[Project] Neural: Use C version of scatter matrix producing

diff --git a/src/plugins/lua/neural.lua b/src/plugins/lua/neural.lua
index 10e49901fe890de233475be3d3139e4616ec0eb8..5d2e9bfd555f0df7382d7ca0c9abd6a5e6c081f0 100644 (file)
--- a/src/plugins/lua/neural.lua
+++ b/src/plugins/lua/neural.lua
@@ -641,40 +641,10 @@ local function register_lock_extender(rule, set, ev_base, ann_key)
   )
 end
 
--- This is an utility function for PCA training
-local function fill_scatter(inputs)
-  local scatter_matrix = rspamd_tensor.new(2, #inputs[1], #inputs[1])
-  local nsamples = #inputs
-
-  -- Convert to a tensor where each row is an input dimension
-  inputs = rspamd_tensor.fromtable(inputs):transpose()
-
-  local meanv = inputs:mean()
-  lua_util.debugm(N, 'means: %s', meanv)
-
-  for i=1,nsamples do
-    local col = rspamd_tensor.new(1, #inputs)
-    for j=1,#inputs do
-      local x = inputs[j][i] - meanv[j]
-      col[j] = x
-    end
-    local prod = col:mul(col, false, true)
-    for ii=1,#prod do
-      for jj=1,#prod[1] do
-        scatter_matrix[ii][jj] = scatter_matrix[ii][jj] + prod[ii][jj]
-      end
-    end
-  end
-
-  lua_util.debugm(N, 'scatter matrix: %s', scatter_matrix)
-
-  return scatter_matrix
-end
-
 -- This function takes all inputs, applies PCA transformation and returns the final
 -- PCA matrix as rspamd_tensor
 local function learn_pca(inputs, max_inputs)
-  local scatter_matrix = fill_scatter(inputs)
+  local scatter_matrix = rspamd_tensor.scatter_matrix(rspamd_tensor.fromtable(inputs))
   local eigenvals = scatter_matrix:eigen()
   -- scatter matrix is not filled with eigenvectors
   lua_util.debugm(N, 'eigenvalues: %s', eigenvals)