tulipy_ind support, multioutput scale and gzip cache support #106 #115 #114 #112 #107

v2realbot/controller/services.py

@@ -14,7 +14,7 @@ from v2realbot.common.PrescribedTradeModel import Trade, TradeDirection, TradeSt
 from datetime import datetime
 from v2realbot.loader.trade_offline_streamer import Trade_Offline_Streamer
 from threading import Thread, current_thread, Event, enumerate
-from v2realbot.config import STRATVARS_UNCHANGEABLES, ACCOUNT1_PAPER_API_KEY, ACCOUNT1_PAPER_SECRET_KEY, ACCOUNT1_LIVE_API_KEY, ACCOUNT1_LIVE_SECRET_KEY, DATA_DIR,BT_FILL_CONS_TRADES_REQUIRED,BT_FILL_LOG_SURROUNDING_TRADES,BT_FILL_CONDITION_BUY_LIMIT,BT_FILL_CONDITION_SELL_LIMIT, GROUP_TRADES_WITH_TIMESTAMP_LESS_THAN, MEDIA_DIRECTORY, RUNNER_DETAIL_DIRECTORY
+from v2realbot.config import STRATVARS_UNCHANGEABLES, ACCOUNT1_PAPER_API_KEY, ACCOUNT1_PAPER_SECRET_KEY, ACCOUNT1_LIVE_API_KEY, ACCOUNT1_LIVE_SECRET_KEY, DATA_DIR,BT_FILL_CONS_TRADES_REQUIRED,BT_FILL_LOG_SURROUNDING_TRADES,BT_FILL_CONDITION_BUY_LIMIT,BT_FILL_CONDITION_SELL_LIMIT, GROUP_TRADES_WITH_TIMESTAMP_LESS_THAN, MEDIA_DIRECTORY, RUNNER_DETAIL_DIRECTORY, OFFLINE_MODE
 import importlib
 from alpaca.trading.requests import GetCalendarRequest
 from alpaca.trading.client import TradingClient
@@ -1609,6 +1609,11 @@ def preview_indicator_byTOML(id: UUID, indicator: InstantIndicator, save: bool =
 
 
             #print("Done", state.indicators[indicator.name])
+
+            #print pro multioutput
+            for ind_name in returns:
+                print("Done",ind_name, state.indicators[indicator.name])                
+
             output_dict = {}
             new_inds[indicator.name] = state.indicators[indicator.name]
             
@@ -1664,7 +1669,7 @@ def preview_indicator_byTOML(id: UUID, indicator: InstantIndicator, save: bool =
                 except Exception as e:
                     print(str(e) + format_exc())
 
-            #print("Done", state.indicators[indicator.name])
+            #print("Done", state.cbar_indicators[indicator.name])
             output_dict = {}
             new_tick_inds[indicator.name] = state.cbar_indicators[indicator.name]
             
@@ -1884,8 +1889,10 @@ def get_alpaca_history_bars(symbol: str, datetime_object_from: datetime, datetim
         return 0, result
     except Exception as e:
         print(str(e) + format_exc())
+        if OFFLINE_MODE:
+            print("OFFLINE MODE ENABLED")
+            return 0, []
         return -2, str(e)    
-
 # change_archived_runner
 # delete_archived_runner_details
 

v2realbot/loader/aggregator.py

@@ -3,7 +3,7 @@
 """
 from v2realbot.enums.enums import RecordType, StartBarAlign
 from datetime import datetime, timedelta
-from v2realbot.utils.utils import parse_alpaca_timestamp, ltp, Queue,is_open_hours,zoneNY
+from v2realbot.utils.utils import parse_alpaca_timestamp, ltp, Queue,is_open_hours,zoneNY, zoneUTC
 from queue import Queue
 from rich import print
 from v2realbot.enums.enums import Mode
@@ -14,6 +14,7 @@ import os
 from v2realbot.config import DATA_DIR, GROUP_TRADES_WITH_TIMESTAMP_LESS_THAN, AGG_EXCLUDED_TRADES
 import pickle
 import dill
+import gzip
 
 class TradeAggregator:  
     def __init__(self,
@@ -149,7 +150,7 @@ class TradeAggregator:
         # else:
         data['t'] = parse_alpaca_timestamp(data['t'])
 
-        if not is_open_hours(datetime.fromtimestamp(data['t'])) and self.exthours is False:
+        if not is_open_hours(datetime.fromtimestamp(data['t'], tz=zoneUTC)) and self.exthours is False:
             #print("AGG: trade not in open hours skipping", datetime.fromtimestamp(data['t']).astimezone(zoneNY))
             return []
 
