Sausage Links

Implementation of the Sausage Links algorithm base Swinging Door in Python.

Example of usage

>>> from datetime import datetime
>>> from pandas import read_csv, DataFrame

>>> df = DataFrame(
...     [
...         {
...             "Date": datetime.strptime(date, "%Y-%m-%d"),
...             "Price": value
...         }
...         for date, value in read_csv(
...             "https://datahub.io/core/oil-prices/r/wti-daily.csv"
...         ).values.tolist()
...     ]
... )

>>> print(len(df))
9895

>>> df.plot(x="Date", y="Price")

>>> from sausage_links import sausage_links

>>> compress = DataFrame(
...      list(
...         {
...             "Date": datetime.fromtimestamp(date),
...             "Price": value
...         }
...         for date, value in sausage_links(
...             iter(
...                 (date.timestamp(), value)
...                 for date, value in df.values.tolist()
...             ), deviation=(1, .5), max_len=604_800,
...             auto_dev_factor=200_000, ema_alpha=0.5
...         )
...     )
... )

>>> print(len(compress))
4177

>>> compress.plot(x="Date", y="Price")

Source

sausage_links

Implementation of the Sausage Links algorithm base Swinging Door in Python.

sausage_links._sloping_calc(stretch: Stretch, deviation: Deviation) → Slopings

Calculate slopings upper and lower.

Parameters:

• stretch (Stretch) – stretch;

• deviation (Deviation) – compression deflection.

Return type:

Slopings

Returns:

Slopings upper and lower.

Raises:

ValueError – if division by 0 occurs during slope calculation.

>>> _sloping_calc(((1, 6), (2, 6.5)), 1)
(1.5, -0.5)

>>> _sloping_calc(((1, 6), (2, 6.5)), (1, 1))
(1.5, -0.5)

>>> _sloping_calc(((1, 6), (1, 6.5)), 1)
Traceback (most recent call last):
    ...
ValueError: The division by 0 occurs during the calculation of the slope.

sausage_links.sausage_links(source: Source, deviation: Deviation = 0.1, max_len: Number = 0, auto_dev_factor: Number = 0, ema_alpha: Number = 0.3) → Generator[Point, None, None]

Implementation of the Sausage Links algorithm base Swinging Door in Python.

Parameters:

• source (Source) – source data;

• deviation (Deviation) – compression deflection;

• max_len (Number) – maximum corridor length;

• auto_dev_factor (Number) – multiplier for EMA;

• ema_alpha (Number) – smoothing coefficient for EMA. [0; 1].

Return type:

Generator[Point, None, None]

Returns:

Compressed data.

>>> list(sausage_links(iter([
...     (1, 6), (2, 6.5), (3, 5.5),
...     (4, 6.5), (5, 8), (6, 7.5),
...     (7, 8), (8, 9.5),
... ]), 1))
[(1, 6), (7, 8), (8, 9.5)]

>>> list(sausage_links(iter([
...     (1, 6), (2, 6.5), (3, 5.5),
...     (4, 6.5), (5, 8), (6, 7.5),
...     (7, 8), (8, 6),
... ]), 1))
[(1, 6), (7, 8), (8, 6)]

>>> list(sausage_links(iter([
...     (1, 6), (2, 6.5), (3, 5.5),
...     (4, 6.5), (5, 8), (6, 7.5),
...     (7, 8),
... ]), 1, 1))
[(1, 6), (2, 6.5), (3, 5.5), (4, 6.5), (5, 8), (6, 7.5), (7, 8)]

>>> list(sausage_links(iter([
...     (1, 6), (2, 6.5), (3, 5.5),
...     (4, 6.5), (5, 8), (6, 7.5),
...     (7, 9.5), (8, 8),
... ]), 1, auto_dev_factor=1))
[(1, 6), (4, 6.5), (5, 8), (8, 8)]

>>> list(sausage_links(iter([
...     (1, 6), (2, 6.5), (3, 5.5),
...     (4, 6.5), (5, 8), (6, 7.5),
...     (7, 3), (8, 8),
... ]), 1, auto_dev_factor=1))
[(1, 6), (4, 6.5), (5, 8), (6, 7.5), (7, 3), (8, 8)]

>>> list(sausage_links(iter([
...     (1, 6), (2, 6.5), (3, 5.5),
...     (4, 6.5), (5, 8), (6, 7.5),
...     (7, 8),
... ]), 0))
[(1, 6), (2, 6.5), (3, 5.5), (4, 6.5), (5, 8), (6, 7.5), (7, 8)]

>>> list(sausage_links(iter([]), 1))
[]

>>> list(sausage_links(iter([(1, 6),]), 1))
[(1, 6)]

>>> list(sausage_links(iter([(1, 6),(1, 6.5),]), 1))
Traceback (most recent call last):
    ...
ValueError: The division by 0 occurs during the calculation of the slope.

Indices and tables

• Index

• Module Index

• Search Page