Merge pull request #4 from stephenhky/develop

deployed LPPL bubble crash prediction

finportbotutil/syminfo.py

@@ -80,3 +80,17 @@ async def get_ma_plots_info(symbol, startdate, enddate, dayswindow, api_url, tit
 
     response = requests.request("GET", api_url, headers=headers, data=payload)
     return json.loads(response.text)
+
+
+async def fit_lppl(symbol, startdate, enddate, api_url):
+    payload = json.dumps({
+        'symbol': symbol,
+        'startdate': startdate,
+        'enddate': enddate
+    })
+    headers = {
+        'Content-Type': 'application/json'
+    }
+
+    response = requests.request("GET", api_url, headers=headers, data=payload)
+    return json.loads(response.text)

main.py

@@ -3,7 +3,7 @@
 import re
 import os
 import logging
-from datetime import date, datetime
+from datetime import date, datetime, timedelta
 from dateutil.relativedelta import relativedelta
 import asyncio
 import urllib
@@ -17,7 +17,7 @@
 import boto3
 
 from finportbotutil.tipcalc import calculate_tips
-from finportbotutil.syminfo import get_symbol_inference, get_symbols_correlation, get_plots_infos, search_symbols, get_ma_plots_info
+from finportbotutil.syminfo import get_symbol_inference, get_symbols_correlation, get_plots_infos, search_symbols, get_ma_plots_info, fit_lppl
 
 logging.basicConfig(level=logging.INFO)
 
@@ -40,14 +40,17 @@
 
 # Search API
 search_api_url = os.getenv('SEARCH')
-modelloadretry = int(os.getenv('MODELLOADRETRY', 5))
+modelloadretry = int(os.getenv('SEARCHMODELLOADRETRY', 5))
 
 # Add or Modify User ARN
 addmodifyuser_arn = os.environ.get('ADDUSERARN')
 
 # Symbol Search Wrapper ARN
 searchwrappwer_arn = os.environ.get('SEARCHWRAPPERARN')
 
+# fit LPPL model URL
+fit_lppl_url = os.environ.get('FITLPPL')
+
 # commands
 CMD_START = ['start']
 CMD_HELP = ['help']
@@ -64,6 +67,7 @@
 CMD_NASDAQ_MA = ['nasdaqma']
 CMD_DJI_MA = ['djima']
 CMD_MAPLOT = ['maplot']
+CMD_FITLPPL = ['predictcrash']
 
 
 # polling flag
@@ -483,6 +487,31 @@ def handle_maplot_callback_query(call):
         logging.error('Unknown error!')
         print('Unknown error!', file=sys.stderr)
 
+@bot.message_handler(commands=CMD_FITLPPL)
+def fit_lppl_bubble_burst(message):
+    logging.info(message)
+    print(message)
+
+    stringlists = re.sub('\s+', ' ', message.text).split(' ')[1:]
+    symbol = '^GSPC' if len(stringlists) < 1 else stringlists[0]
+    startdate = (datetime.today() - timedelta(days=365)).strftime('%Y-%m-%d') if len(stringlists) < 2 else stringlists[1]
+    enddate = datetime.today().strftime('%Y-%m-%d') if len(stringlists) < 3 else stringlists[2]
+
+    results = asyncio.run(fit_lppl(symbol, startdate, enddate, fit_lppl_url))
+    logging.info(results)
+    print(results)
+
+    if 'message' in results.keys() and results['message'] == 'Internal server error':
+        message_text = 'API internal error.'
+        bot.reply_to(message, message_text)
+        return {'message': message_text}
+    else:
+        message_text = open(os.path.join('messagetemplates', 'crash.txt')).read().format(
+            crashdate=results['estimated_crash_date']
+        )
+        bot.reply_to(message, message_text)
+        return {'message': message_text,}
+
 
 def lambda_handler(event, context):
     message = json.loads(event['body'])

messagetemplates/crash.txt

@@ -0,0 +1,3 @@
+From the model and the information given, the projected crash date is {crashdate}.
+
+(Disclaimer: prediction can change due to stochasticity and other events. The estimation is by no means accurate.)