Order execution open source (#1447)

* Waiting for bin data

* Complete readme

* CI

* Add inst filter by time

* Update qlib/data/dataset/processor.py

* typo

* Fix time filter bug

* Add Filter and set Universe

* Complete data pipeline

* Fix Provider Logger Info Args

* Add DQN; a minor bugfix in ppo reward.

* update readme. modify assertion logic in strategy check.

* Fix Doc issues and fix black

* Fix pylint Error

---------

Co-authored-by: Young <afe.young@gmail.com>
Co-authored-by: you-n-g <you-n-g@users.noreply.github.com>

.gitignore

@@ -27,6 +27,8 @@ examples/estimator/estimator_example/
 examples/rl/data/
 examples/rl/checkpoints/
 examples/rl/outputs/
+examples/rl_order_execution/data/
+examples/rl_order_execution/outputs/
 
 *.egg-info/
 

examples/benchmarks/LightGBM/multi_freq_handler.py

@@ -29,13 +29,13 @@ class Avg15minHandler(DataHandlerLP):
         fit_end_time=None,
         process_type=DataHandlerLP.PTYPE_A,
         filter_pipe=None,
-        inst_processor=None,
+        inst_processors=None,
         **kwargs,
     ):
         infer_processors = check_transform_proc(infer_processors, fit_start_time, fit_end_time)
         learn_processors = check_transform_proc(learn_processors, fit_start_time, fit_end_time)
         data_loader = Avg15minLoader(
-            config=self.loader_config(), filter_pipe=filter_pipe, freq=freq, inst_processor=inst_processor
+            config=self.loader_config(), filter_pipe=filter_pipe, freq=freq, inst_processors=inst_processors
         )
         super().__init__(
             instruments=instruments,

examples/benchmarks/LightGBM/workflow_config_lightgbm_Alpha158_multi_freq.yaml

@@ -18,7 +18,7 @@ data_handler_config: &data_handler_config
         label: day
         feature: 1min
     # with label as reference
-    inst_processor:
+    inst_processors:
         feature:
             - class: Resample1minProcessor
               module_path: features_sample.py

examples/benchmarks/LightGBM/workflow_config_lightgbm_multi_freq.yaml

@@ -19,7 +19,7 @@ data_handler_config: &data_handler_config
         feature_15min: 1min
         feature_day: day
     # with label as reference
-    inst_processor:
+    inst_processors:
         feature_15min:
             - class: ResampleNProcessor
               module_path: features_resample_N.py

examples/rl_order_execution/README.md

@@ -0,0 +1,100 @@
+# RL Example for Order Execution
+
+This folder comprises an example of Reinforcement Learning (RL) workflows for order execution scenario, including both training workflows and backtest workflows.
+
+## Data Processing
+
+### Get Data
+
+```
+python -m qlib.run.get_data qlib_data qlib_data --target_dir ./data/bin --region hs300 --interval 5min
+```
+
+### Generate Pickle-Style Data
+
+To run codes in this example, we need data in pickle format. To achieve this, run following commands (might need a few minutes to finish):
+
+```
+python scripts/gen_pickle_data.py -c scripts/pickle_data_config.yml
+python scripts/collect_pickle_dataframe.py
+python scripts/gen_training_orders.py
+python scripts/merge_orders.py
+```
+
+When finished, the structure under `data/` should be:
+
+```
+data
+├── bin
+├── orders
+├── pickle
+└── pickle_dataframe
+```
+
+## Training
+
+Each training task is specified by a config file. The config file for task `TASKNAME` is `exp_configs/train_TASKNAME.yml`. This example provides two training tasks:
+
+- **PPO**: Method proposed by IJCAL 2020 paper "[An End-to-End Optimal Trade Execution Framework based on Proximal Policy Optimization](https://www.ijcai.org/proceedings/2020/0627.pdf)".
+- **OPDS**: Method proposed by AAAI 2021 paper "[Universal Trading for Order Execution with Oracle Policy Distillation](https://arxiv.org/abs/2103.10860)".
+
+The main differece between these two methods is their reward functions. Please see their config files for details.
+
+Take OPDS as an example, to run the training workflow, run:
+
+```
+python -m qlib.rl.contrib.train_onpolicy --config_path exp_configs/train_opds.yml --run_backtest
+```
+
+Metrics, logs, and checkpoints will be stored under `outputs/opds` (configured by `exp_configs/train_opds.yml`). 
+
+## Backtest
+
+Once the training workflow has completed, the trained model can be used for the backtesting workflow. Still taking OPDS as an example, once training is finished, the latest checkpoint of the model can be found at `outputs/opds/checkpoints/latest.pth`. To run backtest workflow:
+
+1. Uncomment the `weight_file` parameter in `exp_configs/train_opds.yml` (it is commented by default). While it is possible to run the backtesting workflow without setting a checkpoint, this will lead to randomly initialized model results, thus making them meaningless.
+2. Run `python -m qlib.rl.contrib.backtest --config_path exp_configs/backtest_opds.yml`.
+
+The backtest result is stored in `outputs/checkpoints/backtest_result.csv`.
+
+In addition to OPDS and PPO, we also provide TWAP ([Time-weighted average price](https://en.wikipedia.org/wiki/Time-weighted_average_price)) as a weak baseline. The config file for TWAP is `exp_configs/backtest_twap.yml`.
+
+### Gap between backtest and training pipeline's testing
+
+It is worthy to notice that the results of the backtesting process may differ from the results of the testing process used during training.
+This is because different simulators are used to simulate market conditions during training and backtesting.
+In training pipeline, the simplified simulator called `SingleAssetOrderExecutionSimple` is used for efficiency reasons. 
+`SingleAssetOrderExecutionSimple` makes no restriction to trading amounts. 
+No matter what the amount of the order is, it can be completely executed.
+However, during backtesting, a more realistic simulator called `SingleAssetOrderExecution` is used. 
+It takes into account practical constraints in more real-world scenarios (for example, the trading volume must be a multiple of the smallest trading unit).
+As a result, the amount of an order that is actually executed during backtesting may differ from the amount expected to be executed.
+
+If you would like to obtain results that are exactly the same as those obtained during testing in the training pipeline, you could run training pipeline with only backtest phrase.
+In order to do this:
+- Modify the training config. Add the path of the checkpoint you want to use (see following for an example).
+- Run `python -m qlib.rl.contrib.train_onpolicy --config_path PATH/TO/CONFIG --run_backtest --no_training`
+
+```yaml
+...
+policy:
+  class: PPO  # PPO, DQN
+  kwargs:
+    lr: 0.0001
+    weight_file: PATH/TO/CHECKPOINT
+  module_path: qlib.rl.order_execution.policy
+...
+```
+
+## Benchmarks (TBD)
+
+To accurately evaluate the performance of models using Reinforcement Learning algorithms, it's best to run experiments multiple times and compute the average performance across all trials. However, given the time-consuming nature of model training, this is not always feasible. An alternative approach is to run each training task only once, selecting the 10 checkpoints with the highest validation performance to simulate multiple trials. In this example, we use "Price Advantage (PA)" as the metric for selecting these checkpoints. The average performance of these 10 checkpoints on the testing set is as follows:
+
+| **Model**                   | **PA mean with std.** |
+|-----------------------------|-----------------------|
+| OPDS (with PPO policy)      |  0.4785 ± 0.7815      |
+| OPDS (with DQN policy)      | -0.0114 ± 0.5780      |
+| PPO                         | -1.0935 ± 0.0922      |
+| TWAP                        |   ≈ 0.0 ± 0.0         |
+
+The table above also includes TWAP as a rule-based baseline. The ideal PA of TWAP should be 0.0, however, in this example, the order execution is divided into two steps: first, the order is split equally among each half hour, and then each five minutes within each half hour. Since trading is forbidden during the last five minutes of the day, this approach may slightly differ from traditional TWAP over the course of a full day (as there are 5 minutes missing in the last "half hour"). Therefore, the PA of TWAP can be considered as a number that is close to 0.0. To verify this, you may run a TWAP backtest and check the results.