@@ -442,7 +443,7 @@ class TradeAggregator:
                     "trades": 1,
                     "hlcc4": data['p'],
                     "confirmed": 0,
-                    "time": datetime.fromtimestamp(data['t']),
+                    "time": datetime.fromtimestamp(data['t'], tz=zoneUTC),
                     "updated": data['t'],
                     "vwap": data['p'],
                     "index": self.barindex,
@@ -476,7 +477,7 @@ class TradeAggregator:
                     "trades": 1,
                     "hlcc4":data['p'],
                     "confirmed": 1,
-                    "time": datetime.fromtimestamp(data['t']),
+                    "time": datetime.fromtimestamp(data['t'], tz=zoneUTC),
                     "updated": data['t'],
                     "vwap": data['p'],
                     "index": self.barindex,
@@ -608,7 +609,7 @@ class TradeAggregator:
                     "trades": 1,
                     "hlcc4": data['p'],
                     "confirmed": 0,
-                    "time": datetime.fromtimestamp(data['t']),
+                    "time": datetime.fromtimestamp(data['t'], tz=zoneUTC),
                     "updated": data['t'],
                     "vwap": data['p'],
                     "index": self.barindex,
@@ -642,7 +643,7 @@ class TradeAggregator:
                     "trades": 1,
                     "hlcc4":data['p'],
                     "confirmed": 1,
-                    "time": datetime.fromtimestamp(data['t']),
+                    "time": datetime.fromtimestamp(data['t'], tz=zoneUTC),
                     "updated": data['t'],
                     "vwap": data['p'],
                     "index": self.barindex,
@@ -787,7 +788,7 @@ class TradeAggregator:
                     "trades": 1,
                     "hlcc4": data['p'],
                     "confirmed": 0,
-                    "time": datetime.fromtimestamp(data['t']),
+                    "time": datetime.fromtimestamp(data['t'], tz=zoneUTC),
                     "updated": data['t'],
                     "vwap": data['p'],
                     "index": self.barindex,
@@ -822,7 +823,7 @@ class TradeAggregator:
                     "trades": 1,
                     "hlcc4":data['p'],
                     "confirmed": 1,
-                    "time": datetime.fromtimestamp(data['t']),
+                    "time": datetime.fromtimestamp(data['t'], tz=zoneUTC),
                     "updated": data['t'],
                     "vwap": data['p'],
                     "index": self.barindex,
@@ -898,7 +899,7 @@ class TradeAggregator:
         #a take excludes result = ''.join(self.excludes.sort())
         self.excludes.sort()  # Sorts the list in place
         excludes_str = ''.join(map(str, self.excludes))  # Joins the sorted elements after converting them to strings
-        cache_file = self.__class__.__name__ + '-' + self.symbol + '-' + str(int(date_from.timestamp())) + '-' + str(int(date_to.timestamp())) + '-' + str(self.rectype) + "-" + str(self.resolution) + "-" + str(self.minsize) + "-" + str(self.align) + '-' + str(self.mintick) + str(self.exthours) + excludes_str + '.cache'
+        cache_file = self.__class__.__name__ + '-' + self.symbol + '-' + str(int(date_from.timestamp())) + '-' + str(int(date_to.timestamp())) + '-' + str(self.rectype) + "-" + str(self.resolution) + "-" + str(self.minsize) + "-" + str(self.align) + '-' + str(self.mintick) + str(self.exthours) + excludes_str + '.cache.gz'
         file_path = DATA_DIR + "/aggcache/" + cache_file
         #print(file_path)
         return file_path
@@ -908,7 +909,7 @@ class TradeAggregator:
         file_path = self.populate_file_name(date_from, date_to)
         if self.skip_cache is False and os.path.exists(file_path):
             ##daily aggregated file exists
-            with open (file_path, 'rb') as fp:
+            with gzip.open (file_path, 'rb') as fp:
                 cachedobject = dill.load(fp)
                 print("AGG CACHE loaded ", file_path)
 
@@ -941,7 +942,7 @@ class TradeAggregator:
         
         file_path = self.populate_file_name(self.cache_from, self.cache_to)
 
-        with open(file_path, 'wb') as fp:
+        with gzip.open(file_path, 'wb') as fp:
             dill.dump(self.cached_object, fp)
         print(f"AGG CACHE stored ({num}) :{file_path}")
         print(f"DATES from:{self.cache_from.strftime('%d.%m.%Y %H:%M')} to:{self.cache_to.strftime('%d.%m.%Y %H:%M')}")

v2realbot/loader/trade_offline_streamer.py

@@ -16,6 +16,7 @@ import asyncio
 from msgpack.ext import Timestamp
 from msgpack import packb
 from pandas import to_datetime
+import gzip
 import pickle
 import os
 from rich import print
@@ -197,7 +198,7 @@ class Trade_Offline_Streamer(Thread):
                         stream_main.enable_cache_output(day.open, day.close)
 
             #trade daily file
-            daily_file = str(symbpole[0]) + '-' + str(int(day.open.timestamp())) + '-' + str(int(day.close.timestamp())) + '.cache'
+            daily_file = str(symbpole[0]) + '-' + str(int(day.open.timestamp())) + '-' + str(int(day.close.timestamp())) + '.cache.gz'
             print(daily_file)
             file_path = DATA_DIR + "/tradecache/"+daily_file
         
@@ -207,7 +208,7 @@ class Trade_Offline_Streamer(Thread):
                 #pokud je start_time < trade < end_time 
                     #odesíláme do queue
                     #jinak pass
-                with open (file_path, 'rb') as fp:
+                with gzip.open (file_path, 'rb') as fp:
                     tradesResponse = pickle.load(fp)
                     print("Loading from Trade CACHE", file_path)
             #daily file doesnt exist
@@ -223,7 +224,7 @@ class Trade_Offline_Streamer(Thread):
                     #ic(datetime.now().astimezone(zoneNY))
                     #ic(day.open, day.close)
                 else:
-                    with open(file_path, 'wb') as fp:
+                    with gzip.open(file_path, 'wb') as fp:
                         pickle.dump(tradesResponse, fp)
 
             #zde už máme daily data

v2realbot/static/js/archivechart.js

@@ -387,6 +387,7 @@ function chart_indicators(data, visible, offset) {
         //indicatory
         //console.log("indicatory TOML", stratvars_toml.stratvars.indicators)
         indId = 1
+        var multiOutsCnf = {}
         indicatorList.forEach((indicators, index, array) => {
 
             //var indicators = data.indicators
@@ -440,10 +441,28 @@ function chart_indicators(data, visible, offset) {
                             //       }
                             // } 
 
+                            //pro multioutput childs dotahneme scale z parenta
+                            if (multiOutsCnf.hasOwnProperty(key)) {
+                                scale = multiOutsCnf[key];
+                            }
+
                             //initialize indicator and store reference to array
                             var obj = {name: key, type: index, series: null, cnf:cnf, instant: instant, returns: returns, indId:indId++}
     
