# screamer function index (name, short description)

# Functors
AD, Chaikin Accumulation/Distribution Line.
ADOSC, Difference of fast and slow EMA of the Accumulation/Distribution line.
ADX, Wilder's ADX with +DI and -DI (3 inputs -> 3 outputs).
ADXR, Wilder's ADXR, the mean of ADX now and ADX one window ago.
ATR, Wilder-smoothed average of TrueRange.
Abs, Absolute value of each element.
Acos, Inverse cosine of each element (radians, input in [-1, 1]).
Add, Elementwise sum of two aligned streams (x + y).
AmihudIlliquidity, Rolling mean of |return| / notional (Amihud 2002 illiquidity ratio).
And, Returns 1.0 if both inputs are nonzero, else 0.0. NaN in either input yields NaN.
Asin, Inverse sine of each element (radians, input in [-1, 1]).
Atan, Inverse tangent of each element (radians).
Atan2, Signed angle of (x, y) from the positive x-axis (numpy.arctan2 order).
BOP, (close - open) / (high - low) per bar. No smoothing.
BacktestCarry, Charge perpetual funding, borrow, or financing on the position of any backtest engine's output.
BacktestL1Orders, Backtest a two-sided market maker against a top-of-book (L1) quote stream.
BacktestL1Target, Backtest a target-position strategy against the L1 book, taking each quote update as a market fill to reach the target immediately.
BacktestL1TradesOrders, Backtest a two-sided market maker against top-of-book quotes with a trade tape driving the fills, into a costed equity curve.
BacktestOHLCOrders, Backtest a two-sided order poster on OHLC bars, posting resting bids and asks that fill when the bar's range reaches them.
BacktestOHLCTarget, Backtest a target-position strategy on OHLC bars, executing market orders at the next bar's open or another bar price, with slippage (causal, no manual lag).
BacktestOHLCVTarget, Backtest a target-position strategy on OHLCV bars, filling each bar at most a set share of its volume.
BacktestPortfolioOHLCTarget, Backtest K assets traded toward per-asset targets from one shared capital account on OHLC bars, executed delay bars later at a chosen bar price.
BacktestPortfolioTarget, Backtest K assets traded toward per-asset target weights, notionals, or positions from one shared capital account, filled at each step's price.
BacktestPriceTarget, Backtest a target position against a value series (price/mark) into a costed mark-to-market equity curve.
BacktestReport, Turn a backtest engine's [equity, pnl, position, cost] into running drawdown, cost, turnover, trades, and Sharpe.
BacktestStats, Turn a backtest engine's [equity, pnl, position, cost] into running exposure, round-trip trade statistics, Sortino, and Calmar.
BacktestTradesOrders, Backtest a two-sided order poster against the trade tape, filling resting quotes when prints cross them, into a costed equity curve.
BacktestTradesTarget, Backtest a target-position strategy against the trade tape, taking each print as a market fill to reach the target immediately.
BandSignal, Three-state mean-reversion position from a z-score: short above +entry, long below -entry, flat again inside the exit level, with an optional stop and lockout.
BayesianRegression, Online Bayesian linear regression with a forgetting factor: the current slope and intercept, and a one-step-ahead predictive mean and standard deviation.
BollingerBands, Mean +/- num_std rolling standard deviations.
BulkVolumeClassifier, Buy-initiated share of a bar's volume estimated as the normal CDF of return / trailing-window volatility.
Butter, General-order IIR Butterworth low-pass filter.
ButterBandpass, General-order IIR Butterworth band-pass filter.
ButterBandstop, General-order IIR Butterworth band-stop (notch) filter.
ButterHighpass, General-order IIR Butterworth high-pass filter (rejects low frequencies).
CCI, Commodity Channel Index over typical price.
Cart2Polar, Convert (x, y) to (r, theta).
Ceil, Round each element toward positive infinity.
Clip, Bound each element below and/or above.
CombineLatest, As-of latest-value join of N streams: one row per distinct index (same-index events coalesce).
