bump libnDPI to 8ea0eaa0d0c4a3be05f67ef7fa1d22c2579cf7d1

 * added build fix for Gitlab CI
 * added friendly C11 check
 * set required libnDPI versionto 4.7
   (ArchLinux ndpi-git sets version to 4.7, which is not released yet)
 * reduced sklearn-random-forest memory consumption by adjusting min. sample leaf

Signed-off-by: Toni Uhlig <matzeton@googlemail.com>

3 files changed, 35 insertions, 3 deletions

diff --git a/examples/c-collectd/c-collectd.c b/examples/c-collectd/c-collectd.c
index 1d470d4ef..624672ef3 100644
--- a/examples/c-collectd/c-collectd.c
+++ b/examples/c-collectd/c-collectd.c
@@ -114,6 +114,11 @@ static struct
     uint64_t flow_category_mining_count;
     uint64_t flow_category_malware_count;
     uint64_t flow_category_advertisment_count;
+    uint64_t flow_category_banned_site_count;
+    uint64_t flow_category_site_unavail_count;
+    uint64_t flow_category_allowed_site_count;
+    uint64_t flow_category_antimalware_count;
+    uint64_t flow_category_crypto_currency_count;
     uint64_t flow_category_unknown_count;
 
     uint64_t flow_l3_ip4_count;
@@ -193,6 +198,11 @@ static struct json_stat_map const categories_map[] = {
     {"Mining", &collectd_statistics.flow_category_mining_count},
     {"Malware", &collectd_statistics.flow_category_malware_count},
     {"Advertisement", &collectd_statistics.flow_category_advertisment_count},
+    {"Banned_Site", &collectd_statistics.flow_category_banned_site_count},
+    {"Site_Unavailable", &collectd_statistics.flow_category_site_unavail_count},
+    {"Allowed_Site", &collectd_statistics.flow_category_allowed_site_count},
+    {"Antimalware", &collectd_statistics.flow_category_antimalware_count},
+    {"Crypto_Currency", &collectd_statistics.flow_category_crypto_currency_count},
     {NULL, &collectd_statistics.flow_category_unknown_count}};
 
 #ifdef ENABLE_MEMORY_PROFILING
@@ -431,7 +441,9 @@ static void print_collectd_exec_output(void)
                                    COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT()
                                        COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT()
                                            COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT()
-                                               COLLECTD_PUTVAL_N_FORMAT(),
+                                               COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT()
+                                                   COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT()
+                                                       COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT(),
 
            COLLECTD_PUTVAL_N(flow_category_media_count),
            COLLECTD_PUTVAL_N(flow_category_vpn_count),
@@ -460,6 +472,11 @@ static void print_collectd_exec_output(void)
            COLLECTD_PUTVAL_N(flow_category_mining_count),
            COLLECTD_PUTVAL_N(flow_category_malware_count),
            COLLECTD_PUTVAL_N(flow_category_advertisment_count),
+           COLLECTD_PUTVAL_N(flow_category_banned_site_count),
+           COLLECTD_PUTVAL_N(flow_category_site_unavail_count),
+           COLLECTD_PUTVAL_N(flow_category_allowed_site_count),
+           COLLECTD_PUTVAL_N(flow_category_antimalware_count),
+           COLLECTD_PUTVAL_N(flow_category_crypto_currency_count),
            COLLECTD_PUTVAL_N(flow_category_unknown_count));
 
     printf(COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT() COLLECTD_PUTVAL_N_FORMAT()
diff --git a/examples/c-collectd/rrdgraph.sh b/examples/c-collectd/rrdgraph.sh
index 6b0a445b5..022ea5ebc 100755
--- a/examples/c-collectd/rrdgraph.sh
+++ b/examples/c-collectd/rrdgraph.sh
@@ -200,6 +200,11 @@ rrdtool_graph Flow-Categories 'Amount(SUM)' "${OUTDIR}/categories" \
 	DEF:cat_voip=${RRDDIR}/gauge-flow_category_voip_count.rrd:value:AVERAGE \
 	DEF:cat_vpn=${RRDDIR}/gauge-flow_category_vpn_count.rrd:value:AVERAGE \
 	DEF:cat_web=${RRDDIR}/gauge-flow_category_web_count.rrd:value:AVERAGE \
+	DEF:cat_banned=${RRDDIR}/gauge-flow_category_web_count.rrd:value:AVERAGE \
+	DEF:cat_unavail=${RRDDIR}/gauge-flow_category_web_count.rrd:value:AVERAGE \
+	DEF:cat_allowed=${RRDDIR}/gauge-flow_category_web_count.rrd:value:AVERAGE \
+	DEF:cat_antimal=${RRDDIR}/gauge-flow_category_web_count.rrd:value:AVERAGE \
+	DEF:cat_crypto=${RRDDIR}/gauge-flow_category_web_count.rrd:value:AVERAGE \
 	$(rrdtool_graph_colorize_missing_data cat_ads) \
 	AREA:cat_ads#f1c232:"Advertisment..........." \
 	$(rrdtool_graph_print_cur_min_max_avg cat_ads) \
@@ -256,7 +261,17 @@ rrdtool_graph Flow-Categories 'Amount(SUM)' "${OUTDIR}/categories" \
 	STACK:cat_vpn#378035:"Virtual-Private-Network" \
 	$(rrdtool_graph_print_cur_min_max_avg cat_vpn) \
 	STACK:cat_web#00fffb:"Web...................." \
-	$(rrdtool_graph_print_cur_min_max_avg cat_web)
+	$(rrdtool_graph_print_cur_min_max_avg cat_web) \
+	STACK:cat_banned#ff1010:"Banned Sites..........." \
+	$(rrdtool_graph_print_cur_min_max_avg cat_banned) \
+	STACK:cat_unavail#ff1010:"Sites Unavailable......" \
+	$(rrdtool_graph_print_cur_min_max_avg cat_unavail) \
+	STACK:cat_allowed#ff1010:"Allowed Sites.........." \
+	$(rrdtool_graph_print_cur_min_max_avg cat_allowed) \
+	STACK:cat_antimal#ff1010:"Antimalware............" \
+	$(rrdtool_graph_print_cur_min_max_avg cat_antimal) \
+	STACK:cat_crypto#afaf00:"Crypto Currency........" \
+	$(rrdtool_graph_print_cur_min_max_avg cat_crypto)
 rrdtool_graph JSON 'Lines' "${OUTDIR}/json_lines" \
 	DEF:json_lines=${RRDDIR}/gauge-json_lines.rrd:value:AVERAGE \
 	$(rrdtool_graph_colorize_missing_data json_lines) \
diff --git a/examples/py-machine-learning/sklearn-random-forest.py b/examples/py-machine-learning/sklearn-random-forest.py
index 275376ba2..9716f2b83 100755
--- a/examples/py-machine-learning/sklearn-random-forest.py
+++ b/examples/py-machine-learning/sklearn-random-forest.py
@@ -206,7 +206,7 @@ if __name__ == '__main__':
                            help='Number of sklearn processes during training.')
     argparser.add_argument('--sklearn-estimators', action='store', type=int, default=1000,
                            help='Number of trees in the forest.')
-    argparser.add_argument('--sklearn-min-samples-leaf', action='store', type=int, default=5,
+    argparser.add_argument('--sklearn-min-samples-leaf', action='store', type=int, default=0.0001,
                            help='The minimum number of samples required to be at a leaf node.')
     argparser.add_argument('--sklearn-class-weight', default='balanced', const='balanced', nargs='?',
                            choices=['balanced', 'balanced_subsample'],