examples/rl_order_execution/exp_configs/backtest_opds.yml

@@ -0,0 +1,59 @@
+order_file: ./data/orders/test_orders.pkl
+start_time: "9:30"
+end_time: "14:54"
+qlib:
+  provider_uri_5min: ./data/bin/
+  feature_root_dir: ./data/pickle/
+  feature_columns_today: [
+    "$open", "$high", "$low", "$close", "$vwap", "$bid", "$ask", "$volume",
+    "$bidV", "$bidV1", "$bidV3", "$bidV5", "$askV", "$askV1", "$askV3", "$askV5"
+  ]
+  feature_columns_yesterday: [
+    "$open_1", "$high_1", "$low_1", "$close_1", "$vwap_1", "$bid_1", "$ask_1", "$volume_1",
+    "$bidV_1", "$bidV1_1", "$bidV3_1", "$bidV5_1", "$askV_1", "$askV1_1", "$askV3_1", "$askV5_1"
+  ]
+exchange:
+  limit_threshold: null
+  deal_price: ["$close", "$close"]
+  volume_threshold: null
+strategies:
+  1day:
+    class: SAOEIntStrategy
+    kwargs:
+      data_granularity: 5
+      action_interpreter:
+        class: CategoricalActionInterpreter
+        kwargs:
+          max_step: 8
+          values: 4
+        module_path: qlib.rl.order_execution.interpreter
+      network:
+        class: Recurrent
+        kwargs: {}
+        module_path: qlib.rl.order_execution.network
+      policy:
+        class: PPO  # PPO, DQN
+        kwargs:
+          lr: 0.0001
+          # Restore `weight_file` once the training workflow finishes. You can change the checkpoint file you want to use.
+          # weight_file: outputs/opds/checkpoints/latest.pth
+        module_path: qlib.rl.order_execution.policy
+      state_interpreter:
+        class: FullHistoryStateInterpreter
+        kwargs:
+          data_dim: 5
+          data_ticks: 48
+          max_step: 8
+          processed_data_provider:
+            class: PickleProcessedDataProvider
+            kwargs:
+              data_dir: ./data/pickle_dataframe/feature
+            module_path: qlib.rl.data.pickle_styled
+        module_path: qlib.rl.order_execution.interpreter
+    module_path: qlib.rl.order_execution.strategy
+  30min:
+    class: TWAPStrategy
+    kwargs: {}
+    module_path: qlib.contrib.strategy.rule_strategy
+concurrency: 16
+output_dir: outputs/opds/