-                            //start
+                            //pokud jde o multioutput parenta ukladam scale parenta pro children
+                            //varianty -  scale je jeden, ukladam jako scale pro vsechny parenty
+                            //         -  scale je list - pouzijeme pro kazdy output scale v listu na stejnem indexu jako output
+                            if (returns) {
+                                returns.forEach((returned, index, array) => {
+                                    //
+                                    if (Array.isArray(scale)) {
+                                        multiOutsCnf[returned] = scale[index]
+                                    }
+                                    else {
+                                        multiOutsCnf[returned] = scale
+                                    }
+                                })
+                            }                        //start
                             //console.log(key)
                             //get configuation of indicator to display
                             conf = get_ind_config(key, index)
@@ -602,6 +621,11 @@ function chart_indicators(data, visible, offset) {
                                 active = true
                             }
                             else {active = false}
+                            
+                            //pro main s multioutputem nezobrazujeme
+                            if (returns) {
+                                active = false
+                            }
                             //add options
                             obj.series.applyOptions({
                                 visible: active?active:visible,

v2realbot/static/js/utils/utils.js

@@ -702,7 +702,8 @@ function create_multioutput_button(item, def, active) {
   multiOutEl.classList.add('multiOut');
   multiOutEl.classList.add('switcher-item');
   //pouze def - u main indikatoru nepamatujeme stav a pozadujeme noaction pro leftclick
-  itemEl = create_indicator_button(item, def, true);
+  //def||active - ani def
+  itemEl = create_indicator_button(item, false, true);
   //hlavni button ridi expand/collapse
   itemEl.setAttribute('data-bs-toggle', 'collapse');
   itemEl.setAttribute('data-bs-target', '.'+item.name);
@@ -744,7 +745,6 @@ function create_multioutput_button(item, def, active) {
     }
   });
 
-
   return multiOutEl
 }
 

v2realbot/strategyblocks/indicators/custom/classes/SimpleMovingAverage.py

@@ -0,0 +1,22 @@
+from v2realbot.strategyblocks.indicators.custom.classes.indicatorbase import IndicatorBase
+from collections import deque
+
+class SimpleMovingAverage(IndicatorBase):
+    """
+    Calculates the Simple Moving Average (SMA) of a given data list.
+    The SMA is calculated over a specified window size.
+    If there are insufficient data points for the full window, the behavior
+    can be controlled by `return_last_if_insufficient`: if True, the last data point is returned,
+    otherwise, 0 is returned.
+    """
+    def __init__(self, period, return_last_if_insufficient=False, state=None):
+        super().__init__(state)
+        self.window_size = period
+        self.return_last_if_insufficient = return_last_if_insufficient
+        self.data_points = deque(maxlen=period)
+
+    def next(self, data):
+        self.data_points.append(data[-1])
+        if len(self.data_points) < self.window_size:
+            return data[-1] if self.return_last_if_insufficient else 0
+        return sum(self.data_points) / len(self.data_points)

v2realbot/strategyblocks/indicators/custom/classes/SuperTrend.py

@@ -0,0 +1,64 @@
+from v2realbot.strategyblocks.indicators.custom.classes.indicatorbase import IndicatorBase
+from collections import deque
+import numpy as np
+
+class SuperTrend(IndicatorBase):
+    """
+    Advanced implementation of the SuperTrend indicator, dynamically calculating the trend direction.
+    This approach considers the previous trend when determining the new trend, making the indicator 
+    more responsive to price changes. Returns [up, dn, is_trend] with is_trend being 0 (no trend), 
+    1 (uptrend), or -1 (downtrend). If there aren't enough values for ATR, it returns the last close 
+    for up and dn, and 0 for is_trend.
+    Generated with Indicator Plugin Builder.
+    Link: [https://chat.openai.com/g/g-aCKuSmbIe-indicator-plugin-builder/c/8cf9ec38-31e0-4577-8331-22919ae149ab]
+    
+    Zajimavá advanced verze - detaily viz link výše
+    """
+    def __init__(self, atr_period=14, multiplier=3, state=None):
+        super().__init__(state)
+        self.atr_period = atr_period
+        self.multiplier = multiplier
+        self.tr_queue = deque(maxlen=atr_period)
+        self.final_upperband = None
+        self.final_lowerband = None
+        self.is_trend = 0
+
+    def next(self, high, low, close):
+        if len(close) < self.atr_period:
+            return [close[-1], close[-1],close[-1], close[-1], 0]
+
+        # True Range calculation
+        current_high = high[-1]
+        current_low = low[-1]
+        previous_close = close[-2] if len(close) > 1 else close[-1]
+        true_range = max(current_high - current_low, abs(current_high - previous_close), abs(current_low - previous_close))
+
+        # Updating the True Range queue
+        self.tr_queue.append(true_range)
+        if len(self.tr_queue) < self.atr_period:
+            return [close[-1], close[-1],close[-1], close[-1], 0]
+
+        # ATR calculation
+        atr = sum(self.tr_queue) / self.atr_period
+
+        # Basic upper and lower bands
+        basic_upperband = (current_high + current_low) / 2 + self.multiplier * atr
+        basic_lowerband = (current_high + current_low) / 2 - self.multiplier * atr
+
+        # Final upper and lower bands
+        if self.final_upperband is None or self.final_lowerband is None:
+            self.final_upperband = basic_upperband
+            self.final_lowerband = basic_lowerband
+        else:
+            if close[-1] <= self.final_upperband:
+                self.final_upperband = basic_upperband
+            if close[-1] >= self.final_lowerband:
+                self.final_lowerband = basic_lowerband
+
+        # Trend determination
+        if close[-1] > self.final_upperband:
+            self.is_trend = 1
+        elif close[-1] < self.final_lowerband:
+            self.is_trend = -1
+
+        return [basic_upperband,basic_lowerband,self.final_upperband, self.final_lowerband, self.is_trend]

