added method for generation data at microsec frequency

Author: wegar-2

moddata/__init__.py

@@ -1,5 +1,7 @@
 from moddata._utils import load_data
+from moddata.dg.dg import make_milisec_data
 
 __all__ = [
-    "load_data"
+    "load_data",
+    "make_milisec_data"
 ]

moddata/dg/__init__.py

@@ -0,0 +1,46 @@
+from datetime import date, datetime, time, timedelta
+
+import numpy as np
+import pandas as pd
+
+from moddata.src.constants import MICRO_SECS_PER_DAY
+
+
+def make_milisec_data(
+        inter_arrival_time_milisecs: float = 50,
+        start_value: float = 5.00,
+        daily_sigma: float = 0.02,
+        day: date = date(2021, 1, 1),
+        seed: int = 123_456
+):
+    """
+    Generate 1 day of data using GBM without drift.
+    Arrivals have exponential distribution.
+
+    :param inter_arrival_time_milisecs:
+    :param start_value: starting value of the GBM process
+    :param daily_sigma: standard deviation of the GBM for 1d period
+    :param day: date to use in the simulation
+    :param seed: random seed
+    :return: pd.DataFrame with columns quote_time and price
+    """
+    dt: datetime = datetime.combine(day, time(0, 0, 0, 0))
+    time_grid: list[datetime] = [dt]
+    prices: list[float] = [start_value]
+    np.random.seed(seed)
+
+    while dt.date() == day:
+        time_diff = int(np.random.exponential(
+            scale=inter_arrival_time_milisecs)) + 1
+        dt += timedelta(milliseconds=time_diff)
+        time_grid.append(dt)
+        prices.append(
+            prices[-1] * (
+                    1 +
+                    np.random.standard_normal(size=1)[0] *
+                    np.sqrt(time_diff / MICRO_SECS_PER_DAY) *
+                    daily_sigma
+            )
+        )
+
+    return pd.DataFrame(data={"quote_time": time_grid, "price": prices})

moddata/dg/dg.py

@@ -5,17 +5,20 @@
 """
 
 
-from typing import TypeAlias, Literal
+from typing import TypeAlias, Literal, Final
 
 import pandas as pd
 
 EncodingAndScalingModelType: TypeAlias = Literal[
     "tree_like",
-    "other"
+    "other",
+    "MICRO_SECS_PER_DAY"
 ]
 
 XyDataFrames: TypeAlias = tuple[pd.DataFrame, pd.DataFrame]
 
 TrainTestXyDataFrames: TypeAlias = (
     tuple[pd.DataFrame, pd.DataFrame, pd.DataFrame, pd.DataFrame]
 )
+
+MICRO_SECS_PER_DAY: Final[int] = 24 * 60 * 60 * 1_000