A few optimizations.

Browse Source
This commit is contained in:
retoor 2024-12-01 19:40:01 +01:00
parent 3ae1979f26
commit 6be2d372f2
2 changed files with 16 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

BIN
isspam
View File

Binary file not shown.

41
retoor_c/isspam.c
View File

@ -174,22 +174,14 @@ bool is_fully_capitalized_word(char *word) {
return true; return true;
} }
sl *get_capitalized_words(char *content) { sl *get_capitalized_words(sl *all_words) {
sl *capitalized_words = sln(); sl *capitalized_words = sln();
sl *sentences = get_sentences(content);
for (uint j = 0; j < sentences->count; j++) {
char *sentence = sentences->strings[j];
sl *all_words = get_words(sentence);
// Always skip the first word since sentences start with
for (uint i = 0; i < all_words->count; i++) { for (uint i = 0; i < all_words->count; i++) {
if (is_fully_capitalized_word(all_words->strings[i])) { if (is_fully_capitalized_word(all_words->strings[i])) {
rstring_list_add(capitalized_words, all_words->strings[i]); rstring_list_add(capitalized_words, all_words->strings[i]);
} }
} }
slf(all_words);
}
slf(sentences);
return capitalized_words; return capitalized_words;
} }
@ -251,16 +243,13 @@ bool containswordi(sl *words, char *word) {
return false; return false;
} }
sl *get_forbidden_words(char *content) { sl *get_forbidden_words(sl *words) {
sl *words = get_words(content);
sl *found = sln(); sl *found = sln();
for (int j = 0; forbidden_words[j] != NULL; j++) { for (int j = 0; forbidden_words[j] != NULL; j++) {
if (containswordi(words, forbidden_words[j])) { if (containswordi(words, forbidden_words[j])) {
rstring_list_add(found, forbidden_words[j]); rstring_list_add(found, forbidden_words[j]);
} }
} }
slf(words);
return found; return found;
} }
unsigned int total = 0; unsigned int total = 0;
@ -278,8 +267,18 @@ void analyze(FILE *f) {
free(clean_data); free(clean_data);
sl *words = get_words(data);
// All words
printf("Words: %llu\n", words->count);
if(show_words)
sld(words);
sbuf = slds(words);
stra(all, sbuf);
free(sbuf);
// All capitalized words // All capitalized words
sl *capitalized_words = get_capitalized_words(data); sl *capitalized_words = get_capitalized_words(words);
ulonglong capitalized_words_count = capitalized_words->count; ulonglong capitalized_words_count = capitalized_words->count;
printf("Capitalized words: %llu\n", capitalized_words_count); printf("Capitalized words: %llu\n", capitalized_words_count);
if(show_capitalized) if(show_capitalized)
@ -298,15 +297,7 @@ void analyze(FILE *f) {
stra(all, sbuf); stra(all, sbuf);
free(sbuf); free(sbuf);
sl *words = get_words(data);
// All words
printf("Words: %llu\n", words->count);
if(show_words)
sld(words);
sbuf = slds(words);
stra(all, sbuf);
free(sbuf);
// Numbers // Numbers
sl *numbers = get_numbers(data); sl *numbers = get_numbers(data);
@ -318,7 +309,7 @@ void analyze(FILE *f) {
free(sbuf); free(sbuf);
// Forbidden words // Forbidden words
sl *fw = get_forbidden_words(data); sl *fw = get_forbidden_words(words);
printf("Forbidden words: %llu\n", fw->count); printf("Forbidden words: %llu\n", fw->count);
if(show_forbidden_words) if(show_forbidden_words)
sld(fw); sld(fw);