lh-l4v/lib/Time_Methods_Cmd_Test.thy

211 lines
8.5 KiB
Plaintext

(*
* Copyright 2018, Data61, CSIRO
*
* This software may be distributed and modified according to the terms of
* the BSD 2-Clause license. Note that NO WARRANTY is provided.
* See "LICENSE_BSD2.txt" for details.
*
* @TAG(DATA61_BSD)
*)
theory Time_Methods_Cmd_Test imports
Lib.Time_Methods_Cmd
Lib.Eisbach_Methods
"HOL-Library.Sublist"
begin
text \<open>
@{method time_methods} is a utility that runs several methods on the same
proof goal, printing the running time of each one.
Usage:
apply (time_methods [(no_check)] [(skip_fail)]
[name1:] \<open>method1\<close>
[name2:] \<open>method2\<close>
...)
Options:
no_check: Don't check that the outputs of each method match.
skip_fail: Don't print anything for failed methods.
\<close>
experiment begin
section \<open>Basic examples\<close>
lemma
"A \<Longrightarrow> B \<Longrightarrow> A \<or> B"
apply (time_methods
\<open>erule disjI1\<close>
\<open>erule disjI2\<close>)
done
text \<open>Give labels to methods:\<close>
lemma
"A \<Longrightarrow> B \<Longrightarrow> A \<or> B"
apply (time_methods
label: \<open>erule disjI1\<close>
"another label": \<open>erule disjI2\<close>)
done
text \<open>no_check prevents failing even if the method results differ.\<close>
lemma
"A \<Longrightarrow> B \<Longrightarrow> A \<or> B"
apply (time_methods (no_check)
\<open>rule disjI1\<close>
\<open>rule disjI2\<close>)
apply assumption
done
text \<open>
Fast and slow list reversals.
\<close>
lemma list_eval_rev_append:
"rev xs = rev xs @ []"
"rev [] @ ys = ys"
"rev (x # xs) @ ys = rev xs @ (x # ys)"
by auto
lemma "rev [0..100] = map ((-) 100) [0..100]"
"rev [0..200] = map ((-) 200) [0..200]"
text \<open>evaluate everything but @{term rev}\<close>
apply (all \<open>match conclusion in "rev x = y" for x y \<Rightarrow>
\<open>rule subst[where t = x], simp add: upto.simps\<close>\<close>)
apply (all \<open>match conclusion in "rev x = y" for x y \<Rightarrow>
\<open>rule subst[where t = y], simp add: upto.simps\<close>\<close>)
text \<open>evaluate @{term rev}\<close>
apply (time_methods
naive100: \<open>simp\<close>
slow100: \<open>simp only: rev.simps append.simps\<close>
fast100: \<open>subst list_eval_rev_append(1), simp only: list_eval_rev_append(2-3)\<close>
)
apply (time_methods
naive200: \<open>simp\<close>
slow200: \<open>simp only: rev.simps append.simps\<close>
fast200: \<open>subst list_eval_rev_append(1), simp only: list_eval_rev_append(2-3)\<close>
)
done
text \<open>
Fast and slow subsequence testing.
\<close>
lemma
"subseq (map ((*) 2) [1 .. 5]) [1 .. 10]"
"subseq (map ((*) 2) [1 .. 6]) [1 .. 12]"
"subseq (map ((*) 2) [1 .. 7]) [1 .. 14]"
"subseq (map ((*) 2) [1 .. 8]) [1 .. 16]"
apply (all \<open>match conclusion in "subseq x y" for x y \<Rightarrow>
\<open>rule subst[where t = x], simp add: upto.simps,
rule subst[where t = y], simp add: upto.simps\<close>\<close>)
apply (time_methods
"HOL simp": \<open>simp\<close>
"l4v simp": \<open>simp cong: if_cong cong del: if_weak_cong\<close>
\<comment> \<open>exponential time!\<close>
)+
done
text \<open>
Which method is a better SAT solver?
Instance 01 from the uf20-91 dataset at http://www.cs.ubc.ca/~hoos/SATLIB/benchm.html
\<close>
lemma "\<exists>x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 x11 x12 x13 x14 x15 x16 x17 x18 x19 x20.
