2020-05-18 15:59:07 +03:00
|
|
|
module Data.Vect
|
|
|
|
|
|
|
|
import Data.List
|
|
|
|
import Data.Nat
|
|
|
|
import public Data.Fin
|
2021-01-27 21:23:08 +03:00
|
|
|
import public Data.Zippable
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
import Decidable.Equality
|
|
|
|
|
2020-07-08 02:55:43 +03:00
|
|
|
%default total
|
|
|
|
|
2020-05-18 15:59:07 +03:00
|
|
|
public export
|
|
|
|
data Vect : (len : Nat) -> (elem : Type) -> Type where
|
|
|
|
||| Empty vector
|
|
|
|
Nil : Vect Z elem
|
|
|
|
||| A non-empty vector of length `S len`, consisting of a head element and
|
|
|
|
||| the rest of the list, of length `len`.
|
Remove linearity subtyping
It's disappointing to have to do this, but I think necessary because
various issue reports have shown it to be unsound (at least as far as
inference goes) and, at the very least, confusing. This patch brings us
back to the basic rules of QTT.
On the one hand, this makes the 1 multiplicity less useful, because it
means we can't flag arguments as being used exactly once which would be
useful for optimisation purposes as well as precision in the type. On
the other hand, it removes some complexity (and a hack) from
unification, and has the advantage of being correct! Also, I still
consider the 1 multiplicity an experiment.
We can still do interesting things like protocol state tracking, which
is my primary motivation at least.
Ideally, if the 1 multiplicity is going to be more generall useful,
we'll need some kind of way of doing multiplicity polymorphism in the
future. I don't think subtyping is the way (I've pretty much always come
to regret adding some form of subtyping).
Fixes #73 (and maybe some others).
2020-12-27 22:58:35 +03:00
|
|
|
(::) : (x : elem) -> (xs : Vect len elem) -> Vect (S len) elem
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
-- Hints for interactive editing
|
2021-01-27 21:23:08 +03:00
|
|
|
%name Vect xs, ys, zs, ws
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
public export
|
|
|
|
length : (xs : Vect len elem) -> Nat
|
|
|
|
length [] = 0
|
2020-07-08 02:55:43 +03:00
|
|
|
length (_::xs) = 1 + length xs
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
||| Show that the length function on vectors in fact calculates the length
|
2020-06-27 17:54:35 +03:00
|
|
|
export
|
2020-07-08 02:55:43 +03:00
|
|
|
lengthCorrect : (xs : Vect len elem) -> length xs = len
|
|
|
|
lengthCorrect [] = Refl
|
|
|
|
lengthCorrect (_ :: xs) = rewrite lengthCorrect xs in Refl
|
2020-05-18 15:59:07 +03:00
|
|
|
|
2020-07-04 13:02:04 +03:00
|
|
|
||| If two vectors are equal, their heads and tails are equal
|
|
|
|
export
|
|
|
|
vectInjective : {0 xs : Vect n a} -> {0 ys : Vect m b} -> x::xs = y::ys -> (x = y, xs = ys)
|
|
|
|
vectInjective Refl = (Refl, Refl)
|
|
|
|
|
2020-05-18 15:59:07 +03:00
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Indexing into vectors
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
||| All but the first element of the vector
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| tail [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
tail : Vect (S len) elem -> Vect len elem
|
2020-07-08 02:55:43 +03:00
|
|
|
tail (_::xs) = xs
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
||| Only the first element of the vector
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| head [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
head : Vect (S len) elem -> elem
|
2020-07-08 02:55:43 +03:00
|
|
|
head (x::_) = x
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
||| The last element of the vector
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| last [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
last : Vect (S len) elem -> elem
|
2020-07-08 02:55:43 +03:00
|
|
|
last [x] = x
|
|
|
|
last (_::y::ys) = last $ y::ys
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
||| All but the last element of the vector
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| init [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
init : Vect (S len) elem -> Vect len elem
|
2020-07-08 02:55:43 +03:00
|
|
|
init [_] = []
|
2020-05-18 15:59:07 +03:00
|
|
|
init (x::y::ys) = x :: init (y::ys)
|
|
|
|
|
2020-07-03 17:25:30 +03:00
|
|
|
||| Extract the first `n` elements of a Vect.
|
|
|
|
public export
|
2020-12-27 23:11:06 +03:00
|
|
|
take : (n : Nat)
|
2020-07-03 17:25:30 +03:00
|
|
|
-> ( xs : Vect (n + m) type)
|
|
|
|
-> Vect n type
|
|
|
|
take 0 xs = Nil
|
|
|
|
take (S k) (x :: xs) = x :: take k xs
|
|
|
|
|
2020-05-18 15:59:07 +03:00
|
|
|
||| Extract a particular element from a vector
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| index 1 [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
index : Fin len -> Vect len elem -> elem
|
2020-07-08 02:55:43 +03:00
|
|
|
index FZ (x::_) = x
|
|
|
|
index (FS k) (_::xs) = index k xs
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
||| Insert an element at a particular index
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| insertAt 1 8 [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
Remove linearity subtyping
It's disappointing to have to do this, but I think necessary because
various issue reports have shown it to be unsound (at least as far as
inference goes) and, at the very least, confusing. This patch brings us
back to the basic rules of QTT.
On the one hand, this makes the 1 multiplicity less useful, because it
means we can't flag arguments as being used exactly once which would be
useful for optimisation purposes as well as precision in the type. On
the other hand, it removes some complexity (and a hack) from
unification, and has the advantage of being correct! Also, I still
consider the 1 multiplicity an experiment.
We can still do interesting things like protocol state tracking, which
is my primary motivation at least.
