Spaces:
Sleeping
Sleeping
File size: 60,407 Bytes
31adc25 139d9d7 6da16d5 df534c5 a88ab54 a697707 31adc25 a697707 65a44fe dff71a5 31adc25 dff71a5 31adc25 dff71a5 31adc25 dff71a5 31adc25 dff71a5 31adc25 8e6c67f df534c5 8e6c67f df534c5 8e6c67f df534c5 8e6c67f df534c5 8e6c67f df534c5 8e6c67f df534c5 8e6c67f df534c5 8e6c67f df534c5 8e6c67f df534c5 8e6c67f a88ab54 8e6c67f df534c5 8e6c67f df534c5 8e6c67f df534c5 8e6c67f df534c5 8e6c67f df534c5 8e6c67f a88ab54 df534c5 8e6c67f a697707 79476b6 31adc25 b92bfe9 31adc25 b92bfe9 139d9d7 a697707 139d9d7 31adc25 139d9d7 b92bfe9 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 a88ab54 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 b92bfe9 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 139d9d7 a697707 8e6c67f a697707 8e6c67f 139d9d7 79476b6 139d9d7 31adc25 dff71a5 65a44fe dff71a5 65a44fe dff71a5 301ae3f dff71a5 301ae3f dff71a5 301ae3f 65a44fe dff71a5 65a44fe dff71a5 79476b6 31adc25 139d9d7 31adc25 139d9d7 31adc25 6da16d5 31adc25 a88ab54 31adc25 139d9d7 31adc25 139d9d7 31adc25 dff71a5 139d9d7 dff71a5 90732c0 dff71a5 90732c0 dff71a5 90732c0 dff71a5 90732c0 dff71a5 31adc25 8e6c67f 31adc25 8e6c67f d1ac8aa 8e6c67f d1ac8aa 8e6c67f 31adc25 8e6c67f d1ac8aa 8e6c67f 139d9d7 cbe3f57 139d9d7 31adc25 8e6c67f 31adc25 139d9d7 31adc25 79476b6 139d9d7 31adc25 79476b6 31adc25 139d9d7 31adc25 139d9d7 97db5ba 139d9d7 97db5ba 139d9d7 dff71a5 139d9d7 dff71a5 139d9d7 dff71a5 139d9d7 79476b6 139d9d7 31adc25 |
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 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 |
import streamlit as st
import asyncio
import re
import os
import plotly.graph_objects as go
import yfinance as yf
import time
from datetime import timedelta
import gnews
from bs4 import BeautifulSoup
import requests
import holidays
import pandas as pd
import numpy as np
from sklearn.metrics import mean_squared_error, r2_score
from sklearn.linear_model import Ridge
from sklearn.model_selection import GridSearchCV
from dotenv import load_dotenv
from sklearn.preprocessing import StandardScaler
from langchain_mcp_adapters.client import MultiServerMCPClient
from langgraph.prebuilt import create_react_agent
from langchain_groq import ChatGroq
# Load environment variables
load_dotenv()
# Check if API key exists - support both .env and Streamlit secrets
os.environ["GROQ_API_KEY"] = os.getenv("GROQ_API_KEY") or st.secrets.get("GROQ_API_KEY")
model_name = os.getenv("MODEL") or st.secrets.get("MODEL")
if not os.environ["GROQ_API_KEY"]:
st.error(
"β Error: GROQ_API_KEY not found. Please set it in your environment variables or Streamlit secrets."
)
st.stop()
if not model_name:
st.error(
"β Error: MODEL not found. Please set it in your environment variables or Streamlit secrets."
)
st.stop()
@st.cache_data(ttl=3600) # Cache for 1 hour
def get_available_tickers():
"""Fetch available tickers using multiple APIs and sources."""
try:
print("Fetching stock tickers from multiple sources...")
tickers_dict = {}
# Method 1: Try to get stocks from a free API
try:
print("Fetching stocks from API...")
# Try to get stocks from a free API endpoint
api_url = "https://api.polygon.io/v3/reference/tickers?market=stocks&active=true&limit=1000"
# Try alternative free APIs
apis_to_try = [
"https://api.twelvedata.com/stocks?country=US&exchange=NASDAQ",
"https://api.twelvedata.com/stocks?country=US&exchange=NYSE",
"https://api.twelvedata.com/stocks?country=US&exchange=AMEX",
]
for api_url in apis_to_try:
try:
response = requests.get(api_url, timeout=10)
if response.status_code == 200:
data = response.json()
if "data" in data:
for item in data["data"]:
ticker = item.get("symbol", "")
name = item.get("name", ticker)
if (
ticker and name and len(ticker) <= 5
): # Filter for likely stock tickers
tickers_dict[ticker] = name
print(f"Loaded {len(tickers_dict)} stocks from {api_url}")
break
except Exception as e:
print(f"Error with API {api_url}: {e}")
continue
except Exception as e:
print(f"Error fetching from APIs: {e}")
# Method 2: Try additional free APIs for more stocks
if len(tickers_dict) < 100: # Only if we didn't get enough from first APIs
try:
print("Fetching additional stocks from more APIs...")
# Try more free APIs
additional_apis = [
"https://api.twelvedata.com/stocks?country=US&exchange=NASDAQ&limit=500",
"https://api.twelvedata.com/stocks?country=US&exchange=NYSE&limit=500",
"https://api.twelvedata.com/stocks?country=US&exchange=AMEX&limit=500",
"https://api.twelvedata.com/stocks?country=CA&exchange=TSX&limit=200",
"https://api.twelvedata.com/stocks?country=GB&exchange=LSE&limit=200",
]
for api_url in additional_apis:
try:
response = requests.get(api_url, timeout=10)
if response.status_code == 200:
data = response.json()
if "data" in data:
for item in data["data"]:
ticker = item.get("symbol", "")
name = item.get("name", ticker)
if (
ticker and name and len(ticker) <= 5
): # Filter for likely stock tickers
if (
ticker not in tickers_dict
): # Avoid duplicates
tickers_dict[ticker] = name
print(f"Loaded additional stocks from {api_url}")
except Exception as e:
print(f"Error with additional API {api_url}: {e}")
continue
print(f"Loaded {len(tickers_dict)} total stocks from all APIs")
except Exception as e:
print(f"Error fetching from additional APIs: {e}")
# Method 3: Try to get stocks from Yahoo Finance screener (if available)
if len(tickers_dict) < 200: # Only if we need more
try:
print("Trying Yahoo Finance screener...")
