Datasets:
Tasks:
Text Generation
Modalities:
Text
Sub-tasks:
language-modeling
Languages:
English
Size:
100K - 1M
License:
File size: 9,771 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 |
/-
Copyright (c) 2020 Simon Hudon. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Simon Hudon
-/
import control.monad.basic
import data.int.basic
import data.stream.defs
import control.uliftable
import tactic.norm_num
import data.bitvec.basic
/-!
# Rand Monad and Random Class
This module provides tools for formulating computations guided by randomness and for
defining objects that can be created randomly.
## Main definitions
* `rand` monad for computations guided by randomness;
* `random` class for objects that can be generated randomly;
* `random` to generate one object;
* `random_r` to generate one object inside a range;
* `random_series` to generate an infinite series of objects;
* `random_series_r` to generate an infinite series of objects inside a range;
* `io.mk_generator` to create a new random number generator;
* `io.run_rand` to run a randomized computation inside the `io` monad;
* `tactic.run_rand` to run a randomized computation inside the `tactic` monad
## Local notation
* `i .. j` : `Icc i j`, the set of values between `i` and `j` inclusively;
## Tags
random monad io
## References
* Similar library in Haskell: https://hackage.haskell.org/package/MonadRandom
-/
open list io applicative
universes u v w
/-- A monad to generate random objects using the generator type `g` -/
@[reducible]
def rand_g (g : Type) (α : Type u) : Type u := state (ulift.{u} g) α
/-- A monad to generate random objects using the generator type `std_gen` -/
@[reducible]
def rand := rand_g std_gen
instance (g : Type) : uliftable (rand_g.{u} g) (rand_g.{v} g) :=
@state_t.uliftable' _ _ _ _ _ (equiv.ulift.trans.{u u u u u} equiv.ulift.symm)
open ulift (hiding inhabited)
/-- Generate one more `ℕ` -/
def rand_g.next {g : Type} [random_gen g] : rand_g g ℕ :=
⟨ prod.map id up ∘ random_gen.next ∘ down ⟩
local infix ` .. `:41 := set.Icc
open stream
/-- `bounded_random α` gives us machinery to generate values of type `α` between certain bounds -/
class bounded_random (α : Type u) [preorder α] :=
(random_r : Π g [random_gen g] (x y : α),
(x ≤ y) → rand_g g (x .. y))
/-- `random α` gives us machinery to generate values of type `α` -/
class random (α : Type u) :=
(random [] : Π (g : Type) [random_gen g], rand_g g α)
/-- shift_31_left = 2^31; multiplying by it shifts the binary
representation of a number left by 31 bits, dividing by it shifts it
right by 31 bits -/
def shift_31_left : ℕ :=
by apply_normed 2^31
namespace rand
open stream
variables (α : Type u)
variables (g : Type) [random_gen g]
/-- create a new random number generator distinct from the one stored in the state -/
def split : rand_g g g := ⟨ prod.map id up ∘ random_gen.split ∘ down ⟩
variables {g}
section random
variables [random α]
export random (random)
/-- Generate a random value of type `α`. -/
def random : rand_g g α :=
random.random α g
/-- generate an infinite series of random values of type `α` -/
def random_series : rand_g g (stream α) :=
do gen ← uliftable.up (split g),
pure $ stream.corec_state (random.random α g) gen
end random
variables {α}
/-- Generate a random value between `x` and `y` inclusive. -/
def random_r [preorder α] [bounded_random α] (x y : α) (h : x ≤ y) : rand_g g (x .. y) :=
bounded_random.random_r g x y h
/-- generate an infinite series of random values of type `α` between `x` and `y` inclusive. -/
def random_series_r [preorder α] [bounded_random α] (x y : α) (h : x ≤ y) :
rand_g g (stream (x .. y)) :=
do gen ← uliftable.up (split g),
pure $ corec_state (bounded_random.random_r g x y h) gen
end rand
namespace io
private def accum_char (w : ℕ) (c : char) : ℕ :=
c.to_nat + 256 * w
/-- create and a seed a random number generator -/
def mk_generator : io std_gen := do
seed ← io.rand 0 shift_31_left,
return $ mk_std_gen seed
variables {α : Type}
/-- Run `cmd` using a randomly seeded random number generator -/
def run_rand (cmd : _root_.rand α) : io α :=
do g ← io.mk_generator,
return $ (cmd.run ⟨g⟩).1
/-- Run `cmd` using the provided seed. -/
def run_rand_with (seed : ℕ) (cmd : _root_.rand α) : io α :=
return $ (cmd.run ⟨mk_std_gen seed⟩).1
section random
variables [random α]
/-- randomly generate a value of type α -/
def random : io α :=
io.run_rand (rand.random α)
/-- randomly generate an infinite series of value of type α -/
def random_series : io (stream α) :=
io.run_rand (rand.random_series α)
end random
section bounded_random
variables [preorder α] [bounded_random α]
/-- randomly generate a value of type α between `x` and `y` -/
def random_r (x y : α) (p : x ≤ y) : io (x .. y) :=
