Class: Musa::Markov::Markov

Inherits:
Object show all
Includes:
Series::Serie::Base
Defined in:
lib/musa-dsl/generative/markov.rb

Overview

Markov chain serie generator.

Generates sequences of states following probabilistic transition rules. Implements Series::Serie interface for integration with series operations.

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(transitions:, start:, finish: nil, random: nil) ⇒ void

Note:

A generator passed in is used, not copied, so several series sharing one draw from the same stream -- next_value advances it. to_a does not, because it consumes a deep copy of the serie.

Creates Markov chain generator.

Examples:

markov = Markov.new(
  transitions: { a: [:b, :c], b: [:a, :c], c: [:a, :b, :x] },
  start: :a,
  finish: :x
)

A seed and a generator built from it are the same thing

transitions = { a: [:b, :c], b: [:a, :c], c: [:a, :b, :x] }

from_seed = Markov.new(start: :a, finish: :x, transitions: transitions, random: 7)
from_generator = Markov.new(start: :a, finish: :x, transitions: transitions,
                            random: Random.new(7))

from_seed.i.to_a == from_generator.i.to_a  # => true

Parameters:

  • transitions (Hash) —

    state transition rules Keys are states, values are next state definitions (Array, Hash, or Proc)

  • start (Object) —

    initial state

  • finish (Object, nil) (defaults to: nil) —

    terminal state (nil for infinite)

  • random (Random, Integer, nil) (defaults to: nil) —

    a generator, or a seed to build one with, or nothing for an unseeded one. A seed and a Random built from that same seed give the same chain.



153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
 
# File 'lib/musa-dsl/generative/markov.rb', line 153

def initialize(transitions:, start:, finish: nil, random: nil)
  @transitions = transitions.clone.freeze

  @start = start
  @finish = finish

  # A seed is turned into a generator; a generator is kept as it is. The
  # second line used to overwrite the ivar the first had not set, so a
  # `Random` passed in was replaced by an unseeded one and the chain came
  # out different on every run while the caller believed it was pinned
  # (issue #79).
  @random = random.is_a?(Integer) ? Random.new(random) : random
  @random ||= Random.new

  @procedure_binders = {}

  mark_as_prototype!
  init
end

Instance Attribute Details

#finish ⇒ Object?

Returns finishing state (nil for infinite).

Returns:

  • (Object, nil) —

    finishing state (nil for infinite)



177
178
179
 
# File 'lib/musa-dsl/generative/markov.rb', line 177

def finish
  @finish
end

#random ⇒ Random

Returns random number generator.

Returns:

  • (Random) —

    random number generator



180
181
182
 
# File 'lib/musa-dsl/generative/markov.rb', line 180

def random
  @random
end

#start ⇒ Object

Returns starting state.

Returns:



174
175
176
 
# File 'lib/musa-dsl/generative/markov.rb', line 174

def start
  @start
end

#transitions ⇒ Hash

Returns transition rules (frozen).

Returns:

  • (Hash) —

    transition rules (frozen)



183
184
185
 
# File 'lib/musa-dsl/generative/markov.rb', line 183

def transitions
  @transitions
end

Instance Method Details

#+(other) ⇒ Sequence Originally defined in module Series::Operations

Appends another serie (operator alias for after).

Examples:

Concatenate

s = S(1, 2) + S(3, 4)
s.i.to_a  # => [1, 2, 3, 4]

Parameters:

  • other (Serie) —

    serie to append

Returns:

  • (Sequence) —

    sequential combination

#after(*series) ⇒ Sequence Originally defined in module Series::Operations

Appends series sequentially.

Alias for MERGE - plays this serie, then others in sequence.

Examples:

Append

s = S(1, 2).after(S(3, 4), S(5, 6))
s.i.to_a  # => [1, 2, 3, 4, 5, 6]

Parameters:

Returns:

  • (Sequence) —

    sequential combination

#anticipate {|current, next_value| ... } ⇒ Anticipate Originally defined in module Series::Operations

Evaluates block one step ahead (anticipate).

Block receives current value and NEXT value (peeked). Enables look-ahead transformations and transitions.

The block gets THREE values -- the previous one, the current one and the next -- and either end is nil where there is nothing.