ContOFI, Cont-Kukanov-Stoikov (2014) order-flow imbalance from L1 book events.
Cos, Cosine of each element (radians).
CrossDemean, Subtract the mean across assets at each step, removing the common component.
CrossNormalize, Scale each row of scores to portfolio weights with a chosen gross exposure, and optionally a net exposure.
CrossRank, Rank each asset against the others at the same step, scaled to [-1, 1] or [0, 1].
CrossTopBottom, +1 for the highest values of each row, -1 for the lowest, 0 for the rest.
CrossWinsorize, Clip each asset's value to the lower and upper quantiles of its row.
CrossZscore, Standardize each asset against the others at the same step: (x - mean) / std across the row.
Cube, Cube of each element, computed as x * x * x.
CumMax, Running maximum from t=0.
CumMin, Running minimum from t=0.
CumProd, Running product from t=0.
CumSum, Running sum from t=0.
CycleAmplitude, Instantaneous amplitude (envelope) of the dominant cycle, scaled to the input amplitude.
CycleFrequency, Instantaneous frequency, in cycles per sample, via Ehlers' homodyne discriminator.
CyclePhase, Instantaneous phase, in degrees, of Ehlers' in-phase and quadrature components.
CycleSine, Ehlers' sinewave indicator (sine, leadsine) from the phase of the dominant cycle.
DEMA, Mulloy's Double EMA: 2*EMA - EMA(EMA).
DX, Wilder directional index, the pre-average input to ADX.
Decycler, Ehlers trend estimate with short cycles removed.
Delay, Re-stamp each event's index by a fixed time offset (a latency line).
Detrend, x[t] minus its rolling mean.
Diff, x[t] - x[t-k] (first difference at lag k).
Diff2, Second-order finite difference (discrete second derivative).
Div, Elementwise quotient of two aligned streams (x / y).
DominantCycle, Dominant cycle period via Ehlers' homodyne discriminator.
DonchianChannels, Trend-following envelope: rolling max(high), rolling min(low), and midline.
Drawdown, Running drawdown from the cumulative peak.
Dropna, Drop events whose value is NaN.
EffectiveSpread, Effective spread, 2*|price - mid|: the round-trip cost paid relative to the mid.
Elu, Exponential linear unit.
Equal, Returns 1.0 if a == b, else 0.0. NaN in either input yields NaN.
Erf, Gauss error function.
Erfc, Complementary error function (1 - erf).
EwAlpha, EW regression intercept, ewmean(x) - EwBeta(x,y) * ewmean(y).
EwBeta, EW CAPM beta: cov(target, regressor) / var(regressor).
EwCorr, EW Pearson correlation of two parallel streams.
EwCov, EW covariance of two parallel streams.
EwGarmanKlassVar, Exponentially weighted Garman-Klass range-based variance estimator (OHLC).
EwGarmanKlassVol, Exponentially weighted Garman-Klass range-based volatility estimator (OHLC).
EwKurt, EW excess kurtosis.
EwKyleLambda, Exponentially-weighted price-impact slope of return on signed order flow.
EwMean, Exponentially-weighted moving average (pandas adjust=True).
EwParkinsonVar, Exponentially weighted Parkinson range-based variance estimator (high, low).
EwParkinsonVol, Exponentially weighted Parkinson range-based volatility estimator (high, low).
EwResidualStd, EW std of the spread x - EwBeta(x, y) * y.
EwRms, EW root-mean-square.
EwRogersSatchellVar, Exponentially weighted Rogers-Satchell drift-robust range-based variance estimator (OHLC).
EwRogersSatchellVol, Exponentially weighted Rogers-Satchell drift-robust range-based volatility estimator (OHLC).
EwSkew, EW skewness.
EwSpread, EW hedge-adjusted residual x - EwBeta(x, y) * y.
EwStd, EW standard deviation.
EwVar, EW variance (pandas adjust=True bias-corrected).
