[Feature] Try to avoid FP for low contrast fonts detection

diff --git a/rules/html.lua b/rules/html.lua
index 955d5a42e79442309eee30a3d89445f4d435d53f..1c4c86cdc0f0665956b764feff0343fceeff7e62 100644 (file)
--- a/rules/html.lua
+++ b/rules/html.lua
@@ -170,6 +170,8 @@ rspamd_config.R_WHITE_ON_WHITE = {
     local tp = task:get_text_parts() -- get text parts in a message
     local ret = false
     local diff = 0.0
+    local normal_len = 0
+    local transp_tags = {}
 
     for _,p in ipairs(tp) do -- iterate over text parts array using `ipairs`
       if p:is_html() then -- if the current part is html part
@@ -189,7 +191,9 @@ rspamd_config.R_WHITE_ON_WHITE = {
 
               if diff < 0.2 then
                 ret = true
-                return true
+                table.insert(transp_tags, tag)
+              else
+                normal_len = normal_len + tag:get_content_length()
               end
             end
           end
@@ -201,7 +205,27 @@ rspamd_config.R_WHITE_ON_WHITE = {
     end
 
     if ret then
-      return true,(0.2 - diff) * 5.0,tostring(diff * 100.0)
+      local transp_len = 0
+      local arg
+
+      for _,t in ipairs(transp_tags) do
+        local bl = t:get_extra()
+        local color = bl['color']
+        local bgcolor = bl['bgcolor']
+        transp_len = transp_len + t:get_content_length()
+
+        if not arg then
+          arg = string.format('%s color #%x%x%x bgcolor #%x%x%x',
+            tostring(t:get_type()),
+            color[1], color[2], color[3],
+            bgcolor[1], bgcolor[2], bgcolor[3])
+        end
+      end
+
+      local transp_rate = transp_len / (normal_len + transp_len)
+      if transp_rate > 0.1 then
+        return true,(transp_rate * 2.0),arg
+      end
     end
 
     return false