Examples:

Smooth transitions

S(1, 5, 3, 8).anticipate { |_previous, current, following|
  following ? (current + following) / 2.0 : current
}.i.to_a
# => [3.0, 4.0, 5.5, 8]

Add interval information

S(60, 64, 67, 72).anticipate { |_previous, pitch, next_pitch|
  { pitch: pitch, interval: next_pitch ? next_pitch - pitch : nil }
}.i.to_a
# => [{ pitch: 60, interval: 4 }, { pitch: 64, interval: 3 },
#     { pitch: 67, interval: 5 }, { pitch: 72, interval: nil }]

Yields:

  • anticipation block

Yield Parameters:

  • current (Object) —

    current value

  • next_value (Object, nil) —

    next value (nil if last)

Yield Returns:

  • (Object) —

    transformed value

Returns:

  • (Anticipate) —

    anticipating serie

#autorestart ⇒ Autorestart Originally defined in module Series::Operations

Note:

The serie ENDS at each pass -- yielding nil -- and restarts on the next call. That nil is the signal that a pass finished, which is what makes the repetition countable; it also means infinite? is false and that to_a stops at the end of the first pass. For a serie with no seam in it, use repeat.

Auto-restarts serie when exhausted.

Creates an infinite serie from an original serie that automatically restarts from beginning when it reaches the end.

Examples:

Restarting after each pass

pattern = S(1, 2, 3).autorestart
inst = pattern.i
10.times.map { inst.next_value }  # => [1, 2, 3, nil, 1, 2, 3, nil, 1, 2]

Returns:

  • (Autorestart) —

    auto-restarting serie

#buffered(sync: false) ⇒ BufferSerie, SyncBufferSerie Originally defined in module Series::Operations

Creates a buffered serie allowing multiple independent iterations over same source.

Provides buffering mechanism enabling multiple "readers" to independently iterate over the same serie source without interfering with each other.

Buffering Modes

  • Async (default): Buffers fill independently, each progresses at own pace
  • Sync: All buffers synchronized, restart affects all

Use Cases

  • Multiple voices reading same melodic sequence at different speeds
  • Polyphonic playback from single source
  • Canonic structures (rounds, fugues)
  • Independent transformations of same base material

Memory Management

History is automatically cleaned when all buffers have progressed past old values, preventing unbounded memory growth.

Examples:

Create buffered serie

buffered = S(1, 2, 3, 4).buffered

# `.buffer` gives a PROTOTYPE, so instantiate it once and keep it: a
# second `.i` would be a second reading, starting over.
reader1 = buffered.buffer.i
reader2 = buffered.buffer.i

reader1.next_value  # => 1
reader2.next_value  # => 1

Multiple independent readers

source = S(1, 2, 3, 4).buffered
reader1 = source.buffer.i
reader2 = source.buffer.i

reader1.next_value  # => 1
reader2.next_value  # => 1 (independent)
reader1.next_value  # => 2
reader2.next_value  # => 2

Canon structure

melody = S(60, 64, 67, 72).buffered
voice1 = melody.buffer
voice2 = melody.buffer
# Play voice2 delayed by N beats

Parameters:

  • sync (Boolean) (defaults to: false) —

    synchronized mode (default: false)

Returns:

  • (BufferSerie, SyncBufferSerie) —

    buffered serie

#compact_timed ⇒ TimedCompacter Originally defined in module Series::Operations

Removes timed events where all values are nil.

Filters out temporal "gaps" where no sources have active values. Useful after union operations that create nil placeholders, or for cleaning sparse sequences.

Removal logic:

  • Direct nil: { time: t, value: nil } → removed
  • All-nil Hash: { time: t, value: { a: nil, b: nil } } → removed
  • Partial Hash: { time: t, value: { a: 1, b: nil } } → kept (has non-nil)
  • All-nil Array: { time: t, value: [nil, nil] } → removed
  • Partial Array: { time: t, value: [1, nil] } → kept (has non-nil)

Examples:

Remove direct nil events

s = S({ time: 0r, value: 1 },
      { time: 1r, value: nil },
      { time: 2r, value: 3 })

s.compact_timed.i.to_a
# => [{ time: 0r, value: 1 },
#     { time: 2r, value: 3 }]

Remove all-nil hash events

s = S({ time: 0r, value: { a: 1, b: 2 } },
      { time: 1r, value: { a: nil, b: nil } },
      { time: 2r, value: { a: 3, b: nil } })

s.compact_timed.i.to_a
# => [{ time: 0r, value: { a: 1, b: 2 } },
#     { time: 2r, value: { a: 3, b: nil } }]  # Kept: has non-nil 'a'

Clean sparse union results

s1 = S({ time: 0r, value: 1 }, { time: 2r, value: 3 })
s2 = S({ time: 1r, value: 10 })

union = TIMED_UNION(melody: s1, bass: s2).i.to_a
# => [{ time: 0r, value: { melody: 1, bass: nil } },
#     { time: 1r, value: { melody: nil, bass: 10 } },
#     { time: 2r, value: { melody: 3, bass: nil } }]

# All events have at least one non-nil, so none removed

Returns:

  • (TimedCompacter) —

    compacted serie

See Also:

#composer { ... } ⇒ ComposerAsOperationSerie Originally defined in module Series::Operations

Creates a composer transformation pipeline for complex multi-stage transformations.