Ideally, if the 1 multiplicity is going to be more generall useful,
we'll need some kind of way of doing multiplicity polymorphism in the
future. I don't think subtyping is the way (I've pretty much always come
to regret adding some form of subtyping).
Fixes #73 (and maybe some others).
2020-12-27 22:58:35 +03:00
|
|
|
insertAt : (idx : Fin (S len)) -> (x : elem) -> (xs : Vect len elem) -> Vect (S len) elem
|
2020-05-18 15:59:07 +03:00
|
|
|
insertAt FZ y xs = y :: xs
|
|
|
|
insertAt (FS k) y (x::xs) = x :: insertAt k y xs
|
|
|
|
|
|
|
|
||| Construct a new vector consisting of all but the indicated element
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| deleteAt 1 [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
2020-07-14 09:25:46 +03:00
|
|
|
deleteAt : Fin (S len) -> Vect (S len) elem -> Vect len elem
|
|
|
|
deleteAt FZ (_::xs) = xs
|
|
|
|
deleteAt (FS k) [x] = absurd k
|
|
|
|
deleteAt (FS k) (x::xs@(_::_)) = x :: deleteAt k xs
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
||| Replace an element at a particlar index with another
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| replaceAt 1 8 [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
replaceAt : Fin len -> elem -> Vect len elem -> Vect len elem
|
2020-07-08 02:55:43 +03:00
|
|
|
replaceAt FZ y (_::xs) = y :: xs
|
2020-05-18 15:59:07 +03:00
|
|
|
replaceAt (FS k) y (x::xs) = x :: replaceAt k y xs
|
|
|
|
|
|
|
|
||| Replace the element at a particular index with the result of applying a function to it
|
|
|
|
||| @ i the index to replace at
|
|
|
|
||| @ f the update function
|
|
|
|
||| @ xs the vector to replace in
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| updateAt 1 (+10) [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
updateAt : (i : Fin len) -> (f : elem -> elem) -> (xs : Vect len elem) -> Vect len elem
|
|
|
|
updateAt FZ f (x::xs) = f x :: xs
|
|
|
|
updateAt (FS k) f (x::xs) = x :: updateAt k f xs
|
|
|
|
|
|
|
|
||| Append two vectors
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| [1,2,3,4] ++ [5,6]
|
|
|
|
||| ```
|
|
|
|
public export
|
Remove linearity subtyping
It's disappointing to have to do this, but I think necessary because
various issue reports have shown it to be unsound (at least as far as
inference goes) and, at the very least, confusing. This patch brings us
back to the basic rules of QTT.
On the one hand, this makes the 1 multiplicity less useful, because it
means we can't flag arguments as being used exactly once which would be
useful for optimisation purposes as well as precision in the type. On
the other hand, it removes some complexity (and a hack) from
unification, and has the advantage of being correct! Also, I still
consider the 1 multiplicity an experiment.
We can still do interesting things like protocol state tracking, which
is my primary motivation at least.
Ideally, if the 1 multiplicity is going to be more generall useful,
we'll need some kind of way of doing multiplicity polymorphism in the
future. I don't think subtyping is the way (I've pretty much always come
to regret adding some form of subtyping).
Fixes #73 (and maybe some others).
2020-12-27 22:58:35 +03:00
|
|
|
(++) : (xs : Vect m elem) -> (ys : Vect n elem) -> Vect (m + n) elem
|
2020-05-18 15:59:07 +03:00
|
|
|
(++) [] ys = ys
|
|
|
|
(++) (x::xs) ys = x :: xs ++ ys
|
|
|
|
|
2021-01-03 17:46:33 +03:00
|
|
|
||| Add an element at the end of the vector.
|
|
|
|
||| The main use case for it is to get the expected type signature
|
|
|
|
||| `Vect n a -> a -> Vect (S n) a` instead of
|
|
|
|
||| `Vect n a -> a -> Vect (n + 1) a` which you get by using `++ [x]`
|
|
|
|
|||
|
|
|
|
||| Snoc gets its name by reversing `cons`, indicating we are
|
|
|
|
||| tacking on the element at the end rather than the begining.
|
|
|
|
||| `append` would also be a suitable name.
|
|
|
|
|||
|
|
|
|
||| @ xs The vector to be appended
|
|
|
|
||| @ v The value to append
|
|
|
|
public export
|
|
|
|
snoc : (xs : Vect n a) -> (v : a) -> Vect (S n) a
|
|
|
|
snoc [] v = [v]
|
|
|
|
snoc (x :: xs) v = x :: snoc xs v
|
|
|
|
|
2020-05-18 15:59:07 +03:00
|
|
|
||| Repeate some value some number of times.