v2realbot/strategyblocks/indicators/custom/classes/SuperTrend1.py

@@ -0,0 +1,57 @@
+from v2realbot.strategyblocks.indicators.custom.classes.indicatorbase import IndicatorBase
+from collections import deque
+
+class SuperTrend1(IndicatorBase):
+    """
+    The SuperTrend indicator is a trend following indicator which uses ATR to calculate its values.
+    It returns a list with three elements: [up, dn, is_trend].
+    `is_trend` can be 1 (uptrend), -1 (downtrend), or 0 (no trend or not enough data).
+    If there are not enough values for ATR, it returns close[-1] for both `up` and `dn`, and 0 for `is_trend`.
+    
+    Note: Code generated with Indicator Plugin Builder.
+    Link: [Indicator Plugin Builder Conversation](https://openai.com/chat/)
+    """
+
+    def __init__(self, multiplier=3, period=14, state=None):
+        super().__init__(state)
+        self.multiplier = multiplier
+        self.period = period
+        self.atr_values = deque(maxlen=period)
+        self.previous_supertrend = None
+        self.previous_close = None
+        self.previous_trend = 0
+
+    def next(self, high, low, close):
+        if len(high) < self.period or len(low) < self.period or len(close) < self.period:
+            return [close[-1], close[-1], 0]
+
+        # Calculate True Range
+        tr = max(high[-1] - low[-1], abs(high[-1] - close[-2]), abs(low[-1] - close[-2]))
+        self.atr_values.append(tr)
+
+        if len(self.atr_values) < self.period:
+            return [close[-1], close[-1], 0]
+
+        # Calculate ATR
+        atr = sum(self.atr_values) / self.period
+
+        # Calculate Supertrend
+        up = close[-1] - (self.multiplier * atr)
+        dn = close[-1] + (self.multiplier * atr)
+
+        if self.previous_supertrend is None:
+            self.previous_supertrend = up
+
+        trend = 0
+        if close[-1] > self.previous_supertrend:
+            trend = 1
+            up = max(up, self.previous_supertrend)
+        elif close[-1] < self.previous_supertrend:
+            trend = -1
+            dn = min(dn, self.previous_supertrend)
+
+        self.previous_supertrend = up if trend == 1 else dn
+        self.previous_close = close[-1]
+        self.previous_trend = trend
+
+        return [up, dn, trend]

v2realbot/strategyblocks/indicators/custom/classes/SuperTrendBard.py

@@ -0,0 +1,61 @@
+from v2realbot.strategyblocks.indicators.custom.classes.indicatorbase import IndicatorBase
+from collections import deque
+
+class SuperTrendBard(IndicatorBase):
+    """
+    Calculates Supertrend indicator values.
+    """
+
+    def __init__(self, period=10, multiplier=3, state=None):
+        super().__init__(state)
+        self.atr_period = period
+        self.multiplier = multiplier
+        self.highs = deque(maxlen=period)
+        self.lows = deque(maxlen=period)
+        self.atr = 0
+        self.upbound = None  # Can set a default value if desired
+        self.downbound = None  # Can set a default value if desired
+        self.is_trend = None
+
+    def next(self, high, low, close):
+        high = high[-1]
+        low = low[-1]
+        close = close[-1]
+
+    # Update ATR calculation
+        self.highs.append(high)
+        self.lows.append(low)
+
+        # Check for sufficient data
+        if len(self.highs) < self.atr_period or len(self.lows) < self.atr_period:
+            return [close, close, 0]
+
+        if len(self.highs) == self.atr_period:
+            true_range = max(high - low, abs(high - self.highs[0]), abs(low - self.lows[0]))
+            self.atr = (self.atr * (self.atr_period - 1) + true_range) / self.atr_period
+
+        # Calculate Supertrend
+        if self.upbound is None:
+            self.upbound = close - (self.multiplier * self.atr)
+            self.downbound = close + (self.multiplier * self.atr)
+            self.is_trend = None  # Set initial trend state to unknown
+        else:
+            # Determine trend based on previous trend and current price
+            if self.is_trend == 1:
+                # Uptrend continuation
+                self.upbound = max(self.upbound, close - (self.multiplier * self.atr))  # Adjust upbound dynamically
+                self.is_trend = 1 if close > self.upbound else 0  # Recalculate trend if needed
+            elif self.is_trend == -1 and close < self.downbound:
+                # Downtrend continues
+                self.downbound = min(self.downbound, low + (self.multiplier * self.atr))
+                self.is_trend = -1
+            else:
+                # Recalculate trend based on current price
+                self.is_trend = 1 if close > self.upbound else -1 if close < self.downbound else 0
+                # Update Supertrend lines based on new trend
+                if self.is_trend == 1:
+                    self.upbound = max(self.upbound, close - (self.multiplier * self.atr))
+                else:
+                    self.downbound = min(self.downbound, low + (self.multiplier * self.atr))
+
+        return [self.upbound, self.downbound, self.is_trend]

v2realbot/strategyblocks/indicators/custom/classes/SuperTrendTV.py

@@ -0,0 +1,50 @@
+from v2realbot.strategyblocks.indicators.custom.classes.indicatorbase import IndicatorBase
+from collections import deque
+import numpy as np
+
+class SuperTrendTV(IndicatorBase):
+    """
+    The SuperTrend indicator is a trend following indicator that is used to identify the direction of the price relative to its historical volatility. 
+    It combines the Average True Range (ATR) with the moving average to determine trend direction and reversal points.
+    This implementation was generated with Indicator Plugin Builder.
+    See conversation: [Indicator Plugin Builder](https://chat.openai.com/g/g-aCKuSmbIe-indicator-plugin-builder/c/1ad650dc-05f1-4cf6-b936-772c0ea86ffa)
+    inspirace https://www.tradingview.com/script/r6dAP7yi/
+    """
+    def __init__(self, atr_period=10, atr_multiplier=3.0, state=None):
+        super().__init__(state)
+        self.atr_period = atr_period
+        self.atr_multiplier = atr_multiplier
+        self.highs = deque(maxlen=atr_period)
+        self.lows = deque(maxlen=atr_period)
+        self.closes = deque(maxlen=atr_period)
+        self.up = None
+        self.down = None
+        self.trend = 1
+
+    def next(self, high, low, close):
+        self.highs.append(high[-1])
+        self.lows.append(low[-1])
+        self.closes.append(close[-1])
+
+        if len(self.highs) < self.atr_period:
+            return [close[-1], close[-1], 0]
+
+        tr = [max(hi - lo, abs(hi - cl), abs(lo - cl))
+              for hi, lo, cl in zip(self.highs, self.lows, self.closes)]
+        atr = np.mean(tr[-self.atr_period:])
+
+        src = (high[-1] + low[-1]) / 2
+        up = src - (self.atr_multiplier * atr)
+        dn = src + (self.atr_multiplier * atr)
+
+        if self.up is None:
+            self.up = up
+            self.down = dn
+        else:
+            self.up = max(up, self.up) if close[-2] > self.up else up
+            self.down = min(dn, self.down) if close[-2] < self.down else dn
+
+        previous_trend = self.trend
+        self.trend = 1 if (self.trend == -1 and close[-1] > self.down) else -1 if (self.trend == 1 and close[-1] < self.up) else self.trend
+
+        return [self.up, self.down, self.trend]

