Datasets:
Tasks:
Text Generation
Modalities:
Text
Sub-tasks:
language-modeling
Languages:
English
Size:
100K - 1M
License:
(* ========================================================================= *) | |
(* Prove that a congruence x^3 + a * x + b == 0 (mod p) has no solution *) | |
(* (fails with an exception if it does). This is essentually done via a *) | |
(* computation of gcd(x^p - x,x^3 + a * x + b) over F_p[x]. *) | |
(* *) | |
(* This can be used to show that a Weierstrass curve y^2 = x^3 + a * x + b *) | |
(* has no 2-torsion points, since such points would necessarily have y = 0 *) | |
(* because the negation in the curve group of (x,y) is (x,-y). *) | |
(* ========================================================================= *) | |
needs "Library/pocklington.ml";; | |
let num_modinv = | |
let rec gcdex(m,n) = | |
if n </ m then let (x,y) = gcdex(n,m) in (y,x) | |
else if m =/ Int 0 then (Int 0,Int 1) else | |
let q = quo_num n m in | |
let r = n -/ q */ m in | |
let (x,y) = gcdex(r,m) in (y -/ q */ x,x) in | |
fun p n -> let (x,y) = gcdex(n,p) in | |
if mod_num (x */ n) p =/ num_1 then mod_num x p | |
else failwith "num_modinv";; | |
let EXCLUDE_MODULAR_CUBIC_ROOTS = | |
let lemma_flt = prove | |
(`((x pow 3 + a * x + b:int == &0) (mod &p) | |
==> (x pow p == c2 * x pow 2 + c1 * x + c0) (mod &p)) | |
==> prime p | |
==> ~(b rem &p = &0) | |
==> ((x pow 3 + a * x + b == &0) (mod &p) | |
==> (&0 == c2 * x pow 2 + (c1 - &1) * x + c0) (mod &p))`, | |
DISCH_THEN(fun th -> REPEAT DISCH_TAC THEN MP_TAC th) THEN | |
ASM_REWRITE_TAC[] THEN MATCH_MP_TAC(INTEGER_RULE | |
`(xp == x) (mod p) | |
==> (xp:int == c2 * x pow 2 + c1 * x + c0) (mod p) | |
==> (&0 == c2 * x pow 2 + (c1 - &1) * x + c0) (mod p)`) THEN | |
FIRST_X_ASSUM(MP_TAC o MATCH_MP (INTEGER_RULE | |
`(x pow 3 + a * x + b:int == &0) (mod p) | |
==> coprime(b,p) ==> coprime(x,p)`)) THEN | |
ANTS_TAC THENL | |
[MATCH_MP_TAC(INTEGER_RULE | |
`(b rem p == b) (mod p) /\ coprime(b rem p,p) ==> coprime(b,p)`) THEN | |
REWRITE_TAC[INT_REM_MOD_SELF] THEN UNDISCH_TAC `~(b rem &p = &0)` THEN | |
SUBGOAL_THEN `&0 <= b rem &p /\ b rem &p < &p` MP_TAC THENL | |
[ASM_REWRITE_TAC[INT_REM_POS_EQ; INT_LT_REM_EQ] THEN | |
ASM_SIMP_TAC[INT_OF_NUM_EQ; PRIME_IMP_NZ; INT_OF_NUM_LT; LE_1]; | |
SPEC_TAC(`b rem &p`,`x:int`)] THEN | |
REWRITE_TAC[GSYM INT_FORALL_POS; IMP_CONJ] THEN | |
X_GEN_TAC `n:num` THEN REWRITE_TAC[INT_OF_NUM_EQ; INT_OF_NUM_LT] THEN | |
REPEAT DISCH_TAC THEN REWRITE_TAC[GSYM num_coprime] THEN | |
ONCE_REWRITE_TAC[COPRIME_SYM] THEN ASM_SIMP_TAC[PRIME_COPRIME_EQ] THEN | |
DISCH_THEN(MP_TAC o MATCH_MP DIVIDES_LE) THEN ASM_ARITH_TAC; | |
