Skip to content

Quick Start

Build your first Flow, back it, and round-trip it to disk. Everything here runs offline against the built-in memory source - no API keys.

New to the vocabulary?

The Concepts page explains the tier stack and the invariants; the Glossary defines each term.

Idea -> first backtest

import signalflow as sf

ds = sf.data("memory", pairs=["BTCUSDT"], start="2023-01-01", interval="1h")

model = sf.ForecastModel(target=sf.FixedHorizon(bars=12),
                         features=sf.FeaturePipe(sf.SMA(10), sf.SMA(20), sf.SMA(50)))
model.fit(ds)                                          # train tier-1 forecaster

flow = sf.Flow(name="sma_rise",
               forecasts={"rise": model},
               detectors=[sf.ThresholdDetector(forecast="rise", p_min=0.6)],
               strategy=sf.RulesStrategy())
run = flow.backtest(ds, capital=50_000)
print(run.scorecard())                                 # total_return, sharpe, max_drawdown, ...

What each step does:

  1. sf.data("memory", ...) builds a Dataset - one lazy, immutable container.
  2. sf.ForecastModel(...) pairs a Target (FixedHorizon) with a FeaturePipe; fit trains it out-of-fold.
  3. sf.Flow(...) assembles the deployable stack: forecasts -> detectors -> strategy.
  4. flow.backtest(...) returns a Run; run.scorecard() is the standard metric dict.

Deploy is data

flow.save serializes the whole stack - config plus trained artifacts - to YAML plus a model directory. sf.Flow.load restores a byte-identical backtest, so promoting a strategy is moving a file.

flow.save("flows/rsi_rise.yaml", model_dir="flows/models")   # yaml + trained artifacts
same = sf.Flow.load("flows/rsi_rise.yaml")
assert same.backtest(ds, capital=50_000).final_equity == run.final_equity

Model artifacts can live on the local filesystem, MLflow, or the Hugging Face Hub (model.save("mlflow://..."), model.save("hf://...")).

Backtest -> paper -> live

One decision core drives all three modes:

flow.paper(ds, capital=50_000)                         # sim fills over a Dataset

feed = sf.PollingFeed(sf.BinanceSource(), pairs=["BTCUSDT"], interval="1m")
flow.live(feed, capital=50_000)                        # live data, SimBroker (paper)

Next steps