examples/rl_order_execution/exp_configs/backtest_ppo.yml

@@ -0,0 +1,59 @@
+order_file: ./data/orders/test_orders.pkl
+start_time: "9:30"
+end_time: "14:54"
+qlib:
+  provider_uri_5min: ./data/bin/
+  feature_root_dir: ./data/pickle/
+  feature_columns_today: [
+    "$open", "$high", "$low", "$close", "$vwap", "$bid", "$ask", "$volume",
+    "$bidV", "$bidV1", "$bidV3", "$bidV5", "$askV", "$askV1", "$askV3", "$askV5"
+  ]
+  feature_columns_yesterday: [
+    "$open_1", "$high_1", "$low_1", "$close_1", "$vwap_1", "$bid_1", "$ask_1", "$volume_1",
+    "$bidV_1", "$bidV1_1", "$bidV3_1", "$bidV5_1", "$askV_1", "$askV1_1", "$askV3_1", "$askV5_1"
+  ]
+exchange:
+  limit_threshold: null
+  deal_price: ["$close", "$close"]
+  volume_threshold: null
+strategies:
+  1day:
+    class: SAOEIntStrategy
+    kwargs:
+      data_granularity: 5
+      action_interpreter:
+        class: CategoricalActionInterpreter
+        kwargs:
+          max_step: 8
+          values: 4
+        module_path: qlib.rl.order_execution.interpreter
+      network:
+        class: Recurrent
+        kwargs: {}
+        module_path: qlib.rl.order_execution.network
+      policy:
+        class: PPO  # PPO, DQN
+        kwargs:
+          lr: 0.0001
+          # Restore `weight_file` once the training workflow finishes. You can change the checkpoint file you want to use.
+          # weight_file: outputs/ppo/checkpoints/latest.pth
+        module_path: qlib.rl.order_execution.policy
+      state_interpreter:
+        class: FullHistoryStateInterpreter
+        kwargs:
+          data_dim: 5
+          data_ticks: 48
+          max_step: 8
+          processed_data_provider:
+            class: PickleProcessedDataProvider
+            kwargs:
+              data_dir: ./data/pickle_dataframe/feature
+            module_path: qlib.rl.data.pickle_styled
+        module_path: qlib.rl.order_execution.interpreter
+    module_path: qlib.rl.order_execution.strategy
+  30min:
+    class: TWAPStrategy
+    kwargs: {}
+    module_path: qlib.contrib.strategy.rule_strategy
+concurrency: 16
+output_dir: outputs/ppo/

examples/rl_order_execution/exp_configs/backtest_twap.yml

@@ -0,0 +1,29 @@
+order_file: ./data/orders/test_orders.pkl
+start_time: "9:30"
+end_time: "14:54"
+qlib:
+  provider_uri_5min: ./data/bin/
+  feature_root_dir: ./data/pickle/
+  feature_columns_today: [
+    "$open", "$high", "$low", "$close", "$vwap", "$bid", "$ask", "$volume",
+    "$bidV", "$bidV1", "$bidV3", "$bidV5", "$askV", "$askV1", "$askV3", "$askV5"
+  ]
+  feature_columns_yesterday: [
+    "$open_1", "$high_1", "$low_1", "$close_1", "$vwap_1", "$bid_1", "$ask_1", "$volume_1",
+    "$bidV_1", "$bidV1_1", "$bidV3_1", "$bidV5_1", "$askV_1", "$askV1_1", "$askV3_1", "$askV5_1"
+  ]
+exchange:
+  limit_threshold: null
+  deal_price: ["$close", "$close"]
+  volume_threshold: null
+strategies:
+  1day:
+    class: TWAPStrategy
+    kwargs: {}
+    module_path: qlib.contrib.strategy.rule_strategy
+  30min:
+    class: TWAPStrategy
+    kwargs: {}
+    module_path: qlib.contrib.strategy.rule_strategy
+concurrency: 16
+output_dir: outputs/twap/

examples/rl_order_execution/exp_configs/train_opds.yml

@@ -0,0 +1,61 @@
+simulator:
+  data_granularity: 5
+  time_per_step: 30
+  vol_limit: null
+env:
+  concurrency: 48
+  parallel_mode: shmem
+action_interpreter:
+  class: CategoricalActionInterpreter
+  kwargs:
+    values: 4
+    max_step: 8
+  module_path: qlib.rl.order_execution.interpreter
+state_interpreter:
+  class: FullHistoryStateInterpreter
+  kwargs:
+    data_dim: 5
+    data_ticks: 48  # 48 = 240 min / 5 min
+    max_step: 8
+    processed_data_provider:
+      class: PickleProcessedDataProvider
+      module_path: qlib.rl.data.pickle_styled
+      kwargs:
+        data_dir: ./data/pickle_dataframe/feature
+  module_path: qlib.rl.order_execution.interpreter
+reward:
+  class: PAPenaltyReward
+  kwargs:
+    penalty: 4.0
+    scale: 0.01
+  module_path: qlib.rl.order_execution.reward
+data:
+  source:
+    order_dir: ./data/orders
+    data_dir: ./data/pickle_dataframe/backtest
+    total_time: 240
+    default_start_time_index: 0
+    default_end_time_index: 235
+    proc_data_dim: 5
+  num_workers: 0
+  queue_size: 20
+network:
+  class: Recurrent
+  module_path: qlib.rl.order_execution.network
+policy:
+  class: PPO  # PPO, DQN
+  kwargs:
+    lr: 0.0001
+  module_path: qlib.rl.order_execution.policy
+runtime:
+  seed: 42
+  use_cuda: false
+trainer:
+  max_epoch: 500
+  repeat_per_collect: 25
+  earlystop_patience: 50
+  episode_per_collect: 10000
+  batch_size: 1024
+  val_every_n_epoch: 4
+  checkpoint_path: ./outputs/opds
+  checkpoint_every_n_iters: 1

examples/rl_order_execution/exp_configs/train_ppo.yml

@@ -0,0 +1,62 @@
+simulator:
+  data_granularity: 5
+  time_per_step: 30
+  vol_limit: null
+env:
+  concurrency: 48
+  parallel_mode: shmem
+action_interpreter:
+  class: CategoricalActionInterpreter
+  kwargs:
+    values: 4
+    max_step: 8
+  module_path: qlib.rl.order_execution.interpreter
+state_interpreter:
+  class: FullHistoryStateInterpreter
+  kwargs:
+    data_dim: 5
+    data_ticks: 48  # 48 = 240 min / 5 min
+    max_step: 8
+    processed_data_provider:
+      class: PickleProcessedDataProvider
+      module_path: qlib.rl.data.pickle_styled
+      kwargs:
+        data_dir: ./data/pickle_dataframe/feature
+  module_path: qlib.rl.order_execution.interpreter
+reward:
+  class: PPOReward
+  kwargs:
+    max_step: 8
+    start_time_index: 0
+    end_time_index: 46  # 46 = (240 - 5) min / 5 min - 1
+  module_path: qlib.rl.order_execution.reward
+data:
+  source:
+    order_dir: ./data/orders
+    data_dir: ./data/pickle_dataframe/backtest
+    total_time: 240
+    default_start_time_index: 0
+    default_end_time_index: 235
+    proc_data_dim: 5
+  num_workers: 0
+  queue_size: 20
+network:
+  class: Recurrent
+  module_path: qlib.rl.order_execution.network
+policy:
+  class: PPO  # PPO, DQN
+  kwargs:
+    lr: 0.0001
+  module_path: qlib.rl.order_execution.policy
+runtime:
+  seed: 42
+  use_cuda: false
+trainer:
+  max_epoch: 500
+  repeat_per_collect: 25
+  earlystop_patience: 50
+  episode_per_collect: 10000
+  batch_size: 1024
+  val_every_n_epoch: 4
+  checkpoint_path: ./outputs/ppo
+  checkpoint_every_n_iters: 1