Composer provides declarative DSL for building transformation pipelines with multiple inputs, outputs, and intermediate processing stages.

Composer Concepts

  • Pipelines: Named transformation chains
  • Inputs: Named input series (proxied and buffered)
  • Outputs: Named output series
  • Operations: Transformation steps in pipeline
  • Auto-commit: Automatic finalization of pipelines

DSL Structure

A pipeline is declared by NAMING it and listing its operations; the names are then wired together with route. Operations are symbols (reverse) or hashes ({ map: ... }, { skip: 1 }, { S: [...] }). input and output are pipelines like any other — they are simply the ones the Composer creates for you, and can be renamed or multiplied through the inputs: and outputs: keywords.

composer do
  stage1 operation1, operation2
  stage2 operation3

  route input, to: stage1
  route stage1, to: stage2
  route stage2, to: output
end

Musical Applications

  • Complex multi-voice processing
  • Effect chains and routing
  • Algorithmic composition pipelines
  • Multi-stage transformations
  • Modular synthesis-style routing

Examples:

Basic composer

s = S(1, 2, 3).composer do
  doubled({ map: ->(x) { x * 2 } })

  route input, to: doubled
  route doubled, to: output
end
s.i.to_a  # => [2, 4, 6]

Several stages in a row

s = S(1, 2, 3, 4, 5).composer do
  backwards reverse
  tail ({ skip: 1 })

  route input, to: backwards
  route backwards, to: tail
  route tail, to: output
end
s.i.to_a  # => [4, 3, 2, 1]

Several outputs

composer = Musa::Series::Composer::Composer.new(input: Musa::Series::Constructors.S(1, 2, 3),
                                                outputs: [:doubled, :tripled]) do
  step_d({ map: ->(x) { x * 2 } })
  step_t({ map: ->(x) { x * 3 } })

  route input, to: step_d
  route input, to: step_t
  route step_d, to: doubled
  route step_t, to: tripled
end
composer.output(:doubled).i.to_a  # => [2, 4, 6]
composer.output(:tripled).i.to_a  # => [3, 6, 9]

Yields:

  • composer DSL block

Returns:

#cut(length) ⇒ Cutter Originally defined in module Series::Operations

Cuts serie into chunks of specified length.

Returns a serie of SERIES, each yielding length values.

The chunks are series and not arrays, which is what keeps the whole thing lazy -- nothing is materialised until each chunk is consumed. Ask for the arrays explicitly with to_a(recursive: true).

Examples:

Cut into pairs

s = S(1, 2, 3, 4, 5, 6).cut(2)
s.i.to_a(recursive: true)  # => [[1, 2], [3, 4], [5, 6]]

Parameters:

  • length (Integer) —

    chunk size

Returns:

  • (Cutter) —

    chunked serie

#defined? ⇒ Boolean Originally defined in module Series::Serie::Prototyping

Checks if serie state is defined (not undefined).

Returns:

  • (Boolean) —

    true if prototype or instance, false if undefined

#flatten ⇒ Flattener Originally defined in module Series::Operations

Flattens nested series into single level.

Recursively consumes series elements that are themselves series.

Examples:

Flatten nested

s = S(S(1, 2), S(3, 4), 5).flatten
s.i.to_a  # => [1, 2, 3, 4, 5]

Returns:

  • (Flattener) —

    flattened serie

#flatten_timed ⇒ TimedFlattener Originally defined in module Series::Operations

Splits compound timed values into individual timed events.

Converts events with Hash or Array values into separate timed events per element, preserving time and extra attributes. Direct values pass through unchanged.