v2realbot/strategyblocks/indicators/custom/talib_ind.py

@@ -11,36 +11,30 @@ from v2realbot.strategyblocks.indicators.helpers import value_or_indicator
 import talib
 
 
-# příklad toml pro indikátor ATR(high, low, close, timeperiod=14)
-# POSITION MATTERS
-# params.series.high = "high"   //series key určuje, že jde o série
-# params.series.low = "low"
-# params.series.low = "close"
-# params.val.timeperiod = 14  //val key určuje, že jde o konkrétní hodnotu, tzn. buď hodnotu nebo název série, ze které vezmu poslední hodnotu (v tom případě by to byl string)
-
-
+# příklad toml pro indikátor ATR(high, low, close, timeperiod=14) -
 # params.series = ["high","low","close"] #pozicni parametry
 # params.keys.timeperiod = 14 #keyword argumenty
-
 #TA-lib prijma positional arguments (zejmena teda ty series)m tzn. series musi byt pozicni
-
-# lookback se aplikuje na vsechy ?
+# lookback se aplikuje na vsechy
 
 
 #IMPLEMENTS usiong of any indicator from TA-lib library
 def talib_ind(state, params, name, returns):
-    funcName = "ma"
-    type = safe_get(params, "type", "SMA")
+    funcName = "talib_ind"
+    type = safe_get(params, "type", None)
+    if type is None:
+        return -2, "type is required"
     #ßsource = safe_get(params, "source", None)
     lookback = safe_get(params, "lookback",None) #celkovy lookback pro vsechny vstupni serie
+    if lookback is not None:
+        #lookback muze byt odkaz na indikator, pak berem jeho hodnotu
+        lookback = int(value_or_indicator(state, lookback))  
+
     start = safe_get(params, "start","linear") #linear/sharp
     defval = safe_get(params, "defval",0)
 
     params = safe_get(params, "params", dict(series=[], keys=[]))
-    #lookback muze byt odkaz na indikator, pak berem jeho hodnotu
-    lookback = int(value_or_indicator(state, lookback))
-    defval = int(value_or_indicator(state, defval))
-
+    defval = float(value_or_indicator(state, defval))
 
     #TODO dopracovat caching, tzn. jen jednou pri inicializaci (linkuje se list) nicmene pri kazde iteraci musime prevest na numpy
     #NOTE doresit, kdyz je val indiaktor, aby se i po inicializaci bral z indikatoru (doresit az pokud bude treba)
@@ -55,7 +49,7 @@ def talib_ind(state, params, name, returns):
             if akt_pocet < lookback and start == "linear":
                 lookback = akt_pocet
 
-        series_list.append(np.array(source_series[-lookback:] if lookback is not None else source_series))
+        series_list.append(np.array(source_series[-lookback:] if lookback is not None else source_series, dtype=np.float64))
 
     for key, val in params.get("keys",{}).items():
          keyArgs[key] = int(value_or_indicator(state, val))
@@ -65,6 +59,21 @@ def talib_ind(state, params, name, returns):
 
     ma_value = talib_function(*series_list, **keyArgs)
 
+    #jde o multioutput, dostavame tuple a prevedeme na list (odpovida poradi v returns)
+    #TODO zapracovat sem def val a isfinite
+    if isinstance(ma_value, tuple):
+        ma_value = list(ma_value)
+        for index, res in enumerate(ma_value):
+            if not np.isfinite(res[-1]):
+                ma_value[index] = defval
+            else:
+                ma_value[index] = round(res[-1],5)
+
+            if res[-1] == 0:
+                ma_value[index] = defval 
+        val = ma_value           
+    #single output
+    else:
         if not np.isfinite(ma_value[-1]):
             val = defval
         else:
@@ -73,5 +82,5 @@ def talib_ind(state, params, name, returns):
         if val == 0:
             val = defval
 
-    state.ilog(lvl=1,e=f"INSIDE {name}:{funcName} {val} {type=} {lookback=}", **params)
+    state.ilog(lvl=1,e=f"INSIDE {name}:{funcName} {str(val)} {type=} {lookback=}", **params)
     return 0, val