examples/rl_order_execution/scripts/collect_pickle_dataframe.py

@@ -0,0 +1,26 @@
+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+
+import os
+import pickle
+import pandas as pd
+from joblib import Parallel, delayed
+
+os.makedirs(os.path.join("data", "pickle_dataframe"), exist_ok=True)
+
+
+def _collect(df: pd.DataFrame, instrument: str, tag: str) -> None:
+    cur = df[df["instrument"] == instrument].sort_values(by=["datetime"])
+    cur = cur.set_index(["instrument", "datetime", "date"])
+    pickle.dump(cur, open(os.path.join("data", "pickle_dataframe", tag, f"{instrument}.pkl"), "wb"))
+
+
+for tag in ("backtest", "feature"):
+    df = pickle.load(open(os.path.join("data", "pickle", f"{tag}.pkl"), "rb"))
+    df = pd.concat(list(df.values())).reset_index()
+    df["date"] = df["datetime"].dt.date.astype("datetime64")
+    instruments = sorted(set(df["instrument"]))
+
+    os.makedirs(os.path.join("data", "pickle_dataframe", tag), exist_ok=True)
+
+    Parallel(n_jobs=-1, verbose=10)(delayed(_collect)(df, instrument, tag) for instrument in instruments)

examples/rl_order_execution/scripts/gen_pickle_data.py

@@ -0,0 +1,46 @@
+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+
+import yaml
+import argparse
+import os
+import shutil
+from copy import deepcopy
+
+from qlib.contrib.data.highfreq_provider import HighFreqProvider
+
+loader = yaml.FullLoader
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("-c", "--config", type=str, default="config.yml")
+    parser.add_argument("-d", "--dest", type=str, default=".")
+    parser.add_argument("-s", "--split", type=str, choices=["none", "date", "stock", "both"], default="stock")
+    args = parser.parse_args()
+
+    conf = yaml.load(open(args.config), Loader=loader)
+
+    for k, v in conf.items():
+        if isinstance(v, dict) and "path" in v:
+            v["path"] = os.path.join(args.dest, v["path"])
+    provider = HighFreqProvider(**conf)
+
+    # Gen dataframe
+    if "feature_conf" in conf:
+        feature = provider._gen_dataframe(deepcopy(provider.feature_conf))
+    if "backtest_conf" in conf:
+        backtest = provider._gen_dataframe(deepcopy(provider.backtest_conf))
+
+    provider.feature_conf["path"] = os.path.splitext(provider.feature_conf["path"])[0] + "/"
+    provider.backtest_conf["path"] = os.path.splitext(provider.backtest_conf["path"])[0] + "/"
+    # Split by date
+    if args.split == "date" or args.split == "both":
+        provider._gen_day_dataset(deepcopy(provider.feature_conf), "feature")
+        provider._gen_day_dataset(deepcopy(provider.backtest_conf), "backtest")
+
+    # Split by stock
+    if args.split == "stock" or args.split == "both":
+        provider._gen_stock_dataset(deepcopy(provider.feature_conf), "feature")
+        provider._gen_stock_dataset(deepcopy(provider.backtest_conf), "backtest")
+
+    shutil.rmtree("stat/", ignore_errors=True)

examples/rl_order_execution/scripts/gen_training_orders.py

@@ -0,0 +1,42 @@
+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+
+import os
+import numpy as np
+import pandas as pd
+from tqdm import tqdm
+from pathlib import Path
+
+DATA_PATH = Path(os.path.join("data", "pickle_dataframe", "backtest"))
+OUTPUT_PATH = Path(os.path.join("data", "orders"))
+
+
+def generate_order(stock: str, start_idx: int, end_idx: int) -> None:
+    df = pd.read_pickle(DATA_PATH / f"{stock}.pkl")
+    df = df.groupby("date").take(range(start_idx, end_idx)).droplevel(level=0)
+    div = df["$volume0"].rolling((end_idx - start_idx) * 60).mean().shift(1).groupby(level="date").transform("first")
+
+    order_all = pd.DataFrame(df.groupby(level=(2, 0)).mean().dropna())
+    order_all["amount"] = np.random.lognormal(-3.28, 1.14) * order_all["$volume0"]
+    order_all = order_all[order_all["amount"] > 0.0]
+    order_all["order_type"] = 0
+    order_all = order_all.drop(columns=["$volume0"])
+
+    order_train = order_all[order_all.index.get_level_values(0) <= pd.Timestamp("2021-06-30")]
+    order_test = order_all[order_all.index.get_level_values(0) > pd.Timestamp("2021-06-30")]
+    order_valid = order_test[order_test.index.get_level_values(0) <= pd.Timestamp("2021-09-30")]
+    order_test = order_test[order_test.index.get_level_values(0) > pd.Timestamp("2021-09-30")]
+
+    for order, tag in zip((order_train, order_valid, order_test, order_all), ("train", "valid", "test", "all")):
+        path = OUTPUT_PATH / tag
+        os.makedirs(path, exist_ok=True)
+        if len(order) > 0:
+            order.to_pickle(path / f"{stock}.pkl.target")
+
+
+np.random.seed(1234)
+file_list = sorted(os.listdir(DATA_PATH))
+stocks = [f.replace(".pkl", "") for f in file_list]
+stocks = sorted(np.random.choice(stocks, size=100, replace=False))
+for stock in tqdm(stocks):
+    generate_order(stock, 0, 240 // 5 - 1)

