container_width=True)
# Recent activity with summarized content
st.subheader("π Recent Activity")
if conversations:
for conv in conversations[:5]:
timestamp = conv['timestamp'].strftime('%Y-%m-%d %H:%M')
conv_type = conv.get('type', 'unknown')
agent_used = conv.get('agent_used', 'unknown')
with st.expander(f"{conv_type.replace('_', ' ').title()} ({agent_used.title()}) - {timestamp}"):
if conv_type == 'health_analysis':
st.write(f"**Risk Level:** {conv.get('risk_level', 'N/A')}/10")
# Use summarized assessment if available, otherwise original
assessment_display = conv.get('summarized_assessment', conv.get('assessment', 'N/A'))
st.markdown(f"""
<div style="color: black;">
<strong>Assessment:</strong>
<div class="bullet-list">
{assessment_display}
</div>
</div>
""", unsafe_allow_html=True)
elif conv_type == 'wellness_tips':
# Use summarized tips if available
tips_display = conv.get('summarized_tips', conv.get('tips', 'N/A'))
st.markdown(f"""
<div style="color: black;">
<strong>Tips:</strong>
<div class="bullet-list">
{tips_display}
</div>
</div>
""", unsafe_allow_html=True)
elif conv_type in ['chat_interaction', 'granite_chat_interaction']:
st.write(f"**Question:** {conv.get('user_question', 'N/A')}")
# Use summarized response if available
response_display = conv.get('summarized_response', conv.get('ai_response', 'N/A'))
st.markdown(f"""
<div style="color: black;">
<strong>Response:</strong>
<div class="bullet-list">
{response_display}
</div>
</div>
""", unsafe_allow_html=True)
# Show conversation metadata if available
if conv.get('conversation_length'):
st.write(f"**Conversation Context:** {conv['conversation_length']} turns in memory")
if conv.get('has_medical_disclaimer'):
st.write("β οΈ **Contains Medical Disclaimer**")
else:
st.info("Your recent interactions with the AI coach will appear here!")
def main():
"""Main application function"""
config, db_manager, groq_agent, granite_agent, granite_chat_agent, reward_system, summarizer = initialize_services()
# Check if user profile exists in session
if "user_profile" not in st.session_state:
st.session_state.user_profile = None
# If no user profile, show profile collection form
if st.session_state.user_profile is None:
st.markdown('<h1 class="main-title">π Welcome to Your Dynamic Wellness Platform</h1>', unsafe_allow_html=True)
st.markdown("*Get personalized AI health coaching, dynamic task assignments, and earn rewards for your wellness journey*")
# Show features overview
col1, col2, col3 = st.columns(3)
with col1:
st.markdown("""
### π€ AI Health Coach
- Advanced risk calculation
- Groq AI analyzes your profile
- Summarized, bullet-point insights
- Smart risk assessment
""")
with col2:
st.markdown("""
### π― Dynamic Tasks
- Granite AI assigns wellness tasks
- Based on calculated risk levels
- Earn coins for completion
- Personalized difficulty levels
""")
with col3:
st.markdown("""
### π¬ Interactive Chat
- **IBM Granite Chat AI**
- Conversation memory & context
- Get personalized health advice
- Multiple personality modes
""")
st.markdown("---")
# Show enhanced features
st.info("""
π **Enhanced Features with Granite Chat AI**:
- **IBM Granite Conversational AI**: Advanced chat capabilities with conversation memory
- **Context-Aware Responses**: Remembers previous conversations for better continuity
- **Enhanced Risk Assessment**: Multi-factor risk calculation based on sleep, stress, work-life balance, and lifestyle
- **Smart Response Processing**: Clean, readable responses with medical disclaimer detection
- **Contextual Advice**: Personalized recommendations based on your specific profile and conversation history
- **Multiple Personality Modes**: Supportive, professional, casual, or direct conversation styles
- **Advanced Chat Features**: Wellness advice, Q&A, support functions with conversation memory
- **Progress Tracking**: Comprehensive conversation history and achievement summaries
""")
# Collect user profile
user_profile = collect_user_profile()
if user_profile:
# Save to database
if db_manager.save_user_profile(user_profile):
st.session_state.user_profile = user_profile
# Show initial risk calculation
risk_level = calculate_risk_level(user_profile)
st.markdown(get_risk_indicator(risk_level), unsafe_allow_html=True)
st.success("β
Profile saved successfully! Redirecting to your dashboard...")
st.rerun()
else:
st.error("β Error saving profile. Please try again.")
else:
# Display main dashboard with granite chat agent
display_user_dashboard(
st.session_state.user_profile,
db_manager,
groq_agent,
granite_agent,
granite_chat_agent, # Added granite_chat_agent parameter
reward_system,
summarizer
)
# Add a reset button in sidebar
with st.sidebar:
st.header("βοΈ Settings")
user_profile = st.session_state.user_profile
reward_summary = reward_system.get_reward_summary(user_profile['user_id'])
risk_level = calculate_risk_level(user_profile)
st.metric("π° Total Coins", reward_summary['total_coins'])
st.metric("π Tasks Completed", reward_summary['completed_tasks'])
st.metric("β³ Pending Tasks", reward_summary['pending_tasks'])
st.metric("β οΈ Risk Level", f"{risk_level}/10")
st.markdown("---")
# Health scores breakdown
st.subheader("π Health Scores")
health_scores = calculate_health_scores(user_profile)
for factor, score in health_scores.items():
color = "π’" if score >= 7 else "π‘" if score >= 5 else "π΄"
st.write(f"{color} {factor}: {score}/10")
st.markdown("---")
if st.button("π New User Profile", type="secondary"):
st.session_state.user_profile = None
st.rerun()
st.markdown("---")
st.markdown("""
### π― Enhanced Features:
1. **IBM Granite Chat AI**: Advanced conversational health coaching with memory
2. **Context-Aware Conversations**: Remembers your chat history for better responses
3. **Smart Risk Analysis**: Multi-factor health assessment with personalized insights
4. **Dynamic Task Assignment**: AI creates tasks based on your specific risk profile
5. **Interactive Chat Features**: Wellness advice, Q&A, and support with conversation continuity
6. **Progress Tracking**: View comprehensive achievements and progress highlights
7. **Reward System**: Earn coins for completing wellness activities
### π Latest Updates:
- **Granite Chat Integration**: Advanced IBM Granite Chat AI with conversation memory
- **Enhanced Context Awareness**: AI remembers your previous conversations
- **Improved Response Quality**: Better cleaning and processing of AI responses
- **Personality Modes**: Choose between supportive, professional, casual, or direct styles
- **Conversation Statistics**: Track your chat interactions and AI memory usage
- **Medical Disclaimer Detection**: Automatic detection and highlighting of medical advice
""")
# Footer
st.markdown("---")
st.markdown("""
<div style='text-align: center; color: #666; padding: 20px;'>
<p>π <strong>Enhanced Dynamic Wellness Platform with IBM Granite Chat AI</strong></p>
<p><small>IBM Granite Chat AI with Memory β’ Advanced Risk Assessment β’ Personalized Task Assignment β’ Smart Health Coaching</small></p>
<p><small>β οΈ This tool provides wellness guidance. Consult healthcare professionals for medical advice.</small></p>
</div>
""", unsafe_allow_html=True)
if __name__ == "__main__":
main() |