fastmm::research Namespace Reference

Classes
struct	EvalConfig
struct	FeatureConfig
struct	FeatureRows
struct	FeatureTable
struct	HorizonCoverage
	Per horizon, how many rows got a forward mid and why the rest did not. More...
struct	HorizonEval
struct	SignalBucket
struct	SignalEval

Enumerations
enum class	Feature : std::uint8_t { Imbalance = 0 , MicropriceEdge = 1 , Spread = 2 }
	A built-in column, so the common signals need no array from the caller. More...

Functions
std::string	horizon_label (std::int64_t ns)
	"100ms", "1s", "1m": the shortest exact unit, the spelling MarkoutHorizon::label() uses.
std::vector< Duration >	default_horizons ()
	kDefaultHorizonsNs as Durations.
FeatureTable	extract_features (MdSource &source, const FeatureConfig &cfg={})
	Reads source from its current position to the end.
std::string_view	to_string (Feature f) noexcept
std::vector< double >	feature_values (const FeatureTable &t, Feature f)
	The feature of every row, in row order, in the units above.
SignalEval	evaluate_signal (const FeatureTable &t, std::span< const double > values, std::string_view name, const EvalConfig &cfg={})
	values must hold one signal value per row of t, in row order.
SignalEval	evaluate_feature (const FeatureTable &t, Feature f, const EvalConfig &cfg={})
	evaluate_signal() over feature_values(t, f).

Variables
constexpr std::int64_t	kDefaultHorizonsNs []
	Horizons measured by default, in nanoseconds of event time: 100 ms, 1 s, 10 s, 1 minute.

Enumeration Type Documentation

Feature

enum class fastmm::research::Feature : std::uint8_t

strong

A built-in column, so the common signals need no array from the caller.

Enumerator
Imbalance	(bid qty - ask qty) / (bid qty + ask qty) at the touch, in [-1, 1]
MicropriceEdge	(microprice - mid) / mid, in basis points
Spread	(best ask - best bid) / mid, in basis points

Function Documentation

horizon_label()

std::string fastmm::research::horizon_label

(

std::int64_t

ns

)

"100ms", "1s", "1m": the shortest exact unit, the spelling MarkoutHorizon::label() uses.

default_horizons()

std::vector< Duration > fastmm::research::default_horizons

(

)

kDefaultHorizonsNs as Durations.

extract_features()

FeatureTable fastmm::research::extract_features	(	MdSource &	source,
		const FeatureConfig &	cfg = {}
	)

Reads source from its current position to the end.

Throws std::invalid_argument when instruments is 0 or imbalance_levels is 0.

to_string()

std::string_view fastmm::research::to_string ( Feature  f )

noexcept

feature_values()

std::vector< double > fastmm::research::feature_values	(	const FeatureTable &	t,
		Feature	f
	)

The feature of every row, in row order, in the units above.

evaluate_signal()

SignalEval fastmm::research::evaluate_signal	(	const FeatureTable &	t,
		std::span< const double >	values,
		std::string_view	name,
		const EvalConfig &	cfg = {}
	)

values must hold one signal value per row of t, in row order.

Throws std::invalid_argument on a length mismatch or a bucket/block count of 0.

evaluate_feature()

SignalEval fastmm::research::evaluate_feature	(	const FeatureTable &	t,
		Feature	f,
		const EvalConfig &	cfg = {}
	)

evaluate_signal() over feature_values(t, f).

Variable Documentation

kDefaultHorizonsNs

constexpr std::int64_t fastmm::research::kDefaultHorizonsNs[]

inlineconstexpr

Initial value:

= {
    100'000'000LL, 1'000'000'000LL, 10'000'000'000LL, 60'000'000'000LL}

Horizons measured by default, in nanoseconds of event time: 100 ms, 1 s, 10 s, 1 minute.