# This is a fallback that doesn't hardcode tickers
# We'll try to get some popular stocks dynamically
popular_keywords = [
"technology",
"finance",
"healthcare",
"energy",
"consumer",
]
for keyword in popular_keywords:
try:
# Try to search for stocks by sector
search_url = f"https://api.twelvedata.com/stocks?search={keyword}&limit=50"
response = requests.get(search_url, timeout=10)
if response.status_code == 200:
data = response.json()
if "data" in data:
for item in data["data"]:
ticker = item.get("symbol", "")
name = item.get("name", ticker)
if (
ticker and name and len(ticker) <= 5
): # Filter for likely stock tickers
if (
ticker not in tickers_dict
): # Avoid duplicates
tickers_dict[ticker] = name
except Exception as e:
print(f"Error searching for {keyword}: {e}")
continue
print(
f"Loaded {len(tickers_dict)} total stocks (including sector searches)"
)
except Exception as e:
print(f"Error fetching from sector searches: {e}")
if len(tickers_dict) > 0:
print(
f"Successfully loaded {len(tickers_dict)} valid tickers from multiple sources"
)
return tickers_dict
else:
print("No tickers loaded from APIs, using fallback list")
except Exception as e:
print(f"Error in main ticker fetching: {e}")
# Fallback to comprehensive list if all APIs fail
try:
print("Using comprehensive fallback list...")
fallback_tickers = {}
# Comprehensive list of major stocks across sectors
fallback_ticker_list = [
"AAPL",
"MSFT",
"GOOG",
"AMZN",
"META",
"NVDA",
"TSLA",
"NFLX",
"ADBE",
]
print(f"Loading {len(fallback_ticker_list)} fallback tickers...")
# Get company names for each ticker
for ticker in fallback_ticker_list:
try:
ticker_obj = yf.Ticker(ticker)
info = ticker_obj.info
if info and (info.get("longName") or info.get("shortName")):
company_name = info.get("longName", info.get("shortName", ticker))
fallback_tickers[ticker] = company_name
except Exception as e:
# Skip tickers that cause errors
continue
print(f"Successfully loaded {len(fallback_tickers)} tickers from fallback")
return fallback_tickers
except Exception as e:
st.error(f"Error fetching available tickers: {e}")
# Final fallback to basic tickers if there's an error
return {
"AAPL": "Apple Inc.",
"TSLA": "Tesla Inc.",
"MSFT": "Microsoft Corporation",
"GOOG": "Alphabet Inc. (Google)",
"AMZN": "Amazon.com Inc.",
"META": "Meta Platforms Inc.",
"NVDA": "NVIDIA Corporation",
"JPM": "JPMorgan Chase & Co.",
"JNJ": "Johnson & Johnson",
"PG": "Procter & Gamble Co.",
}
@st.cache_data(ttl=3600) # Cache for 1 hour
def search_ticker(ticker_symbol):
"""Search for a ticker symbol and get its company name using yfinance."""
try:
ticker = yf.Ticker(ticker_symbol)
info = ticker.info
company_name = info.get("longName", info.get("shortName", ticker_symbol))
return company_name
except Exception as e:
return None
def calculate_rsi(data, window):
"""Calculate RSI (Relative Strength Index) for the given data."""
delta = data.diff()
gain = delta.where(delta > 0, 0)
loss = -delta.where(delta < 0, 0)
avg_gain = gain.rolling(window=window, min_periods=1).mean()
avg_loss = loss.rolling(window=window, min_periods=1).mean()
rs = avg_gain / avg_loss
rsi = 100 - (100 / (1 + rs))
return rsi
@st.cache_data(ttl=3600) # Cache for 1 hour
def create_stock_chart(ticker: str):
"""Create an interactive stock price chart with Ridge Regression predictions for the given ticker."""
try:
# Get stock data - 5 years for training Ridge Regression
with st.spinner(f"π Fetching stock data for {ticker}..."):
stock = yf.Ticker(ticker)
hist_data = stock.history(period="5y")
if hist_data.empty:
st.warning(f"No data available for {ticker}")
return None
# Prepare data for Ridge Regression with technical indicators
df = hist_data.reset_index()
# Flatten the multi-level column index if it exists
if isinstance(df.columns, pd.MultiIndex):
df.columns = df.columns.get_level_values(0)
# Calculate technical indicators (same as in the notebook)
# Moving averages
df["SMA_20"] = df["Close"].rolling(window=20).mean()
df["SMA_50"] = df["Close"].rolling(window=50).mean()
# RSI
df["RSI"] = calculate_rsi(df["Close"], window=14)
# MACD
exp12 = df["Close"].ewm(span=12, adjust=False).mean()
exp26 = df["Close"].ewm(span=26, adjust=False).mean()
df["MACD"] = exp12 - exp26
df["MACD_Signal"] = df["MACD"].ewm(span=9, adjust=False).mean()
# Bollinger Band component
df["BB_StdDev"] = df["Close"].rolling(window=20).std()
# Volume moving average
df["Volume_Avg"] = df["Volume"].rolling(window=20).mean()
# Price momentum and volatility
df["Price_Change"] = df["Close"].pct_change()
df["Price_Change_5d"] = df["Close"].pct_change(periods=5)
df["Price_Change_20d"] = df["Close"].pct_change(periods=20)
df["Price_Volatility"] = df["Close"].rolling(window=20).std()
df["Price_Range"] = (df["High"] - df["Low"]) / df["Close"] # Daily range
# Volume-Based Features
df["Volume_Change"] = df["Volume"].pct_change()
df["Volume_Price_Trend"] = df["Volume"] * df["Price_Change"]
df["Volume_SMA_Ratio"] = df["Volume"] / df["Volume"].rolling(window=20).mean()
df["Volume_StdDev"] = df["Volume"].rolling(window=20).std()
# Advanced Technical Indicators
# Stochastic Oscillator
def calculate_stochastic(df, window=14):