Hash values → Hash of timed events (keyed by original keys):

{ time: 0, value: { a: 1, b: 2 }, velocity: { a: 80, b: 90 } }
# becomes:
{ a: { time: 0, value: 1, velocity: 80 },
  b: { time: 0, value: 2, velocity: 90 } }

Array values → Array of timed events (indexed):

{ time: 0, value: [1, 2], velocity: [80, 90] }
# becomes:
[{ time: 0, value: 1, velocity: 80 },
 { time: 0, value: 2, velocity: 90 }]

Direct values → Pass through unchanged (already flat)

Use Cases

  • Separate polyphonic events into individual voices
  • Split multi-track sequences for independent processing
  • Prepare for voice-specific routing via split
  • Enable per-voice filtering with compact_timed

Examples:

Hash values to individual voices

s = S({ time: 0r, value: { a: 60, b: 64 }, velocity: { a: 80, b: 90 } })

flat = s.flatten_timed.i
flat.next_value
# => { a: { time: 0r, value: 60, velocity: 80 },
#      b: { time: 0r, value: 64, velocity: 90 } }

Array values to indexed events

s = S({ time: 0r, value: [60, 64], velocity: [80, 90] })

flat = s.flatten_timed.i
flat.next_value
# => [{ time: 0r, value: 60, velocity: 80 },
#     { time: 0r, value: 64, velocity: 90 }]

Direct values pass through

s = S({ time: 0r, value: 60, velocity: 80 })

flat = s.flatten_timed.i
flat.next_value  # => { time: 0r, value: 60, velocity: 80 }

Returns:

  • (TimedFlattener) —

    flattened timed serie

See Also:

#hashify(*keys) ⇒ HashFromSeriesArray Originally defined in module Series::Operations

Converts array values to hash with specified keys.

Takes array-valued serie and converts to hash using provided keys.

Examples:

Array to hash

s = S([60, 96], [64, 80]).hashify(:pitch, :velocity)
s.i.next_value  # => {pitch: 60, velocity: 96}

Parameters:

  • keys (Array) —

    hash keys for array elements

Returns:

  • (HashFromSeriesArray) —

    hashified serie

#infinite? ⇒ Boolean

Checks if Markov chain is infinite.

Returns:

  • (Boolean) —

    true if no finish state defined



250
251
252
 
# File 'lib/musa-dsl/generative/markov.rb', line 250

def infinite?
  @finish.nil?
end

#instance(built = nil) ⇒ Serie Also known as: i Originally defined in module Series::Serie::Prototyping

Creates or returns instance of serie.

  • If already instance, returns self
  • If prototype, creates new instance by cloning
  • If undefined, raises PrototypingError

Cloning Process

  1. Clones serie structure
  2. Marks clone as :instance
  3. Propagates instance creation to sources
  4. Calls init if defined

Each call creates independent instance with separate state.

Examples:

Create instances

proto = S(1, 2, 3)
a = proto.instance
b = proto.instance  # Different instance

a.next_value  # => 1
b.next_value  # => 1 (independent)

Returns:

  • (Serie) —

    instance serie

Raises:

#instance? ⇒ Boolean Originally defined in module Series::Serie::Prototyping

Checks if serie is in instance state.

Returns:

  • (Boolean) —

    true if instance, false otherwise

#lazy {|previous| ... } ⇒ LazySerieEval Originally defined in module Series::Operations

Delays evaluation to next step (lazy evaluation).

Block receives previous value and evaluates for current step. Enables state-dependent transformations.

Examples:

Cumulative sum

s = S(1, 2, 3, 4).lazy { |prev| (prev || 0) + value }

Yields:

  • lazy evaluation block

Yield Parameters:

  • previous (Object, nil) —

    previous value

Yield Returns:

Returns:

  • (LazySerieEval) —

    lazy-evaluated serie

#lock ⇒ Locker Originally defined in module Series::Operations

Locks serie preventing further modifications.

Returns locked copy that cannot be transformed further.

Returns:

  • (Locker) —

    locked serie

#map(isolate_values: nil) {|value| ... } ⇒ ProcessWith Originally defined in module Series::Operations

Maps values via transformation block.

Simplest and most common transformation. Applies block to each value. Shorthand for with without additional series.

Examples:

Transpose notes

notes = S(60, 64, 67).map { |n| n + 12 }
notes.i.to_a  # => [72, 76, 79]

Transform to hash

s = S(1, 2, 3).map { |n| {value: n, squared: n**2} }

Parameters:

  • isolate_values (Boolean, nil) (defaults to: nil) —

    clone values to prevent mutation (default: false)

Yields:

  • transformation block

Yield Parameters:

  • value (Object) —

    current value

Yield Returns:

  • (Object) —

    transformed value

Returns:

  • (ProcessWith) —

    mapped serie

#max_size(length) ⇒ LengthLimiter Originally defined in module Series::Operations

Limits serie to maximum number of values.

