Ferric Alpha

Ferric Alpha is a high-performance Rust implementation of factor analysis workflows inspired by alphalens. It uses Polars as its core table engine and provides both Rust and Python APIs.

The project is in pre-alpha. The 0.1.x releases are intended for early factor research, API feedback, and compatibility validation.

Current scope

Ferric Alpha currently supports factor data preparation, core performance metrics, turnover and portfolio series, serializable tear-sheet data, and native report rendering.

Performance

The benchmark below compares the Python-facing ferric-alpha API with alphalens on the same deterministic long/short factor fixture. It measures function execution time only; fixture construction is outside the timed block.

Benchmark setup:

• Machine: Apple silicon macOS 26.3.1
• Dataset: 250 assets, 252 sessions, 63,000 factor rows
• Forward returns: 1D and 5D
• Iterations: 5; result is median wall-clock time
• Rank autocorrelation recheck: median of 7 runs, 15 iterations per run
• ferric-alpha: 0.1.0, release build, Python 3.13.4, Polars 1.42.1
• alphalens: 0.4.0, Python 3.9.25, Pandas 1.5.3, NumPy 1.23.5

Results where ferric-alpha is faster:

Metric	ferric-alpha	alphalens	Result
factor_information_coefficient	13.616 ms	124.205 ms	9.1x
factor_returns	11.778 ms	46.425 ms	3.9x
factor_weights	15.153 ms	43.330 ms	2.9x
mean_return_by_quantile	21.022 ms	60.900 ms	2.9x
quantile_turnover	3.018 ms	6.136 ms	2.0x
factor_rank_autocorrelation	6.823 ms	15.497 ms	2.3x

factor_rank_autocorrelation automatically uses a dense matrix for complete date-by-asset panels and retains the compatibility path for sparse or changing universes. A direct alphalens oracle comparison over shuffled data with ties and lags 1, 3, and 10 matched null positions exactly; the maximum absolute numeric difference was 1.11e-16.

Reproduce the benchmark from this repository:

.venv/bin/maturin develop --release
.venv/bin/python scripts/benchmark_alphalens_comparison.py --engine ferric

python3.9 -m venv /tmp/alphalens-bench
/tmp/alphalens-bench/bin/python -m pip install \
  'alphalens==0.4.0' 'pandas==1.5.3' 'numpy==1.23.5' \
  'scipy==1.10.1' 'statsmodels==0.13.5' 'empyrical==0.5.5'
PYTHONWARNINGS=ignore /tmp/alphalens-bench/bin/python \
  scripts/benchmark_alphalens_comparison.py --engine alphalens

Data model

Core APIs use explicit long-form columns such as date, asset, and factor. They do not emulate Pandas MultiIndex behavior.

Required columns:

• date: Polars Datetime
• asset: Polars String
• factor: Polars Float64; null values are preserved

The optional group column must be Polars String. The (date, asset) key must be non-null and unique.

Quickstart

Run the complete factor-analysis example after building from source:

python examples/factor_quickstart.py --output-dir quickstart-output

It prepares a known-signal dataset, computes IC, quantile returns, factor returns, alpha/beta, turnover, and rank autocorrelation, then writes a native HTML tear sheet. See the factor analysis quickstart for the input schema and an API walkthrough.

Development

Prerequisites are Python 3.10 or newer, Rust 1.91 or newer, and make.

Create the local environment and build the extension:

make develop

Run all formatting checks, linters, Rust tests, and Python tests:

make verify

The Python package has Polars as its only default runtime dependency.

import ferric_alpha

validated = ferric_alpha.validate_factor_frame(factor_frame)

Rendering

Native HTML is the default plotting path and does not require Matplotlib, Pandas, Seaborn, SciPy, or Statsmodels.

import ferric_alpha as fa

report = fa.tears.create_full_tear_sheet_data(factor_data)
rendered = fa.plotting.render(report)
rendered.save("factor-report.html")

Native SVG and PNG are available directly:

svg = fa.plotting.render_svg(report)
png = fa.plotting.render_png(report)

Notebook display uses the same native HTML object:

fa.plotting.display(report)

The optional Matplotlib backend is installed separately:

pip install 'ferric-alpha[plot]'

figure = fa.plotting.render(report, backend="matplotlib")
figure.savefig("factor-report.png", dpi=144)

The Matplotlib backend returns a matplotlib.figure.Figure and never calls show() implicitly.

License

Ferric Alpha is available under the Zero-Clause BSD (0BSD) license.