lowest_low = df["Low"].rolling(window=window).min()
highest_high = df["High"].rolling(window=window).max()
k_percent = 100 * ((df["Close"] - lowest_low) / (highest_high - lowest_low))
return k_percent
df["Stochastic_K"] = calculate_stochastic(df)
df["Stochastic_D"] = df["Stochastic_K"].rolling(window=3).mean()
# Williams %R
def calculate_williams_r(df, window=14):
highest_high = df["High"].rolling(window=window).max()
lowest_low = df["Low"].rolling(window=window).min()
williams_r = -100 * (
(highest_high - df["Close"]) / (highest_high - lowest_low)
)
return williams_r
df["Williams_R"] = calculate_williams_r(df)
# Commodity Channel Index (CCI)
def calculate_cci(df, window=20):
typical_price = (df["High"] + df["Low"] + df["Close"]) / 3
sma_tp = typical_price.rolling(window=window).mean()
mad = typical_price.rolling(window=window).apply(
lambda x: np.mean(np.abs(x - x.mean()))
)
cci = (typical_price - sma_tp) / (0.015 * mad)
return cci
df["CCI"] = calculate_cci(df)
# Moving Average Crossovers
df["SMA_10"] = df["Close"].rolling(window=10).mean()
df["SMA_20"] = df["Close"].rolling(window=20).mean()
df["SMA_50"] = df["Close"].rolling(window=50).mean()
df["SMA_200"] = df["Close"].rolling(window=200).mean()
# Crossover signals
df["SMA_10_20_Cross"] = (df["SMA_10"] > df["SMA_20"]).astype(int)
df["SMA_20_50_Cross"] = (df["SMA_20"] > df["SMA_50"]).astype(int)
df["SMA_50_200_Cross"] = (df["SMA_50"] > df["SMA_200"]).astype(int)
# Bollinger Bands Components
df["BB_Upper"] = df["SMA_20"] + (df["BB_StdDev"] * 2)
df["BB_Lower"] = df["SMA_20"] - (df["BB_StdDev"] * 2)
df["BB_Position"] = (df["Close"] - df["BB_Lower"]) / (
df["BB_Upper"] - df["BB_Lower"]
)
df["BB_Squeeze"] = (df["BB_Upper"] - df["BB_Lower"]) / df[
"SMA_20"
] # Volatility indicator
# Support and Resistance
df["Resistance_20d"] = df["High"].rolling(window=20).max()
df["Support_20d"] = df["Low"].rolling(window=20).min()
df["Price_to_Resistance"] = df["Close"] / df["Resistance_20d"]
df["Price_to_Support"] = df["Close"] / df["Support_20d"]
# Time-based features
df["Day_of_Week"] = df["Date"].dt.dayofweek
df["Month"] = df["Date"].dt.month
df["Quarter"] = df["Date"].dt.quarter
df["Is_Month_End"] = df["Date"].dt.is_month_end.astype(int)
df["Is_Quarter_End"] = df["Date"].dt.is_quarter_end.astype(int)
# Market Sentiment Features
df["Price_Above_SMA200"] = (df["Close"] > df["SMA_200"]).astype(int)
df["Volume_Spike"] = (
df["Volume"] > df["Volume"].rolling(window=20).mean() * 1.5
).astype(int)
df["Price_Spike"] = (
df["Price_Change"].abs() > df["Price_Change"].rolling(window=20).std() * 2
).astype(int)
# Drop rows with NaN values created by moving averages and new features
df.dropna(inplace=True)
# Define features and target (same as notebook)
features = [
"SMA_10",
"SMA_20",
"SMA_50",
"SMA_200",
"RSI",
"MACD",
"MACD_Signal",
"BB_StdDev",
"BB_Position",
"BB_Squeeze",
"Stochastic_K",
"Stochastic_D",
"Williams_R",
"CCI",
"Price_Change",
"Price_Change_5d",
"Price_Change_20d",
"Price_Volatility",
"Price_Range",
"Volume_Change",
"Volume_Price_Trend",
"Volume_SMA_Ratio",
"Volume_StdDev",
"SMA_10_20_Cross",
"SMA_20_50_Cross",
"SMA_50_200_Cross",
"Price_to_Resistance",
"Price_to_Support",
"Day_of_Week",
"Month",
"Quarter",
"Is_Month_End",
"Is_Quarter_End",
"Price_Above_SMA200",
"Volume_Spike",
"Price_Spike",
"Volume_Avg",
]
target = "Close"
X = df[features]
y = df[target]
# Train on ALL available data (5 years)
X_train = X # Use all available data for training
y_train = y
# Add feature scaling
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
# Train Ridge Regression model with cross-validation
start_time = time.time()
with st.spinner(f"Training Ridge Regression model for {ticker}..."):
# Use Ridge with cross-validation to find optimal alpha
ridge_model = Ridge()
# Grid search for optimal regularization strength
param_grid = {"alpha": [0.001, 0.01, 0.1, 1.0, 10.0, 100.0]}
grid_search = GridSearchCV(ridge_model, param_grid, cv=5, scoring="r2")
grid_search.fit(X_train_scaled, y_train)
# Use the best model
model = grid_search.best_estimator_
# Track training time
training_time = time.time() - start_time
# Get the best alpha value for display
best_alpha = grid_search.best_params_["alpha"]
best_score = grid_search.best_score_
# Create future dates for next 30 days
last_date = df["Date"].max()
future_dates = pd.date_range(
start=last_date + timedelta(days=1), periods=30, freq="D"
)
# Filter for trading days only
future_trading_dates = [date for date in future_dates if is_trading_day(date)]
# Create a more sophisticated future prediction approach
# We'll use a more realistic projection with some randomness and market patterns
future_features = []
# Get the last few values to calculate trends
last_20_prices = df["Close"].tail(20).values
last_50_prices = df["Close"].tail(50).values
last_volumes = df["Volume"].tail(20).values
# Get the last known values for technical indicators
last_values = df.iloc[-1]
# Calculate more sophisticated trends
price_trend = (
df["Close"].iloc[-1] - df["Close"].iloc[-20]
) / 20 # Daily price change
volume_trend = (
df["Volume"].iloc[-1] - df["Volume"].iloc[-20]
) / 20 # Daily volume change
