Datasets:
Tasks:
Text Generation
Modalities:
Text
Sub-tasks:
language-modeling
Languages:
English
Size:
100K - 1M
License:
File size: 7,925 Bytes
4365a98 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 |
/-
Copyright (c) 2019 Microsoft Corporation. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Simon Hudon
The writer monad transformer for passing immutable state.
-/
import algebra.group.defs
import logic.equiv.basic
universes u v w u₀ u₁ v₀ v₁
structure writer_t (ω : Type u) (m : Type u → Type v) (α : Type u) : Type (max u v) :=
(run : m (α × ω))
@[reducible] def writer (ω : Type u) := writer_t ω id
attribute [pp_using_anonymous_constructor] writer_t
namespace writer_t
section
variable {ω : Type u}
variable {m : Type u → Type v}
variable [monad m]
variables {α β : Type u}
open function
@[ext]
protected lemma ext (x x' : writer_t ω m α)
(h : x.run = x'.run) :
x = x' := by cases x; cases x'; congr; apply h
@[inline] protected def tell (w : ω) : writer_t ω m punit :=
⟨pure (punit.star, w)⟩
@[inline] protected def listen : writer_t ω m α → writer_t ω m (α × ω)
| ⟨ cmd ⟩ := ⟨ (λ x : α × ω, ((x.1,x.2),x.2)) <$> cmd ⟩
@[inline] protected def pass : writer_t ω m (α × (ω → ω)) → writer_t ω m α
| ⟨ cmd ⟩ := ⟨ uncurry (uncurry $ λ x (f : ω → ω) w, (x,f w)) <$> cmd ⟩
@[inline] protected def pure [has_one ω] (a : α) : writer_t ω m α :=
⟨ pure (a,1) ⟩
@[inline] protected def bind [has_mul ω] (x : writer_t ω m α) (f : α → writer_t ω m β) :
writer_t ω m β :=
⟨ do x ← x.run,
x' ← (f x.1).run,
pure (x'.1,x.2 * x'.2) ⟩
instance [has_one ω] [has_mul ω] : monad (writer_t ω m) :=
{ pure := λ α, writer_t.pure, bind := λ α β, writer_t.bind }
instance [monoid ω] [is_lawful_monad m] : is_lawful_monad (writer_t ω m) :=
{ id_map := by { intros, cases x, simp [(<$>),writer_t.bind,writer_t.pure] },
pure_bind := by { intros, simp [has_pure.pure,writer_t.pure,(>>=),writer_t.bind], ext; refl },
bind_assoc := by { intros, simp [(>>=),writer_t.bind,mul_assoc] with functor_norm } }
@[inline] protected def lift [has_one ω] (a : m α) : writer_t ω m α :=
⟨ flip prod.mk 1 <$> a ⟩
instance (m) [monad m] [has_one ω] : has_monad_lift m (writer_t ω m) :=
⟨ λ α, writer_t.lift ⟩
@[inline] protected def monad_map {m m'} [monad m] [monad m'] {α} (f : Π {α}, m α → m' α) :
writer_t ω m α → writer_t ω m' α :=
λ x, ⟨ f x.run ⟩
instance (m m') [monad m] [monad m'] : monad_functor m m' (writer_t ω m) (writer_t ω m') :=
⟨@writer_t.monad_map ω m m' _ _⟩
@[inline] protected def adapt {ω' : Type u} {α : Type u} (f : ω → ω') :
writer_t ω m α → writer_t ω' m α :=
λ x, ⟨prod.map id f <$> x.run⟩
instance (ε) [has_one ω] [monad m] [monad_except ε m] : monad_except ε (writer_t ω m) :=
{ throw := λ α, writer_t.lift ∘ throw,
catch := λ α x c, ⟨catch x.run (λ e, (c e).run)⟩ }
end
end writer_t
/--
An implementation of [MonadReader](
https://hackage.haskell.org/package/mtl-2.2.2/docs/Control-Monad-Reader-Class.html#t:MonadReader).
It does not contain `local` because this function cannot be lifted using `monad_lift`.
Instead, the `monad_reader_adapter` class provides the more general `adapt_reader` function.
