One-import ASRQuant API¶

ASRQuant v1.2.0 preserves the one-import public contract introduced in 1.0.0:

import asrquant as asr

Users do not need to import plotting, machine-learning, econometric, numerical or table libraries directly for normal workflows. ASRQuant installs and calls validated engines internally while owning the public names, argument normalization, result objects and reproducibility metadata.

Data¶

lab = asr.open_lab("prices.csv", date_column="Date")
remote = asr.open_lab(provider="yahoo", symbols=["SPY", "QQQ"], start="2020-01-01")
frame = asr.frame({"SPY": [100, 101]}, index=asr.date_range("2026-01-01", periods=2))

Visualizations¶

result = lab.backtest("sma", fast=20, slow=100, costs_bps=5)
asr.show(result, kind="dashboard")
asr.save(result, "equity.png", kind="equity", dpi=180)

asr.visualize(...) returns a PlotHandle with show(), save() and close(). PlotHandle.raw exists only for exceptional advanced backend access.

Machine learning¶

model = asr.models.random_forest(
    task="regression", trees=500, depth=6, seed=7
)
result = lab.ml(
    model, train_size=504, test_size=63, gap=5
)

A model name can replace the model object:

result = lab.ml(
    "ridge",
    train_size=504,
    test_size=63,
    model_params={"alpha": 1.0},
)

Available factories include linear regression, ridge, lasso, elastic net, logistic regression, decision trees, random forests, extra trees, gradient boosting, histogram gradient boosting, KNN, SVM, Gaussian naive Bayes, PCA, k-means and isolation forests.

Numerical functions¶

x = asr.math.linspace(0.1, 5.0, 50)
y = asr.math.normal_cdf(x)
rng = asr.math.random_generator(7)

The namespace includes arrays, grids, exponentials, logarithms, trigonometric functions, reductions, quantiles, normal distribution functions and stable logsumexp.

Scientific namespaces¶

asr.stats: econometrics and statistical inference;
asr.portfolio: portfolio construction and optimization;
asr.options: derivatives and Greeks;
asr.stochastic: stochastic processes and Monte Carlo;
asr.rates: fixed income;
asr.vol: volatility;
asr.visuals: lower-level visualization catalog.

Design boundary¶

ASRQuant does not claim to reproduce from scratch every algorithm implemented by mature scientific libraries. That would increase numerical risk and maintenance burden. Instead, it provides a stable, finance-oriented facade and keeps the engines replaceable behind the public ASRQuant contract.

Literature-to-decision workflow¶

import asrquant as asr

project = asr.research.from_pdfs("papers/", topic="rates and equity styles")
registry = project.discover_hypotheses()
project.select_hypothesis("H001", predictor="US10Y", expected_sign="positive")

The same namespace provides data planning, feature construction, econometric testing, backtesting, robustness, decision governance and paper trading. asr.trading exposes broker-neutral order primitives and the safe built-in paper broker.