EwYangZhangVar, Exponentially weighted Yang-Zhang range-based variance estimator (OHLC).
EwYangZhangVol, Exponentially weighted Yang-Zhang range-based volatility estimator (OHLC).
EwZscore, Latest sample standardized by the EW mean and std.
Exp, e to the power of each element.
ExpandingKurt, Running bias-corrected excess kurtosis (Fisher) over the whole history.
ExpandingMax, Running maximum since the start of the stream.
ExpandingMean, Running mean over the whole history since the last reset.
ExpandingMin, Running minimum since the start of the stream.
ExpandingProd, Running product since the start of the stream.
ExpandingSkew, Running bias-corrected sample skewness (G1) over the whole history.
ExpandingSlope, Running OLS slope of the series against time over the whole history.
ExpandingStd, Running sample standard deviation (ddof=1) over the whole history.
ExpandingSum, Running sum since the start of the stream.
ExpandingVar, Running sample variance (ddof=1) over the whole history.
Ffill, Replace NaN with the most recent non-NaN value.
FillNa, Replace NaN with a user-specified scalar.
Filter, Two-input mask gate that keeps each data value whose aligned mask is nonzero.
First, Latch the first finite value seen since reset.
Floor, Round each element toward negative infinity.
FracDiff, Fractional differencing filter. Removes the trend while keeping more of the series' memory than an integer difference.
GreaterEqual, Returns 1.0 if a >= b, else 0.0. NaN in either input yields NaN.
GreaterThan, Returns 1.0 if a > b, else 0.0. NaN in either input yields NaN.
Hampel, Robust Hampel despiker, replace samples far from the window median (in MAD units).
HawkesIntensity, Conditional intensity of an exponential-kernel Hawkes process: lambda_t = mu + kappa_t.
HilbertPhasor, Ehlers' in-phase and quadrature components (TA-Lib HT_PHASOR), scaled to the input amplitude.
Hold, Time-latch operator. Latches a nonzero finite input and holds it for n steps; returns the release value once the hold expires.
HullMA, Hull's responsive MA: WMA(2*WMA(n/2) - WMA(n), sqrt(n)).
Hypot, Euclidean distance sqrt(x^2 + y^2), numerically stable.
Identity, Pass-through (y = x).
ImpulseClip, Causal impulse remover, detects spikes on the trend-free first difference.
Input, Create a source Node - a named placeholder for a timed stream.
InstantaneousTrendline, Ehlers' instantaneous trendline, the mean over one measured dominant cycle.
IsFinite, Returns 1.0 for finite values, 0.0 for NaN or inf. Does not propagate NaN.
IsNan, Returns 1.0 if the input is NaN, else 0.0. Does not propagate NaN.
KAMA, Adaptive MA whose smoothing constant responds to the efficiency ratio.
KalmanFilter, Scalar 1-D Kalman filter for a noisy random-walk model.
KalmanRegression, Regression with a random-walk intercept and slope, estimated by a Kalman filter: the dynamic hedge ratio of a pairs trade, its spread, and the spread's predicted deviation.
KeltnerChannels, Volatility-adapted envelope: EMA midline plus/minus a multiple of ATR.
Lag, Output is the input delayed by k samples.
Last, Return the most recent finite value seen since reset.
LeeReadySign, Trade sign by the Lee-Ready (1991) rule: quote test with tick-rule fallback.
LessEqual, Returns 1.0 if a <= b, else 0.0. NaN in either input yields NaN.
LessThan, Returns 1.0 if a < b, else 0.0. NaN in either input yields NaN.
Linear, Affine transform: scale * x + shift.
Linear2, Two-input affine combination: a*x + b*y + c.
Log, Natural logarithm of each element.
LogReturn, Log return log(x[t] / x[t-k]) at lag k.
MACD, MACD line, signal line, and histogram (3 outputs).
MAMA, Ehlers' adaptive moving average pair (MAMA, FAMA), smoothing set by the rate of the cycle phase.
MFI, Volume-weighted analogue of RSI on the typical price.