DISCH_THEN(MP_TAC o MATCH_MP (INTEGER_RULE | |
`coprime(b,p) ==> (b rem p == b) (mod p) ==> coprime(b rem p,p)`)) THEN | |
REWRITE_TAC[INT_REM_MOD_SELF]] THEN | |
REWRITE_TAC[GSYM INT_REM_EQ] THEN ONCE_REWRITE_TAC[GSYM INT_POW_REM] THEN | |
SUBGOAL_THEN `&0 <= x rem &p /\ x rem &p < &p` MP_TAC THENL | |
[ASM_REWRITE_TAC[INT_REM_POS_EQ; INT_LT_REM_EQ] THEN | |
ASM_SIMP_TAC[INT_OF_NUM_EQ; PRIME_IMP_NZ; INT_OF_NUM_LT; LE_1]; | |
SPEC_TAC(`x rem &p`,`x:int`)] THEN | |
REWRITE_TAC[GSYM INT_FORALL_POS; IMP_CONJ] THEN | |
REWRITE_TAC[INT_POW_REM; GSYM num_coprime; INT_OF_NUM_LT] THEN | |
X_GEN_TAC `n:num` THEN REPEAT STRIP_TAC THEN | |
REWRITE_TAC[INT_OF_NUM_POW; INT_OF_NUM_REM; INT_OF_NUM_EQ] THEN | |
MP_TAC(SPECL [`n:num`; `p:num`] FERMAT_LITTLE_PRIME) THEN | |
ASM_REWRITE_TAC[] THEN DISCH_THEN(MP_TAC o MATCH_MP (NUMBER_RULE | |
`(x EXP a == 1) (mod p) ==> (x * x EXP a == x) (mod p)`)) THEN | |
ASM_SIMP_TAC[GSYM(CONJUNCT2 EXP); CONG; MOD_LT] THEN | |
ASM_SIMP_TAC[PRIME_IMP_NZ; ARITH_RULE `~(p = 0) ==> SUC(p - 1) = p`]) | |
and lemma_tolin = prove | |
(`(x pow 3 + a * x + b == &0) (mod p) /\ | |
(&0 == &1 * x pow 2 + c1 * x + c0) (mod p) | |
==> (&0:int == &0 * x pow 2 + | |
(a + c1 pow 2 - c0) * x + (b + c0 * c1)) (mod p)`, | |
INTEGER_TAC) | |
and lemma_toconst = prove | |
(`(x pow 3 + a * x + b == &0) (mod p) /\ | |
(&0:int == &0 * x pow 2 + &1 * x + c0) (mod p) | |
==> (&0:int == --c0 pow 3 - a * c0 + b) (mod p)`, | |
INTEGER_TAC) | |
and lemma_3_2 = prove | |
(`(x pow 3 + a * x + b:int == &0) (mod p) | |
==> (y == c3 * x pow 3 + c2 * x pow 2 + c1 * x + c0) (mod p) | |
==> (y == c2 * x pow 2 + (c1 - a * c3) * x + (c0 - b * c3)) (mod p)`, | |
INTEGER_TAC) | |
and lemma_4_3 = prove | |
(`(x pow 3 + a * x + b:int == &0) (mod p) | |
==> (y == c4 * x pow 4 + c3 * x pow 3 + c2 * x pow 2 + c1 * x + c0) (mod p) | |
==> (y == c3 * x pow 3 + (c2 - a * c4) * x pow 2 + | |
(c1 - b * c4) * x + c0) (mod p)`, | |
INTEGER_TAC) | |
and lemma_rem = prove | |
(`(y:int == c2 * x pow 2 + c1 * x + c0) (mod p) | |
==> (y == c2 rem p * x pow 2 + c1 rem p * x + c0 rem p) (mod p)`, | |
MATCH_MP_TAC(INTEGER_RULE | |
`(c2:int == c2') (mod p) /\ (c1 == c1') (mod p) /\ (c0 == c0') (mod p) | |
==> (y:int == c2 * x pow 2 + c1 * x + c0) (mod p) | |
==> (y:int == c2' * x pow 2 + c1' * x + c0') (mod p)`) THEN | |
REWRITE_TAC[INT_CONG_RREM] THEN CONV_TAC INTEGER_RULE) | |
and lemma_mulx = prove | |
(`((x:int) pow n == c2 * x pow 2 + c1 * x + c0) (mod p) | |
==> (x pow (n + 1) == c2 * x pow 3 + c1 * x pow 2 + c0 * x + &0) (mod p)`, | |
REWRITE_TAC[INT_POW_ADD] THEN INTEGER_TAC) | |
and lemma_sqr = prove | |
(`((x:int) pow n == c2 * x pow 2 + c1 * x + c0) (mod p) | |
