X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=binaries%2Fsrc%2Fmafft%2Fcore%2FpartSalignmm.c;fp=binaries%2Fsrc%2Fmafft%2Fcore%2FpartSalignmm.c;h=b670b73d22105ff9991208587bd20935d6fb9e62;hb=063b30bb5e8161134ae764742636ab538e10eea7;hp=0000000000000000000000000000000000000000;hpb=6e0ce943f09b5ac30f3eb8dc0f20bc75114669ce;p=jabaws.git diff --git a/binaries/src/mafft/core/partSalignmm.c b/binaries/src/mafft/core/partSalignmm.c new file mode 100644 index 0000000..b670b73 --- /dev/null +++ b/binaries/src/mafft/core/partSalignmm.c @@ -0,0 +1,1266 @@ +#include "mltaln.h" +#include "dp.h" + +#define MACHIGAI 0 +#define OUTGAP0TRY 1 +#define DEBUG 0 +#define XXXXXXX 0 +#define USE_PENALTY_EX 0 +#define FASTMATCHCALC 1 + +#if 0 +static void st_OpeningGapCount( float *ogcp, int clus, char **seq, double *eff, int len ) +{ + int i, j, gc, gb; + float feff; + + for( i=0; i impmtx=%f\n", i1, j1, impmtx[i1][j1] ); + return( impmtx[i1][j1] ); +#if 0 + if( i1 == l1 || j1 == l2 ) return( 0.0 ); + return( impmtx[i1+start1][j1+start2] ); +#endif +} +static void part_imp_match_out_vead_gapmap( float *imp, int i1, int lgth2, int start2, int *gapmap2 ) +{ +#if FASTMACHCALC + float *pt = imp; + int *gapmappt = gapmap2; + while( lgth2-- ) + *pt++ += impmtx[i1][start2+*gapmappt++]; +#else + int j; + for( j=0; jstart1 ); +// fprintf( stderr, "end1 = %d\n", tmpptr->end1 ); +// fprintf( stderr, "i = %d, seq1 = \n%s\n", i, seq1[i] ); +// fprintf( stderr, "j = %d, seq2 = \n%s\n", j, seq2[j] ); + pt = seq1[i]; + tmpint = -1; + while( *pt != 0 ) + { + if( *pt++ != '-' ) tmpint++; + if( tmpint == tmpptr->start1 ) break; + } + start1 = (int)( pt - seq1[i] ) - 1; + + if( tmpptr->start1 == tmpptr->end1 ) end1 = start1; + else + { +#if MACHIGAI + while( *pt != 0 ) + { + if( tmpint == tmpptr->end1 ) break; + if( *pt++ != '-' ) tmpint++; + } + end1 = (int)( pt - seq1[i] ) - 1; +#else + while( *pt != 0 ) + { +// fprintf( stderr, "tmpint = %d, end1 = %d pos = %d\n", tmpint, tmpptr->end1, pt-seq1[i] ); + if( *pt++ != '-' ) tmpint++; + if( tmpint == tmpptr->end1 ) break; + } + end1 = (int)( pt - seq1[i] ) - 1; +#endif + } + + pt = seq2[j]; + tmpint = -1; + while( *pt != 0 ) + { + if( *pt++ != '-' ) tmpint++; + if( tmpint == tmpptr->start2 ) break; + } + start2 = (int)( pt - seq2[j] ) - 1; + if( tmpptr->start2 == tmpptr->end2 ) end2 = start2; + else + { +#if MACHIGAI + while( *pt != 0 ) + { + if( tmpint == tmpptr->end2 ) break; + if( *pt++ != '-' ) tmpint++; + } + end2 = (int)( pt - seq2[j] ) - 1; +#else + while( *pt != 0 ) + { + if( *pt++ != '-' ) tmpint++; + if( tmpint == tmpptr->end2 ) break; + } + end2 = (int)( pt - seq2[j] ) - 1; +#endif + } +// fprintf( stderr, "start1 = %d (%c), end1 = %d (%c), start2 = %d (%c), end2 = %d (%c)\n", start1, seq1[i][start1], end1, seq1[i][end1], start2, seq2[j][start2], end2, seq2[j][end2] ); +// fprintf( stderr, "step 0\n" ); + if( end1 - start1 != end2 - start2 ) + { +// fprintf( stderr, "CHUUI!!, start1 = %d, end1 = %d, start2 = %d, end2 = %d\n", start1, end1, start2, end2 ); + } + + k1 = start1; k2 = start2; + pt1 = seq1[i] + k1; + pt2 = seq2[j] + k2; + while( *pt1 && *pt2 ) + { + if( *pt1 != '-' && *pt2 != '-' ) + { +// ½Å¤ß¤òÆó½Å¤Ë¤«¤±¤Ê¤¤¤è¤¦¤ËÃí°Õ¤·¤Æ²¼¤µ¤¤¡£ +// impmtx[k1][k2] += tmpptr->wimportance * fastathreshold; +// impmtx[k1][k2] += tmpptr->importance * effij; +// impmtx[k1][k2] += tmpptr->fimportance * effij; + if( tmpptr->korh == 'k' ) + impmtx[k1][k2] += tmpptr->fimportance * effij_kozo; + else + impmtx[k1][k2] += tmpptr->fimportance * effij; +// fprintf( stderr, "k1=%d, k2=%d, impalloclen=%d\n", k1, k2, impalloclen ); +// fprintf( stderr, "mark, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 ); + k1++; k2++; + pt1++; pt2++; + } + else if( *pt1 != '-' && *pt2 == '-' ) + { +// fprintf( stderr, "skip, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 ); + k2++; pt2++; + } + else if( *pt1 == '-' && *pt2 != '-' ) + { +// fprintf( stderr, "skip, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 ); + k1++; pt1++; + } + else if( *pt1 == '-' && *pt2 == '-' ) + { +// fprintf( stderr, "skip, %d (%c) - %d (%c) \n", k1, *pt1, k2, *pt2 ); + k1++; pt1++; + k2++; pt2++; + } + if( k1 > end1 || k2 > end2 ) break; + } + tmpptr = tmpptr->next; + } + } + } +#if 0 + fprintf( stderr, "impmtx = \n" ); + for( k2=0; k2start1 ); + fprintf( stderr, "end1 = %d\n", localhom[i][j]->end1 ); + fprintf( stderr, "j = %d, seq2 = %s\n", j, seq2[j] ); + pt = seq1[i]; + tmpint = -1; + while( *pt != 0 ) + { + if( *pt++ != '-' ) tmpint++; + if( tmpint == localhom[i][j]->start1 ) break; + } + start1 = pt - seq1[i] - 1; + + while( *pt != 0 ) + { +// fprintf( stderr, "tmpint = %d, end1 = %d pos = %d\n", tmpint, localhom[i][j].end1, pt-seq1[i] ); + if( *pt++ != '-' ) tmpint++; + if( tmpint == localhom[i][j]->end1 ) break; + } + end1 = pt - seq1[i] - 1; + + pt = seq2[j]; + tmpint = -1; + while( *pt != 0 ) + { + if( *pt++ != '-' ) tmpint++; + if( tmpint == localhom[i][j]->start2 ) break; + } + start2 = pt - seq2[j] - 1; + while( *pt != 0 ) + { + if( *pt++ != '-' ) tmpint++; + if( tmpint == localhom[i][j]->end2 ) break; + } + end2 = pt - seq2[j] - 1; +// fprintf( stderr, "start1 = %d, end1 = %d, start2 = %d, end2 = %d\n", start1, end1, start2, end2 ); + k1 = start1; + k2 = start2; + fprintf( stderr, "step 0\n" ); + while( k1 <= end1 && k2 <= end2 ) + { +#if 0 + if( !nocount1[k1] && !nocount2[k2] ) + impmtx[k1][k2] += localhom[i][j].wimportance * eff1[i] * eff2[j]; + k1++; k2++; +#else + if( !nocount1[k1] && !nocount2[k2] ) + impmtx[k1][k2] += localhom[i][j]->wimportance * eff1[i] * eff2[j]; + k1++; k2++; +#endif + } + + dif = ( end1 - start1 ) - ( end2 - start2 ); + fprintf( stderr, "dif = %d\n", dif ); + if( dif > 0 ) + { + do + { + fprintf( stderr, "dif = %d\n", dif ); + k1 = start1; + k2 = start2 - dif; + while( k1 <= end1 && k2 <= end2 ) + { + if( 0 <= k2 && start2 <= k2 && !nocount1[k1] && !nocount2[k2] ) + impmtx[k1][k2] = localhom[i][j]->wimportance * eff1[i] * eff2[j]; + k1++; k2++; + } + } + while( dif-- ); + } + else + { + do + { + k1 = start1 + dif; + k2 = start2; + while( k1 <= end1 ) + { + if( k1 >= 0 && k1 >= start1 && !nocount1[k1] && !nocount2[k2] ) + impmtx[k1][k2] = localhom[i][j]->wimportance * eff1[i] * eff2[j]; + k1++; k2++; + } + } + while( dif++ ); + } + } + } +#if 0 + fprintf( stderr, "impmtx = \n" ); + for( k2=0; k2-1 ) + *matchpt += scarr[*cpmxpdnpt++] * *cpmxpdpt++; + matchpt++; + } + } +#else + int j, k, l; + float scarr[26]; + float **cpmxpd = floatwork; + int **cpmxpdn = intwork; + // simple + if( initialize ) + { + int count = 0; + for( j=0; j-1; k++ ) + match[j] += scarr[cpmxpdn[k][j]] * cpmxpd[k][j]; + } +#endif +} + +static void Atracking_localhom( float *impwmpt, float *lasthorizontalw, float *lastverticalw, + char **seq1, char **seq2, + char **mseq1, char **mseq2, + float **cpmx1, float **cpmx2, + int **ijp, int icyc, int jcyc, + int start1, int end1, int start2, int end2, + int *gapmap1, int *gapmap2 ) +{ + int i, j, l, iin, jin, ifi, jfi, lgth1, lgth2, k; +// char gap[] = "-"; + char *gap; + float wm; + gap = newgapstr; + lgth1 = strlen( seq1[0] ); + lgth2 = strlen( seq2[0] ); + +#if 0 + for( i=0; i= wm ) + { + wm = lastverticalw[i]; + iin = i; jin = lgth2-1; + ijp[lgth1][lgth2] = +( lgth1 - i ); + } + } + for( j=0; j= wm ) + { + wm = lasthorizontalw[j]; + iin = lgth1-1; jin = j; + ijp[lgth1][lgth2] = -( lgth2 - j ); + } + } + } + + for( i=0; i 0 ) + { + ifi = iin-ijp[iin][jin]; jfi = jin-1; + } + else + { + ifi = iin-1; jfi = jin-1; + } + l = iin - ifi; + while( --l ) + { + for( i=0; i= wm ) + { + wm = lastverticalw[i]; + iin = i; jin = lgth2-1; + ijp[lgth1][lgth2] = +( lgth1 - i ); + } + } + for( j=0; j= wm ) + { + wm = lasthorizontalw[j]; + iin = lgth1-1; jin = j; + ijp[lgth1][lgth2] = -( lgth2 - j ); + } + } + } + + for( i=0; i 0 ) + { + ifi = iin-ijp[iin][jin]; jfi = jin-1; + } + else + { + ifi = iin-1; jfi = jin-1; + } + l = iin - ifi; + while( --l ) + { + for( i=0; i lgth1, outgap == 1 -> lgth1+1 */ + int lgth1, lgth2; + int resultlen; + float wm = 0.0; /* int ?????? */ + float g; + float *currentw, *previousw; +#if 1 + float *wtmp; + int *ijppt; + float *mjpt, *prept, *curpt; + int *mpjpt; +#endif + static TLS float mi, *m; + static TLS int **ijp; + static TLS int mpi, *mp; + static TLS float *w1, *w2; + static TLS float *match; + static TLS float *initverticalw; /* kufuu