v2realbot/strategyblocks/indicators/custom/tulipy_ind.py

@@ -0,0 +1,100 @@
+from v2realbot.utils.utils import isrising, isfalling,zoneNY, price2dec, print, safe_get, is_still, is_window_open, eval_cond_dict, crossed_down, crossed_up, crossed, is_pivot, json_serial, pct_diff, create_new_bars, slice_dict_lists
+from v2realbot.strategy.base import StrategyState
+import v2realbot.indicators.moving_averages as mi
+from v2realbot.strategyblocks.indicators.helpers import get_source_series
+from rich import print as printanyway
+from traceback import format_exc
+import numpy as np
+from collections import defaultdict
+from v2realbot.strategyblocks.indicators.helpers import value_or_indicator
+# from talib import BBANDS, MACD, RSI, MA_Type
+import tulipy
+
+#NOTE if Exception is raised by the plugin - the system stores previous value for the indicator
+#plugin returns configurable default value when exception happens
+#this overrides default behaviour which is when plugin raises exception the custom_hub will store previous value of the indicator as next value
+
+#IMPLEMENTS usiong of any indicator from tulipy library
+def tulipy_ind(state, params, name, returns):
+    funcName = "tulipy_ind"
+    type = safe_get(params, "type", None)
+    if type is None:
+        return -2, "type is required"
+    #ßsource = safe_get(params, "source", None)
+    lookback = safe_get(params, "lookback",None) #celkovy lookback pro vsechny vstupni serie
+    if lookback is not None:
+        #lookback muze byt odkaz na indikator, pak berem jeho hodnotu
+        lookback = int(value_or_indicator(state, lookback))  
+
+    start = safe_get(params, "start","sharp") #linear/sharp
+    defval = safe_get(params, "defval",0)
+
+    params = safe_get(params, "params", dict(series=[], keys=[]))
+    defval = float(value_or_indicator(state, defval))
+
+    try:
+        #TODO dopracovat caching, tzn. jen jednou pri inicializaci (linkuje se list) nicmene pri kazde iteraci musime prevest na numpy
+        #NOTE doresit, kdyz je val indiaktor, aby se i po inicializaci bral z indikatoru (doresit az pokud bude treba)
+        #NOTE doresit lookback, zda se aplikuje na vsechny pred volanim funkce nebo kdy?
+        series_list = []
+        keyArgs = {}
+        for index, item in enumerate(params.get("series",[])):
+            source_series = get_source_series(state, item)
+            #upravujeme lookback pokud not enough values (staci jen pro prvni - jsou vsechny stejne)
+            #REMOVED: neupravujeme lookback, ale az options nize
+            # if index == 0 and lookback is not None:
+            #     akt_pocet = len(source_series)
+            #     if akt_pocet < lookback and start == "linear":
+            #         lookback = akt_pocet
+
+            #to same pokud mame nejake op
+
+            series_list.append(np.array(source_series[-lookback:] if lookback is not None else source_series, dtype=np.float64))
+
+        for key, val in params.get("keys",{}).items():
+            keyArgs[key] = int(value_or_indicator(state, val))
+
+            #pokud jsou zde nejake options s period nebo timeperiodou a mame nastaveny linear, pak zkracujeme pro lepsi rozjezd
+            #zatim porovnavame s prvni serii - bereme ze jsou vsechny ze stejne skupiny
+            #zatim pri linearu davame vzdy akt. - 1 (napr. RSI potrebuje extra datapoint)
+            if key in ["period", "timeperiod"]:
+                akt_pocet = len(series_list[0])
+                if akt_pocet < keyArgs[key] and start == "linear" and akt_pocet != 1:
+                    keyArgs[key] = akt_pocet - 1
+                    printanyway(f"zkracujeme na rozjezd celkem v serii {akt_pocet} nastavujeme period na {keyArgs[key]}")       
+
+        type = "tulipy."+type
+        talib_function = eval(type)
+
+        ma_value = talib_function(*series_list, **keyArgs)
+
+        #jde o multioutput, dostavame tuple a prevedeme na list (odpovida poradi v returns)
+        #TODO zapracovat sem def val a isfinite
+        if isinstance(ma_value, tuple):
+            ma_value = list(ma_value)
+            for index, res in enumerate(ma_value):
+                if not np.isfinite(res[-1]):
+                    ma_value[index] = defval
+                else:
+                    ma_value[index] = round(res[-1],5)
+
+                if res[-1] == 0:
+                    ma_value[index] = defval 
+            val = ma_value           
+        #single output
+        else:
+            if not np.isfinite(ma_value[-1]):
+                val = defval
+            else:
+                val = round(ma_value[-1],4)
+
+            if val == 0:
+                val = defval
+
+        state.ilog(lvl=1,e=f"INSIDE {name}:{funcName} {str(val)} {type=} {lookback=}", **params)
+    #pri Exceptione vracime default value (pokud raisneme Exception do custom_hubu o patro vys, tak ten pouzije posledni hodnotu)
+    except Exception as e:
+        state.ilog(lvl=1,e=f"IND ERROR {name} {funcName} vracime default {defval}", message=str(e)+format_exc())
+        val = defval if defval is not None else 0
+    finally:
+        return 0, val

v2realbot/strategyblocks/indicators/custom_hub.py

@@ -7,6 +7,7 @@ from traceback import format_exc
 import importlib
 import v2realbot.strategyblocks.indicators.custom as ci
 from v2realbot.strategyblocks.indicators.helpers import find_index_optimized
+import numpy as np
 
 
 def populate_dynamic_custom_indicator(data, state: StrategyState, name):
@@ -172,6 +173,9 @@ def populate_dynamic_custom_indicator(data, state: StrategyState, name):
                     ret_val = dict(zip(returns, ret_val))
                 #pokud je to neco jineho nez dict (float,int..) jde o puvodni single output udelame z toho dict s hlavnim jmenem as key
                 elif not isinstance(ret_val, dict):
+                    #checkneme jestli nejde o numpy typ (napr. posledni clen z numpy), prevedem na python basic typ 
+                    if isinstance(ret_val, (np.ndarray, np.generic)):
+                        ret_val = ret_val.item()
                     ret_val = {name: ret_val}
                 #v ostatnich pripadech predpokladame jiz dict
                     

v2realbot/strategyblocks/indicators/helpers.py

@@ -28,6 +28,7 @@ def find_index_optimized(time_list, seconds):
 #podle toho jak se osvedci se zakl.indikatory to s state
 #zatim se mi to moc nezda
 