|
|
|
|
|||
|
|
|
|
||| @ len the number of times to repeat it
|
|
|
|
||| @ x the value to repeat
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| replicate 4 1
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
replicate : (len : Nat) -> (x : elem) -> Vect len elem
|
2020-07-08 02:55:43 +03:00
|
|
|
replicate Z _ = []
|
2020-05-18 15:59:07 +03:00
|
|
|
replicate (S k) x = x :: replicate k x
|
|
|
|
|
|
|
|
||| Merge two ordered vectors
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| mergeBy compare (fromList [1,3,5]) (fromList [2,3,4,5,6])
|
|
|
|
||| ```
|
|
|
|
export
|
|
|
|
mergeBy : (elem -> elem -> Ordering) -> (xs : Vect n elem) -> (ys : Vect m elem) -> Vect (n + m) elem
|
2021-02-16 14:56:43 +03:00
|
|
|
mergeBy _ [] ys = ys
|
|
|
|
mergeBy _ xs [] = rewrite plusZeroRightNeutral n in xs
|
2020-05-18 15:59:07 +03:00
|
|
|
mergeBy {n = S k} {m = S k'} order (x :: xs) (y :: ys)
|
|
|
|
= case order x y of
|
|
|
|
LT => x :: mergeBy order xs (y :: ys)
|
|
|
|
_ => rewrite sym (plusSuccRightSucc k k') in
|
|
|
|
y :: mergeBy order (x :: xs) ys
|
|
|
|
|
|
|
|
export
|
|
|
|
merge : Ord elem => Vect n elem -> Vect m elem -> Vect (n + m) elem
|
|
|
|
merge = mergeBy compare
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Transformations
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
||| Reverse the order of the elements of a vector
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| reverse [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
Remove linearity subtyping
It's disappointing to have to do this, but I think necessary because
various issue reports have shown it to be unsound (at least as far as
inference goes) and, at the very least, confusing. This patch brings us
back to the basic rules of QTT.
On the one hand, this makes the 1 multiplicity less useful, because it
means we can't flag arguments as being used exactly once which would be
useful for optimisation purposes as well as precision in the type. On
the other hand, it removes some complexity (and a hack) from
unification, and has the advantage of being correct! Also, I still
consider the 1 multiplicity an experiment.
We can still do interesting things like protocol state tracking, which
is my primary motivation at least.
Ideally, if the 1 multiplicity is going to be more generall useful,
we'll need some kind of way of doing multiplicity polymorphism in the
future. I don't think subtyping is the way (I've pretty much always come
to regret adding some form of subtyping).
Fixes #73 (and maybe some others).
2020-12-27 22:58:35 +03:00
|
|
|
reverse : (xs : Vect len elem) -> Vect len elem
|
2020-05-18 15:59:07 +03:00
|
|
|
reverse xs = go [] xs
|
Remove linearity subtyping
It's disappointing to have to do this, but I think necessary because
various issue reports have shown it to be unsound (at least as far as
inference goes) and, at the very least, confusing. This patch brings us
back to the basic rules of QTT.
On the one hand, this makes the 1 multiplicity less useful, because it
means we can't flag arguments as being used exactly once which would be
useful for optimisation purposes as well as precision in the type. On
the other hand, it removes some complexity (and a hack) from
unification, and has the advantage of being correct! Also, I still
consider the 1 multiplicity an experiment.
We can still do interesting things like protocol state tracking, which
is my primary motivation at least.
Ideally, if the 1 multiplicity is going to be more generall useful,
we'll need some kind of way of doing multiplicity polymorphism in the
future. I don't think subtyping is the way (I've pretty much always come
to regret adding some form of subtyping).
Fixes #73 (and maybe some others).
2020-12-27 22:58:35 +03:00
|
|
|
where go : Vect n elem -> Vect m elem -> Vect (n+m) elem
|
2020-05-18 15:59:07 +03:00
|
|
|
go {n} acc [] = rewrite plusZeroRightNeutral n in acc
|
|
|
|
go {n} {m=S m} acc (x :: xs) = rewrite sym $ plusSuccRightSucc n m
|
|
|
|
in go (x::acc) xs
|
|
|
|
|
|
|
|
||| Alternate an element between the other elements of a vector
|
|
|
|
||| @ sep the element to intersperse
|
|
|
|
||| @ xs the vector to separate with `sep`
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| intersperse 0 [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
export
|
|
|
|
intersperse : (sep : elem) -> (xs : Vect len elem) -> Vect (len + pred len) elem
|
|
|
|
intersperse sep [] = []
|
|
|
|
intersperse sep (x::xs) = x :: intersperse' sep xs
|
|
|
|
where
|
|
|
|
intersperse' : elem -> Vect n elem -> Vect (n + n) elem
|
|
|
|
intersperse' sep [] = []
|
|
|
|
intersperse' {n=S n} sep (x::xs) = rewrite sym $ plusSuccRightSucc n n
|
|
|
|
in sep :: x :: intersperse' sep xs
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Conversion from list (toList is provided by Foldable)
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
2020-09-05 11:41:31 +03:00
|
|
|
public export
|
|
|
|
toVect : (n : Nat) -> List a -> Maybe (Vect n a)
|
|
|
|
toVect Z [] = Just []
|
|
|
|
toVect (S k) (x :: xs)
|
|
|
|
= do xs' <- toVect k xs
|
|
|
|
pure (x :: xs')
|
|
|
|
toVect _ _ = Nothing
|
|
|
|
|
2020-05-18 15:59:07 +03:00
|
|
|
public export
|
Remove linearity subtyping
It's disappointing to have to do this, but I think necessary because
various issue reports have shown it to be unsound (at least as far as
inference goes) and, at the very least, confusing. This patch brings us
back to the basic rules of QTT.
On the one hand, this makes the 1 multiplicity less useful, because it
means we can't flag arguments as being used exactly once which would be
useful for optimisation purposes as well as precision in the type. On
the other hand, it removes some complexity (and a hack) from
unification, and has the advantage of being correct! Also, I still
consider the 1 multiplicity an experiment.
We can still do interesting things like protocol state tracking, which
is my primary motivation at least.
Ideally, if the 1 multiplicity is going to be more generall useful,
we'll need some kind of way of doing multiplicity polymorphism in the
future. I don't think subtyping is the way (I've pretty much always come
to regret adding some form of subtyping).
Fixes #73 (and maybe some others).