MaxDrawdown, Worst drawdown experienced so far (since reset).
Merge, Merge N value streams into one index-sorted (values, sources, index).
MicroPrice, Imbalance-weighted mid (Stoikov 2018, first-order): fair value that leans toward the thinner side of the book.
MinusDI, Wilder negative directional indicator.
MinusDM, Wilder smoothed negative directional movement.
Momentum, x[t] - x[t-k], TA-Lib's MOM. Mathematically identical to Diff(k).
MovingAverage, Finite-impulse-response filter with user-supplied taps.
Mul, Elementwise product of two aligned streams (x * y).
NATR, ATR scaled to a percentage of the current close.
NegPart, Negative part of x: max(-x, 0).
Node, An immutable handle for a stream inside a pipeline.
Not, Returns 0.0 if input is nonzero, 1.0 if zero. NaN propagates.
NotEqual, Returns 1.0 if a != b, else 0.0. NaN in either input yields NaN.
OBV, On-Balance Volume: signed cumulative volume by close-direction (Granville, 1963).
OFI, Normalized signed order flow, (buy - sell) / (buy + sell).
Or, Returns 1.0 if either input is nonzero, else 0.0. NaN in either input yields NaN.
Pipeline, A reusable N-in / M-out function you define once and call on stored or live data.
PlusDI, Wilder positive directional indicator.
PlusDM, Wilder smoothed positive directional movement.
Polar2Cart, Convert (r, theta) to (x, y).
PortfolioReport, Reduce a multi-asset backtest engine output into causal portfolio-level report columns.
PosPart, Positive part of x: max(x, 0).
Power, x raised to a fixed exponent p.
Propagator, Bouchaud (2004) propagator model: price impact as a decaying-kernel convolution over past signed order flow.
QueueImbalance, Normalized L1 book imbalance, (bid_size - ask_size) / (bid_size + ask_size).
ROC, Rate of change 100 * (x[t] / x[t-k] - 1), in percent (TA-Lib ROC).
ROCP, Rate of change x[t] / x[t-k] - 1, as a fraction (TA-Lib ROCP). Identical to Return.
ROCR, Rate of change ratio x[t] / x[t-k] (TA-Lib ROCR).
RealizedSpread, Realized spread, 2*D*(trade price - mid lag steps later): the part of the effective spread the liquidity provider keeps.
Relu, Rectified linear unit: max(0, x).
Resample, Causal windowed downsample of a 1-D or multi-column value stream.
Return, Simple return (x[t] - x[t-k]) / x[t-k] at lag k.
RollSpread, Roll (1984) effective spread from trade prices alone: 2*sqrt(-cov(dP_t, dP_{t-1})) over a trailing window.
RollingADF, Rolling augmented Dickey-Fuller t-statistic: how strongly a series, such as a pairs spread, reverts to a mean over the last window.
RollingAlpha, Rolling OLS intercept of a target regressed on a regressor (companion to RollingBeta).
RollingArgmax, Position of the maximum within the trailing window (0 is the oldest sample).
RollingArgmin, Position of the minimum within the trailing window (0 is the oldest sample).
RollingBeta, Rolling regression slope cov(x, y) / var(y) of the first input x (dependent) on the second input y (regressor).
RollingCVaR, Historical Conditional Value-at-Risk (Expected Shortfall): the mean loss in the worst alpha tail over a window.
RollingCalmar, Annualized mean return divided by the worst drawdown in a trailing window.
RollingCorr, Rolling Pearson correlation of two parallel streams.
RollingCov, Rolling sample covariance of two parallel streams.
RollingDownsideDeviation, Trailing-window downside semideviation: the RMS of returns falling below a minimum acceptable return.
RollingGarmanKlassVar, Var form of the Garman-Klass range-based volatility estimator (OHLC).
RollingGarmanKlassVol, Vol form of the Garman-Klass range-based volatility estimator (OHLC).
RollingHitRate, Fraction of strictly positive values in a trailing window.