# Calculate volatility for more realistic projections
price_volatility = df["Close"].pct_change().std()
volume_volatility = df["Volume"].pct_change().std()
for i, date in enumerate(future_trading_dates):
# Add some randomness to make predictions more realistic
# Use a smaller random component to avoid extreme outliers
random_factor = np.random.normal(0, price_volatility * 0.1)
# Project prices forward using the trend with some randomness
projected_price = (
df["Close"].iloc[-1] + (price_trend * (i + 1)) + random_factor
)
# Ensure projected price doesn't go negative
projected_price = max(projected_price, df["Close"].iloc[-1] * 0.5)
# Update the price arrays for calculating moving averages
if i < 20:
# For first 20 days, use historical data + projected
current_20_prices = np.append(
last_20_prices[-(20 - i - 1) :], [projected_price] * (i + 1)
)
else:
# After 20 days, use only projected prices
current_20_prices = np.array([projected_price] * 20)
if i < 50:
# For first 50 days, use historical data + projected
current_50_prices = np.append(
last_50_prices[-(50 - i - 1) :], [projected_price] * (i + 1)
)
else:
# After 50 days, use only projected prices
current_50_prices = np.array([projected_price] * 50)
# Calculate projected technical indicators
sma_20 = np.mean(current_20_prices)
sma_50 = np.mean(current_50_prices)
# Project volume with some randomness
volume_random_factor = np.random.normal(0, volume_volatility * 0.1)
projected_volume = (
df["Volume"].iloc[-1] + (volume_trend * (i + 1)) + volume_random_factor
)
projected_volume = max(
projected_volume, df["Volume"].iloc[-1] * 0.3
) # Don't go too low
volume_avg = np.mean(
np.append(
last_volumes[-(20 - i - 1) :], [projected_volume] * min(i + 1, 20)
)
)
# Add some variation to RSI and MACD instead of keeping them constant
# RSI typically oscillates between 30-70, so add small random changes
rsi_variation = np.random.normal(0, 2) # Small random change
new_rsi = last_values["RSI"] + rsi_variation
new_rsi = max(10, min(90, new_rsi)) # Keep RSI in reasonable bounds
# MACD variation
macd_variation = np.random.normal(0, abs(last_values["MACD"]) * 0.1)
new_macd = last_values["MACD"] + macd_variation
new_macd_signal = last_values["MACD_Signal"] + macd_variation * 0.5
# Bollinger Band variation
bb_variation = np.random.normal(0, last_values["BB_StdDev"] * 0.1)
new_bb_std = last_values["BB_StdDev"] + bb_variation
new_bb_std = max(
new_bb_std, last_values["BB_StdDev"] * 0.5
) # Don't go too low
# Calculate additional features for future predictions
# Use the last known values and add small variations
new_stochastic_k = last_values.get("Stochastic_K", 50) + np.random.normal(
0, 5
)
new_stochastic_k = max(0, min(100, new_stochastic_k))
new_stochastic_d = last_values.get("Stochastic_D", 50) + np.random.normal(
0, 5
)
new_stochastic_d = max(0, min(100, new_stochastic_d))
new_williams_r = last_values.get("Williams_R", -50) + np.random.normal(0, 5)
new_williams_r = max(-100, min(0, new_williams_r))
new_cci = last_values.get("CCI", 0) + np.random.normal(0, 20)
# Calculate BB position and squeeze
bb_upper = sma_20 + (new_bb_std * 2)
bb_lower = sma_20 - (new_bb_std * 2)
bb_position = (
(projected_price - bb_lower) / (bb_upper - bb_lower)
if (bb_upper - bb_lower) > 0
else 0.5
)
bb_squeeze = (bb_upper - bb_lower) / sma_20 if sma_20 > 0 else 0
# Price changes
price_change = (projected_price - df["Close"].iloc[-1]) / df["Close"].iloc[
-1
]
price_change_5d = price_change * 0.8 # Approximate
price_change_20d = price_change * 0.6 # Approximate
# Volume changes
volume_change = (projected_volume - df["Volume"].iloc[-1]) / df[
"Volume"
].iloc[-1]
volume_price_trend = projected_volume * price_change
volume_sma_ratio = projected_volume / volume_avg if volume_avg > 0 else 1
# Moving average crossovers
sma_10 = (
np.mean(current_20_prices[-10:])
if len(current_20_prices) >= 10
else sma_20
)
sma_200 = sma_50 # Approximate for future
sma_10_20_cross = 1 if sma_10 > sma_20 else 0
sma_20_50_cross = 1 if sma_20 > sma_50 else 0
sma_50_200_cross = 1 if sma_50 > sma_200 else 0
# Support and resistance
resistance_20d = projected_price * 1.05 # Approximate
support_20d = projected_price * 0.95 # Approximate
price_to_resistance = projected_price / resistance_20d
price_to_support = projected_price / support_20d
# Time-based features (use the actual future date)
day_of_week = date.weekday()
month = date.month
quarter = (month - 1) // 3 + 1
is_month_end = 1 if date.day >= 25 else 0 # Approximate
is_quarter_end = 1 if month in [3, 6, 9, 12] and date.day >= 25 else 0
# Market sentiment
price_above_sma200 = 1 if projected_price > sma_200 else 0
volume_spike = 1 if projected_volume > volume_avg * 1.5 else 0
price_spike = 1 if abs(price_change) > price_volatility * 2 else 0
future_row = {
"SMA_10": sma_10,
"SMA_20": sma_20,
"SMA_50": sma_50,
"SMA_200": sma_200,
"RSI": new_rsi,
"MACD": new_macd,
"MACD_Signal": new_macd_signal,
"BB_StdDev": new_bb_std,
"BB_Position": bb_position,
"BB_Squeeze": bb_squeeze,
"Stochastic_K": new_stochastic_k,
"Stochastic_D": new_stochastic_d,
"Williams_R": new_williams_r,
"CCI": new_cci,
"Price_Change": price_change,
"Price_Change_5d": price_change_5d,
"Price_Change_20d": price_change_20d,
