From d658ac9e61a78122c9fe0cfba3797dead382ba3b Mon Sep 17 00:00:00 2001 From: Luca Deri Date: Thu, 9 Jul 2020 00:15:21 +0200 Subject: Added -b to ndpiReader to evaluate bins --- src/include/ndpi_api.h.in | 3 ++- src/lib/ndpi_analyze.c | 34 ++++++++++++++++++++-------------- 2 files changed, 22 insertions(+), 15 deletions(-) (limited to 'src') diff --git a/src/include/ndpi_api.h.in b/src/include/ndpi_api.h.in index 94f5f54fe..b21bcd3fe 100644 --- a/src/include/ndpi_api.h.in +++ b/src/include/ndpi_api.h.in @@ -1078,7 +1078,8 @@ extern "C" { char* ndpi_print_bin(struct ndpi_bin *b, u_int8_t normalize_first, char *out_buf, u_int out_buf_len); float ndpi_bin_similarity(struct ndpi_bin *b1, struct ndpi_bin *b2, u_int8_t normalize_first); int ndpi_cluster_bins(struct ndpi_bin *bins, u_int16_t num_bins, - u_int8_t num_clusters, u_int16_t *cluster_ids); + u_int8_t num_clusters, u_int16_t *cluster_ids, + struct ndpi_bin *centroids); #ifdef __cplusplus } diff --git a/src/lib/ndpi_analyze.c b/src/lib/ndpi_analyze.c index 8d4e939f0..8c21c0e0e 100644 --- a/src/lib/ndpi_analyze.c +++ b/src/lib/ndpi_analyze.c @@ -476,25 +476,29 @@ float ndpi_bin_similarity(struct ndpi_bin *b1, struct ndpi_bin *b2, u_int8_t nor - (in) bins: a vection 'num_bins' long of bins to cluster - (in) 'num_clusters': number of desired clusters 0...(num_clusters-1) - (out) 'cluster_ids': a vector 'num_bins' long containing the id's of each clustered bin - + - (out) 'centroids': an optional 'num_clusters' long vector of (centroid) bins See - https://en.wikipedia.org/wiki/K-means_clustering */ int ndpi_cluster_bins(struct ndpi_bin *bins, u_int16_t num_bins, - u_int8_t num_clusters, u_int16_t *cluster_ids) { + u_int8_t num_clusters, u_int16_t *cluster_ids, + struct ndpi_bin *centroids) { u_int16_t i, j, max_iterations = 100, num_iterations = 0, num_moves; - struct ndpi_bin *centroids; - u_int8_t verbose = 0; + u_int8_t verbose = 0, alloc_centroids = 0; if(num_clusters > num_bins) return(-1); - if((centroids = (struct ndpi_bin*)ndpi_malloc(sizeof(struct ndpi_bin)*num_clusters)) == NULL) - return(-2); - else { - for(i=0; i