/*
 * CK2PHILIP.C
 *
 *    Version 1.0:  Christopher Putnam 
 *                  cdputnam@scripps.edu
 *                  http://www.scripps.edu/~cdputnam
 *
 *    Source released under GPL
 */
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#ifndef TRUE
#define TRUE (1)
#define FALSE (!TRUE)
#endif

typedef struct seqs {
    char name[10];
    char *seq;
    unsigned long length;
    unsigned long dim;
} seqs;

void seqs_free( seqs *seqlist, unsigned long numseq )
{
    unsigned long i;
    if (seqlist==NULL) return;
    for (i=0; i<numseq; ++i) {
        if (seqlist[i].seq != NULL) free(seqlist[i].seq);
    }
    free(seqlist);
}

unsigned long seqs_init ( seqs **seqlistptr )
{
    seqs *seqlist;
    const unsigned int minsize = 100;
    seqlist = (seqs *) malloc (sizeof(seqs)*minsize);
    if (seqlist==NULL) {
       fprintf(stderr,"ck2phylip: Cannot allocate memory in seqs_init (%ld elements)\n",minsize);
       exit(EXIT_FAILURE);
    }
    *seqlistptr = seqlist;
    return minsize;
}
unsigned long seqs_getmore( seqs **seqlistptr, unsigned long olddimseq )
{
    unsigned int newsize;
    seqs *seqlist;
    newsize = olddimseq*2;
    seqlist = *seqlistptr;
    seqlist = (seqs *) realloc (seqlist, sizeof(seqs)*newsize);
    if (seqlist==NULL) {
       fprintf(stderr,"ck2phylip: Cannot reallocate memory in seqs_getmore (%ld elements)\n",newsize);
       exit(EXIT_FAILURE);
    }
    *seqlistptr = seqlist;
    return newsize;
}    

char *nonws (char *q)
{
    char *p;
    p = q;
    while (*p==' ' || *p=='\t' || *p=='\r' || *p=='\n') p++;
    return p;
}

unsigned long normalize (seqs *seqlist, unsigned long numseq)
{
    unsigned long i,j,max;
    char *newptr;
    if (seqlist==NULL || numseq==0) return 0;
    max = seqlist[0].length;
    /* get maxlength */
    for (i=1; i<numseq; ++i) {
       if (seqlist[i].length>max) max=seqlist[i].length;
    }
    /* normalize length */
    for (i=0; i<numseq; ++i) {
       if (seqlist[i].length < max) {
          if (seqlist[i].dim < max) {
                newptr = (char *) realloc(seqlist[i].seq,sizeof(char)*max);
                if (newptr==NULL) { fprintf(stderr,"ERROR REALLOC.\n"); exit(EXIT_FAILURE); }
                seqlist[i].seq = newptr;
                for (j=seqlist[i].length;j<max; ++j) seqlist[i].seq[j]='-';
          }
       }
    }
    return max; 
}

void write_phylip (seqs *seqlist, unsigned long numseq, unsigned long maxseqlength)
{
    unsigned long i,j;
    if (seqlist==NULL) return;
    printf(" %ld %ld\n",numseq,maxseqlength);
    for (i=0; i<numseq; ++i) {
         for (j=0; j<10; ++j) {
             printf("%c",seqlist[i].name[j]);
         }
         for (j=0; j<maxseqlength; ++j) {
             printf("%c",seqlist[i].seq[j]);
         }
         printf("\n");
    }
    printf("\n"); 
}

void read_ckwhence(char *filename, seqs **seqlistptr, unsigned long *numseqptr, unsigned long *dimseqptr)
{
    FILE *fp;
    seqs *seqlist=NULL;
    char buf[10000];
    unsigned long numseq, dimseq, pos, currseq, len;
    int eos,i;  /* end of string */
    char notlessthan = '*', *p;

    fp = fopen(filename,"r");
    if (fp==NULL) {
       fprintf(stderr,"Cannot open file %s\n",filename);
       return;
    }

