Mercurial > repos > bitlab > imsame

comparison IMSAME/src/combine_reads.c @ 0:762009a91895 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author bitlab
date Sat, 15 Dec 2018 18:04:10 -0500
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:762009a91895
1 #include <stdio.h>
2 #include <stdlib.h>
3 #include <string.h>
4 #include <ctype.h>
5 #include "structs.h"
6 #include "commonFunctions.h"
7
8 #define MAX(x, y) (((x) > (y)) ? (x) : (y))
9 #define MIN(x, y) (((x) <= (y)) ? (x) : (y))
10 #define STARTING_SEQS 1000
11 #define PIECE_OF_DB_REALLOC 3200000 //half a gigabyte if divided by 8 bytes
12 #define MATVAL 101
13
14 // void terror(char *s) {
15 // fprintf(stdout, "ERR**** %s ****\n", s);
16 // exit(-1);
17 // }
18
19
20 int main(int argc, char ** av){
21
22 if(argc != 2) terror("USE:: combine_reads <file>");
23
24
25 char names[10][100] = {"14np", "47np", "BR0716C", "BR11044P", "CA01044C", "CA01067C", "CA11149P", "CA11154P", "MA07066C", "MA09032P"};
26 uint64_t llen[10] = {1758451, 959020, 1616163, 884193, 934256, 1007436, 1670521, 584151, 671010, 516632};
27 uint64_t nseqs = 10;
28
29 FILE * results, * data;
30 data = fopen(av[1], "rt");
31 if(data == NULL) terror("Could not open input file");
32
33 uint64_t * mat = (uint64_t *) calloc(MATVAL, sizeof(uint64_t));
34 if(mat == NULL) terror("Could not allocate matrix array");
35 long double * mat_e = (long double *) calloc(MATVAL, sizeof(long double));
36 if(mat_e == NULL) terror("Could not allocate float table");
37
38 char current_file[100];
39 uint64_t i=5, idx=0;
40 while(av[1][i] != '_' && av[1][i+1] != 'g'){
41 current_file[idx++] = av[1][i];
42 i++;
43 }
44 current_file[idx] = '\0';
45 fprintf(stdout, "Current file is %s\n", current_file);
46
47 char buffer[MAXLID];
48 if ((results = fopen("accu.log", "r")) == NULL){
49 results = fopen("accu.log", "wt");
50 }else{
51 // Load the matrix
52
53
54 for(i=0;i<100;i++){
55 if(0 == fgets(buffer, MAXLID, results)) terror("Missing number on load");
56
57 //fprintf(stdout, "Have %s\n", buffer);
58 buffer[strlen(buffer)-1] = '\0';
59 //mat[i] = asciiToUint64(buffer);
60 mat_e[i] = (long double) atof(buffer);
61 //fprintf(stdout, "%"PRIu64"\n", mat[i]);
62 //getchar();
63 }
64 fclose(results);
65 results = fopen("accu.log", "wt"); // Re open
66 }
67
68 // Read file
69 uint64_t read_id_1, read_id_2, coverage, identity, length, current, currmax, j, last = 100000, lastmax = 0xFFFFFFFFFFFFFFFF;
70 while(!feof(data)){
71 if(0 == fgets(buffer, MAXLID, data) && !feof(data)) terror("Missing values");
72 // 2 77277 89 64 213
73 sscanf(buffer, "%"PRIu64" %"PRIu64" %"PRIu64" %"PRIu64" %"PRIu64, &read_id_1, &read_id_2, &coverage, &identity, &length);
74 //fprintf(stdout, "%"PRIu64" %"PRIu64" %"PRIu64" %"PRIu64" %"PRIu64"\n", read_id_1, read_id_2, coverage, identity, length);
75 currmax = MIN(identity, coverage);
76 //fprintf(stdout, "%"PRIu64"\n", currmax);
77 current = read_id_1;
78 /*
79 for(j=currmax; j > 1; j--){
80 if(current != lasts[j]){
81 mat[j]++;
82 lasts[j] = current;
83 }
84 }
85 */
86 if(lastmax == 0xFFFFFFFFFFFFFFFF){
87 last = current;
88 lastmax = 0;
89 }
90 if(current == last){
91 //mat[currmax] += 1;
92 if(lastmax < currmax){
93 lastmax = currmax;
94 }
95 }else{
96 mat[lastmax]++;
97 last = current;
98 lastmax = 0;
99 }
100
101 }
102
103 uint64_t divider = 0;
104 for(i=0; i < nseqs; i++){
105
106 if(strcmp(current_file, names[i]) == 0){
107 divider = llen[i];
108 }
109 }
110 if(divider == 0) terror("Could not find file for divider"); else fprintf(stdout, "The divider is %"PRIu64"\n", divider);
111
112 for(j=99; j>0; j--){
113 mat[j] += mat[j+1];
114 }
115 mat[0] = mat[1];
116
117 for(j=0; j<100; j++){
118 fprintf(stdout, "%"PRIu64" --> %.5f\n", mat[j], (float)((100*(long double)mat[j])/(long double)divider));
119
120 fprintf(results, "%.5f\n", (float)(mat_e[j] + ((100*(long double)mat[j])/(long double)divider))/2);
121 }
122
123
124 fclose(results);
125 fclose(data);
126 free(mat);
127 return 0;
128 }