decomm ml, target algorithm a dalsi upravy

v2realbot/strategyblocks/activetrade/helpers.py

@@ -2,7 +2,7 @@ from v2realbot.enums.enums import RecordType, StartBarAlign, Mode, Account, Foll
 from v2realbot.common.PrescribedTradeModel import Trade, TradeDirection, TradeStatus
 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.utils.directive_utils import get_conditions_from_configuration
-from v2realbot.ml.mlutils import load_model
+# from v2realbot.ml.mlutils import load_model
 from v2realbot.common.model import SLHistory
 from v2realbot.config import KW
 from uuid import uuid4

v2realbot/strategyblocks/helpers.py

@@ -1,6 +1,6 @@
 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.utils.directive_utils import get_conditions_from_configuration
-from v2realbot.ml.mlutils import load_model
+# from v2realbot.ml.mlutils import load_model
 from v2realbot.common.model import SLHistory
 from v2realbot.config import KW
 from uuid import uuid4

v2realbot/strategyblocks/indicators/cbar_price.py

@@ -5,7 +5,7 @@ def populate_cbar_tick_price_indicator(data, state: StrategyState):
     conf_bar = data['confirmed']
 
     #specifická sekce pro CBARVOLUME, kde vzdy máme nova data v confirmation baru (tzn. tickprice pocitame jak pri potvrzenem tak nepotvrzenem)
-    if state.rectype in (RecordType.CBARVOLUME, RecordType.CBARRENKO):
+    if state.rectype in (RecordType.CBARVOLUME, RecordType.CBARDOLLAR, RecordType.CBARRENKO):
         try:
             tick_price = data['close']
             tick_delta_volume = data['volume'] - state.vars.last_tick_volume

v2realbot/strategyblocks/indicators/custom/divergence.py

@@ -0,0 +1,42 @@
+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
+from v2realbot.indicators.indicators import ema, natr, roc
+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
+
+#abs/rel divergence of two indicators
+def divergence(state, params, name):
+    funcName = "indicatorDivergence"
+    source1 = safe_get(params, "source1", None)
+    source1_series = get_source_series(state, source1)
+    source2 = safe_get(params, "source2", None)
+    source2_series = get_source_series(state, source2)
+    mode = safe_get(params, "type")
+    state.ilog(lvl=0,e=f"INSIDE {name} {funcName} {source1=} {source2=} {mode=}", **params)
+    val = 0
+    if mode == "abs":
+        val =  round(abs(float(source1_series[-1]) - float(source2_series[-1])),4)
+    elif mode == "absn":
+        val =  round((abs(float(source1_series[-1]) - float(source2_series[-1])))/float(source1_series[-1]),4)
+    elif mode == "rel":
+        val =  round(float(source1_series[-1]) - float(source2_series[-1]),4)
+    elif mode == "reln":
+        val =  round((float(source1_series[-1]) - float(source2_series[-1]))/float(source1_series[-1]),4)
+    elif mode == "pctabs":
+        val = pct_diff(num1=float(source1_series[-1]),num2=float(source2_series[-1]), absolute=True)
+    elif mode == "pct":
+        val = pct_diff(num1=float(source1_series[-1]),num2=float(source2_series[-1]))
+    return 0, val
+
+#model - naloadovana instance modelu
+#seq - sekvence pro vstup
+
+
+
+
+
+
+