2020-12-27 22:58:35 +03:00
|
|
|
fromList' : (xs : Vect len elem) -> (l : List elem) -> Vect (length l + len) elem
|
2020-05-18 15:59:07 +03:00
|
|
|
fromList' ys [] = ys
|
2021-02-16 14:56:43 +03:00
|
|
|
fromList' ys (x::xs) =
|
2020-05-18 15:59:07 +03:00
|
|
|
rewrite (plusSuccRightSucc (length xs) len) in
|
|
|
|
fromList' (x::ys) xs
|
|
|
|
|
|
|
|
||| Convert a list to a vector.
|
|
|
|
|||
|
|
|
|
||| The length of the list should be statically known.
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| fromList [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
Remove linearity subtyping
It's disappointing to have to do this, but I think necessary because
various issue reports have shown it to be unsound (at least as far as
inference goes) and, at the very least, confusing. This patch brings us
back to the basic rules of QTT.
On the one hand, this makes the 1 multiplicity less useful, because it
means we can't flag arguments as being used exactly once which would be
useful for optimisation purposes as well as precision in the type. On
the other hand, it removes some complexity (and a hack) from
unification, and has the advantage of being correct! Also, I still
consider the 1 multiplicity an experiment.
We can still do interesting things like protocol state tracking, which
is my primary motivation at least.
Ideally, if the 1 multiplicity is going to be more generall useful,
we'll need some kind of way of doing multiplicity polymorphism in the
future. I don't think subtyping is the way (I've pretty much always come
to regret adding some form of subtyping).
Fixes #73 (and maybe some others).
2020-12-27 22:58:35 +03:00
|
|
|
fromList : (xs : List elem) -> Vect (length xs) elem
|
2020-05-18 15:59:07 +03:00
|
|
|
fromList l =
|
|
|
|
rewrite (sym $ plusZeroRightNeutral (length l)) in
|
|
|
|
reverse $ fromList' [] l
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Equality
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
public export
|
2020-07-04 13:02:04 +03:00
|
|
|
Eq a => Eq (Vect n a) where
|
2020-05-18 15:59:07 +03:00
|
|
|
(==) [] [] = True
|
|
|
|
(==) (x::xs) (y::ys) = x == y && xs == ys
|
|
|
|
|
2020-07-04 13:02:04 +03:00
|
|
|
export
|
|
|
|
DecEq a => DecEq (Vect n a) where
|
|
|
|
decEq [] [] = Yes Refl
|
|
|
|
decEq (x::xs) (y::ys) with (decEq x y, decEq xs ys)
|
|
|
|
decEq (x::xs) (x::xs) | (Yes Refl, Yes Refl) = Yes Refl
|
|
|
|
decEq (x::xs) (y::ys) | (No nhd, _) = No $ nhd . fst . vectInjective
|
|
|
|
decEq (x::xs) (y::ys) | (_, No ntl) = No $ ntl . snd . vectInjective
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Order
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
public export
|
|
|
|
implementation Ord elem => Ord (Vect len elem) where
|
|
|
|
compare [] [] = EQ
|
|
|
|
compare (x::xs) (y::ys)
|
|
|
|
= case compare x y of
|
|
|
|
EQ => compare xs ys
|
|
|
|
x => x
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Maps
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
public export
|
|
|
|
implementation Functor (Vect n) where
|
|
|
|
map f [] = []
|
|
|
|
map f (x::xs) = f x :: map f xs
|
|
|
|
|
|
|
|
||| Map a partial function across a vector, returning those elements for which
|
|
|
|
||| the function had a value.
|
|
|
|
|||
|
|
|
|
||| The first projection of the resulting pair (ie the length) will always be
|
|
|
|
||| at most the length of the input vector. This is not, however, guaranteed
|
|
|
|
||| by the type.
|
|
|
|
|||
|
|
|
|
||| @ f the partial function (expressed by returning `Maybe`)
|
|
|
|
||| @ xs the vector to check for results
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| mapMaybe ((find (=='a')) . unpack) (fromList ["abc","ade","bgh","xyz"])
|
|
|
|
||| ```
|
|
|
|
export
|
|
|
|
mapMaybe : (f : a -> Maybe b) -> (xs : Vect len a) -> (m : Nat ** Vect m b)
|
|
|
|
mapMaybe f [] = (_ ** [])
|
|
|
|
mapMaybe f (x::xs) =
|
|
|
|
let (len ** ys) = mapMaybe f xs
|
|
|
|
in case f x of
|
|
|
|
Just y => (S len ** y :: ys)
|
|
|
|
Nothing => ( len ** ys)
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Folds
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
public export
|
|
|
|
foldrImpl : (t -> acc -> acc) -> acc -> (acc -> acc) -> Vect n t -> acc
|
|
|
|
foldrImpl f e go [] = go e
|
|
|
|
foldrImpl f e go (x::xs) = foldrImpl f e (go . (f x)) xs
|
|
|
|
|
|
|
|
public export
|
|
|
|
implementation Foldable (Vect n) where
|
|
|
|
foldr f e xs = foldrImpl f e id xs
|
|
|
|
|
2020-12-10 21:04:23 +03:00
|
|
|
null [] = True
|
|
|
|
null _ = False
|
|
|
|
|
2020-05-18 15:59:07 +03:00
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Special folds
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
||| Flatten a vector of equal-length vectors
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| concat [[1,2,3], [4,5,6]]
|
|
|
|
||| ```
|
|
|
|
public export
|
Remove linearity subtyping
It's disappointing to have to do this, but I think necessary because
various issue reports have shown it to be unsound (at least as far as
inference goes) and, at the very least, confusing. This patch brings us
back to the basic rules of QTT.