"Price_Volatility": price_volatility,
"Price_Range": abs(price_change) * 0.02, # Approximate
"Volume_Change": volume_change,
"Volume_Price_Trend": volume_price_trend,
"Volume_SMA_Ratio": volume_sma_ratio,
"Volume_StdDev": volume_volatility,
"SMA_10_20_Cross": sma_10_20_cross,
"SMA_20_50_Cross": sma_20_50_cross,
"SMA_50_200_Cross": sma_50_200_cross,
"Price_to_Resistance": price_to_resistance,
"Price_to_Support": price_to_support,
"Day_of_Week": day_of_week,
"Month": month,
"Quarter": quarter,
"Is_Month_End": is_month_end,
"Is_Quarter_End": is_quarter_end,
"Price_Above_SMA200": price_above_sma200,
"Volume_Spike": volume_spike,
"Price_Spike": price_spike,
"Volume_Avg": volume_avg,
}
future_features.append(future_row)
# Create X_future AFTER future_features is populated
X_future = pd.DataFrame(future_features)
X_future_scaled = scaler.transform(X_future)
# Make predictions for the next 30 trading days
future_predictions = model.predict(X_future_scaled)
# Create interactive chart with historical data and future predictions
fig = go.Figure()
# Filter data to show only the last 1 year for display
one_year_ago = last_date - timedelta(days=365)
df_display = df[df["Date"] >= one_year_ago]
# Add historical price data (last 1 year only)
fig.add_trace(
go.Scatter(
x=df_display["Date"],
y=df_display["Close"],
mode="lines+markers",
name=f"{ticker} Historical Price (Last Year)",
line=dict(color="#1f77b4", width=2),
marker=dict(size=4),
)
)
# Add future predictions
fig.add_trace(
go.Scatter(
x=future_trading_dates,
y=future_predictions,
mode="lines+markers",
name=f"{ticker} Future Predictions (Next 30 Days)",
line=dict(color="#ff7f0e", width=2, dash="dash"),
marker=dict(size=4),
)
)
# Update layout
fig.update_layout(
title=f"{ticker} Stock Price with Next 30-Day Ridge Regression Predictions",
xaxis_title="Date",
yaxis_title="Price ($)",
height=500,
hovermode="x unified",
legend=dict(
orientation="h", yanchor="bottom", y=1.02, xanchor="right", x=1
),
)
# Update axes
fig.update_xaxes(
title_text="Date",
tickformat="%b %d",
tickangle=45,
)
fig.update_yaxes(title_text="Price ($)")
# Display prediction summary
current_price = df["Close"].iloc[-1]
predicted_price_30d = (
future_predictions[-1] if len(future_predictions) > 0 else current_price
)
price_change = predicted_price_30d - current_price
price_change_pct = (price_change / current_price) * 100
# Calculate model performance on historical data (for reference)
y_pred_historical = model.predict(
X_train_scaled
) # Use scaled data for historical fit
r2_historical = r2_score(y_train, y_pred_historical)
mse_historical = mean_squared_error(y_train, y_pred_historical)
# Display detailed prediction information
col1, col2, col3 = st.columns([1, 1, 1])
with col1:
st.metric(
"Current Price",
f"${current_price:.2f}",
)
with col2:
st.metric(
"30-Day Prediction",
f"${predicted_price_30d:.2f}",
delta=f"{price_change_pct:+.2f}%",
)
with col3:
st.metric(
"Expected Change",
f"${price_change:.2f} ({price_change_pct:+.2f}%)",
)
# Additional prediction details
st.info(
f"""
**π 30-Day Ridge Regression Prediction for {ticker}:**
- **Model Performance (Historical Fit):**
- RΒ² Score: {r2_historical:.4f} ({r2_historical*100:.2f}% accuracy)
- Mean Squared Error: {mse_historical:.4f}
- Best Alpha (Regularization): {best_alpha}
- Cross-Validation Score: {best_score:.4f}
- **Model Training Time:** {training_time:.2f}s
- **Training Data:** 5 years of historical data
β οΈ **Disclaimer**: Stock predictions have approximately 70% accuracy.
These forecasts are for informational purposes only and should not be used as
the sole basis for investment decisions. Always conduct your own research
and consider consulting with financial advisors.
"""
)
return fig
except Exception as e:
st.error(f"Error creating chart for {ticker}: {e}")
return create_basic_stock_chart(ticker)
@st.cache_data(ttl=3600) # Cache for 1 hour
def create_basic_stock_chart(ticker: str):
"""Create a basic stock price chart without Prophet predictions."""
try:
# Get stock data with loading state
with st.spinner(f"π Fetching basic stock data for {ticker}..."):
stock = yf.Ticker(ticker)
hist_data = stock.history(period="30d")
if hist_data.empty:
st.warning(f"No data available for {ticker}")
return None
# Create simple line chart
fig = go.Figure()
# Add price line chart
fig.add_trace(
go.Scatter(
x=hist_data.index,
y=hist_data["Close"],
mode="lines+markers",
name=f"{ticker} Price",
line=dict(color="#1f77b4", width=2),
marker=dict(size=4),
)
)
# Update layout
fig.update_layout(
title=f"{ticker} Stock Price (30 Days)",
xaxis_title="Date",
yaxis_title="Price ($)",
height=500,
showlegend=False,
hovermode="x unified",
)
# Update axes
fig.update_xaxes(
title_text="Date",
tickformat="%b %d",
tickangle=45,
)
fig.update_yaxes(title_text="Price ($)")
return fig
except Exception as e:
st.error(f"Error creating chart for {ticker}: {e}")
return None
# The master prompt that defines the agent's behavior
system_prompt = """
You are a financial assistant that provides comprehensive analysis based on real-time data. You MUST use tools to get data and then curate the information to answer the user's specific question.