+#vraci skalar nebo posledni hodnoty indikatoru
 def value_or_indicator(state,value):
     #preklad direktivy podle typu, pokud je int anebo float - je to primo hodnota
     #pokud je str, jde o indikator a dotahujeme posledni hodnotu z nej
@@ -46,6 +47,27 @@ def value_or_indicator(state,value):
                 state.ilog(lvl=1,e=f"Neexistuje indikator s nazvem {value} vracime 0" + str(e) + format_exc())
             return ret
 
+#vraci skalar nebo pokud je indikator tak cely list indikatoru
+def value_or_indicator_list(state,value):
+    #preklad direktivy podle typu, pokud je int anebo float - je to primo hodnota
+    #pokud je str, jde o indikator a dotahujeme posledni hodnotu z nej
+        if isinstance(value, (float, int)):
+            return value
+        elif isinstance(value, str):
+            try:
+                #pokud existuje v indikatoru MA bereme MA jinak indikator, pokud neexistuje bereme bar 
+                ret = get_source_or_MA(state, indicator=value)
+                lvl = 0
+                #TODO tento len dat pryc po overeni
+                delka = len(ret)
+                if delka == 0:
+                    lvl = 1
+                state.ilog(lvl=1,e=f"Pro porovnani bereme cely list s delkou {delka} z indikatoru {value}")
+            except Exception as e   :
+                ret = 0
+                state.ilog(lvl=1,e=f"Neexistuje indikator s nazvem {value} vracime 0" + str(e) + format_exc())
+            return ret
+
 #OPTIMALIZOVANO CHATGPT
 #funkce vytvori podminky (bud pro AND/OR) z pracovniho dict
 def evaluate_directive_conditions(state, work_dict, cond_type):
@@ -67,9 +89,9 @@ def evaluate_directive_conditions(state, work_dict, cond_type):
         "risingc": lambda ind, val: isrisingc(get_source_or_MA(state, ind), val),
         "falling": lambda ind, val: isfalling(get_source_or_MA(state, ind), val),
         "rising": lambda ind, val: isrising(get_source_or_MA(state, ind), val),
-        "crossed_down": lambda ind, val: buy_if_crossed_down(state, ind, value_or_indicator(state,val)),
-        "crossed_up": lambda ind, val: buy_if_crossed_up(state, ind, value_or_indicator(state,val)),
-        "crossed": lambda ind, val: buy_if_crossed_down(state, ind, value_or_indicator(state,val)) or buy_if_crossed_up(state, ind, value_or_indicator(state,val)),
+        "crossed_down": lambda ind, val: buy_if_crossed_down(state, ind, value_or_indicator_list(state,val)),
+        "crossed_up": lambda ind, val: buy_if_crossed_up(state, ind, value_or_indicator_list(state,val)),
+        "crossed": lambda ind, val: buy_if_crossed_down(state, ind, value_or_indicator_list(state,val)) or buy_if_crossed_up(state, ind, value_or_indicator_list(state,val)),
         "pivot_a": lambda ind, val: is_pivot(source=get_source_or_MA(state, ind), leg_number=val, type="A"),
         "pivot_v": lambda ind, val: is_pivot(source=get_source_or_MA(state, ind), leg_number=val, type="V"),
         "still_for": lambda ind, val: is_still(get_source_or_MA(state, ind), val, 2),
@@ -129,13 +151,13 @@ def get_source_series(state, source: str, numpy: bool = False):
 #TYTO NEJSPIS DAT do util
 #vrati true pokud dany indikator prekrocil threshold dolu
 def buy_if_crossed_down(state, indicator, value):
-    res = crossed_down(threshold=value, list=get_source_or_MA(state, indicator))
+    res = crossed_down(value=value, primary_line=get_source_or_MA(state, indicator))
     #state.ilog(lvl=0,e=f"signal_if_crossed_down {indicator} {value} {res}")
     return res
 
 #vrati true pokud dany indikator prekrocil threshold nahoru
 def buy_if_crossed_up(state, indicator, value):
-    res = crossed_up(threshold=value, list=get_source_or_MA(state, indicator))
+    res = crossed_up(value=value, primary_line=get_source_or_MA(state, indicator))
     #state.ilog(lvl=0,e=f"signal_if_crossed_up {indicator} {value} {res}")
     return res    
    

v2realbot/utils/utils.py

@@ -175,8 +175,162 @@ def is_pivot(source: list, leg_number: int, type: str = "A"):
         print("Unknown type")
         return False
     
+#upravene a rozsirene o potencialne vetsi confrm body
+#puvodni verze odpovida confirm_points = 1
+#https://chat.openai.com/c/0e614d96-6af4-40db-a6ec-a8c57ce481b8
+# def crossed_up(threshold, list, confirm_points=2):
+#     """
+#     Check if the threshold has crossed up in the last few values in price_list.
+#     A crossover is confirmed if the threshold is below the earlier prices and then crosses above in the later prices.
+#     The number of confirmation points can be specified; the default is 2.
+#     """
+#     try:
+#         if len(list) < confirm_points * 2:
+#             # Not enough data to confirm crossover
+#             return False
 