Note: This class can be seen as a simplification of the more "principled" definition
```
class monad_reader (ρ : out_param (Type u)) (n : Type u → Type u) :=
(lift {α : Type u} : (∀ {m : Type u → Type u} [monad m], reader_t ρ m α) → n α)
```
-/
class monad_writer (ω : out_param (Type u)) (m : Type u → Type v) :=
(tell (w : ω) : m punit)
(listen {α} : m α → m (α × ω))
(pass {α : Type u} : m (α × (ω → ω)) → m α)
export monad_writer
instance {ω : Type u} {m : Type u → Type v} [monad m] : monad_writer ω (writer_t ω m) :=
{ tell := writer_t.tell,
listen := λ α, writer_t.listen,
pass := λ α, writer_t.pass }
instance {ω ρ : Type u} {m : Type u → Type v} [monad m] [monad_writer ω m] :
monad_writer ω (reader_t ρ m) :=
{ tell := λ x, monad_lift (tell x : m punit),
listen := λ α ⟨ cmd ⟩, ⟨ λ r, listen (cmd r) ⟩,
pass := λ α ⟨ cmd ⟩, ⟨ λ r, pass (cmd r) ⟩ }
def swap_right {α β γ} : (α × β) × γ → (α × γ) × β
| ⟨⟨x,y⟩,z⟩ := ((x,z),y)
instance {ω σ : Type u} {m : Type u → Type v} [monad m] [monad_writer ω m] :
monad_writer ω (state_t σ m) :=
{ tell := λ x, monad_lift (tell x : m punit),
listen := λ α ⟨ cmd ⟩, ⟨ λ r, swap_right <$> listen (cmd r) ⟩,
pass := λ α ⟨ cmd ⟩, ⟨ λ r, pass (swap_right <$> cmd r) ⟩ }
open function
def except_t.pass_aux {ε α ω} : except ε (α × (ω → ω)) → except ε α × (ω → ω)
| (except.error a) := (except.error a,id)
| (except.ok (x,y)) := (except.ok x,y)
instance {ω ε : Type u} {m : Type u → Type v} [monad m] [monad_writer ω m] :
monad_writer ω (except_t ε m) :=
{ tell := λ x, monad_lift (tell x : m punit),
listen := λ α ⟨ cmd ⟩, ⟨ uncurry (λ x y, flip prod.mk y <$> x) <$> listen cmd ⟩,
pass := λ α ⟨ cmd ⟩, ⟨ pass (except_t.pass_aux <$> cmd) ⟩ }
def option_t.pass_aux {α ω} : option (α × (ω → ω)) → option α × (ω → ω)
| none := (none ,id)
| (some (x,y)) := (some x,y)
instance {ω : Type u} {m : Type u → Type v} [monad m] [monad_writer ω m] :
monad_writer ω (option_t m) :=
{ tell := λ x, monad_lift (tell x : m punit),
listen := λ α ⟨ cmd ⟩, ⟨ uncurry (λ x y, flip prod.mk y <$> x) <$> listen cmd ⟩,
pass := λ α ⟨ cmd ⟩, ⟨ pass (option_t.pass_aux <$> cmd) ⟩ }
/-- Adapt a monad stack, changing the type of its top-most environment.
This class is comparable to
[Control.Lens.Magnify](https://hackage.haskell.org/package/lens-4.15.4/docs/Control-Lens-Zoom.html#t:Magnify),
but does not use lenses (why would it), and is derived automatically for any transformer
implementing `monad_functor`.
Note: This class can be seen as a simplification of the more "principled" definition
```
class monad_reader_functor (ρ ρ' : out_param (Type u)) (n n' : Type u → Type u) :=
(map {α : Type u} :
(∀ {m : Type u → Type u} [monad m], reader_t ρ m α → reader_t ρ' m α) → n α → n' α)
```
-/
class monad_writer_adapter (ω ω' : out_param (Type u)) (m m' : Type u → Type v) :=
(adapt_writer {α : Type u} : (ω → ω') → m α → m' α)
export monad_writer_adapter (adapt_writer)
section
variables {ω ω' : Type u} {m m' : Type u → Type v}
/-- Transitivity.
This instance generates the type-class problem with a metavariable argument (which is why this
is marked as `[nolint dangerous_instance]`).
Currently that is not a problem, as there are almost no instances of `monad_functor` or
`monad_writer_adapter`.
see Note [lower instance priority] -/
@[nolint dangerous_instance, priority 100]
instance monad_writer_adapter_trans {n n' : Type u → Type v} [monad_writer_adapter ω ω' m m']
[monad_functor m m' n n'] : monad_writer_adapter ω ω' n n' :=
⟨λ α f, monad_map (λ α, (adapt_writer f : m α → m' α))⟩
instance [monad m] : monad_writer_adapter ω ω' (writer_t ω m) (writer_t ω' m) :=
⟨λ α, writer_t.adapt⟩
end
instance (ω : Type u) (m out) [monad_run out m] : monad_run (λ α, out (α × ω)) (writer_t ω m) :=
⟨λ α x, run $ x.run ⟩
/-- reduce the equivalence between two writer monads to the equivalence between
their underlying monad -/
def writer_t.equiv {m₁ : Type u₀ → Type v₀} {m₂ : Type u₁ → Type v₁}
{α₁ ω₁ : Type u₀} {α₂ ω₂ : Type u₁} (F : (m₁ (α₁ × ω₁)) ≃ (m₂ (α₂ × ω₂))) :
writer_t ω₁ m₁ α₁ ≃ writer_t ω₂ m₂ α₂ :=
{ to_fun := λ ⟨f⟩, ⟨F f⟩,
inv_fun := λ ⟨f⟩, ⟨F.symm f⟩,
left_inv := λ ⟨f⟩, congr_arg writer_t.mk $ F.left_inv _,
right_inv := λ ⟨f⟩, congr_arg writer_t.mk $ F.right_inv _ }
|