On the one hand, this makes the 1 multiplicity less useful, because it
means we can't flag arguments as being used exactly once which would be
useful for optimisation purposes as well as precision in the type. On
the other hand, it removes some complexity (and a hack) from
unification, and has the advantage of being correct! Also, I still
consider the 1 multiplicity an experiment.
We can still do interesting things like protocol state tracking, which
is my primary motivation at least.
Ideally, if the 1 multiplicity is going to be more generall useful,
we'll need some kind of way of doing multiplicity polymorphism in the
future. I don't think subtyping is the way (I've pretty much always come
to regret adding some form of subtyping).
Fixes #73 (and maybe some others).
2020-12-27 22:58:35 +03:00
|
|
|
concat : (xss : Vect m (Vect n elem)) -> Vect (m * n) elem
|
2020-05-18 15:59:07 +03:00
|
|
|
concat [] = []
|
|
|
|
concat (v::vs) = v ++ Vect.concat vs
|
|
|
|
|
|
|
|
||| Foldr without seeding the accumulator
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| foldr1 (-) (fromList [1,2,3])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
foldr1 : (t -> t -> t) -> Vect (S n) t -> t
|
|
|
|
foldr1 f [x] = x
|
|
|
|
foldr1 f (x::y::xs) = f x (foldr1 f (y::xs))
|
|
|
|
|
|
|
|
||| Foldl without seeding the accumulator
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| foldl1 (-) (fromList [1,2,3])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
foldl1 : (t -> t -> t) -> Vect (S n) t -> t
|
|
|
|
foldl1 f (x::xs) = foldl f x xs
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Scans
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
||| The scanl function is similar to foldl, but returns all the intermediate
|
|
|
|
||| accumulator states in the form of a vector.
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| scanl (-) 0 (fromList [1,2,3])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
scanl : (res -> elem -> res) -> res -> Vect len elem -> Vect (S len) res
|
|
|
|
scanl f q [] = [q]
|
|
|
|
scanl f q (x::xs) = q :: scanl f (f q x) xs
|
|
|
|
|
|
|
|
||| The scanl1 function is a variant of scanl that doesn't require an explicit
|
|
|
|
||| starting value.
|
|
|
|
||| It assumes the first element of the vector to be the starting value and then
|
|
|
|
||| starts the fold with the element following it.
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| scanl1 (-) (fromList [1,2,3])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
scanl1 : (elem -> elem -> elem) -> Vect len elem -> Vect len elem
|
|
|
|
scanl1 f [] = []
|
|
|
|
scanl1 f (x::xs) = scanl f x xs
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Membership tests
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
||| Search for an item using a user-provided test
|
|
|
|
||| @ p the equality test
|
|
|
|
||| @ e the item to search for
|
|
|
|
||| @ xs the vector to search in
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| elemBy (==) 2 [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
elemBy : (p : elem -> elem -> Bool) -> (e : elem) -> (xs : Vect len elem) -> Bool
|
|
|
|
elemBy p e [] = False
|
|
|
|
elemBy p e (x::xs) = p e x || elemBy p e xs
|
|
|
|
|
|
|
|
||| Use the default Boolean equality on elements to search for an item
|
|
|
|
||| @ x what to search for
|
|
|
|
||| @ xs where to search
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| elem 3 [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
elem : Eq elem => (x : elem) -> (xs : Vect len elem) -> Bool
|
|
|
|
elem = elemBy (==)
|
|
|
|
|
|
|
|
||| Find the association of some key with a user-provided comparison
|
|
|
|
||| @ p the comparison operator for keys (True if they match)
|
|
|
|
||| @ e the key to look for
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| lookupBy (==) 2 [(1, 'a'), (2, 'b'), (3, 'c')]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
lookupBy : (p : key -> key -> Bool) -> (e : key) -> (xs : Vect n (key, val)) -> Maybe val
|
|
|
|
lookupBy p e [] = Nothing
|
|
|
|
lookupBy p e ((l, r)::xs) = if p e l then Just r else lookupBy p e xs
|
|
|
|
|
|
|
|
||| Find the assocation of some key using the default Boolean equality test
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| lookup 3 [(1, 'a'), (2, 'b'), (3, 'c')]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
lookup : Eq key => key -> Vect n (key, val) -> Maybe val
|
|
|
|
lookup = lookupBy (==)
|
|
|
|
|
|
|
|
||| Check if any element of xs is found in elems by a user-provided comparison
|
|
|
|
||| @ p the comparison operator
|
|
|
|
||| @ elems the vector to search
|
|
|
|
||| @ xs what to search for
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| hasAnyBy (==) [2,5] [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
hasAnyBy : (p : elem -> elem -> Bool) -> (elems : Vect m elem) -> (xs : Vect len elem) -> Bool
|
|
|
|
hasAnyBy p elems [] = False
|
|
|
|
hasAnyBy p elems (x::xs) = elemBy p x elems || hasAnyBy p elems xs
|
|
|
|
|
|
|
|
||| Check if any element of xs is found in elems using the default Boolean equality test
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| hasAny [2,5] [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
hasAny : Eq elem => Vect m elem -> Vect len elem -> Bool
|
|
|
|
hasAny = hasAnyBy (==)
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Searching with a predicate
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
||| Find the first element of the vector that satisfies some test
|
|
|
|
||| @ p the test to satisfy
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| find (== 3) [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
find : (p : elem -> Bool) -> (xs : Vect len elem) -> Maybe elem
|
|
|
|
find p [] = Nothing
|
|
|
|
find p (x::xs) = if p x then Just x else find p xs
|
|
|
|
|
|
|
|
||| Find the index of the first element of the vector that satisfies some test
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| findIndex (== 3) [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
findIndex : (elem -> Bool) -> Vect len elem -> Maybe (Fin len)
|
|
|
|
findIndex p [] = Nothing
|
|
|
|
findIndex p (x :: xs) = if p x then Just FZ else map FS (findIndex p xs)
|
|
|
|
|
|
|
|
||| Find the indices of all elements that satisfy some test
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| findIndices (< 3) [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
findIndices : (elem -> Bool) -> Vect m elem -> List (Fin m)
|
|
|
|
findIndices p [] = []
|
|
|
|
findIndices p (x :: xs)
|
|
|
|
= let is = map FS $ findIndices p xs in
|
|
|
|
if p x then FZ :: is else is
|
|
|
|
|
|
|
|
||| Find the index of the first element of the vector that satisfies some test
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| elemIndexBy (==) 3 [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
elemIndexBy : (elem -> elem -> Bool) -> elem -> Vect m elem -> Maybe (Fin m)
|
|
|
|
elemIndexBy p e = findIndex $ p e
|
|
|
|
|
|
|
|
||| Find the index of the first element of the vector equal to the given one.
