src/bfile_heuristics.c

Mon, 27 Jul 2020 17:19:56 +0200

author
Mike Becker <universe@uap-core.de>
date
Mon, 27 Jul 2020 17:19:56 +0200
changeset 61
9c8d768f0244
parent 57
68018eac46c3
permissions
-rw-r--r--

adds option to compute individual sums

universe@20 1 /*
universe@34 2 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS HEADER.
universe@57 3 * Copyright 2018 Mike Becker. All rights reserved.
universe@34 4 *
universe@34 5 * Redistribution and use in source and binary forms, with or without
universe@34 6 * modification, are permitted provided that the following conditions are met:
universe@34 7 *
universe@34 8 * 1. Redistributions of source code must retain the above copyright
universe@34 9 * notice, this list of conditions and the following disclaimer.
universe@34 10 *
universe@34 11 * 2. Redistributions in binary form must reproduce the above copyright
universe@34 12 * notice, this list of conditions and the following disclaimer in the
universe@34 13 * documentation and/or other materials provided with the distribution.
universe@34 14 *
universe@34 15 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
universe@34 16 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
universe@34 17 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
universe@34 18 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
universe@34 19 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
universe@34 20 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
universe@34 21 * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
universe@34 22 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
universe@34 23 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
universe@57 24 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
universe@20 25 */
universe@20 26
universe@20 27 #include "bfile_heuristics.h"
universe@22 28 #include <ctype.h>
universe@20 29
universe@21 30 bfile_heuristics_t *new_bfile_heuristics_t() {
universe@21 31 bfile_heuristics_t *ret = malloc(sizeof(bfile_heuristics_t));
universe@21 32 ret->level = BFILE_MEDIUM_ACCURACY;
universe@22 33 bfile_reset(ret);
universe@21 34 return ret;
universe@20 35 }
universe@20 36
universe@21 37 void destroy_bfile_heuristics_t(bfile_heuristics_t *def) {
universe@21 38 free(def);
universe@20 39 }
universe@21 40
universe@22 41 void bfile_reset(bfile_heuristics_t *def) {
universe@22 42 def->bcount = 0;
universe@22 43 def->tcount = 0;
universe@22 44 }
universe@22 45
universe@21 46 bool bfile_check(bfile_heuristics_t *def, int next_char) {
universe@21 47 bool ret = false;
universe@22 48 if (def->level != BFILE_IGNORE) {
universe@22 49 def->tcount++;
universe@22 50 if (!isprint(next_char) && !isspace(next_char)) {
universe@22 51 def->bcount++;
universe@22 52 }
universe@22 53
universe@23 54 if (def->tcount > 1) { /* empty files are text files */
universe@23 55 switch (def->level) {
universe@23 56 case BFILE_LOW_ACCURACY:
universe@23 57 if (def->tcount > 15 || next_char == EOF) {
universe@23 58 ret = (1.0*def->bcount)/def->tcount > 0.32;
universe@23 59 }
universe@23 60 break;
universe@23 61 case BFILE_HIGH_ACCURACY:
universe@23 62 if (def->tcount > 500 || next_char == EOF) {
universe@23 63 ret = (1.0*def->bcount)/def->tcount > 0.1;
universe@23 64 }
universe@23 65 break;
universe@23 66 default: /* BFILE_MEDIUM_ACCURACY */
universe@23 67 if (def->tcount > 100 || next_char == EOF) {
universe@23 68 ret = (1.0*def->bcount)/def->tcount > 0.1;
universe@23 69 }
universe@23 70 break;
universe@22 71 }
universe@22 72 }
universe@22 73 }
universe@21 74
universe@21 75 return ret;
universe@21 76 }

mercurial