Stops after N values regardless of source length.

Examples:

Limit to 5

s = FOR(from: 0, step: 1).max_size(5)
s.i.to_a  # => [0, 1, 2, 3, 4]

Parameters:

  • length (Integer) —

    maximum number of values

Returns:

  • (LengthLimiter) —

    length-limited serie

#merge ⇒ MergeSerieOfSeries Originally defined in module Series::Operations

Merges serie of series into single serie.

Flattens one level: consumes serie where each element is itself a serie, merging them sequentially.

Examples:

Merge phrases

phrases = S(S(1, 2, 3), S(4, 5, 6))
merged = phrases.merge
merged.i.to_a  # => [1, 2, 3, 4, 5, 6]

Returns:

  • (MergeSerieOfSeries) —

    merged serie

#multiplex(*indexed_series, **hash_series) ⇒ MultiplexSelector Originally defined in module Series::Operations

Advances every serie and gives back the value of the one selected.

All of them move, all of the time. What the selector chooses is which one is heard, not which one runs -- so this is a window onto several simultaneous streams, and looking away from one does not pause it. A crossfade, a texture where one layer surfaces at a time, an instrument that changes what it is doubling.

Compare #switch, where the unselected series wait.

Examples:

The same selection as switch, and what it costs

a = S(:a1, :a2, :a3)
b = S(:b1, :b2)

S(0, 1, 0, 1, 0).multiplex(a, b).i.to_a
# => [:a1, :b2, :a3]

# Three values where switch gave five: every step spent one of each,
# so by the fourth b had nothing left -- and it is the step that
# SELECTS an exhausted serie that ends everything. While nobody looks
# at it, an exhausted source costs nothing.

Which layer is heard, by name

S(:high, :low, :high).multiplex(high: S(72, 74, 76),
                                low: S(48, 50, 52)).i.to_a
# => [72, 50, 76]

Parameters:

  • indexed_series (Array) —

    series chosen by integer position

  • hash_series (Hash) —

    series chosen by name

Returns:

  • (MultiplexSelector) —

    multiplexed serie

#process_with(**parameters) {|value, parameters| ... } ⇒ Processor Originally defined in module Series::Operations

Transforms each value with a block whose parameters can change later.

When this is the answer, and not .map

.map { |v| v + 7 } freezes the 7 in the block. process_with keeps it on the serie, where it can be changed while the serie is running -- so the transformation is under control instead of being decided once. A transposition that moves mid-phrase, a scaling driven from outside, a depth that a controller sets: anything where the amount is itself a parameter of the piece.

If the amount is known and fixed, .map says it with less.

Examples:

Transposition as a parameter, not as a constant

line = S(60, 62, 64, 65).process_with(interval: 0) { |v, interval:| v + interval }

i = line.i
i.next_value        # => 60
i.parameters = { interval: 12 }
i.next_value        # => 74
i.next_value        # => 76

Parameters:

  • parameters (Hash) —

    parameters passed to the block, and settable later

Yields:

  • processor block

Yield Parameters:

  • value (Object) —

    current value

  • parameters (Hash) —

    the current parameters

Returns:

#prototype ⇒ Serie Also known as: p Originally defined in module Series::Serie::Prototyping

Returns prototype of serie.

  • If already prototype, returns self
  • If instance, returns original prototype (if available)
  • If undefined, raises PrototypingError

Examples:

Get prototype

proto = S(1, 2, 3)
inst = proto.instance
inst.prototype  # => proto

Returns:

  • (Serie) —

    prototype serie

Raises:

#prototype? ⇒ Boolean Originally defined in module Series::Serie::Prototyping

Checks if serie is in prototype state.

Returns:

  • (Boolean) —

    true if prototype, false otherwise

#proxy ⇒ Object Originally defined in module Series::Operations

TODO add test case

#quantize(reference: nil, step: nil, value_attribute: nil, stops: nil, predictive: nil, left_open: nil, right_open: nil) ⇒ RawQuantizer, PredictiveQuantizer Originally defined in module Series::Operations

Quantizes time-value serie to discrete steps.