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| elemIndex 3 [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
elemIndex : Eq elem => elem -> Vect m elem -> Maybe (Fin m)
|
|
|
|
elemIndex = elemIndexBy (==)
|
|
|
|
|
|
|
|
||| Find the indices of all elements that satisfy some test
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| elemIndicesBy (<=) 3 [1,2,3,4]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
elemIndicesBy : (elem -> elem -> Bool) -> elem -> Vect m elem -> List (Fin m)
|
|
|
|
elemIndicesBy p e = findIndices $ p e
|
|
|
|
|
|
|
|
||| Find the indices of all elements uquals to the given one
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| elemIndices 3 [1,2,3,4,3]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
elemIndices : Eq elem => elem -> Vect m elem -> List (Fin m)
|
|
|
|
elemIndices = elemIndicesBy (==)
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Filters
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
||| Find all elements of a vector that satisfy some test
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| filter (< 3) (fromList [1,2,3,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
filter : (elem -> Bool) -> Vect len elem -> (p ** Vect p elem)
|
|
|
|
filter p [] = ( _ ** [] )
|
|
|
|
filter p (x::xs) =
|
|
|
|
let (_ ** tail) = filter p xs
|
|
|
|
in if p x then
|
|
|
|
(_ ** x::tail)
|
|
|
|
else
|
|
|
|
(_ ** tail)
|
|
|
|
|
|
|
|
||| Make the elements of some vector unique by some test
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| nubBy (==) (fromList [1,2,2,3,4,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
nubBy : (elem -> elem -> Bool) -> Vect len elem -> (p ** Vect p elem)
|
|
|
|
nubBy = nubBy' []
|
|
|
|
where
|
|
|
|
nubBy' : forall len . Vect m elem -> (elem -> elem -> Bool) -> Vect len elem -> (p ** Vect p elem)
|
|
|
|
nubBy' acc p [] = (_ ** [])
|
|
|
|
nubBy' acc p (x::xs) with (elemBy p x acc)
|
|
|
|
nubBy' acc p (x :: xs) | True = nubBy' acc p xs
|
|
|
|
nubBy' acc p (x :: xs) | False with (nubBy' (x::acc) p xs)
|
|
|
|
nubBy' acc p (x :: xs) | False | (_ ** tail) = (_ ** x::tail)
|
|
|
|
|
|
|
|
||| Make the elements of some vector unique by the default Boolean equality
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| nub (fromList [1,2,2,3,4,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
nub : Eq elem => Vect len elem -> (p ** Vect p elem)
|
|
|
|
nub = nubBy (==)
|
|
|
|
|
|
|
|
||| Delete first element from list according to some test
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| deleteBy (<) 3 (fromList [1,2,2,3,4,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
deleteBy : {len : _} -> -- needed for the dependent pair
|
|
|
|
(elem -> elem -> Bool) -> elem -> Vect len elem -> (p ** Vect p elem)
|
|
|
|
deleteBy _ _ [] = (_ ** [])
|
|
|
|
deleteBy eq x (y::ys) =
|
|
|
|
let (len ** zs) = deleteBy eq x ys
|
|
|
|
in if x `eq` y then (_ ** ys) else (S len ** y ::zs)
|
|
|
|
|
|
|
|
||| Delete first element from list equal to the given one
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| delete 2 (fromList [1,2,2,3,4,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
delete : {len : _} ->
|
|
|
|
Eq elem => elem -> Vect len elem -> (p ** Vect p elem)
|
|
|
|
delete = deleteBy (==)
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Splitting and breaking lists
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
||| A tuple where the first element is a `Vect` of the `n` first elements and
|
|
|
|
||| the second element is a `Vect` of the remaining elements of the original.
|
|
|
|
||| It is equivalent to `(take n xs, drop n xs)` (`splitAtTakeDrop`),
|
|
|
|
||| but is more efficient.
|
|
|
|
|||
|
|
|
|
||| @ n the index to split at
|
|
|
|
||| @ xs the `Vect` to split in two
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| splitAt 2 (fromList [1,2,3,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
splitAt : (n : Nat) -> (xs : Vect (n + m) elem) -> (Vect n elem, Vect m elem)
|
|
|
|
splitAt Z xs = ([], xs)
|
|
|
|
splitAt (S k) (x :: xs) with (splitAt k {m} xs)
|
|
|
|
splitAt (S k) (x :: xs) | (tk, dr) = (x :: tk, dr)
|
|
|
|
|
|
|
|
||| A tuple where the first element is a `Vect` of the `n` elements passing given test
|
|
|
|
||| and the second element is a `Vect` of the remaining elements of the original.