examples/rl_order_execution/scripts/merge_orders.py

@@ -0,0 +1,15 @@
+import pickle
+import os
+import pandas as pd
+from tqdm import tqdm
+
+for tag in ["test", "valid"]:
+    files = os.listdir(os.path.join("data/orders/", tag))
+    dfs = []
+    for f in tqdm(files):
+        df = pickle.load(open(os.path.join("data/orders/", tag, f), "rb"))
+        df = df.drop(["$close0"], axis=1)
+        dfs.append(df)
+
+    total_df = pd.concat(dfs)
+    pickle.dump(total_df, open(os.path.join("data", "orders", f"{tag}_orders.pkl"), "wb"))

examples/rl_order_execution/scripts/pickle_data_config.yml

@@ -0,0 +1,77 @@
+# start & end time for training/validation/test datasets
+start_time: !!str &start 2020-01-01
+end_time: !!str &end 2021-12-31
+train_end_time: !!str &tend 2021-06-30
+valid_start_time: !!str &vstart 2021-07-01
+valid_end_time: !!str &vend 2021-09-30
+test_start_time: !!str &tstart 2021-10-01
+# the instrument set
+instruments: &ins csi300s19_22
+# qlib related configuration
+qlib_conf:
+    provider_uri: 
+        5min: ./data/bin # path to generated qlib bin
+    redis_port: 233
+feature_conf:
+    path: ./data/pickle/feature.pkl # output path of feature
+    class: DatasetH
+    module_path: qlib.data.dataset
+    kwargs:
+        handler:
+            class: HighFreqGeneralHandler
+            module_path: qlib.contrib.data.highfreq_handler
+            kwargs:
+                start_time: *start
+                end_time: *end
+                fit_start_time: *start
+                fit_end_time: *tend
+                instruments: *ins
+                day_length: 240 # how many minutes in one trading day
+                freq: 5min
+                columns: ["$open", "$high", "$low", "$close"]
+                infer_processors:
+                - class: HighFreqNorm
+                  module_path: qlib.contrib.data.highfreq_processor
+                  kwargs:
+                    feature_save_dir: ./stat/  #  output path of statistics of features (for feature normalization)
+                    norm_groups: 
+                        price: 8
+                        volume: 2
+                inst_processors:
+                - class: TimeRangeFlt
+                  module_path: qlib.data.dataset.processor
+                  kwargs:
+                    start_time: "2020-01-01"
+                    end_time: "2021-12-31"
+                    freq: 5min
+        segments:
+            train: !!python/tuple [*start, *tend]
+            valid: !!python/tuple [*vstart, *vend]
+            test: !!python/tuple [*tstart, *end]
+backtest_conf:
+    path: ./data/pickle/backtest.pkl # output path of backtest
+    class: DatasetH
+    module_path: qlib.data.dataset
+    kwargs:
+        handler:
+            class: HighFreqGeneralBacktestHandler
+            module_path: qlib.contrib.data.highfreq_handler
+            kwargs:
+                start_time: *start
+                end_time: *end
+                instruments: *ins
+                day_length: 240
+                freq: 5min
+                columns: ["$close", "$volume"]
+                inst_processors:
+                - class: TimeRangeFlt
+                  module_path: qlib.data.dataset.processor
+                  kwargs:
+                    start_time: "2020-01-01"
+                    end_time: "2021-12-31"
+                    freq: 5min
+        segments:
+            train: !!python/tuple [*start, *tend]
+            valid: !!python/tuple [*vstart, *vend]
+            test: !!python/tuple [*tstart, *end]
+freq: 5min

qlib/contrib/data/handler.py

@@ -56,7 +56,7 @@ class Alpha360(DataHandlerLP):
         fit_start_time=None,
         fit_end_time=None,
         filter_pipe=None,
-        inst_processor=None,
+        inst_processors=None,
         **kwargs
     ):
         infer_processors = check_transform_proc(infer_processors, fit_start_time, fit_end_time)
@@ -71,7 +71,7 @@ class Alpha360(DataHandlerLP):
                 },
                 "filter_pipe": filter_pipe,
                 "freq": freq,
-                "inst_processor": inst_processor,
+                "inst_processors": inst_processors,
             },
         }
 
@@ -152,7 +152,7 @@ class Alpha158(DataHandlerLP):
         fit_end_time=None,
         process_type=DataHandlerLP.PTYPE_A,
         filter_pipe=None,
-        inst_processor=None,
+        inst_processors=None,
         **kwargs
     ):
         infer_processors = check_transform_proc(infer_processors, fit_start_time, fit_end_time)
@@ -167,7 +167,7 @@ class Alpha158(DataHandlerLP):
                 },
                 "filter_pipe": filter_pipe,
                 "freq": freq,
-                "inst_processor": inst_processor,
+                "inst_processors": inst_processors,
             },
         }
         super().__init__(

