Types

Shared data types used across all pymlfinance modules.

Type	Description
TickData	A single market tick with timestamp, price, and volume
OhlcvBar	An OHLCV bar with VWAP
TripleBarrierConfig	Configuration for the triple-barrier labeling method
Event	A labeled event from the triple-barrier method
TrendScanResult	Result of trend scanning label detection
DrawdownResult	Result of drawdown analysis
CscvResult	Result of Combinatorially Symmetric Cross-Validation
KMeansResult	Result of K-means clustering
OncResult	Result of Optimal Number of Clusters analysis
AllocationComparison	Comparison of HRP, CLA, and IVP allocations
BootstrapComparison	Comparison of sequential vs. standard bootstrap uniqueness
FoldIndices	Train/test split indices for a cross-validation fold

All types are available at the root level:

import pymlfinance as ml

tick = ml.TickData(timestamp=1000, price=100.0, volume=50.0)
config = ml.TripleBarrierConfig(profit_taking=2.0, stop_loss=2.0, max_holding=20)

pymlfinance

TickData

A single market tick with timestamp, price, and volume.

Parameters:

Name	Type	Description	Default
timestamp	float	Unix timestamp in seconds (fractional seconds supported).	required
price	float	Trade price.	required
volume	float	Trade volume.	required

Examples:

>>> tick = TickData(1609459200.0, 100.5, 10.0)
>>> tick.price
100.5

OhlcvBar

An OHLCV bar (Open, High, Low, Close, Volume) with VWAP.

Parameters:

Name	Type	Description	Default
timestamp	float	Unix timestamp of the bar open.	required
open	float	Opening price.	required
high	float	Highest price during the bar.	required
low	float	Lowest price during the bar.	required
close	float	Closing price.	required
volume	float	Total volume traded.	required
vwap	float	Volume-weighted average price (default 0.0).	None

Attributes:

Name	Type	Description
mid_price()	float	(high + low) / 2
typical_price()	float	(high + low + close) / 3
dollar_volume()	float	vwap * volume

dollar_volume

dollar_volume()

Return the dollar volume: vwap * volume.

mid_price

mid_price()

Return the mid price: (high + low) / 2.

typical_price

typical_price()

Return the typical price: (high + low + close) / 3.

TripleBarrierConfig

Configuration for the triple-barrier labeling method.

Defines the upper profit-taking barrier, lower stop-loss barrier, and maximum holding period for event labeling (AFML Ch. 3).

Parameters:

Name	Type	Description	Default
upper_barrier	float	Profit-taking threshold as a fraction of daily volatility.	None
lower_barrier	float	Stop-loss threshold as a fraction of daily volatility.	None
max_holding_period	int	Maximum number of bars before forced exit (vertical barrier).	None

Examples:

>>> config = TripleBarrierConfig(upper_barrier=2.0, lower_barrier=2.0, max_holding_period=10)

Event

A labeled event from the triple-barrier method.

Attributes:

Name	Type	Description
entry_idx	int	Index of the entry bar.
exit_idx	int	Index of the exit bar.
touch_type	str	Which barrier was touched first: "upper", "lower", or "vertical".
return_value	float	Return from entry to exit.

TrendScanResult

Result of trend scanning label detection.

Attributes:

Name	Type	Description
t_stat	float	t-statistic of the best linear fit.
label	int	Trend direction: +1 (up), -1 (down), or 0 (no trend).
best_window	int	Look-ahead window that produced the highest |t-stat|.
r_squared	float	R-squared of the best linear fit.

DrawdownResult

Result of drawdown analysis on a return series.

Attributes:

Name	Type	Description
max_drawdown	float	Maximum peak-to-trough decline.
max_drawdown_duration	int	Longest drawdown duration in bars.
drawdown_series	list[float]	Per-bar drawdown values.
time_under_water	list[int]	Per-bar count of consecutive bars in drawdown.

CscvResult

Result of Combinatorially Symmetric Cross-Validation (CSCV).

Attributes:

Name	Type	Description
pbo	float	Probability of Backtest Overfitting.
rank_logits	list[float]	Log-odds of each strategy's rank degradation.

KMeansResult

Result of K-means clustering.

Attributes:

Name	Type	Description
labels	list[int]	Cluster assignment for each data point.
n_iterations	int	Number of iterations until convergence.
centroids	list[list[float]]	Cluster centroids (k x n_features).

OncResult

Result of Optimal Number of Clusters (ONC) analysis.

Attributes:

Name	Type	Description
labels	list[int]	Cluster assignment for each feature/variable.
silhouette	float	Silhouette score of the optimal clustering.
n_clusters	int	Optimal number of clusters found.

AllocationComparison

Comparison of HRP, CLA, and IVP portfolio allocations.

Attributes:

Name	Type	Description
hrp_sharpe	float	Out-of-sample Sharpe ratio for HRP.
cla_sharpe	float	Out-of-sample Sharpe ratio for CLA (min-variance).
ivp_sharpe	float	Out-of-sample Sharpe ratio for Inverse Variance.
hrp_variance	float	Out-of-sample portfolio variance for HRP.
cla_variance	float	Out-of-sample portfolio variance for CLA.
ivp_variance	float	Out-of-sample portfolio variance for IVP.

BootstrapComparison

Comparison of sequential vs. standard bootstrap uniqueness.

Attributes:

Name	Type	Description
seq_uniqueness	float	Average uniqueness from sequential bootstrap.
std_uniqueness	float	Average uniqueness from standard (IID) bootstrap.

FoldIndices

Train/test split indices for a single cross-validation fold.

Attributes:

Name	Type	Description
train	list[int]	Indices of training samples.
test	list[int]	Indices of test samples.