io.run_rand (bounded_random.random_r _ x y p)
/-- randomly generate an infinite series of value of type α between `x` and `y` -/
def random_series_r (x y : α) (h : x ≤ y) : io (stream $ x .. y) :=
io.run_rand (rand.random_series_r x y h)
end bounded_random
end io
namespace tactic
/-- create a seeded random number generator in the `tactic` monad -/
meta def mk_generator : tactic std_gen := do
tactic.unsafe_run_io @io.mk_generator
/-- run `cmd` using the a randomly seeded random number generator
in the tactic monad -/
meta def run_rand {α : Type u} (cmd : rand α) : tactic α := do
⟨g⟩ ← tactic.up mk_generator,
return (cmd.run ⟨g⟩).1
variables {α : Type u}
section bounded_random
variables [preorder α] [bounded_random α]
/-- Generate a random value between `x` and `y` inclusive. -/
meta def random_r (x y : α) (h : x ≤ y) : tactic (x .. y) :=
run_rand (rand.random_r x y h)
/-- Generate an infinite series of random values of type `α` between `x` and `y` inclusive. -/
meta def random_series_r (x y : α) (h : x ≤ y) : tactic (stream $ x .. y) :=
run_rand (rand.random_series_r x y h)
end bounded_random
section random
variables [random α]
/-- randomly generate a value of type α -/
meta def random : tactic α :=
run_rand (rand.random α)
/-- randomly generate an infinite series of value of type α -/
meta def random_series : tactic (stream α) :=
run_rand (rand.random_series α)
end random
end tactic
open nat (succ one_add mod_eq_of_lt zero_lt_succ add_one succ_le_succ)
variables {g : Type} [random_gen g]
open nat
namespace fin
variables {n : ℕ} [fact (0 < n)]
/-- generate a `fin` randomly -/
protected def random : rand_g g (fin n) :=
⟨ λ ⟨g⟩, prod.map of_nat' up $ rand_nat g 0 n ⟩
end fin
open nat
instance nat_bounded_random : bounded_random ℕ :=
{ random_r := λ g inst x y hxy,
do z ← @fin.random g inst (succ $ y - x) _,
pure ⟨z.val + x, nat.le_add_left _ _,
by rw ← le_tsub_iff_right hxy; apply le_of_succ_le_succ z.is_lt⟩ }
/-- This `bounded_random` interval generates integers between `x` and
`y` by first generating a natural number between `0` and `y - x` and
shifting the result appropriately. -/
instance int_bounded_random : bounded_random ℤ :=
{ random_r := λ g inst x y hxy,
do ⟨z,h₀,h₁⟩ ← @bounded_random.random_r ℕ _ _ g inst 0 (int.nat_abs $ y - x) dec_trivial,
pure ⟨z + x,
int.le_add_of_nonneg_left (int.coe_nat_nonneg _),
int.add_le_of_le_sub_right $ le_trans
(int.coe_nat_le_coe_nat_of_le h₁)
(le_of_eq $ int.of_nat_nat_abs_eq_of_nonneg (int.sub_nonneg_of_le hxy)) ⟩ }
instance fin_random (n : ℕ) [fact (0 < n)] : random (fin n) :=
{ random := λ g inst, @fin.random g inst _ _ }
instance fin_bounded_random (n : ℕ) : bounded_random (fin n) :=
{ random_r := λ g inst (x y : fin n) p,
do ⟨r, h, h'⟩ ← @rand.random_r ℕ g inst _ _ x.val y.val p,
pure ⟨⟨r,lt_of_le_of_lt h' y.is_lt⟩, h, h'⟩ }
/-- A shortcut for creating a `random (fin n)` instance from
a proof that `0 < n` rather than on matching on `fin (succ n)` -/
def random_fin_of_pos : ∀ {n : ℕ} (h : 0 < n), random (fin n)
| (succ n) _ := fin_random _
| 0 h := false.elim (nat.not_lt_zero _ h)
lemma bool_of_nat_mem_Icc_of_mem_Icc_to_nat (x y : bool) (n : ℕ) :
n ∈ (x.to_nat .. y.to_nat) → bool.of_nat n ∈ (x .. y) :=
begin
simp only [and_imp, set.mem_Icc], intros h₀ h₁,
split;
[ have h₂ := bool.of_nat_le_of_nat h₀, have h₂ := bool.of_nat_le_of_nat h₁ ];
rw bool.of_nat_to_nat at h₂; exact h₂,
end
instance : random bool :=
{ random := λ g inst,
(bool.of_nat ∘ subtype.val) <$> @bounded_random.random_r ℕ _ _ g inst 0 1 (nat.zero_le _) }
instance : bounded_random bool :=
{ random_r := λ g _inst x y p,
subtype.map bool.of_nat (bool_of_nat_mem_Icc_of_mem_Icc_to_nat x y) <$>
@bounded_random.random_r ℕ _ _ g _inst x.to_nat y.to_nat (bool.to_nat_le_to_nat p) }
open_locale fin_fact
/-- generate a random bit vector of length `n` -/
def bitvec.random (n : ℕ) : rand_g g (bitvec n) :=
bitvec.of_fin <$> rand.random (fin $ 2^n)
/-- generate a random bit vector of length `n` -/
def bitvec.random_r {n : ℕ} (x y : bitvec n) (h : x ≤ y) : rand_g g (x .. y) :=
have h' : ∀ (a : fin (2 ^ n)), a ∈ (x.to_fin .. y.to_fin) → bitvec.of_fin a ∈ (x .. y),
begin
simp only [and_imp, set.mem_Icc], intros z h₀ h₁,
replace h₀ := bitvec.of_fin_le_of_fin_of_le h₀,
replace h₁ := bitvec.of_fin_le_of_fin_of_le h₁,
rw bitvec.of_fin_to_fin at h₀ h₁, split; assumption,
end,
subtype.map bitvec.of_fin h' <$> rand.random_r x.to_fin y.to_fin (bitvec.to_fin_le_to_fin_of_le h)
open nat
instance random_bitvec (n : ℕ) : random (bitvec n) :=
{ random := λ _ inst, @bitvec.random _ inst n }
instance bounded_random_bitvec (n : ℕ) : bounded_random (bitvec n) :=
{ random_r := λ _ inst x y p, @bitvec.random_r _ inst _ _ _ p }
|