qlib/contrib/data/highfreq_handler.py

@@ -115,6 +115,7 @@ class HighFreqGeneralHandler(DataHandlerLP):
         day_length=240,
         freq="1min",
         columns=["$open", "$high", "$low", "$close", "$vwap"],
+        inst_processors=None,
     ):
         self.day_length = day_length
         self.columns = columns
@@ -128,6 +129,7 @@ class HighFreqGeneralHandler(DataHandlerLP):
                 "config": self.get_feature_config(),
                 "swap_level": False,
                 "freq": freq,
+                "inst_processors": inst_processors,
             },
         }
         super().__init__(
@@ -257,6 +259,7 @@ class HighFreqGeneralBacktestHandler(DataHandler):
         day_length=240,
         freq="1min",
         columns=["$close", "$vwap", "$volume"],
+        inst_processors=None,
     ):
         self.day_length = day_length
         self.columns = set(columns)
@@ -266,6 +269,7 @@ class HighFreqGeneralBacktestHandler(DataHandler):
                 "config": self.get_feature_config(),
                 "swap_level": False,
                 "freq": freq,
+                "inst_processors": inst_processors,
             },
         }
         super().__init__(
@@ -311,6 +315,7 @@ class HighFreqOrderHandler(DataHandlerLP):
         learn_processors=[],
         fit_start_time=None,
         fit_end_time=None,
+        inst_processors=None,
         drop_raw=True,
     ):
 
@@ -323,6 +328,7 @@ class HighFreqOrderHandler(DataHandlerLP):
                 "config": self.get_feature_config(),
                 "swap_level": False,
                 "freq": "1min",
+                "inst_processors": inst_processors,
             },
         }
         super().__init__(

qlib/contrib/data/highfreq_provider.py

@@ -128,7 +128,7 @@ class HighFreqProvider:
             raise ValueError("Must specify the path to save the dataset.") from e
         if os.path.isfile(path):
             start = time.time()
-            self.logger.info("Dataset exists, load from disk.", __name__)
+            self.logger.info(f"[{__name__}]Dataset exists, load from disk.")
 
             # res = dataset.prepare(['train', 'valid', 'test'])
             with open(path, "rb") as f:
@@ -137,11 +137,11 @@ class HighFreqProvider:
                 res = [data[i] for i in datasets]
             else:
                 res = data.prepare(datasets)
-            self.logger.info(f"Data loaded, time cost: {time.time() - start:.2f}", __name__)
+            self.logger.info(f"[{__name__}]Data loaded, time cost: {time.time() - start:.2f}")
         else:
             if not os.path.exists(os.path.dirname(path)):
                 os.makedirs(os.path.dirname(path))
-            self.logger.info("Generating dataset", __name__)
+            self.logger.info(f"[{__name__}]Generating dataset")
             start_time = time.time()
             self._prepare_calender_cache()
             dataset = init_instance_by_config(config)
@@ -160,7 +160,7 @@ class HighFreqProvider:
             with open(path[:-4] + "test.pkl", "wb") as f:
                 pkl.dump(testset, f)
             res = [data[i] for i in datasets]
-            self.logger.info(f"Data generated, time cost: {(time.time() - start_time):.2f}", __name__)
+            self.logger.info(f"[{__name__}]Data generated, time cost: {(time.time() - start_time):.2f}")
         return res
 
     def _gen_data(self, config, datasets=["train", "valid", "test"]):
@@ -170,7 +170,7 @@ class HighFreqProvider:
             raise ValueError("Must specify the path to save the dataset.") from e
         if os.path.isfile(path):
             start = time.time()
-            self.logger.info("Dataset exists, load from disk.", __name__)
+            self.logger.info(f"[{__name__}]Dataset exists, load from disk.")
 
             # res = dataset.prepare(['train', 'valid', 'test'])
             with open(path, "rb") as f:
@@ -179,18 +179,18 @@ class HighFreqProvider:
                 res = [data[i] for i in datasets]
             else:
                 res = data.prepare(datasets)
-            self.logger.info(f"Data loaded, time cost: {time.time() - start:.2f}", __name__)
+            self.logger.info(f"[{__name__}]Data loaded, time cost: {time.time() - start:.2f}")
         else:
             if not os.path.exists(os.path.dirname(path)):
                 os.makedirs(os.path.dirname(path))
-            self.logger.info("Generating dataset", __name__)
+            self.logger.info(f"[{__name__}]Generating dataset")
             start_time = time.time()
             self._prepare_calender_cache()
             dataset = init_instance_by_config(config)
             dataset.config(dump_all=True, recursive=True)
             dataset.to_pickle(path)
             res = dataset.prepare(datasets)
-            self.logger.info(f"Data generated, time cost: {(time.time() - start_time):.2f}", __name__)
+            self.logger.info(f"[{__name__}]Data generated, time cost: {(time.time() - start_time):.2f}")
         return res
 
     def _gen_dataset(self, config):
@@ -200,21 +200,21 @@ class HighFreqProvider:
             raise ValueError("Must specify the path to save the dataset.") from e
         if os.path.isfile(path):
             start = time.time()
-            self.logger.info("Dataset exists, load from disk.", __name__)
+            self.logger.info(f"[{__name__}]Dataset exists, load from disk.")
 
             with open(path, "rb") as f:
                 dataset = pkl.load(f)
-            self.logger.info(f"Data loaded, time cost: {time.time() - start:.2f}", __name__)
+            self.logger.info(f"[{__name__}]Data loaded, time cost: {time.time() - start:.2f}")
         else:
             start = time.time()
             if not os.path.exists(os.path.dirname(path)):
                 os.makedirs(os.path.dirname(path))
-            self.logger.info("Generating dataset", __name__)
+            self.logger.info(f"[{__name__}]Generating dataset")
             self._prepare_calender_cache()
             dataset = init_instance_by_config(config)
-            self.logger.info(f"Dataset init, time cost: {time.time() - start:.2f}", __name__)
+            self.logger.info(f"[{__name__}]Dataset init, time cost: {time.time() - start:.2f}")
             dataset.prepare(["train", "valid", "test"])
-            self.logger.info(f"Dataset prepared, time cost: {time.time() - start:.2f}", __name__)
+            self.logger.info(f"[{__name__}]Dataset prepared, time cost: {time.time() - start:.2f}")
             dataset.config(dump_all=True, recursive=True)
             dataset.to_pickle(path)
         return dataset
@@ -227,15 +227,15 @@ class HighFreqProvider:
 
