/*
 *  Permission to use, copy, modify, distribute, and sell this software
 *  for any purpose and without fee, restriction or acknowledgement is
 *  hereby granted.  The author (James Knight of the Univ. of California,
 *  Davis) places it in the public domain.
 *
 *  This software is provided AS IS with no warranties of any kind.  The
 *  author shall have no liability with respect to the infringement of
 *  copyrights, trade secrets or any patents by this software or any part
 *  thereof.  In no event will the author be liable for any lost revenue
 *  or profits or other special, indirect and consequential damages.
 */

#include <stdio.h>
#include <stdlib.h>
#include "seqio.h"

/*
 * This is a simple version of a `wc' program for biological sequences.
 * It reads the input and counts the number of entries, number of sequences
 * and total length of the sequences.
 *
 * It reads the input a sequence at a time (making it simple to count
 * the number of sequences and their total length), and it determines
 * the number of entries by counting the number of times "seqfseqno"
 * (which gives the position of the sequence in the current entry) is 1.
 *
 * The program also reports the counts for each file of a database, using
 * bioseq_parse to retrieve the files.
 */

int main(int argc, char *argv[])
{
  int i, len, entries, seqs, total, bigentries, bigseqs, bigtotal;
  int multifile_mode, firsttimeflag, is_singleentry;
  char *s, *seq, *file, *filelist, *format;
  SEQFILE *sfp;

  if (argc == 1) {
    if ((sfp = seqfopen2("-")) != NULL) {
      entries = seqs = total = 0;
      while ((seq = seqfgetseq(sfp, &len, 0)) != NULL) {
        if (seqfseqno(sfp) == 1)
          entries++;
        seqs++;
        total += len;
      }
      seqfclose(sfp);

      printf(" %10d %10d %10d (stdin)\n", entries, seqs, total);
    }
  }
  else {
    entries = bigentries = bigseqs = bigtotal = 0;
    file = filelist = NULL;
    for (i=1; i < argc; i++) {
      /*
       * Figure out whether it's an existing file, a database specification
       * or neither.
       *
       * If it's a database specification and is a specification of single
       * entries of a database, only collect and print a total count of the
       * single entries.  If the database specification consists of complete
       * files, then collect and print totals for each database file.
       */
      is_singleentry = 0;
      if (seqfisafile(argv[i])) {
        file = argv[i];
        format = NULL;
        multifile_mode = 0;
      }
      else if ((filelist = bioseq_parse(argv[i])) != NULL) {
        for (s=filelist; *s; s++) {
          if (*s == '\n')
            *s = '\0';
          else if (*s == '@')
            is_singleentry = 1;
        }

        file = filelist;
        format = bioseq_info(argv[i], "Format");
        multifile_mode = 1;
      }
      else {
        fprintf(stderr, "%s:  Not a file or database.\n", argv[i]);
        continue;
      }

      entries = seqs = total = 0;

      firsttimeflag = 1;
      while (firsttimeflag || multifile_mode) {
        firsttimeflag = 0;

        /*
         * Read the file and count the entries and sequences.
         */
        if ((sfp = seqfopen(file, "r", format)) != NULL) {
          if (!is_singleentry)
            entries = seqs = total = 0;

          while ((seq = seqfgetseq(sfp, &len, 0)) != NULL) {
            if (seqfseqno(sfp) == 1)
              entries++;
            seqs++;
            total += len;
          }
          seqfclose(sfp);

          if (!is_singleentry)
            printf(" %10d %10d %10d %s\n", entries, seqs, total, file);

          bigentries += entries;
          bigseqs += seqs;
          bigtotal += total;
        }

        /*
         * If reading a database, advance to the next file in the list.
         */
        if (multifile_mode) {
          while (*file != '\0')
            file++;
          file++;

          if (*file == '\0')
            break;
        }
      }

      if (is_singleentry)
        printf(" %10d %10d %10d %s\n", entries, seqs, total, argv[i]);

      /*
       * If reading a database, free the space allocated by bioseq_parse
       * and bioseq_info.
       */
      if (multifile_mode) {
        if (format != NULL)
          free(format);

        free(filelist);
      }
    }

    if (bigentries != entries)
      printf(" %10d %10d %10d total\n", bigentries, bigseqs, bigtotal);
  }

  return 0;
}