This is the same as SequenceT.

These are meant to suffix a sollu. Since the sollu is considered part of the duration, the number is one higher than the number of rests. E.g. din.__3 is a 3 count, and equivalent to din... __M is the version that doesn't do this.

These are meant to suffix a sollu. Since the sollu is considered part of the duration, the number is one higher than the number of rests. E.g. din.__3 is a 3 count, and equivalent to din... __M is the version that doesn't do this.

These are meant to suffix a sollu. Since the sollu is considered part of the duration, the number is one higher than the number of rests. E.g. din.__3 is a 3 count, and equivalent to din... __M is the version that doesn't do this.

These are meant to suffix a sollu. Since the sollu is considered part of the duration, the number is one higher than the number of rests. E.g. din.__3 is a 3 count, and equivalent to din... __M is the version that doesn't do this.

These are meant to suffix a sollu. Since the sollu is considered part of the duration, the number is one higher than the number of rests. E.g. din.__3 is a 3 count, and equivalent to din... __M is the version that doesn't do this.

These are meant to suffix a sollu. Since the sollu is considered part of the duration, the number is one higher than the number of rests. E.g. din.__3 is a 3 count, and equivalent to din... __M is the version that doesn't do this.

These are meant to suffix a sollu. Since the sollu is considered part of the duration, the number is one higher than the number of rests. E.g. din.__3 is a 3 count, and equivalent to din... __M is the version that doesn't do this.

These are meant to suffix a sollu. Since the sollu is considered part of the duration, the number is one higher than the number of rests. E.g. din.__3 is a 3 count, and equivalent to din... __M is the version that doesn't do this.

Like splitM_, but mark the sequences as groups. This way they remember the sollus which were dropped, and realize using the complete sequence, not the fragment.

TODO the class constraints are unnecessary, but if I want to verify eagerly I'll need them back.

Split the sequence at the given FMatra. Unlike splitM, this directly splits the sequence, it doesn't create a group.

Duration-using variants of the matra functions. These are only valid at the top level, in S.defaultTempo. TODO require Tempo arg?

Duration-using variants of the matra functions. These are only valid at the top level, in S.defaultTempo. TODO require Tempo arg?

Duration-using variants of the matra functions. These are only valid at the top level, in S.defaultTempo. TODO require Tempo arg?

Duration-using variants of the matra functions. These are only valid at the top level, in S.defaultTempo. TODO require Tempo arg?

Drop sollus equal in length to some others. This is intenedd for repeated sequences that get elided away, e.g. tri p7 . sandi p7 (p7.p6.p5).

I considered an annotation that automatically drops stuff from before which matches stuff afterwards, but it seemed more complicated and less reliable than just dropping explicitly.

Repeat thrice, with no karvai.

Repeat thrice, with the given separator. The _nomid variant doesn't add the mid tag, which is useful for nested calls.

Repeat thrice, with the given separator. The _nomid variant doesn't add the mid tag, which is useful for nested calls.

Three different patterns with the same separator.

Tirmanams with a variant final repeat.

tri_ with variable separators.

replicate + mconcat.

Intersperse between each stroke. TODO won't look inside a group

Multiple prefixes on a single suffix.

Succesively accumulate suffixes.

Reduce three times, with a separator.

reduceToL, except mconcat the result.

Reduce by a duration until a final duration.

Like reduceToL, but drop from the end instead of the front.

Start fully reduced, and expand n times by the given duration.

Unlike most other functions that drop from a sequence, this one doesn't make a group. Since these are used to construct a new sequence, it seems more confusing than helpful.

As with (and (==), this is higher precedence than (.), so paretheses are needed: (a.b) replaceStart xyz.

Unlike most other functions that drop from a sequence, this one doesn't make a group. Since these are used to construct a new sequence, it seems more confusing than helpful.

As with (and (==), this is higher precedence than (.), so paretheses are needed: (a.b) replaceStart xyz.

Operators to embed a sequence at the beginning or end of sarvalaghu.

The precedence is such that sequences must be parenthesized: (a.b) <== 8. If it's lower than (.), then a.b <== 4 . c.d <== 4 is still wrong. Since sometimes I want (.) to be tighter and sometimes looser, I go for always looser and require parentheses, just like function calls. But by being at 8, at least I can be below (^) and (§).

Like matrasOf, but throw an error if it's not integral.

I think defaultTempo is ok because these functions are used on fragments.

Set relative speed.

Mnemonic: speed up, slow down.

Mnemonic: speed up, slow down.

Mark a theme group.

Mark a pattern group. These are like patterns, except with a specific realization.

Make a named group.

Infix operator to Solkattu.Tag all of the sollus it applies to.

Solkattu.Middle.

Set if not already set.

Align to the end of the avartanam, with rests.

This should only be used at the top level, since it gets the timing wrong under a tempo change.

Align to the end of the given number of aksharams.

ktknkook -> kt_kn_ko_ok

Append a sequence after a number of syllables. This works across groups, but not tempo changes.

Apply a stateful transformation within groups. Since the transformation is allowed to add or remove notes, this will throw if there is a TempoChange in there, since now we are changing an unknown amount of time.

Return groups, or runs of Notes. Nothing if there's a TempoChange in there.