         if os.path.isfile(path + "tmp_dataset.pkl"):
             start = time.time()
-            self.logger.info("Dataset exists, load from disk.", __name__)
+            self.logger.info(f"[{__name__}]Dataset exists, load from disk.")
         else:
             start = time.time()
             if not os.path.exists(os.path.dirname(path)):
                 os.makedirs(os.path.dirname(path))
-            self.logger.info("Generating dataset", __name__)
+            self.logger.info(f"[{__name__}]Generating dataset")
             self._prepare_calender_cache()
             dataset = init_instance_by_config(config)
-            self.logger.info(f"Dataset init, time cost: {time.time() - start:.2f}", __name__)
+            self.logger.info(f"[{__name__}]Dataset init, time cost: {time.time() - start:.2f}")
             dataset.config(dump_all=False, recursive=True)
             dataset.to_pickle(path + "tmp_dataset.pkl")
 
@@ -268,15 +268,15 @@ class HighFreqProvider:
 
         if os.path.isfile(path + "tmp_dataset.pkl"):
             start = time.time()
-            self.logger.info("Dataset exists, load from disk.", __name__)
+            self.logger.info(f"[{__name__}]Dataset exists, load from disk.")
         else:
             start = time.time()
             if not os.path.exists(os.path.dirname(path)):
                 os.makedirs(os.path.dirname(path))
-            self.logger.info("Generating dataset", __name__)
+            self.logger.info(f"[{__name__}]Generating dataset")
             self._prepare_calender_cache()
             dataset = init_instance_by_config(config)
-            self.logger.info(f"Dataset init, time cost: {time.time() - start:.2f}", __name__)
+            self.logger.info(f"[{__name__}]Dataset init, time cost: {time.time() - start:.2f}")
             dataset.config(dump_all=False, recursive=True)
             dataset.to_pickle(path + "tmp_dataset.pkl")
 

qlib/data/dataset/loader.py

@@ -153,7 +153,7 @@ class QlibDataLoader(DLWParser):
         filter_pipe: List = None,
         swap_level: bool = True,
         freq: Union[str, dict] = "day",
-        inst_processor: dict = None,
+        inst_processors: Union[dict, list] = None,
     ):
         """
         Parameters
@@ -167,16 +167,19 @@ class QlibDataLoader(DLWParser):
         freq:  dict or str
             If type(config) == dict and type(freq) == str, load config data using freq.
             If type(config) == dict and type(freq) == dict, load config[<group_name>] data using freq[<group_name>]
-        inst_processor: dict
-            If inst_processor is not None and type(config) == dict; load config[<group_name>] data using inst_processor[<group_name>]
+        inst_processors: dict | list
+            If inst_processors is not None and type(config) == dict; load config[<group_name>] data using inst_processors[<group_name>]
+            If inst_processors is a list, then it will be applied to all groups.
         """
         self.filter_pipe = filter_pipe
         self.swap_level = swap_level
         self.freq = freq
 
         # sample
-        self.inst_processor = inst_processor if inst_processor is not None else {}
-        assert isinstance(self.inst_processor, dict), f"inst_processor(={self.inst_processor}) must be dict"
+        self.inst_processors = inst_processors if inst_processors is not None else {}
+        assert isinstance(
+            self.inst_processors, (dict, list)
+        ), f"inst_processors(={self.inst_processors}) must be dict or list"
 
         super().__init__(config)
 
@@ -187,8 +190,8 @@ class QlibDataLoader(DLWParser):
                     if _gp not in freq:
                         raise ValueError(f"freq(={freq}) missing group(={_gp})")
                 assert (
-                    self.inst_processor
-                ), f"freq(={self.freq}), inst_processor(={self.inst_processor}) cannot be None/empty"
+                    self.inst_processors
+                ), f"freq(={self.freq}), inst_processors(={self.inst_processors}) cannot be None/empty"
 
     def load_group_df(
         self,
@@ -208,9 +211,10 @@ class QlibDataLoader(DLWParser):
             warnings.warn("`filter_pipe` is not None, but it will not be used with `instruments` as list")
 
         freq = self.freq[gp_name] if isinstance(self.freq, dict) else self.freq
-        df = D.features(
-            instruments, exprs, start_time, end_time, freq=freq, inst_processors=self.inst_processor.get(gp_name, [])
+        inst_processors = (
+            self.inst_processors if isinstance(self.inst_processors, list) else self.inst_processors.get(gp_name, [])
         )
+        df = D.features(instruments, exprs, start_time, end_time, freq=freq, inst_processors=inst_processors)
         df.columns = names
         if self.swap_level:
             df = df.swaplevel().sort_index()  # NOTE: if swaplevel, return <datetime, instrument>

qlib/data/dataset/processor.py

@@ -2,7 +2,7 @@
 # Licensed under the MIT License.
 
 import abc
-from typing import Union, Text
+from typing import Union, Text, Optional
 import numpy as np
 import pandas as pd
 