Quantization Modes

  • Raw: Rounds values to nearest step, interpolates between points
  • Predictive: Predicts crossings of quantization boundaries

Applications

  • Quantize MIDI controller data to discrete values
  • Convert continuous pitch bends to semitones
  • Snap timing to grid
  • Generate stepped automation curves
  • Convert analog input to digital steps

Examples:

Basic quantization

# Quantize to semitones. The events must carry the AbsTimed dataset:
# a bare hash of the same shape is not one and will not be processed.
pitch_bend = S({time: 0r, value: 60.3}.extend(Musa::Datasets::AbsTimed),
               {time: 1r, value: 61.8}.extend(Musa::Datasets::AbsTimed))
quantized = pitch_bend.quantize(step: 1)
quantized.i.to_a
# => [{time: 0r, value: 60, duration: 1/2r}, {time: 1/2r, value: 61, duration: 1/2r}]

Predictive quantization

continuous = S({time: 0r, value: 0}, {time: 4r, value: 10})
pred = continuous.quantize(step: 2, predictive: true)
# Generates crossing points at values 0, 2, 4, 6, 8, 10

Parameters:

  • reference (Numeric, nil) (defaults to: nil) —

    quantization reference (default: 0)

  • step (Numeric, nil) (defaults to: nil) —

    step size (default: 1)

  • value_attribute (Symbol, nil) (defaults to: nil) —

    attribute to quantize (default: :value)

  • stops (Boolean, nil) (defaults to: nil) —

    include stop points (default: false)

  • predictive (Boolean, nil) (defaults to: nil) —

    use predictive mode (default: false)

  • left_open (Boolean, nil) (defaults to: nil) —

    left boundary open (raw mode)

  • right_open (Boolean, nil) (defaults to: nil) —

    right boundary open (raw mode)

Returns:

  • (RawQuantizer, PredictiveQuantizer) —

    quantized serie

#queued ⇒ QueueSerie Originally defined in module Series::Operations

Wraps this serie in a queue.

Returns:

#randomize(random: nil) ⇒ Randomizer Originally defined in module Series::Operations

Randomizes order of values.

Shuffles values randomly. Requires finite serie.

Examples:

Shuffle

s = S(1, 2, 3, 4, 5).randomize
s.i.to_a  # => random permutation

The same seed is the same shuffle

S(1, 2, 3, 4, 5).randomize(random: 7).i.to_a  # => [5, 2, 4, 3, 1]
S(1, 2, 3, 4, 5).randomize(random: 7).i.to_a  # => [5, 2, 4, 3, 1]

Parameters:

  • random (Random, Integer, nil) (defaults to: nil) —

    a generator, or a seed to build one with, or nothing for an unseeded one -- the same three forms Constructors#RND takes, since they are the same idea offered twice.

Returns:

  • (Randomizer) —

    randomized serie

#remove(block = nil) {|value| ... } ⇒ Remover Originally defined in module Series::Operations

Removes values matching condition.

Filters out values where block returns true.

Examples:

Remove odds

s = S(1, 2, 3, 4, 5).remove { |n| n.odd? }
s.i.to_a  # => [2, 4]

Parameters:

  • block (Proc, nil) (defaults to: nil) —

    filter block

Yields:

  • filter condition

Yield Parameters:

  • value (Object) —

    current value

Yield Returns:

  • (Boolean) —

    true to remove value

Returns:

  • (Remover) —

    filtered serie

#repeat(times = nil, condition: nil) { ... } ⇒ Repeater, InfiniteRepeater Originally defined in module Series::Operations

Repeats serie multiple times or conditionally.

Three modes:

  • times: Repeat exact number of times
  • condition: Repeat while condition true
  • neither: Infinite repetition

Examples:

Fixed repetitions

s = S(1, 2, 3).repeat(3)
s.i.to_a  # => [1, 2, 3, 1, 2, 3, 1, 2, 3]

Conditional repeat

count = 0
S(1, 2, 3).repeat { (count += 1) < 3 }.i.to_a
# => [1, 2, 3, 1, 2, 3, 1, 2, 3]

# The block is asked between turns, so three turns run: the third
# answer is what stops a fourth.

Infinite repeat

s = S(1, 2, 3).repeat
s.infinite?  # => true

Parameters:

  • times (Integer, nil) (defaults to: nil) —

    number of repetitions

  • condition (Proc, nil) (defaults to: nil) —

    condition block

Yields:

  • optional condition block

Returns:

  • (Repeater, InfiniteRepeater) —

    repeating serie

#reverse ⇒ Reverser Originally defined in module Series::Operations

Reverses order of values.

Consumes entire serie and returns values in reverse order. Requires finite serie.

Examples:

Retrograde

s = S(1, 2, 3, 4).reverse
s.i.to_a  # => [4, 3, 2, 1]

Returns:

  • (Reverser) —

    reversed serie

#select(block = nil) {|value| ... } ⇒ Selector Originally defined in module Series::Operations

Selects values matching condition.

Keeps only values where block returns true.

Examples:

Select evens

s = S(1, 2, 3, 4, 5).select { |n| n.even? }
s.i.to_a  # => [2, 4]

Parameters:

  • block (Proc, nil) (defaults to: nil) —

    filter block

Yields:

  • filter condition

Yield Parameters:

  • value (Object) —

    current value

Yield Returns:

  • (Boolean) —

    true to keep value

Returns:

  • (Selector) —

    filtered serie

#shift(shift) ⇒ Shifter Originally defined in module Series::Operations

Rotates serie elements circularly.

Performs circular rotation of elements:

  • Negative values rotate left (first elements move to end)
  • Positive values rotate right (last elements move to beginning)
  • Zero performs no rotation

Note: Right rotation (positive values) requires finite series as the entire serie must be loaded into memory for rotation.

Examples:

Rotate left (negative shift)

s = S(60, 64, 67).shift(-1)  # First element moves to end
s.i.to_a  # => [64, 67, 60]

Rotate right (positive shift)

s = S(1, 2, 3, 4, 5).shift(2)  # Last 2 elements move to beginning
s.i.to_a  # => [4, 5, 1, 2, 3]

No rotation

s = S(1, 2, 3).shift(0)
s.i.to_a  # => [1, 2, 3]

Parameters:

  • shift (Integer) —

    rotation amount

    • Negative: rotate left by N positions
    • Positive: rotate right by N positions
    • Zero: no rotation

Returns:

  • (Shifter) —

    rotated serie

#skip(length) ⇒ Skipper Originally defined in module Series::Operations

Skips first N values.

Discards first length values, returns rest.

Examples:

Skip first 2

s = S(1, 2, 3, 4, 5).skip(2)
s.i.to_a  # => [3, 4, 5]

Parameters:

  • length (Integer) —

    number of values to skip

Returns:

  • (Skipper) —

    serie with skipped values

#split ⇒ Splitter Originally defined in module Series::Operations

Serie splitter for decomposing hash/array values into component series.

Splits series of hash or array values into individual component series, enabling independent access to each component.

Splitting Modes

  • Hash mode: Split {pitch: 60, velocity: 96} into separate series for :pitch and :velocity
  • Array mode: Split [60, 96] into separate series for indices 0, 1

Component Access

  • Hash: splitter[:pitch], splitter[:velocity]
  • Array: splitter[0], splitter[1]
  • Enumerable: splitter.each { |component| ... }

Use Cases

  • Separate polyphonic voices from single source
  • Independent processing of musical parameters
  • Extract specific components (pitch, duration, velocity, etc.)
  • Multi-track decomposition

Examples:

Split hash values

notes = S({pitch: 60, vel: 96}, {pitch: 64, vel: 80})
splitter = notes.split.i

pitches = splitter[:pitch]
velocities = splitter[:vel]

pitches.next_value  # => 60
velocities.next_value  # => 96

Split array values

pairs = S([1, 10], [2, 20], [3, 30]).split.i
first = pairs[0]
second = pairs[1]

Returns:

  • (Splitter) —

    hash/array splitter

#state ⇒ Symbol Originally defined in module Series::Serie::Prototyping

Returns current state of serie.

Attempts to resolve undefined state from sources before returning. State is one of: :prototype, :instance, or :undefined.

Returns:

  • (Symbol) —

    current state (:prototype, :instance, :undefined)

#switch(*indexed_series, **hash_series) ⇒ Switcher Originally defined in module Series::Operations

Reads one value from whichever serie the selector names.

The unselected series do not advance. Each one waits where it was and continues from there next time it is chosen, which is what makes this a dialogue between materials rather than a window onto them: two voices taking turns, each picking up its own thread.

Compare #multiplex, where everything advances and only one value is heard.