v2realbot/strategyblocks/indicators/custom/target.py

@@ -0,0 +1,110 @@
+from v2realbot.utils.utils import isrising, isfalling,zoneNY, price2dec, 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
+from v2realbot.indicators.indicators import ema, natr, roc
+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
+
+#target algorithm for ML
+""""
+v teto funkci bere posledni cenu (future price) porovname s cenou X baru zpet(current_price) a vysledek ulozime jako hodnotu X baru zpet
+
+source - target priceline
+window_length_value = 3 #delka okna ve kterem pocitame target (current_price and future_price)
+window_length_unit = "bar", "time" #zatim jen v barech
+no_move_threshold = 0.033 #zmene v procentech povazovana za no move, v % (1tick pri cene 30 je 0.033%), pripadne rozlisit na _upper and _lower
+pct_change_full_scale = #pct change that is considered 1, used in scaler to determine target_value based on price change
+
+
+TODO musi se signal trochu tahnout az kde se opravdu rozjede, aby si to model spojil
+"""""
+def target(state, params, name):
+    funcName = "target"
+    source = safe_get(params, "source", "vwap")
+    source_series = get_source_series(state, source)
+    #zatim podporovano jen bar - dokud nepridame o level vyse do "custom" save_to_past na urovni casu
+    window_length_value = safe_get(params, "window_length_value", None)
+    window_length_unit= safe_get(params, "window_length_unit", "bar")    
+    no_move_threshold= safe_get(params, "no_move_threshold", 0.033)
+    pct_change_full_scale = safe_get(params, "pct_change_full_scale", 0.1) #pct change that is considered for full scaling factor
+    move_base= safe_get(params, "move_base", 0.2)  #base output value for move up/down, to this value scaling factor is added based on magnitude of change
+    scaling_factor= safe_get(params, "scaling_factor", 0.8)# This represents the maximum additional value to be added to 0.5, based on steepness of the move
+
+    if window_length_value is None or source is None:
+        return -2, "window_length_value/source required"
+
+    state.ilog(lvl=0,e=f"INSIDE {name} {funcName}", **params)
+
+    #0 = no move value
+    val = 0
+
+    future_price = float(source_series[-1])
+
+    try:
+        current_price = float(source_series[-window_length_value])
+    except IndexError:
+        return 0, val
+    
+    upper_move_threshold = add_pct(no_move_threshold, current_price)
+    lower_move_threshold = add_pct(-no_move_threshold, current_price)
+    
+    #no move
+    if lower_move_threshold <= future_price <= upper_move_threshold:
+        return 0, val
+
+    #calculates weight based on magnitude of change
+    def calculate_factor(current_price, future_price):
+        #podle vzdalenosti od no_move_thresholdu vratime hodnotu
+        # pct_change_full_scale - pct that is considered maximum for scaling_factor value above
+        current_pct_delta = abs(pct_delta(current_price, future_price)) #aktuální pct gap mezi curr a price
+        magnitude_val = min(current_pct_delta / pct_change_full_scale, 1) * scaling_factor
+        return magnitude_val
+
+    #price is bigger than threshold
+    if upper_move_threshold < future_price:
+        magnitude_val = calculate_factor(current_price, future_price)
+        return 0, move_base + magnitude_val
+
+    #price is bigger than threshold
+    if lower_move_threshold > future_price:
+        magnitude_val = calculate_factor(current_price, future_price)
+        return 0, - move_base - magnitude_val
+
+
+def pct_delta(base, second_number):
+    """
+    Calculate the percentage difference between the base and the second number.
+
+    Parameters:
+    base (float): The base value.
+    second_number (float): The second number to compare against the base.
+
+    Returns:
+    float: The percentage difference (delta) between the second number and the base.
+    """
+    if base == 0:
+        raise ValueError("Base cannot be zero.")
+
+    return ((second_number - base) / base) * 100
+
+def add_pct(pct, value):
+    """
+    Add a percentage to a value. If pct is negative it is subtracted.
+
+    Parameters:
+    pct (float): The percentage to add (e.g., 10 for 10%).
+    value (float): The original value.
+
+    Returns:
+    float: The new value after adding the percentage.
+    """
+    return value * (1 + pct / 100)
+
+
+if __name__ == '__main__':
+    print(add_pct(1,100))
+    print(add_pct(-1,100))
+    print(pct_delta(100,102))
+    print(pct_delta(100,98))

v2realbot/strategyblocks/inits/init_indicators.py

@@ -4,9 +4,9 @@ from v2realbot.enums.enums import RecordType, StartBarAlign, Mode, Account, Foll
 from v2realbot.common.PrescribedTradeModel import Trade, TradeDirection, TradeStatus
 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.utils.directive_utils import get_conditions_from_configuration
-from v2realbot.ml.mlutils import load_model
+import mlroom.utils.mlutils as ml
 from v2realbot.common.model import SLHistory
-from v2realbot.config import KW
+from v2realbot.config import KW, MODEL_DIR
 from uuid import uuid4
 from datetime import datetime
 #import random
@@ -48,7 +48,7 @@ def initialize_dynamic_indicators(state):
                     modelname = safe_get(indsettings["cp"], 'name', None)
                     modelversion = safe_get(indsettings["cp"], 'version', "1")
                     if modelname is not None:
-                        state.vars.loaded_models[modelname] =  load_model(modelname, modelversion)
+                        state.vars.loaded_models[modelname] =  ml.load_model(modelname, modelversion, MODEL_DIR)
                         if state.vars.loaded_models[modelname] is not None:
                             printanyway(f"model {modelname} loaded")
                         else: