* Add learning interface to rspamd (still requires classifier to work)

3 files changed, 35 insertions, 2 deletions

diff --git a/src/tokenizers/osb.c b/src/tokenizers/osb.c
index f78e20992..6799a121b 100644
--- a/src/tokenizers/osb.c
+++ b/src/tokenizers/osb.c
@@ -21,7 +21,7 @@ static const int primes[] = {
 };
 
 token_list_t *
-osb_tokenize_text (memory_pool_t *pool, f_str_t *input)
+osb_tokenize_text (struct tokenizer *tokenizer, memory_pool_t *pool, f_str_t *input)
 {
 	token_list_t *new = NULL, *head = NULL, *last = NULL;
 	f_str_t token = { NULL, 0, 0 };
@@ -33,7 +33,7 @@ osb_tokenize_text (memory_pool_t *pool, f_str_t *input)
 		hashpipe[i] = 0xABCDEF;
 	}
 
-	while (get_next_word (input, &token)) {
+	while (tokenizer->get_next_word (input, &token)) {
 		/* Shift hashpipe */
 		for (i = FEATURE_WINDOW_SIZE - 1; i > 0; i --) {
 			hashpipe[i] = hashpipe[i - 1];
diff --git a/src/tokenizers/tokenizers.c b/src/tokenizers/tokenizers.c
index 132a57ce0..25b13a289 100644
--- a/src/tokenizers/tokenizers.c
+++ b/src/tokenizers/tokenizers.c
@@ -5,6 +5,24 @@
 #include <sys/types.h>
 #include "tokenizers.h"
 
+struct tokenizer tokenizers[] = {
+	{"osb-text", osb_tokenize_text, get_next_word },
+};
+
+struct tokenizer*
+get_tokenizer (char *name)
+{
+	int i;
+
+	for (i = 0; i < sizeof (tokenizers) / sizeof (tokenizers[0]); i ++) {
+		if (strcmp (tokenizers[i].name, name) == 0) {
+			return &tokenizers[i];
+		}
+	}
+
+	return NULL;
+}
+
 /* Get next word from specified f_str_t buf */
 f_str_t *
 get_next_word (f_str_t *buf, f_str_t *token)
diff --git a/src/tokenizers/tokenizers.h b/src/tokenizers/tokenizers.h
index 6b4bff5e0..96a2027a5 100644
--- a/src/tokenizers/tokenizers.h
+++ b/src/tokenizers/tokenizers.h
@@ -20,8 +20,23 @@ typedef struct token_list_s {
 	struct token_list_s *next;
 } token_list_t;
 
+
+/* Common tokenizer structure */
+struct tokenizer {
+	char *name;
+	token_list_t* (*tokenize_func)(struct tokenizer *tokenizer, memory_pool_t *pool, f_str_t *input);
+	f_str_t* (*get_next_word)(f_str_t *buf, f_str_t *token);
+};
+
+/* Get tokenizer structure by name or return NULL if this name is not found */
+struct tokenizer* get_tokenizer (char *name);
 /* Get next word from specified f_str_t buf */
 f_str_t *get_next_word (f_str_t *buf, f_str_t *token);
+/* OSB tokenize function */
+token_list_t* osb_tokenize_text (struct tokenizer *tokenizer, memory_pool_t *pool, f_str_t *input);
+
+/* Array of all defined tokenizers */
+extern struct tokenizer tokenizers[];
 
 #endif
 /*