Examples:

Two materials taking turns, each continuing where it left off

question = S(:q1, :q2, :q3)
answer   = S(:a1, :a2)

S(0, 1, 0, 1, 0).switch(question, answer).i.to_a
# => [:q1, :a1, :q2, :a2, :q3]

Naming them instead of counting them

S(:melody, :bass, :melody).switch(melody: S(60, 62, 64),
                                  bass: S(36, 38)).i.to_a
# => [60, 36, 62]

Parameters:

  • indexed_series (Array) —

    series chosen by integer position

  • hash_series (Hash) —

    series chosen by name

Returns:

  • (Switcher) —

    switching serie

#switch_serie(*indexed_series, **hash_series) ⇒ SwitchFullSerie Originally defined in module Series::Operations

Switches to entirely different series based on selector.

Changes which serie is being consumed entirely.

Parameters:

  • indexed_series (Array) —

    series to switch between

  • hash_series (Hash) —

    series to switch between by key

Returns:

  • (SwitchFullSerie) —

    serie switcher

#undefined? ⇒ Boolean Originally defined in module Series::Serie::Prototyping

Checks if serie is in undefined state.

Returns:

  • (Boolean) —

    true if undefined, false otherwise

#union_timed(*other_timed_series, key: nil, **other_key_timed_series) ⇒ TimedUnionOfArrayOfTimedSeries, TimedUnionOfHashOfTimedSeries Originally defined in module Series::Operations

Combines this timed serie with others via TIMED_UNION.

Convenience method for unioning series, supporting both array and hash modes. Calls Constructors#TIMED_UNION constructor with appropriate parameters.

Array mode: s1.union_timed(s2, s3) Hash mode: s1.union_timed(key: :melody, bass: s2, drums: s3)

Examples:

Array mode

melody = S({ time: 0r, value: 60 })
bass = S({ time: 0r, value: 36 })

melody.union_timed(bass).i.next_value
# => { time: 0r, value: [60, 36] }

Hash mode

melody = S({ time: 0r, value: 60 })
bass = S({ time: 0r, value: 36 })
drums = S({ time: 0r, value: 38 })

melody.union_timed(key: :melody, bass: bass, drums: drums).i.next_value
# => { time: 0r, value: { melody: 60, bass: 36, drums: 38 } }

Parameters:

  • other_timed_series (Array<Serie>) —

    additional series (array mode)

  • key (Symbol, nil) (defaults to: nil) —

    key name for this serie (hash mode)

  • other_key_timed_series (Hash{Symbol => Serie}) —

    named series (hash mode)

Returns:

  • (TimedUnionOfArrayOfTimedSeries, TimedUnionOfHashOfTimedSeries) —

    union

Raises:

  • (ArgumentError) —

    if mixing array and hash modes

See Also:

#with(*with_series, on_restart: nil, isolate_values: nil, **with_key_series) {|main_value, with_values, with_key_values| ... } ⇒ With Also known as: eval Originally defined in module Series::Operations

Combines multiple series for mapping.

Synchronously iterates multiple series, passing all values to block. Enables multi-voice transformations and combinations.

Parameters

  • with_series: Positional series (passed as array to block)
  • with_key_series: Named series (passed as keywords to block)
  • on_restart: Block called on restart
  • isolate_values: Clone values to prevent mutation

Block Parameters

Block receives:

  • Main serie value (first argument)
  • Positional with_series values (array)
  • Keyword with_key_series values (keywords)

Examples:

Combine pitches and velocities

# Positional with-series cannot isolate their values, so say so:
# otherwise prefer the named form below.
pitches = S(60, 64, 67)
velocities = S(96, 80, 64)
notes = pitches.with(velocities, isolate_values: false) { |p, v| {pitch: p, velocity: v} }
notes.i.to_a  # => [{pitch: 60, velocity: 96}, {pitch: 64, velocity: 80}, {pitch: 67, velocity: 64}]

Named series

melody = S(60, 64, 67)
rhythm = S(1r, 1/2r, 1/2r)

melody.with(duration: rhythm) { |pitch, duration:|
  { pitch: pitch, duration: duration }
}.i.to_a
# => [{ pitch: 60, duration: 1r },
#     { pitch: 64, duration: 1/2r },
#     { pitch: 67, duration: 1/2r }]

Parameters:

  • with_series (Array<Serie>) —

    positional series to combine

  • on_restart (Proc, nil) (defaults to: nil) —

    restart callback

  • isolate_values (Boolean, nil) (defaults to: nil) —

    clone values to prevent mutation

  • with_key_series (Hash) —

    keyword series to combine

Yields:

  • combination block

Yield Parameters:

  • main_value (Object) —

    value from main serie

  • with_values (Array) —

    values from positional series

  • with_key_values (Hash) —

    values from keyword series

Yield Returns:

Returns:

  • (With) —

    combined serie