RollingHurst, Rolling-window Hurst exponent via Anis-Lloyd corrected rescaled-range analysis.
RollingInfoRatio, Annualized information ratio: the Sharpe ratio of active returns against a benchmark.
RollingIqr, Q3 minus Q1 over the trailing window.
RollingKurt, Trailing-window excess kurtosis.
RollingKyleLambda, Rolling-window price-impact slope of return on signed order flow (Kyle 1985).
RollingLinearRegression, Rolling OLS fit of y on x, returning slope, intercept, R-squared and the standard error of the fit.
RollingMad, Trailing-window mean absolute deviation from the rolling mean.
RollingMax, Trailing-window maximum (monotonic deque).
RollingMaxDrawdown, Worst peak-to-trough drawdown inside a trailing window.
RollingMean, Trailing-window arithmetic mean (simple moving average).
RollingMedian, Trailing-window median.
RollingMedianAD, Rolling median absolute deviation, median(|x - median|), a robust scale estimate.
RollingMin, Trailing-window minimum (monotonic deque).
RollingMinMax, Trailing-window minimum and maximum, computed in one pass.
RollingOU, Rolling MLE fit of a mean-reverting Ornstein-Uhlenbeck process.
RollingOmega, Omega ratio over a window: total gains above a threshold divided by total losses below it.
RollingOrderImbalance, Trailing-window sum of signed order flow (Chordia-Roll-Subrahmanyam imbalance).
RollingParkinsonVar, Var form of the Parkinson range-based volatility estimator (uses high & low).
RollingParkinsonVol, Vol form of the Parkinson range-based volatility estimator (uses high & low).
RollingPercentile, Percentile rank (rank / window) of the current value within the trailing window.
RollingPoly1, OLS fit y = a + b*t over a trailing window. derivative_order selects value/slope/zero.
RollingPoly2, OLS fit y = a + b*t + c*t^2 over a trailing window.
RollingQuantile, Trailing-window quantile (e.g. 0.25 = lower quartile).
RollingRSI, Relative Strength Index. Wilder's smoothing by default; Cutler's via method='cutler'.
RollingRange, Trailing-window max minus min.
RollingRank, Rank of the current value within the trailing window (1-based, average tie rule).
RollingResidualStd, Standard deviation of the rolling hedge-adjusted residual y - beta * x.
RollingRms, Trailing-window root-mean-square.
RollingRogersSatchellVar, Var form of the Rogers-Satchell drift-robust range-based estimator.
RollingRogersSatchellVol, Vol form of the Rogers-Satchell drift-robust range-based estimator.
RollingSharpe, Annualized Sharpe ratio over a trailing window of returns.
RollingSigmaClip, Clip samples that lie beyond lower or upper standard deviations from the rolling mean, or replace them with NaN.
RollingSkew, Trailing-window skewness.
RollingSortino, Annualized Sortino ratio: the Sharpe ratio with downside deviation in the denominator.
RollingSpread, Rolling hedge-adjusted residual x - beta * y, with beta the rolling regression slope.
RollingStd, Trailing-window sample standard deviation (ddof=1).
RollingSum, Trailing-window sum.
RollingTLSBeta, Rolling total least squares (orthogonal regression) slope of the first input on the second: a hedge ratio that does not depend on which leg is called dependent.
RollingTSF, Linear regression of the input on time, projected one step ahead (TA-Lib TSF).
RollingVWAP, Rolling volume-weighted average price (typical-price weighted).
RollingVar, Trailing-window sample variance (ddof=1).
RollingYangZhangVar, Var form of the Yang-Zhang estimator (drift + gap robust).
RollingYangZhangVol, Vol form of the Yang-Zhang estimator (drift + gap robust).
RollingZscore, Latest sample standardized by the trailing-window mean and standard deviation.
RoofingFilter, Ehlers bandpass, a highpass then a SuperSmoother.
Round, Round each element to the nearest integer (half-to-even).