==> (x pow (2 * n) == | |
(c2 pow 2) * x pow 4 + (&2 * c1 * c2) * x pow 3 + | |
(c1 pow 2 + &2 * c0 * c2) * x pow 2 + | |
(&2 * c0 * c1) * x + | |
c0 pow 2) (mod p)`, | |
REWRITE_TAC[MULT_2; INT_POW_ADD] THEN | |
INTEGER_TAC) | |
and lemma_mulc = prove | |
(`(y:int == c2 * x pow 2 + c1 * x + c0) (mod p) | |
==> !c. (c * y == (c * c2) * x pow 2 + (c * c1) * x + c * c0) (mod p)`, | |
INTEGER_TAC) | |
and pth0 = INTEGER_RULE `!p. ((x:int) pow 0 == &0 * x pow 2 + &0 * x + &1) | |
(mod &p)` in | |
fun tm primeth -> | |
let th = ASSUME tm in | |
let rule_3_2 = MATCH_MP(MATCH_MP lemma_3_2 th) | |
and rule_4_3 = MATCH_MP(MATCH_MP lemma_4_3 th) | |
and rule_mulx = MATCH_MP lemma_mulx | |
and rule_mulc = MATCH_MP lemma_mulc | |
and rule_sqr = MATCH_MP lemma_sqr | |
and rule_rem = MATCH_MP lemma_rem in | |
let qtm = rand(concl primeth) in | |
let q = dest_numeral qtm in | |
let th0 = SPEC qtm pth0 in | |
let rec power_rule n = | |
if n =/ num_0 then th0 else | |
let m = quo_num n num_2 in | |
let th1 = power_rule m in | |
let th2 = (CONV_RULE INT_REDUCE_CONV o | |
CONV_RULE NUM_REDUCE_CONV o | |
rule_rem o rule_3_2 o rule_4_3 o rule_sqr) th1 in | |
if mod_num n num_2 =/ num_0 then th2 else | |
(CONV_RULE INT_REDUCE_CONV o | |
CONV_RULE NUM_REDUCE_CONV o | |
rule_rem o rule_3_2 o rule_mulx) th2 in | |
let th1 = DISCH tm (power_rule q) in | |
let th2 = MP (MATCH_MP lemma_flt th1) primeth in | |
let th3 = MP th2 (EQT_ELIM(INT_REDUCE_CONV(lhand(concl th2)))) in | |
let th4 = (CONV_RULE INT_REDUCE_CONV o rule_rem) (UNDISCH th3) in | |
let hc = dest_intconst(lhand(lhand(rand(rator(concl th4))))) in | |
let th8 = | |
if hc =/ num_0 then th4 else | |
let th5 = SPEC (mk_intconst(num_modinv q hc)) (rule_mulc th4) in | |
let th6 = (CONV_RULE INT_REDUCE_CONV o rule_rem) th5 in | |
let th7 = MATCH_MP lemma_tolin (CONJ (ASSUME tm) th6) in | |
(CONV_RULE INT_REDUCE_CONV o rule_rem) th7 in | |
let lc = dest_intconst(lhand(lhand(rand(rand(rator(concl th8)))))) in | |
let th9 = SPEC (mk_intconst(num_modinv q lc)) (rule_mulc th8) in | |
let tha = (CONV_RULE INT_REDUCE_CONV o rule_rem) th9 in | |
let thb = MATCH_MP lemma_toconst (CONJ (ASSUME tm) tha) in | |
let thc = CONV_RULE INT_REDUCE_CONV (REWRITE_RULE[GSYM INT_REM_EQ] thb) in | |
NOT_INTRO(DISCH tm thc);; | |
let EXCLUDE_MODULAR_CUBIC_ROOTS_TAC primeth defths = | |
X_GEN_TAC `x:int` THEN | |
GEN_REWRITE_TAC ONCE_DEPTH_CONV | |
[INT_ARITH `x pow 3 - &3 * x + b:int = x pow 3 + (-- &3) * x + b /\ | |
x pow 3 + &c = x pow 3 + &0 * x + &c`] THEN | |
REWRITE_TAC defths THEN | |
W(fun (asl,w) -> ACCEPT_TAC | |
(EXCLUDE_MODULAR_CUBIC_ROOTS (rand w) | |
(ONCE_REWRITE_RULE defths primeth)));; | |