improve suspicious heuristic (closes #468)

medialab · Dec 6, 2022 · e525b5e · e525b5e
1 parent 9adc373
commit e525b5e
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/hyphe_frontend/app/js/service_utils.js b/hyphe_frontend/app/js/service_utils.js
@@ -665,10 +665,10 @@ angular.module('hyphe.service_utils', [])
       } else if(job.crawling_status != 'FINISHED'){
         job.globalStatus = job.crawling_status
       } else if(job.indexing_status == 'FINISHED'){
-        if(job.nb_crawled_pages_200 > 2){
+        if(job.nb_crawled_pages_200 > 10){
           job.globalStatus = 'ACHIEVED'
         } else if(job.nb_crawled_pages_200 > 0) {
-          if (job.crawl_arguments.max_depth > 0) {
+          if (job.crawl_arguments.max_depth > 0 && job.nb_links / job.nb_crawled_pages_200 < 10) {
             job.globalStatus = 'SUSPICIOUS'
           } else {
             job.globalStatus = 'ACHIEVED'