@@ -11,6 +11,8 @@ from ...constant import EPS
 from .utils import fetch_df_by_index
 from ...utils.serial import Serializable
 from ...utils.paral import datetime_groupby_apply
+from qlib.data.inst_processor import InstProcessor
+from qlib.data import D
 
 
 def get_group_columns(df: pd.DataFrame, group: Union[Text, None]):
@@ -378,3 +380,42 @@ class HashStockFormat(Processor):
         from .storage import HashingStockStorage  # pylint: disable=C0415
 
         return HashingStockStorage.from_df(df)
+
+
+class TimeRangeFlt(InstProcessor):
+    """
+    This is a filter to filter stock.
+    Only keep the data that exist from start_time to end_time (the existence in the middle is not checked.)
+    WARNING:  It may induce leakage!!!
+    """
+
+    def __init__(
+        self,
+        start_time: Optional[Union[pd.Timestamp, str]] = None,
+        end_time: Optional[Union[pd.Timestamp, str]] = None,
+        freq: str = "day",
+    ):
+        """
+        Parameters
+        ----------
+        start_time : Optional[Union[pd.Timestamp, str]]
+            The data must start earlier (or equal) than `start_time`
+            None indicates data will not be filtered based on `start_time`
+        end_time : Optional[Union[pd.Timestamp, str]]
+            similar to start_time
+        freq : str
+            The frequency of the calendar
+        """
+        # Align to calendar before filtering
+        cal = D.calendar(start_time=start_time, end_time=end_time, freq=freq)
+        self.start_time = None if start_time is None else cal[0]
+        self.end_time = None if end_time is None else cal[-1]
+
+    def __call__(self, df: pd.DataFrame, instrument, *args, **kwargs):
+        if (
+            df.empty
+            or (self.start_time is None or df.index.min() <= self.start_time)
+            and (self.end_time is None or df.index.max() >= self.end_time)
+        ):
+            return df
+        return df.head(0)

qlib/rl/contrib/backtest.py

@@ -357,7 +357,10 @@ def backtest(backtest_config: dict, with_simulator: bool = False) -> pd.DataFram
 
     if not output_path.exists():
         os.makedirs(output_path)
-    res.to_csv(output_path / "summary.csv")
+
+    if "pa" in res.columns:
+        res["pa"] = res["pa"] * 10000.0  # align with training metrics
+    res.to_csv(output_path / "backtest_result.csv")
     return res
 
 

qlib/rl/order_execution/policy.py

@@ -12,11 +12,11 @@ import torch
 import torch.nn as nn
 from gym.spaces import Discrete
 from tianshou.data import Batch, ReplayBuffer, to_torch
-from tianshou.policy import BasePolicy, PPOPolicy
+from tianshou.policy import BasePolicy, PPOPolicy, DQNPolicy
 
 from qlib.rl.trainer.trainer import Trainer
 
-__all__ = ["AllOne", "PPO"]
+__all__ = ["AllOne", "PPO", "DQN"]
 
 
 # baselines #
@@ -158,6 +158,56 @@ class PPO(PPOPolicy):
             set_weight(self, Trainer.get_policy_state_dict(weight_file))
 
 
+DQNModel = PPOActor  # Reuse PPOActor.
+
+
+class DQN(DQNPolicy):
+    """A wrapper of tianshou DQNPolicy.
+
+    Differences:
+
+    - Auto-create model network. Supports discrete action space only.
+    - Support a ``weight_file`` that supports loading checkpoint.
+    """
+
+    def __init__(
+        self,
+        network: nn.Module,
+        obs_space: gym.Space,
+        action_space: gym.Space,
+        lr: float,
+        weight_decay: float = 0.0,
+        discount_factor: float = 0.99,
+        estimation_step: int = 1,
+        target_update_freq: int = 0,
+        reward_normalization: bool = False,
+        is_double: bool = True,
+        clip_loss_grad: bool = False,
+        weight_file: Optional[Path] = None,
+    ) -> None:
+        assert isinstance(action_space, Discrete)
+
+        model = DQNModel(network, action_space.n)
+        optimizer = torch.optim.Adam(
+            model.parameters(),
+            lr=lr,
+            weight_decay=weight_decay,
+        )
+
+        super().__init__(
+            model,
+            optimizer,
+            discount_factor=discount_factor,
+            estimation_step=estimation_step,
+            target_update_freq=target_update_freq,
+            reward_normalization=reward_normalization,
+            is_double=is_double,
+            clip_loss_grad=clip_loss_grad,
+        )
+        if weight_file is not None:
+            set_weight(self, Trainer.get_policy_state_dict(weight_file))
+
+
 # utilities: these should be put in a separate (common) file. #
 
 

qlib/rl/order_execution/reward.py

@@ -70,7 +70,19 @@ class PPOReward(Reward[SAOEState]):
 
     def reward(self, simulator_state: SAOEState) -> float:
         if simulator_state.cur_step == self.max_step - 1 or simulator_state.position < 1e-6:
-            vwap_price = cast(dict, simulator_state.metrics)["trade_price"]
+            if simulator_state.history_exec["deal_amount"].sum() == 0.0:
+                vwap_price = cast(
+                    float,
+                    np.average(simulator_state.history_exec["market_price"]),
+                )
+            else:
+                vwap_price = cast(
+                    float,
+                    np.average(
+                        simulator_state.history_exec["market_price"],
+                        weights=simulator_state.history_exec["deal_amount"],
+                    ),
+                )
             twap_price = simulator_state.backtest_data.get_deal_price().mean()
 
             if simulator_state.order.direction == OrderDir.SELL:

qlib/rl/order_execution/strategy.py

@@ -7,6 +7,7 @@ import collections
 from types import GeneratorType
 from typing import Any, Callable, cast, Dict, Generator, List, Optional, Tuple, Union
 
+import warnings
 import numpy as np
 import pandas as pd
 import torch
@@ -137,7 +138,12 @@ class SAOEStateAdapter:
             exec_vol[idx - last_step_range[0]] = order.deal_amount
 
         if exec_vol.sum() > self.position and exec_vol.sum() > 0.0:
-            assert exec_vol.sum() < self.position + 1, f"{exec_vol} too large"
+            if exec_vol.sum() > self.position + 1.0:
+                warnings.warn(
+                    f"Sum of execution volume is {exec_vol.sum()} which is larger than "
+                    f"position + 1.0 = {self.position} + 1.0 = {self.position + 1.0}. "
+                    f"All execution volume is scaled down linearly to ensure that their sum does not position."
+                )
             exec_vol *= self.position / (exec_vol.sum())
 
         market_volume = cast(