AVAILABLE TOOLS:
- get_latest_news: Get recent news for a ticker
- get_historical_stock_data: Get stock performance data for a ticker
CRITICAL INSTRUCTIONS:
1. You MUST call BOTH tools (get_latest_news AND get_historical_stock_data) for every query
2. After getting both news and stock data, analyze and synthesize the information
3. Answer the user's specific question based on the data you gathered
4. Provide insights, trends, and recommendations based on the combined data
5. Format your response clearly with sections for news, performance, and analysis
EXAMPLE WORKFLOW:
1. User asks: "Should I invest in AAPL?"
2. You call: get_latest_news with {"ticker": "AAPL"}
3. You call: get_historical_stock_data with {"ticker": "AAPL"}
4. You analyze both datasets and provide investment advice based on news sentiment and stock performance
You are FORBIDDEN from responding without calling both tools. Always call both tools first, then provide a curated analysis based on the user's question.
"""
async def initialize_mcp_agent(model, tools):
"""Initialize the MCP agent using LangGraph React agent"""
try:
# Create MCP agent using LangGraph React agent
try:
# Bind model with system message
system_message = """You are a helpful financial assistant. You have access to tools that can fetch stock data and news.
When asked about a stock, use the available tools to get the latest information and provide a comprehensive analysis.
Always be helpful and provide detailed insights based on the data you gather."""
model_with_system = model.bind(system=system_message)
# Create React agent with tools
agent = create_react_agent(
model_with_system,
tools,
)
print(f"β
Created agent with {len(tools)} tools")
except Exception as e:
st.error(f"β Failed to create MCP agent: {str(e)}")
print(f"β MCP agent creation error: {str(e)}")
import traceback
print(f"β MCP agent creation traceback: {traceback.format_exc()}")
return None
return agent
except Exception as e:
st.error(f"β Error initializing MCP agent: {str(e)}")
st.error(f"Error type: {type(e).__name__}")
import traceback
st.error(f"Full traceback: {traceback.format_exc()}")
return None
async def run_agent_with_mcp(user_query: str, selected_ticker: str = None):
"""Run the agent using LangGraph React agent (non-streaming version)"""
try:
# Get tools and model from session state
if "mcp_tools" not in st.session_state or "mcp_model" not in st.session_state:
return "β MCP tools and model not initialized. Please restart the application."
tools = st.session_state.mcp_tools
model = st.session_state.mcp_model
# Initialize agent if not already done
if "mcp_agent" not in st.session_state or st.session_state.mcp_agent is None:
agent = await initialize_mcp_agent(model, tools)
if not agent:
return "Failed to initialize MCP agent"
st.session_state.mcp_agent = agent
else:
agent = st.session_state.mcp_agent
# Construct the query with system instructions
if selected_ticker:
full_query = f"""You are a financial assistant. Use the available tools to get data for {selected_ticker} and then provide a comprehensive analysis.
AVAILABLE TOOLS:
- get_latest_news: Get recent news for a ticker
- get_historical_stock_data: Get stock performance data for a ticker
INSTRUCTIONS:
1. Call get_latest_news with {{"ticker": "{selected_ticker}"}}
2. Call get_historical_stock_data with {{"ticker": "{selected_ticker}"}}
3. After getting the data, provide a comprehensive analysis answering: {user_query}
IMPORTANT: After calling the tools, you MUST provide a final analysis with insights, trends, and recommendations. Do not just show the tool calls.
Question: {user_query} for {selected_ticker}"""
else:
full_query = user_query
# Run the agent with LangGraph
with st.spinner("π€ Processing with MCP agent..."):
try:
# Use LangGraph ainvoke method for async tools
result = await agent.ainvoke(
{"messages": [{"role": "user", "content": full_query}]}
)
# Debug: Print the raw result to see what we're getting
print("π Raw result type:", type(result))
print("π Raw result:", result)
# Try to extract AIMessages from structured data first
final_response = ""
if isinstance(result, dict) and "messages" in result:
# Work with structured data directly
messages = result["messages"]
print(f"π Found {len(messages)} messages in structured data")
# Filter for AIMessage instances
ai_messages = []
for msg in messages:
if (
hasattr(msg, "__class__")
and msg.__class__.__name__ == "AIMessage"
):
ai_messages.append(msg)
elif isinstance(msg, dict) and msg.get("type") == "AIMessage":
ai_messages.append(msg)
print(f"π Found {len(ai_messages)} AIMessages in structured data")
if len(ai_messages) >= 2:
# Get the second AIMessage (comprehensive analysis)
second_ai_message = ai_messages[1]
if hasattr(second_ai_message, "content"):
final_response = second_ai_message.content
elif isinstance(second_ai_message, dict):
final_response = second_ai_message.get("content", "")
print(f"π Selected message 2 (the comprehensive analysis)")
print(f"π Selected message starts with: {final_response[:50]}")
elif len(ai_messages) == 1:
# Fallback to first message if only one found
first_ai_message = ai_messages[0]
if hasattr(first_ai_message, "content"):
final_response = first_ai_message.content
elif isinstance(first_ai_message, dict):
final_response = first_ai_message.get("content", "")
print(f"π Selected message 1 (only one found)")
print(f"π Selected message starts with: {final_response[:50]}")
else:
# Fallback to string processing if no structured AIMessages found
if isinstance(result, dict) and "output" in result:
final_response = result["output"]
elif isinstance(result, str):
final_response = result
else:
final_response = str(result)
# Clean up the final response to remove escaped characters
final_response = (
final_response.replace("\\n", "\n")
.replace("\\'", "'")
.replace('\\"', '"')
)
# Try regex extraction as fallback
if "AIMessage" in final_response:
# Look for AIMessages in JSON format - multiple patterns to catch different formats
ai_messages = []
# Pattern 1: AIMessage with content in double quotes
ai_messages.extend(
re.findall(
r'AIMessage\(content="([^"]*)"',
final_response,
re.DOTALL,
)
)
# Pattern 2: AIMessage with content in single quotes
if not ai_messages:
ai_messages.extend(
re.findall(
r"AIMessage\(content='([^']*)'",
final_response,
re.DOTALL,
)
)
# Pattern 3: AIMessage in JSON format with "content" field
if not ai_messages:
ai_messages.extend(
re.findall(
r'"content":\s*"([^"]*)"',
final_response,
re.DOTALL,
)
)
# Pattern 4: AIMessage in JSON format with 'content' field
if not ai_messages:
ai_messages.extend(
re.findall(
r"'content':\s*'([^']*)'",
final_response,
re.DOTALL,
)
)
if ai_messages:
print(
f"π Found {len(ai_messages)} AIMessages via regex"
)
for i, msg in enumerate(ai_messages):
print(
f"π Message {i+1} length: {len(msg.strip())}"
)
print(
f"π Message {i+1} preview: {msg.strip()[:100]}..."