-def crossed_up(threshold, list):
+#         # Split the list into two parts for comparison
+#         earlier_prices = list[-confirm_points*2:-confirm_points]
+#         later_prices = list[-confirm_points:]
+
+#         # Check if threshold was below earlier prices and then crossed above
+#         was_below = all(threshold < price for price in earlier_prices)
+#         now_above = all(threshold >= price for price in later_prices)
+
+#         return was_below and now_above
+
+#     except IndexError:
+#         # In case of an IndexError, return False
+#         return False
+
+#recent cross up of two arrays (price1 crossed up price2), fallback to standard
+#inputs are numpy arrays
+# def crossed_up_numpy(price1, price2):
+#     if price1.size < 2 or price2.size < 2:
+#         return False  # Not enough data
+
+#     # Calculate slopes for the last two points
+#     x = np.array([price1.size - 2, price1.size - 1])
+#     slope1, intercept1 = np.polyfit(x, price1[-2:], 1)
+#     slope2, intercept2 = np.polyfit(x, price2[-2:], 1)
+
+#     # Check if lines are almost parallel
+#     if np.isclose(slope1, slope2):
+#         return False
+
+#     # Calculate intersection point
+#     x_intersect = (intercept2 - intercept1) / (slope1 - slope2)
+#     y_intersect = slope1 * x_intersect + intercept1
+
+#     # Check if the intersection occurred between the last two points
+#     if x[0] < x_intersect <= x[1]:
+#         # Check if line1 crossed up line2
+#         return price1[-1] > price2[-1] and price1[-2] <= price2[-2]
+    
+#     return False
+
+#same but more efficient approach
+def crossed_up_numpy(price1, price2):
+    if price1.size < 2 or price2.size < 2:
+        return False  # Not enough data
+
+    # Indices for the last two points
+    x1, x2 = price1.size - 2, price1.size - 1
+
+    # Direct calculation of slopes and intercepts
+    slope1 = (price1[-1] - price1[-2]) / (x2 - x1)
+    intercept1 = price1[-1] - slope1 * x2
+    slope2 = (price2[-1] - price2[-2]) / (x2 - x1)
+    intercept2 = price2[-1] - slope2 * x2
+
+    # Check if lines are almost parallel
+    if np.isclose(slope1, slope2):
+        return False
+
+    # Calculate intersection point (x-coordinate only)
+    if slope1 != slope2:  # Avoid division by zero
+        x_intersect = (intercept2 - intercept1) / (slope1 - slope2)
+
+        # Check if the intersection occurred between the last two points
+        if x1 < x_intersect <= x2:
+            # Check if line1 crossed up line2
+            return price1[-1] > price2[-1] and price1[-2] <= price2[-2]
+    
+    return False
+
+
+#recent cross up of two arrays (price1 crossed up price2), fallback to standard
+#inputs are numpy arrays
+# def crossed_down_numpy(price1, price2):
+#     if price1.size < 2 or price2.size < 2:
+#         return False  # Not enough data
+
+#     # Calculate slopes for the last two points
+#     x = np.array([price1.size - 2, price1.size - 1])
+#     slope1, intercept1 = np.polyfit(x, price1[-2:], 1)
+#     slope2, intercept2 = np.polyfit(x, price2[-2:], 1)
+
+#     # Check if lines are almost parallel
+#     if np.isclose(slope1, slope2):
+#         return False
+
+#     # Calculate intersection point
+#     x_intersect = (intercept2 - intercept1) / (slope1 - slope2)
+#     y_intersect = slope1 * x_intersect + intercept1
+
+#     # Check if the intersection occurred between the last two points
+#     if x[0] < x_intersect <= x[1]:
+#         # Check if line1 crossed down line2
+#         return price1[-1] < price2[-1] and price1[-2] >= price2[-2]
+    
+#     return False
+
+#more efficient yet same, price1 - faster, price2 - slower
+def crossed_down_numpy(price1, price2):
+    if price1.size < 2 or price2.size < 2:
+        return False  # Not enough data
+
+    # Indices for the last two points
+    x1, x2 = price1.size - 2, price1.size - 1
+
+    # Direct calculation of slopes and intercepts
+    slope1 = (price1[-1] - price1[-2]) / (x2 - x1)
+    intercept1 = price1[-1] - slope1 * x2
+    slope2 = (price2[-1] - price2[-2]) / (x2 - x1)
+    intercept2 = price2[-1] - slope2 * x2
+
+    # Check if lines are almost parallel
+    if np.isclose(slope1, slope2):
+        return False
+
+    # Calculate intersection point (x-coordinate only)
+    if slope1 != slope2:  # Avoid division by zero
+        x_intersect = (intercept2 - intercept1) / (slope1 - slope2)
+
+        # Check if the intersection occurred between the last two points
+        if x1 < x_intersect <= x2:
+            # Check if line1 crossed down line2
+            return price1[-1] < price2[-1] and price1[-2] >= price2[-2]
+    
+    return False
+
+#obalka pro crossup listu a thresholdu nebo listu a druheho listu
+#value  - svcalar or list, primary_line - usually faster
+def crossed_up(value, primary_line):
+    if isinstance(value, list):
+        return crossed_up_numpy(np.array(primary_line), np.array(value))
+    else:
+        return crossed_up_threshold(threshold=value, list=primary_line)    
+
+#obalka pro crossdown listu a thresholdu nebo listu a druheho listu
+#value  - svcalar or list, primary_line - usually faster
+def crossed_down(value, primary_line):
+    if isinstance(value, list):
+        return crossed_down_numpy(np.array(primary_line), np.array(value))
+    else:
+        return crossed_down_threshold(threshold=value, list=primary_line)    
+
+def crossed_up_threshold(threshold, list):
     """check if threshold has crossed up last thresholdue in list"""
     try:
         if threshold < list[-1] and threshold > list[-2]:
@@ -190,7 +344,7 @@ def crossed_up(threshold, list):
     except IndexError:
         return False
     
-def crossed_down(threshold, list):
+def crossed_down_threshold(threshold, list):
     """check if threshold has crossed down last thresholdue in list"""
     """
     Checks if a threshold has just crossed down a line represented by a list.
@@ -383,6 +537,7 @@ def json_serial(obj):
         datetime: lambda obj: obj.timestamp(),
         UUID: lambda obj: str(obj),
         Enum: lambda obj: str(obj),
+        np.int32: lambda obj: int(obj),
         np.int64: lambda obj: int(obj),
         np.float64: lambda obj: float(obj),
         Order: lambda obj: obj.__dict__,
@@ -445,7 +600,7 @@ qu = Queue()
 
 #zoneNY = tz.gettz('America/New_York')
 zoneNY = pytz.timezone('US/Eastern')
-
+zoneUTC = pytz.utc
 zonePRG = pytz.timezone('Europe/Amsterdam')
 
 def print(*args, **kwargs):