sureba iranai */ + static TLS float *lastverticalw; /* kufuu sureba iranai */ + static TLS char **mseq1; + static TLS char **mseq2; + static TLS char **mseq; + static TLS float *ogcp1; + static TLS float *ogcp2; + static TLS float *fgcp1; + static TLS float *fgcp2; + static TLS float **cpmx1; + static TLS float **cpmx2; + static TLS int **intwork; + static TLS float **floatwork; + static TLS int orlgth1 = 0, orlgth2 = 0; + float fpenalty = (float)penalty; +#if USE_PENALTY_EX + float fpenalty_ex = (float)penalty_ex; +#endif + float *fgcp2pt; + float *ogcp2pt; + float fgcp1va; + float ogcp1va; + + + if( seq1 == NULL ) + { + if( orlgth1 ) + { +// fprintf( stderr, "## Freeing local arrays in A__align\n" ); + orlgth1 = 0; + orlgth2 = 0; + + part_imp_match_init_strict( NULL, 0, 0, 0, 0, NULL, NULL, NULL, NULL, NULL, NULL, NULL, 0 ); + + free( mseq1 ); + free( mseq2 ); + FreeFloatVec( w1 ); + FreeFloatVec( w2 ); + FreeFloatVec( match ); + FreeFloatVec( initverticalw ); + FreeFloatVec( lastverticalw ); + + FreeFloatVec( m ); + FreeIntVec( mp ); + + FreeCharMtx( mseq ); + + FreeFloatVec( ogcp1 ); + FreeFloatVec( ogcp2 ); + FreeFloatVec( fgcp1 ); + FreeFloatVec( fgcp2 ); + + + FreeFloatMtx( cpmx1 ); + FreeFloatMtx( cpmx2 ); + + FreeFloatMtx( floatwork ); + FreeIntMtx( intwork ); + + } + else + { +// fprintf( stderr, "## Not allocated\n" ); + } + return( 0.0 ); + } + + lgth1 = strlen( seq1[0] ); + lgth2 = strlen( seq2[0] ); +#if 1 + if( lgth1 == 0 ) fprintf( stderr, "WARNING: lgth1=0 in partA__align\n" ); + if( lgth2 == 0 ) fprintf( stderr, "WARNING: lgth2=0 in partA__align\n" ); + + if( lgth1 == 0 && lgth2 == 0 ) + return( 0.0 ); + + if( lgth1 == 0 ) + { + for( i=0; i orlgth1 || lgth2 > orlgth2 ) + { + int ll1, ll2; + + if( orlgth1 > 0 && orlgth2 > 0 ) + { + FreeFloatVec( w1 ); + FreeFloatVec( w2 ); + FreeFloatVec( match ); + FreeFloatVec( initverticalw ); + FreeFloatVec( lastverticalw ); + + FreeFloatVec( m ); + FreeIntVec( mp ); + + FreeCharMtx( mseq ); + + FreeFloatVec( ogcp1 ); + FreeFloatVec( ogcp2 ); + FreeFloatVec( fgcp1 ); + FreeFloatVec( fgcp2 ); + + + FreeFloatMtx( cpmx1 ); + FreeFloatMtx( cpmx2 ); + + FreeFloatMtx( floatwork ); + FreeIntMtx( intwork ); + } + + ll1 = MAX( (int)(1.3*lgth1), orlgth1 ) + 100; + ll2 = MAX( (int)(1.3*lgth2), orlgth2 ) + 100; + +#if DEBUG + fprintf( stderr, "\ntrying to allocate (%d+%d)xn matrices ... ", ll1, ll2 ); +#endif + + w1 = AllocateFloatVec( ll2+2 ); + w2 = AllocateFloatVec( ll2+2 ); + match = AllocateFloatVec( ll2+2 ); + + initverticalw = AllocateFloatVec( ll1+2 ); + lastverticalw = AllocateFloatVec( ll1+2 ); + + m = AllocateFloatVec( ll2+2 ); + mp = AllocateIntVec( ll2+2 ); + + mseq = AllocateCharMtx( njob, ll1+ll2 ); + + ogcp1 = AllocateFloatVec( ll1+2 ); + ogcp2 = AllocateFloatVec( ll2+2 ); + fgcp1 = AllocateFloatVec( ll1+2 ); + fgcp2 = AllocateFloatVec( ll2+2 ); + + cpmx1 = AllocateFloatMtx( 26, ll1+2 ); + cpmx2 = AllocateFloatMtx( 26, ll2+2 ); + +#if FASTMATCHCALC + floatwork = AllocateFloatMtx( MAX( ll1, ll2 )+2, 26 ); + intwork = AllocateIntMtx( MAX( ll1, ll2 )+2, 26 ); +#else + floatwork = AllocateFloatMtx( 26, MAX( ll1, ll2 )+2 ); + intwork = AllocateIntMtx( 26, MAX( ll1, ll2 )+2 ); +#endif + +#if DEBUG + fprintf( stderr, "succeeded\n" ); +#endif + + orlgth1 = ll1 - 100; + orlgth2 = ll2 - 100; + } + + + for( i=0; i commonAlloc1 || orlgth2 > commonAlloc2 ) + { + int ll1, ll2; + + if( commonAlloc1 && commonAlloc2 ) + { + FreeIntMtx( commonIP ); + } + + ll1 = MAX( orlgth1, commonAlloc1 ); + ll2 = MAX( orlgth2, commonAlloc2 ); + +#if DEBUG + fprintf( stderr, "\n\ntrying to allocate %dx%d matrices ... ", ll1+1, ll2+1 ); +#endif + + commonIP = AllocateIntMtx( ll1+10, ll2+10 ); + +#if DEBUG + fprintf( stderr, "succeeded\n\n" ); +#endif + + commonAlloc1 = ll1; + commonAlloc2 = ll2; + } + ijp = commonIP; + + cpmx_calc_new( seq1, cpmx1, eff1, lgth1, icyc ); + cpmx_calc_new( seq2, cpmx2, eff2, lgth2, jcyc ); + + if( sgap1 ) + { + new_OpeningGapCount( ogcp1, icyc, seq1, eff1, lgth1, sgap1 ); + new_OpeningGapCount( ogcp2, jcyc, seq2, eff2, lgth2, sgap2 ); + new_FinalGapCount( fgcp1, icyc, seq1, eff1, lgth1, egap1 ); + new_FinalGapCount( fgcp2, jcyc, seq2, eff2, lgth2, egap2 ); + } + else + { + st_OpeningGapCount( ogcp1, icyc, seq1, eff1, lgth1 ); + st_OpeningGapCount( ogcp2, jcyc, seq2, eff2, lgth2 ); + st_FinalGapCount( fgcp1, icyc, seq1, eff1, lgth1 ); + st_FinalGapCount( fgcp2, jcyc, seq2, eff2, lgth2 ); + } + + for( i=0; i tbfast.c + if( localhom ) + imp_match_calc( currentw, icyc, jcyc, lgth1, lgth2, seq1, seq2, eff1, eff2, localhom, 1, 0 ); + +#endif + + if( outgap == 1 ) + { + for( i=1; i", wm ); +#endif + g = mi + *fgcp2pt; +#if 0 + fprintf( stderr, "%5.0f?", g ); +#endif + if( g > wm ) + { + wm = g; + *ijppt = -( j - mpi ); + } + g = *prept + *ogcp2pt; + if( g >= mi ) + { + mi = g; + mpi = j-1; + } +#if USE_PENALTY_EX + mi += fpenalty_ex; +#endif + + g = *mjpt + fgcp1va; +#if 0 + fprintf( stderr, "%5.0f?", g ); +#endif + if( g > wm ) + { + wm = g; + *ijppt = +( i - *mpjpt ); + } + g = *prept + ogcp1va; + if( g >= *mjpt ) + { + *mjpt = g; + *mpjpt = i-1; + } +#if USE_PENALTY_EX + m[j] += fpenalty_ex; +#endif + +#if 0 + fprintf( stderr, "%5.0f ", wm ); +#endif + *curpt += wm; + ijppt++; + mjpt++; + prept++; + mpjpt++; + curpt++; + fgcp2pt++; + ogcp2pt++; + } + lastverticalw[i] = currentw[lgth2-1]; + + } + +#if OUTGAP0TRY + if( !outgap ) + { + for( j=1; j" ); + for( i=0; i N ) + { + fprintf( stderr, "alloclen=%d, resultlen=%d, N=%d\n", alloclen, resultlen, N ); + ErrorExit( "LENGTH OVER!\n" ); + } + + + for( i=0; i