final ass1

[sws1-1415.git] / ass1 / mart / sws1-s4109503-s4202015 / exercise2 / parsegenome.c

diff --git a/ass1/mart/sws1-s4109503-s4202015/exercise2/parsegenome.c b/ass1/mart/sws1-s4109503-s4202015/exercise2/parsegenome.c
index 3faf91c..d561d4f 100644 (file)
--- a/ass1/mart/sws1-s4109503-s4202015/exercise2/parsegenome.c
+++ b/ass1/mart/sws1-s4109503-s4202015/exercise2/parsegenome.c
@@ -1,40 +1,29 @@
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-
+#include <stdint.h> /* - Basically we save the 4 frequencies in a 64bit int  */
+#include <stdio.h>  /* - The polynome fits the following coordinates:        */
+#include <stdlib.h> /* (65,0), (67, 16), (71, 32), (84, 48)                  */
+#include <string.h> /* - Which is the mapping from ascii value and shiftwidth*/
+                    /* - When no input file is given assume stdin            */
 int main(int argc, char* argv[])
 {
        char c;
        char charsperline = 0;
        short numlines = 0;
-       long long unsigned int occs = 0;
-       /* When the filepath is not given or "-" assume stdin */
+       uint64_t occs = 0;
        FILE *f = (argc!=2 || strcmp(argv[1], "-")==0) ? stdin : fopen(argv[1], "r");
        if(f == NULL)
-       {
-               printf("Unable to open file or stdin...\n");
                return 2;
-       }
-       while((c = fgetc(f)) != EOF)
-       {
-               if(c == '\n' && charsperline == 100 && numlines <= 500)
-               {
-                       numlines += 1;
+       while((c=fgetc(f)) != EOF)
+               if(c == '\n' && charsperline == 100 && numlines++ < 500)
                        charsperline = 0;
-               }
-               else if((c == 'A' || c=='C' || c=='G' || c=='T') && charsperline < 100)
-               { /*Fancy polynomial that translates the ascii value to a bitshift*/
-                       occs += 1LL<<(int)(0.026514*c*c*c-6.049*c*c+459.96*c-11621.2);
-                       charsperline += 1;
-               }
+               else if((c=='A' ||c=='C' || c=='G' || c=='T') && charsperline++<100)
+                       occs += 1LL<<(char)(0.026514*c*c*c-6.049*c*c+459.96*c-11621.2);
                else
                        break;
-       }
        if(f != stdout)
                fclose(f);
-       if(numlines != 500)
+       if(numlines!=500)
                return -1;
-       printf("A: %llu\nC: %llu\nG: %llu\nT: %llu\n", occs & 0xFFFF,
-               (occs >> 16) & 0xFFFF, (occs >> 32) & 0xFFFF, (occs >> 48) & 0xFFFF);
+       printf("A:%u\nC:%u\nG:%u\nT:%u\n", (uint16_t)occs, (uint16_t)(occs>>16),
+               (uint16_t)(occs>>32), (uint16_t)(occs>>48));
        return 0;
 }