)
print(
f"π Message {i+1} starts with: {msg.strip()[:20]}"
)
# Select the second AIMessage (index 1) which contains the comprehensive analysis
# The first AIMessage is usually just the tool-calling message
if len(ai_messages) >= 2:
final_response = ai_messages[
1
].strip() # Get the 2nd message
print(
f"π Selected message 2 (the comprehensive analysis)"
)
print(
f"π Selected message starts with: {final_response[:50]}"
)
else:
# Fallback to last message if only one found
final_response = ai_messages[-1].strip()
print(
f"π Selected message {len(ai_messages)} (the last one)"
)
print(
f"π Selected message starts with: {final_response[:50]}"
)
else:
# If no AIMessage found, try to extract from the raw response
final_response = final_response.strip()
else:
# Fallback for non-structured data
if isinstance(result, dict) and "output" in result:
final_response = result["output"]
elif isinstance(result, str):
final_response = result
else:
final_response = str(result)
# Clean up the final response to remove escaped characters
final_response = (
final_response.replace("\\n", "\n")
.replace("\\'", "'")
.replace('\\"', '"')
)
# Remove any remaining tool call artifacts
final_response = re.sub(r"<\|.*?\|>", "", final_response)
final_response = re.sub(
r"functions\.[a-zA-Z_]+:\d+", "", final_response
)
final_response = re.sub(r'\{[^{}]*"ticker"[^{}]*\}', "", final_response)
# Remove LangGraph metadata
final_response = re.sub(
r"\{.*?agent.*?\}", "", final_response, flags=re.DOTALL
)
final_response = re.sub(
r"\{.*?tools.*?\}", "", final_response, flags=re.DOTALL
)
final_response = re.sub(
r"ToolMessage.*?\]", "", final_response, flags=re.DOTALL
)
final_response = re.sub(
r"additional_kwargs.*?usage_metadata.*?\}",
"",
final_response,
flags=re.DOTALL,
)
# Clean up extra whitespace and formatting
final_response = re.sub(r"\n\s*\n", "\n\n", final_response)
final_response = final_response.strip()
print("π Final cleaned response:", final_response)
return final_response
except Exception as e:
st.error(f"β Error during agent execution: {str(e)}")
return f"Error during execution: {str(e)}"
except Exception as e:
st.error(f"β Error running MCP agent: {e}")
return f"Error: {e}"
@st.cache_data(ttl=1800) # Cache for 30 minutes
def display_top_news(ticker: str):
"""Display top news headlines for the given ticker with clickable links."""
try:
# Check if news is already cached
news_cache_key = f"news_data_{ticker}"
if news_cache_key in st.session_state:
articles = st.session_state[news_cache_key]
else:
# Get news data with loading state
with st.spinner(f"π° Loading news for {ticker}..."):
google_news = gnews.GNews(language="en", country="US", period="7d")
search_query = f'"{ticker}" stock market news'
articles = google_news.get_news(search_query)
# Cache the articles
st.session_state[news_cache_key] = articles
if not articles:
st.info(f"No recent news found for {ticker}")
return
# Display top 5 articles
for i, article in enumerate(articles[:5], 1):
# Clean the title text
title = article.get("title", "")
if title:
soup = BeautifulSoup(title, "html.parser")
title = soup.get_text().strip()
url = article.get("url", "")
publisher = article.get("publisher", {}).get("title", "Unknown Source")
# Create a clickable link
if url:
st.markdown(f"[{title}]({url})")
st.caption(f"Source: {publisher}")
else:
st.markdown(f"{title}")
st.caption(f"Source: {publisher}")
# Add some spacing between articles
if i < 5:
st.markdown("---")
except Exception as e:
st.error(f"Error fetching news for {ticker}: {e}")
def is_trading_day(date):
"""Check if a date is a trading day (not weekend or holiday)."""
# Check if it's a weekend
if date.weekday() >= 5: # Saturday = 5, Sunday = 6
return False
# Check if it's a US market holiday
us_holidays = holidays.US()
if date in us_holidays:
return False
return True
def get_next_trading_days(start_date, num_days):
"""Get the next N trading days starting from start_date."""
trading_days = []
current_date = start_date
while len(trading_days) < num_days:
if is_trading_day(current_date):
trading_days.append(current_date)
current_date += timedelta(days=1)
return trading_days
def create_trading_day_future_dataframe(model, periods=30, freq="D"):
"""Create a future dataframe with only trading days."""
# Get the last date from the training data
last_date = model.history["ds"].max()
# Generate trading days
trading_days = []
current_date = last_date + timedelta(days=1)
while len(trading_days) < periods:
if is_trading_day(current_date):
trading_days.append(current_date)
current_date += timedelta(days=1)
# Create future dataframe with only trading days
future_df = pd.DataFrame({"ds": trading_days})
return future_df
def main():
st.set_page_config(page_title="QueryStockAI", page_icon="π", layout="wide")
st.title("π QueryStockAI")
st.markdown(
"Get comprehensive financial analysis and insights for your selected stocks."