(~x4 & x18 & ~x19) | (~x3 & ~x18 & x5) | (x5 & x8 & x15) | (x20 & ~x7 & x16) |
(~x10 & x13 & x7) | (x12 & x9 & ~x17) | (~x17 & ~x19 & ~x5) | (x16 & ~x9 & ~x15) |
(~x11 & x5 & x14) | (~x18 & x10 & ~x13) | (x3 & ~x11 & ~x12) | (x6 & x17 & x8) |
(x18 & ~x14 & ~x1) | (x19 & x15 & ~x10) | (~x12 & ~x18 & x19) | (x8 & ~x4 & ~x7) |
(x8 & x9 & ~x4) | (~x7 & ~x17 & x15) | (~x12 & x7 & x14) | (x10 & x11 & ~x8) |
(~x2 & x15 & x11) | (~x9 & ~x6 & ~x1) | (x11 & ~x20 & x17) | (~x9 & x15 & ~x13) |
(~x12 & x7 & x17) | (x18 & x2 & ~x20) | (~x20 & ~x12 & ~x4) | (~x19 & ~x11 & ~x14) |
(x16 & ~x18 & x4) | (x1 & x17 & x19) | (x13 & ~x15 & ~x10) | (x12 & x14 & x13) |
(~x12 & x14 & x7) | (x7 & ~x16 & ~x10) | (~x6 & ~x10 & ~x7) | (~x20 & ~x14 & x16) |
(x19 & ~x17 & ~x11) | (x7 & ~x1 & x20) | (x5 & ~x12 & ~x15) | (x4 & x9 & x13) |
(~x12 & x11 & x7) | (x5 & ~x19 & x8) | (~x1 & ~x16 & ~x17) | (~x20 & x14 & x15) |
(~x13 & x4 & ~x10) | (~x14 & ~x7 & ~x10) | (x5 & ~x9 & ~x20) | (~x10 & ~x1 & x19) |
(x16 & x15 & x1) | (~x16 & ~x3 & x11) | (x15 & x10 & ~x4) | (~x4 & x15 & x3) |
(x10 & x16 & ~x11) | (x8 & ~x12 & x5) | (~x14 & x6 & ~x12) | (~x1 & ~x6 & ~x11) |
(x13 & x5 & x1) | (x7 & x2 & ~x12) | (~x1 & x20 & ~x19) | (x2 & x13 & x8) |
(~x15 & ~x18 & ~x4) | (x11 & ~x14 & ~x9) | (x6 & x15 & x2) | (~x5 & x12 & x15) |
(x6 & ~x17 & ~x5) | (x13 & ~x5 & x19) | (~x20 & x1 & ~x14) | (~x9 & x17 & ~x15) |
(x5 & ~x19 & x18) | (x12 & ~x8 & x10) | (x18 & ~x14 & x4) | (~x15 & x9 & ~x13) |
(~x9 & x5 & x1) | (~x10 & x19 & x14) | (~x20 & ~x9 & ~x4) | (x9 & x2 & ~x19) |
(x5 & ~x13 & x17) | (~x2 & x10 & x18) | (x18 & ~x3 & ~x11) | (~x7 & x9 & ~x17) |
(x15 & x6 & x3) | (x2 & ~x3 & x13) | (~x12 & ~x3 & x2) | (x2 & x3 & ~x17) |
(~x20 & x15 & x16) | (x5 & x17 & x19) | (x20 & x18 & ~x11) | (x9 & ~x1 & x5) |
(x19 & ~x9 & ~x17) | (~x12 & x2 & ~x17)"
using [[meson_max_clauses=99]]
apply (time_methods
blast: \<open>blast\<close>
metis: \<open>metis\<close>
meson: \<open>meson\<close>
smt: \<open>smt\<close>
force: \<open>force\<close>
fastforce: \<open>fastforce intro: ex_bool_eq[THEN iffD2]\<close>
fastforce: \<open>fastforce simp: ex_bool_eq\<close>
presburger: \<open>use ex_bool_eq[simp] in presburger\<close>
)
done
section \<open>Other tests\<close>
text \<open>Test ML interface\<close>
lemma "True"
apply (tactic \<open>
let val method = SIMPLE_METHOD (simp_tac @{context} 1)
fun dummy_callback _ _ = ()
in (fn st => Time_Methods.time_methods false false dummy_callback [(NONE, method)] [] st
|> (fn ([timing], st') => (tracing (Timing.message timing); st')))
|> Method.NO_CONTEXT_TACTIC @{context}
end\<close>)
done
text \<open>Check that we fail when the methods give different results\<close>
lemma
"A \<or> B"
apply (tactic \<open>
let
fun skip_dummy_state tac = fn st =>
case Thm.prop_of st of
Const ("Pure.prop", _) $ (Const ("Pure.term", _) $ Const ("Pure.dummy_pattern", _)) =>
Seq.succeed st
| _ => tac st;
val methods = @{thms disjI1 disjI2}
|> map (fn rule => (NONE, SIMPLE_METHOD (resolve_tac @{context} [rule] 1)))
fun dummy_callback _ _ = ()
in (fn st => Time_Methods.time_methods false false dummy_callback methods [] st
|> (fn (timing, st') => error "test failed: shouldn't reach here")
handle THM _ => Seq.succeed (Seq.Result st)) (* should reach here *)
|> Method.NO_CONTEXT_TACTIC @{context}
|> skip_dummy_state
end\<close>)
oops
text \<open>Check that we propagate failures from the input methods\<close>
lemma "A"
apply (fails \<open>time_methods \<open>simp\<close>\<close>)
oops
text \<open>Check that we skip failing methods when skip_fail set\<close>
lemma "A \<and> B \<Longrightarrow> A"
apply (
( \<comment> \<open>roughly corresponds to "time_methods (skip_fail) \<open>fail\<close>",
but errors if it calls the output callback\<close>
tactic \<open>
let
fun timing_callback _ _ = error "test failed: shouldn't reach here"
val methods = [(NONE, SIMPLE_METHOD no_tac)]
in
(fn st =>
#2 (Time_Methods.time_methods false true timing_callback methods [] st))
|> Method.NO_CONTEXT_TACTIC @{context}
end\<close>)
| time_methods (skip_fail) good_simp: \<open>simp\<close>)
done
end
end