Documentation

Factor Class Open

Factor Class Open Documentation

Public documentation is available without registration or login. Install the client, activate a license, then use the Python SDK or Codex/Claude Code to drive the factor runtime.

Client Install and Authorization

Install the Python package locally, set the platform URL and license key, then activate the current device.

python3 -m venv .venv
source .venv/bin/activate
python -m pip install -U pip
python -m pip install -e .

export FACTOR_CLASS_OPEN_STATE_DIR="$PWD/.factor-class-open-state"
export FCO_PLATFORM_URL="https://platform.example.com"
export FCO_LICENSE_KEY="<your_fco_license_key>"

factorctl license activate \
  --server-url "$FCO_PLATFORM_URL" \
  --license-key "$FCO_LICENSE_KEY" \
  --release-tag runtime-v0.1.1
factorctl doctor

Python SDK Manual

Engineering users can initialize authorization in Python and call the public SDK for D1, unary, plan, and composable factor workflows.

import os
import factor_class_open as fco

fco.init(
    platform_url=os.environ["FCO_PLATFORM_URL"],
    license_key=os.environ["FCO_LICENSE_KEY"],
)

print(fco.status()["status"])
print(fco.d1_stat([1.0, 2.0, 3.0], "SUM"))
print(fco.unary([1.0, 3.0, 2.0], "hi", 2))
print(fco.plan()["run_id"])

Composable Factor Example

Users can programmatically combine indicators, periods, daily aggregation, unary operations, and backtests into reproducible factor recipes.

spec = (
    fco.factor("RSI")
    .periods([7, 14, 28])
    .timeframe("M30")
    .zscore_bars(12)
    .daily_aggregate(["STD"])
    .unary(["hi", "lo"], windows=[3])
    .name("rsi_std_unary_demo")
)

result = spec.run(bars, returns=forward_returns)
report = result.backtest()
report.save_return_curve("rsi_std_unary_demo.svg")
print(result.summary())
print(report.summary())

Codex / Claude Code Low-Code Manual

Low-code users can give the following task brief to Codex or Claude Code and let it generate factor exploration scripts through the public SDK.

You are working inside the factor-class-open client project.
Use only the public factor_class_open SDK and public examples.
First check FCO_PLATFORM_URL, FCO_LICENSE_KEY, and factorctl doctor.
Then use local OHLCV data to combine indicators, periods, D1 aggregation, and unary operations.
Output factor summaries, IS/OOS metrics, SVG return curves, and the factor recipe.
Do not read private paths, write real license keys, or import sibling project source code.