)
# Initialize MCP client and tools silently
try:
# Initialize MCP client with proper configuration
if MultiServerMCPClient is None:
st.error(
"β MultiServerMCPClient not available. Please install langchain-mcp-adapters"
)
st.stop()
try:
# Pass servers configuration as positional argument
client = MultiServerMCPClient(
{
"news_server": {
"url": "http://localhost:8002/mcp",
"transport": "streamable_http",
},
"stock_server": {
"url": "http://localhost:8001/mcp",
"transport": "streamable_http",
},
}
)
except Exception as e:
# Try with different transport type
try:
client = MultiServerMCPClient(
{
"news_server": {
"url": "http://localhost:8002/mcp",
"transport": "http",
},
"stock_server": {
"url": "http://localhost:8001/mcp",
"transport": "http",
},
}
)
except Exception as e2:
st.error(f"β Failed to initialize MCP client: {str(e2)}")
st.stop()
# Get tools from MCP servers
tools = asyncio.run(client.get_tools())
# Create model with proper configuration
model = ChatGroq(model=model_name, temperature=0.1, max_tokens=4096)
# Store tools and model in session state for later use
st.session_state.mcp_tools = tools
st.session_state.mcp_model = model
st.session_state.mcp_client = client
except Exception as e:
st.error(f"β Failed to initialize MCP client: {str(e)}")
st.stop()
# Available tickers
with st.spinner("π Loading available tickers..."):
available_tickers = get_available_tickers()
# Sidebar for ticker selection
st.sidebar.header("π Stock Selection")
st.sidebar.subheader("π Popular Stocks")
# Only show selectbox if tickers are loaded
if available_tickers and len(available_tickers) > 0:
selected_ticker = st.sidebar.selectbox(
"Choose a stock ticker:",
options=list(available_tickers.keys()),
format_func=lambda x: f"{x} - {available_tickers[x]}",
index=None,
placeholder="Select a ticker...",
)
else:
st.sidebar.error("β Failed to load tickers. Please refresh the page.")
selected_ticker = None
# Add search functionality
st.sidebar.subheader("π Search Custom Ticker")
custom_ticker = st.sidebar.text_input(
"Enter ticker symbol, if not found in above dropdown (e.g., AAPL, TSLA):",
placeholder="Enter ticker symbol...",
key="custom_ticker_input",
)
# Add info button with helpful information
if custom_ticker:
custom_ticker = custom_ticker.upper().strip()
if custom_ticker:
# Search for the custom ticker
company_name = search_ticker(custom_ticker)
if company_name:
st.sidebar.success(
f"β
Found: {custom_ticker} - {company_name} -> Added to dropdown list above."
)
# Add to available tickers temporarily
available_tickers[custom_ticker] = company_name
else:
st.sidebar.error(f"β Could not find ticker: {custom_ticker}")
# Clear cache when ticker changes
if (
"current_ticker" in st.session_state
and st.session_state.current_ticker != selected_ticker
):
# Clear all cached data for the previous ticker
for key in list(st.session_state.keys()):
if key.startswith("chart_") or key.startswith("news_"):
del st.session_state[key]
# Clear chat history when ticker changes
if "messages" in st.session_state:
del st.session_state.messages
# Update current ticker
if selected_ticker:
st.session_state.current_ticker = selected_ticker
# Main content area
if not selected_ticker:
st.info(
"π Please select a stock ticker from the sidebar to view the chart and start chatting."
)
st.markdown(
"""
**How to use:**
1. Select a stock ticker from the sidebar
2. View the interactive stock price chart
3. Ask questions about the stock's performance, news, or investment advice
4. The agent will fetch real-time data and provide comprehensive analysis
**Example questions:**
- "How is this stock performing?"
- "What's the latest news about this company?"
- "Should I invest in this stock?"
- "What are the recent trends?"
"""
)
else:
st.success(
f"β
Selected: {selected_ticker} - {available_tickers[selected_ticker]}"
)
# Add loading state for initial page load
if "page_loaded" not in st.session_state:
with st.spinner("π Loading application..."):
st.session_state.page_loaded = True
# Stock Chart and News Section
st.header("π Stock Analysis")
# Create two columns for chart and news
col1, col2 = st.columns([2, 1])
with col1:
st.subheader("π Stock Price Chart")
# Always create the chart - it's cached by the function itself
chart_fig = create_stock_chart(selected_ticker)
if chart_fig:
st.plotly_chart(chart_fig, use_container_width=True)
else:
st.warning(f"Could not load chart for {selected_ticker}")
with col2:
st.subheader("π° Top News")
# Display news - it's cached by the function
display_top_news(selected_ticker)
# Chat Section
st.header("π¬ Chat with Financial Agent")
# Initialize chat history
if "messages" not in st.session_state:
st.session_state.messages = []
# Display existing chat messages using custom styling
for message in st.session_state.messages:
if message["role"] == "user":
st.markdown(
f"""
<div style="padding: 10px; border-radius: 10px; margin: 5px 0; border: 1px solid #bbdefb;">
<strong>You:</strong> {message["content"]}
</div>
""",
unsafe_allow_html=True,
)
else:
st.markdown(
f"""
<div style="padding: 10px; border-radius: 10px; margin: 5px 0; border: 1px solid #e0e0e0;">
<strong>Agent:</strong>
</div>
""",
unsafe_allow_html=True,
)
# Render the content as markdown for proper formatting with controlled text size
st.markdown(
f"""
<div style="font-size: 13px; line-height: 1.4; padding: 12px; border-radius: 8px; margin: 5px 0; border-left: 4px solid #007bff; max-height: 400px; overflow-y: auto;"
""",
unsafe_allow_html=True,
)
# Clean up the content to remove raw markdown syntax
cleaned_content = (
message["content"].replace("\\n", "\n").replace("\\'", "'")
)
st.markdown(cleaned_content)
st.markdown("</div>", unsafe_allow_html=True)
# Chat input with proper loading state
if prompt := st.chat_input(f"Ask about {selected_ticker}...", key="chat_input"):
# Add user message to chat history
st.session_state.messages.append({"role": "user", "content": prompt})
# Display assistant response with spinner above input
with st.spinner("π€ Analyzing your request..."):
response = asyncio.run(run_agent_with_mcp(prompt, selected_ticker))
# Ensure only a string is appended to chat history
if isinstance(response, dict) and "content" in response:
clean_response = response["content"]
elif (
isinstance(response, list)
and len(response) > 0
and "content" in response[-1]
):
clean_response = response[-1]["content"]
else:
clean_response = str(response)
st.session_state.messages.append(
{"role": "assistant", "content": clean_response}
)
# Rerun to display the new message - the chart and news are cached in session state
st.rerun()
if __name__ == "__main__":
main()
|