    numseq = *numseqptr;
    dimseq = *dimseqptr;
    p = &notlessthan;  /* ensure goes into parse > 1st time */

    while (!feof(fp)) {
       /*
        * Parse till ">"
        */
        while (!feof(fp) && *p!='>') {
          if (fgets(buf,sizeof(buf),fp)!=NULL) {
             p = nonws(buf);    
          }
        }
       /*
        * Grab name (fill with spaces until end, no \0 at end)
        */
        if (seqlist==NULL) seqlist = *seqlistptr;
        if (seqlist==NULL) dimseq = seqs_init(&seqlist); 
        if (*p=='>') {
           if (dimseq == numseq) dimseq = seqs_getmore(&seqlist,dimseq); 
           currseq=numseq;
           numseq++;
           p++;
           eos=FALSE;
           for (i=0; i<10; ++i) {
              if (p[i]=='\0' || p[i]=='\r' || p[i]=='\n' ) eos=TRUE;
              seqlist[currseq].name[i] = ( eos==TRUE ) ? ' ' : p[i];
           }
           seqlist[currseq].seq=NULL;
           seqlist[currseq].dim=0;
           seqlist[currseq].length=0;
        }
        while (!feof(fp) && *p!='>') {
           /* get next line and put p to 1st non ws */
           if (fgets(buf,sizeof(buf),fp)!=NULL) {
             p = nonws(buf);
           }
           /* if couldn't get line or we're at new title go on...*/
           if (!feof(fp) && *p!='>') {
              len = strlen(p) + seqlist[currseq].length + 1;
              if (seqlist[currseq].seq==NULL) {
                  seqlist[currseq].seq = malloc(sizeof(char)*len);
                  if (seqlist[currseq].seq==NULL) {
                     fprintf(stderr,"Cannot alloc to size %ld!!!\n",len);
                     exit(EXIT_FAILURE);
                  }
                  seqlist[currseq].dim = len;
              }
              else if (len>seqlist[currseq].dim) {
                  seqlist[currseq].seq = realloc(seqlist[currseq].seq,sizeof(char)*len);
                  if (seqlist[currseq].seq==NULL) {
                     fprintf(stderr,"Cannot realloc to size %ld!!!\n",len);
                     exit(EXIT_FAILURE);
                  }
                  seqlist[currseq].dim = len;
              }
              pos = seqlist[currseq].length;
              while (*p) { 
                  p = nonws(p);
                  if (*p!='\0') {
                        if (*p!='*') seqlist[currseq].seq[pos++]=*p;
                        *p++;
                  }
              }
              seqlist[currseq].seq[pos]='\0';
              seqlist[currseq].length=pos; 
           }
        } 
    }

    *seqlistptr = seqlist;
    *numseqptr = numseq;
    *dimseqptr = dimseq;

    fclose(fp); 
}

int main (int argc, char *argv[])
{
    int i;
    seqs *seqlist=NULL;
    unsigned long numseq=0,maxlenseq=0,dimseq=0;
    if (argc<2) {
       fprintf(stderr,"Usage:  ck2phylip <sequoiafile> [<sequoiafile>....]\n");
       fprintf(stderr,"\nTakes SEQUOIA/CKWHENCE files and generates ");
       fprintf(stderr,"non-interlaced PHYLIP files.\n");
       fprintf(stderr,"SEQUOIA:  http://www.scripps.edu/~bruns/sequoia.html\n");
       fprintf(stderr,"PHYLIP:   http://evolution.genetics.washington.edu/phylip.html\n");
       fprintf(stderr,"\nVersion 1.0, February 2000\n\n");
       exit(EXIT_FAILURE);
    }
    for (i=1; i<argc; ++i) {
       read_ckwhence(argv[i],&seqlist,&numseq,&dimseq);
    }
    maxlenseq = normalize(seqlist,numseq);
    write_phylip(seqlist,numseq,maxlenseq);
    seqs_free(seqlist,numseq);
}