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| partition (== 2) (fromList [1,2,3,2,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
partition : (elem -> Bool) -> Vect len elem -> ((p ** Vect p elem), (q ** Vect q elem))
|
|
|
|
partition p [] = ((_ ** []), (_ ** []))
|
|
|
|
partition p (x::xs) =
|
|
|
|
let ((leftLen ** lefts), (rightLen ** rights)) = partition p xs in
|
|
|
|
if p x then
|
|
|
|
((S leftLen ** x::lefts), (rightLen ** rights))
|
|
|
|
else
|
|
|
|
((leftLen ** lefts), (S rightLen ** x::rights))
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Predicates
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
||| Verify vector prefix
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| isPrefixOfBy (==) (fromList [1,2]) (fromList [1,2,3,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
isPrefixOfBy : (elem -> elem -> Bool) -> Vect m elem -> Vect len elem -> Bool
|
|
|
|
isPrefixOfBy p [] right = True
|
|
|
|
isPrefixOfBy p left [] = False
|
|
|
|
isPrefixOfBy p (x::xs) (y::ys) = p x y && isPrefixOfBy p xs ys
|
|
|
|
|
|
|
|
||| Verify vector prefix
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| isPrefixOf (fromList [1,2]) (fromList [1,2,3,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
isPrefixOf : Eq elem => Vect m elem -> Vect len elem -> Bool
|
|
|
|
isPrefixOf = isPrefixOfBy (==)
|
|
|
|
|
|
|
|
||| Verify vector suffix
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| isSuffixOfBy (==) (fromList [3,4]) (fromList [1,2,3,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
isSuffixOfBy : (elem -> elem -> Bool) -> Vect m elem -> Vect len elem -> Bool
|
|
|
|
isSuffixOfBy p left right = isPrefixOfBy p (reverse left) (reverse right)
|
|
|
|
|
|
|
|
||| Verify vector suffix
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| isSuffixOf (fromList [3,4]) (fromList [1,2,3,4])
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
isSuffixOf : Eq elem => Vect m elem -> Vect len elem -> Bool
|
|
|
|
isSuffixOf = isSuffixOfBy (==)
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Conversions
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
||| Convert Maybe type into Vect
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| maybeToVect (Just 2)
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
maybeToVect : Maybe elem -> (p ** Vect p elem)
|
|
|
|
maybeToVect Nothing = (_ ** [])
|
|
|
|
maybeToVect (Just j) = (_ ** [j])
|
|
|
|
|
|
|
|
||| Convert first element of Vect (if exists) into Maybe.
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| vectToMaybe [2]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
vectToMaybe : Vect len elem -> Maybe elem
|
|
|
|
vectToMaybe [] = Nothing
|
|
|
|
vectToMaybe (x::xs) = Just x
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Misc
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
2020-07-08 02:55:43 +03:00
|
|
|
||| Filter out Nothings from Vect and unwrap the Justs
|
2020-05-18 15:59:07 +03:00
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| catMaybes [Just 1, Just 2, Nothing, Nothing, Just 5]
|
|
|
|
||| ```
|
|
|
|
public export
|
Remove linearity subtyping
It's disappointing to have to do this, but I think necessary because
various issue reports have shown it to be unsound (at least as far as
inference goes) and, at the very least, confusing. This patch brings us
back to the basic rules of QTT.
On the one hand, this makes the 1 multiplicity less useful, because it
means we can't flag arguments as being used exactly once which would be
useful for optimisation purposes as well as precision in the type. On
the other hand, it removes some complexity (and a hack) from
unification, and has the advantage of being correct! Also, I still
consider the 1 multiplicity an experiment.
We can still do interesting things like protocol state tracking, which
is my primary motivation at least.
Ideally, if the 1 multiplicity is going to be more generall useful,
we'll need some kind of way of doing multiplicity polymorphism in the
future. I don't think subtyping is the way (I've pretty much always come
to regret adding some form of subtyping).
Fixes #73 (and maybe some others).
2020-12-27 22:58:35 +03:00
|
|
|
catMaybes : (xs : Vect n (Maybe elem)) -> (p ** Vect p elem)
|
2020-05-18 15:59:07 +03:00
|
|
|
catMaybes [] = (_ ** [])
|
|
|
|
catMaybes (Nothing::xs) = catMaybes xs
|
|
|
|
catMaybes ((Just j)::xs) =
|
|
|
|
let (_ ** tail) = catMaybes xs
|
|
|
|
in (_ ** j::tail)
|
|
|
|
|
|
|
|
||| Get diagonal elements
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| diag [[1,2,3], [4,5,6], [7,8,9]]
|
|
|
|
||| ```
|
|
|
|
public export
|
|
|
|
diag : Vect len (Vect len elem) -> Vect len elem
|
|
|
|
diag [] = []
|
|
|
|
diag ((x::xs)::xss) = x :: diag (map tail xss)
|
|
|
|
|
|
|
|
public export
|
|
|
|
range : {len : Nat} -> Vect len (Fin len)
|
|
|
|
range {len=Z} = []
|
|
|
|
range {len=S _} = FZ :: map FS range
|
|
|
|
|
2021-01-27 21:23:08 +03:00
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Zippable
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
export
|
|
|
|
Zippable (Vect k) where
|
|
|
|
zipWith _ [] [] = []
|
|
|
|
zipWith f (x :: xs) (y :: ys) = f x y :: zipWith f xs ys
|
|
|
|
|
|
|
|
zipWith3 _ [] [] [] = []
|
|
|
|
zipWith3 f (x :: xs) (y :: ys) (z :: zs) = f x y z :: zipWith3 f xs ys zs
|
|
|
|
|
|
|
|
unzipWith f [] = ([], [])
|
|
|
|
unzipWith f (x :: xs) = let (b, c) = f x
|
|
|
|
(bs, cs) = unzipWith f xs in
|
|
|
|
(b :: bs, c :: cs)
|
|
|
|
|
|
|
|
unzipWith3 f [] = ([], [], [])
|
|
|
|
unzipWith3 f (x :: xs) = let (b, c, d) = f x
|
|
|
|
(bs, cs, ds) = unzipWith3 f xs in
|
|
|
|
(b :: bs, c :: cs, d :: ds)
|
|
|
|
|
2020-07-10 14:47:05 +03:00
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Matrix transposition
|
|
|
|
--------------------------------------------------------------------------------
|
2020-05-18 15:59:07 +03:00
|
|
|
||| Transpose a `Vect` of `Vect`s, turning rows into columns and vice versa.