SchmittTrigger, Hysteresis comparator. Latches 1.0 above the upper threshold, 0.0 below the lower threshold, and retains its previous value in between.
Select, Pick column(s) from a wide (M, N) value stream.
Selu, Scaled exponential linear unit (self-normalizing networks).
Sigmoid, Logistic sigmoid: 1 / (1 + exp(-x)).
Sign, Sign of each element: -1, 0, or +1.
SignedVolume, Aggressor-signed volume, sign * volume.
Sin, Sine of each element (radians).
Softsign, Softsign: x / (1 + |x|).
Sqrt, Square root of each element.
Square, Square of each element, computed as x * x.
Stoch, Stochastic oscillator %K and %D from high, low and close.
StochRSI, Stochastic oscillator applied to the RSI, giving %K and %D.
Sub, Elementwise difference of two aligned streams (x - y).
SuperSmoother, Ehlers 2-pole low-lag lowpass filter.
TEMA, Mulloy's Triple EMA: 3*EMA - 3*EMA(EMA) + EMA(EMA(EMA)).
TRIMA, Triangular moving average: a simple moving average of a simple moving average.
TRIX, ROC of a triple-smoothed EMA.
Tanh, Hyperbolic tangent.
TickRuleSign, Trade sign by the tick rule (+1 up-tick, -1 down-tick, carry on unchanged).
TrendMode, Ehlers' trend mode, 1.0 while the series trends and 0.0 while it cycles.
TrueRange, Per-bar true range accounting for overnight gaps (Wilder, 1978).
UltimateOscillator, Three-period weighted oscillator (Williams, 1976).
VPIN, Volume-synchronized probability of informed trading (Easley-López de Prado-O'Hara 2012): order-flow toxicity.
WMA, Linearly weighted moving average.
Where, Returns a if mask is nonzero, b otherwise. NaN mask yields NaN.
WilliamsR, Inverse stochastic oscillator in [-100, 0].
backtest_report, Running statistics and a summary for a backtest engine's output.
choosing_a_backtest_engine, The engine grid by data model and order definition, with fill rules, sizing, volume limits, and encoding.
circular_shift, Randomly rotated copies of a series, for randomization tests.
deflated_sharpe, Probabilistic Sharpe ratio against the Sharpe ratio the best of your tried variants would reach by luck.
forecast_pairs, Build a forecasting training set by aligning lagged features with a future causal target.
from_pandas, Convert a pandas Series or DataFrame to a (values, index) tuple.
permute_bars, Bar-permuted copies of an OHLC series that keep its return distribution but no serial structure.
permute_returns, Copies of a price series with its returns in random order.
position_test, Test whether a backtest's timing beats randomly timed copies of the same strategy, with a p-value.
probabilistic_sharpe, Probability that a strategy's true Sharpe ratio is above a benchmark, given its sample length and fat tails.
reshape, Name the axes of data with a layout string (T, A, F) and reshape it to match.
segment_shuffle, Copies of a series with its runs of equal value in random order, for randomization tests.
split, Partition a merged tagged stream back into per-source streams.
strategy_test, Test a whole strategy, model included, against the same strategy run on permuted market data or shifted features.
stream-tuple-convention, The (values, index) tuple at every operator boundary.
to_pandas, Convert a (values, index) tuple to a pandas Series or DataFrame.
walk_forward, Pick the best of several strategy variants on each training window and trade it on the next test window, out of sample.

# Stream operators
Merge, Index-sorted N-way merge.
CombineLatest, As-of latest-value join.
Dropna, Drop NaN events.
Filter, 2-input mask gate: keep each data value whose aligned mask is nonzero (zero or NaN drops).
Select, Pick columns from a wide stream.
split, Partition a merged tagged stream back into per-source streams.
Resample, Causal windowed downsample.

# Pipeline
Input, Create a source Node - a named placeholder for a timed stream.
Pipeline, A reusable N-in / M-out function you define once and call on stored or live data.
Node, An immutable handle for a stream inside a pipeline.
