-
Notifications
You must be signed in to change notification settings - Fork 78
/
Copy pathmock2_forecast_system.py
198 lines (174 loc) · 7.43 KB
/
mock2_forecast_system.py
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
"""
Import as:
import dataflow_amp.system.mock2.mock2_forecast_system as dtfasmmfosy
"""
import os
import pandas as pd
import core.config as cconfig
import core.finance.market_data_example as cfmadaex
import dataflow.core as dtfcore
import dataflow.system as dtfsys
import dataflow_amp.pipelines.mock2.mock2_pipeline as dtapmmopi
import dataflow_amp.system.mock1.mock1_builders as dtfasmmobu
import market_data as mdata
import oms
# #############################################################################
# Mock2_NonTime_ForecastSystem
# #############################################################################
class Mock2_NonTime_ForecastSystem(dtfsys.NonTime_ForecastSystem):
"""
Create a System with:
- a HistoricalMarketData
- a non-timed DAG
This is used to run a backtest for a Mock2 pipeline.
"""
def __init__(self):
super().__init__()
# TODO(Grisha): consider exposing to the interface.
self.train_test_mode = "ins"
def _get_system_config_template(self) -> cconfig.Config:
_ = self
dag_builder = dtapmmopi.Mock2_DagBuilder()
system_config = dtfsys.get_SystemConfig_template_from_DagBuilder(
dag_builder
)
return system_config
def _get_market_data(self) -> mdata.ImClientMarketData:
# TODO(Grisha): we should not use Mock1 builders here, better to
# factor out common code, put it under a common dir and re-use here.
market_data = dtfasmmobu.get_Mock1_MarketData_example2(self)
return market_data
def _get_dag(self) -> dtfcore.DAG:
# TODO(Grisha): pass via `System.config` instead.
timestamp_column_name = "end_ts"
# TODO(Grisha): we should not use Mock1 builders here, better to factor
# out common code, put it under a common dir and re-use here.
dag = dtfasmmobu.get_Mock1_HistoricalDag_example1(
self, timestamp_column_name
)
return dag
def _get_dag_runner(self) -> dtfcore.DagRunner:
dag_runner = dtfcore.FitPredictDagRunner(self.dag)
return dag_runner
# #############################################################################
# Mock2_Time_ForecastSystem_with_DataFramePortfolio
# #############################################################################
class Mock2_Time_ForecastSystem_with_DataFramePortfolio(
dtfsys.Time_ForecastSystem_with_DataFramePortfolio
):
"""
Build a System with:
- `ReplayedMarketData`
- a timed DAG
- `DataFramePortfolio`
"""
def _get_system_config_template(self) -> cconfig.Config:
dag_builder = dtapmmopi.Mock2_DagBuilder()
system_config = dtfsys.get_SystemConfig_template_from_DagBuilder(
dag_builder
)
return system_config
def _get_market_data(self) -> mdata.ImClientMarketData:
# Generate random OHLCV bars.
start_datetime = pd.Timestamp(
"2023-08-01 10:00:00", tz="America/New_York"
)
end_datetime = pd.Timestamp("2023-08-31 10:00:00", tz="America/New_York")
n_assets = self.config["market_data_config", "number_of_assets"]
asset_ids = [i for i in range(n_assets)]
bar_duration_in_seconds = self.config["bar_duration_in_seconds"]
bar_duration = f"{int(bar_duration_in_seconds/60)}T"
df = cfmadaex.generate_random_ohlcv_bars(
start_datetime, end_datetime, asset_ids, bar_duration=bar_duration
)
# Specify parameters for `ReplayedMarketData_from_df`.
self.config[
"market_data_config", "replayed_delay_in_mins_or_timestamp"
] = pd.Timestamp("2023-08-15 10:30:00", tz="America/New_York")
self.config["market_data_config", "data"] = df
# We wait 10 seconds in any case, because we add 10 seconds to
# `timestamp_db` column in `generate_random_ohlcv_bars()`.
self.config["market_data_config", "delay_in_secs"] = 0
# This is needed to construct the Portfolio.
self.config["market_data_config", "asset_ids"] = asset_ids
# Build `MarketData`.
market_data = dtfsys.get_ReplayedMarketData_from_df(self)
return market_data
def _get_dag(self) -> dtfcore.DAG:
# Add `RealTimeDataSource`.
bar_duration_in_minutes = int(self.config["bar_duration_in_seconds"] / 60)
# TODO(Grisha): we should remove the resampling piece for this pipeline
# or create another Mock3 pipeline.
self.config[
"dag_config", "resample", "transformer_kwargs", "rule"
] = f"{bar_duration_in_minutes}T"
dag = dtfsys.add_real_time_data_source(self)
# Configure a `ProcessForecastNode`.
log_dir = os.path.join(self.config["system_log_dir"], "process_forecasts")
# Here we configure only `compute_target_positions_kwargs` from the outside.
compute_target_positions_kwargs = self.config[
"process_forecasts_node_dict",
"process_forecasts_dict",
"optimizer_config",
"params",
"kwargs",
]
process_forecasts_dict = {
# Params for `process_forecasts()`.
"order_config": {
"order_type": "price@twap",
"passivity_factor": None,
"order_duration_in_mins": bar_duration_in_minutes,
},
"optimizer_config": {
"backend": "pomo",
"params": {
"style": "cross_sectional",
"kwargs": compute_target_positions_kwargs,
},
},
"execution_mode": "real_time",
"log_dir": log_dir,
}
# Params for `ProcessForecastsNode`.
dag_builder = self.config["dag_builder_object"]
volatility_col = dag_builder.get_column_name("volatility")
prediction_col = dag_builder.get_column_name("prediction")
process_forecasts_node_dict = {
"prediction_col": prediction_col,
"volatility_col": volatility_col,
"spread_col": None,
"portfolio": self.portfolio,
# This configures `process_forecasts()`.
"process_forecasts_dict": process_forecasts_dict,
}
process_forecasts_node_config = cconfig.Config.from_dict(
process_forecasts_node_dict
)
self.config["process_forecasts_node_dict"] = process_forecasts_node_config
self = dtfsys.apply_ProcessForecastsNode_config_for_equities(self)
# Append the `ProcessForecastNode` to the `Dag`.
dag = dtfsys.add_ProcessForecastsNode(self, dag)
return dag
def _get_portfolio(self) -> oms.Portfolio:
# Portfolio config.
bar_duration_in_seconds = self.config["bar_duration_in_seconds"]
self.config["portfolio_config", "mark_to_market_col"] = "close"
self.config[
"portfolio_config", "pricing_method"
] = f"twap.{int(bar_duration_in_seconds/60)}T"
column_remap = {
"bid": "bid",
"ask": "ask",
"midpoint": "midpoint",
"price": "close",
}
self.config[
"portfolio_config", "column_remap"
] = cconfig.Config.from_dict(column_remap)
# Build `Portfolio`.
portfolio = dtfsys.get_DataFramePortfolio_from_System(self)
return portfolio
def _get_dag_runner(self) -> dtfsys.RealTimeDagRunner:
dag_runner = dtfsys.get_RealTimeDagRunner_from_System(self)
return dag_runner