|
|
|
|
|||
|
|
|
|
||| This is like zipping all the inner `Vect`s together and is equivalent to `traverse id` (`transposeTraverse`).
|
|
|
|
|||
|
|
|
|
||| As the types ensure rectangularity, this is an involution, unlike `Prelude.List.transpose`.
|
|
|
|
|||
|
|
|
|
||| ```idris example
|
|
|
|
||| transpose [[1,2], [3,4], [5,6], [7,8]]
|
|
|
|
||| ```
|
|
|
|
public export
|
Remove linearity subtyping
It's disappointing to have to do this, but I think necessary because
various issue reports have shown it to be unsound (at least as far as
inference goes) and, at the very least, confusing. This patch brings us
back to the basic rules of QTT.
On the one hand, this makes the 1 multiplicity less useful, because it
means we can't flag arguments as being used exactly once which would be
useful for optimisation purposes as well as precision in the type. On
the other hand, it removes some complexity (and a hack) from
unification, and has the advantage of being correct! Also, I still
consider the 1 multiplicity an experiment.
We can still do interesting things like protocol state tracking, which
is my primary motivation at least.
Ideally, if the 1 multiplicity is going to be more generall useful,
we'll need some kind of way of doing multiplicity polymorphism in the
future. I don't think subtyping is the way (I've pretty much always come
to regret adding some form of subtyping).
Fixes #73 (and maybe some others).
2020-12-27 22:58:35 +03:00
|
|
|
transpose : {n : _} -> (array : Vect m (Vect n elem)) -> Vect n (Vect m elem)
|
2020-07-10 19:17:19 +03:00
|
|
|
transpose [] = replicate _ [] -- = [| [] |]
|
2020-12-30 00:25:00 +03:00
|
|
|
transpose (x :: xs) = zipWith (::) x (transpose xs) -- = [| x :: xs |]
|
2020-05-18 15:59:07 +03:00
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Applicative/Monad/Traversable
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- These only work if the length is known at run time!
|
|
|
|
|
2020-06-30 02:21:48 +03:00
|
|
|
public export
|
2020-05-18 15:59:07 +03:00
|
|
|
implementation {k : Nat} -> Applicative (Vect k) where
|
|
|
|
pure = replicate _
|
|
|
|
fs <*> vs = zipWith apply fs vs
|
|
|
|
|
|
|
|
-- ||| This monad is different from the List monad, (>>=)
|
|
|
|
-- ||| uses the diagonal.
|
|
|
|
implementation {k : Nat} -> Monad (Vect k) where
|
|
|
|
m >>= f = diag (map f m)
|
|
|
|
|
|
|
|
public export
|
|
|
|
implementation Traversable (Vect k) where
|
|
|
|
traverse f [] = pure []
|
|
|
|
traverse f (x :: xs) = [| f x :: traverse f xs |]
|
|
|
|
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
-- Show
|
|
|
|
--------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
export
|
|
|
|
implementation Show elem => Show (Vect len elem) where
|
|
|
|
show = show . toList
|
|
|
|
|
|
|
|
-- Some convenience functions for testing lengths
|
|
|
|
|
|
|
|
-- Needs to be Maybe rather than Dec, because if 'n' is unequal to m, we
|
|
|
|
-- only know we don't know how to make a Vect n a, not that one can't exist.
|
|
|
|
export
|
|
|
|
exactLength : {m : Nat} -> -- expected at run-time
|
|
|
|
(len : Nat) -> (xs : Vect m a) -> Maybe (Vect len a)
|
|
|
|
exactLength {m} len xs with (decEq m len)
|
|
|
|
exactLength {m = m} m xs | (Yes Refl) = Just xs
|
|
|
|
exactLength {m = m} len xs | (No contra) = Nothing
|
|
|
|
|
|
|
|
||| If the given Vect is at least the required length, return a Vect with
|
|
|
|
||| at least that length in its type, otherwise return Nothing
|
|
|
|
||| @len the required length
|
|
|
|
||| @xs the vector with the desired length
|
|
|
|
export
|
|
|
|
overLength : {m : Nat} -> -- expected at run-time
|
|
|
|
(len : Nat) -> (xs : Vect m a) -> Maybe (p ** Vect (plus p len) a)
|
2021-02-16 14:56:43 +03:00
|
|
|
overLength n xs with (cmp m n)
|
|
|
|
overLength {m} (plus m (S y)) xs | (CmpLT y) = Nothing
|
|
|
|
overLength {m} m xs | CmpEQ = Just (0 ** xs)
|
2020-05-18 15:59:07 +03:00
|
|
|
overLength {m = plus n (S x)} n xs | (CmpGT x)
|
|
|
|
= Just (S x ** rewrite plusCommutative (S x) n in xs)
|