extract hist vec, mark interface ops, code ready for processor interface

src/tree/hist/evaluate_splits.h

@@ -357,7 +357,7 @@ class HistEvaluator {
     // based on global histogram. Other parties will receive the final best split information
     // Hence the below computation is not performed by the passive parties for secure vertical
     // All other cases need it
-    if (! ((is_col_split_) && (is_secure_) && (collective::GetRank() == 0))) {
+    if (!((is_col_split_) && (is_secure_) && (collective::GetRank() == 0))) {
       // Evaluate the splits for each feature
       common::ParallelFor2d(space, n_threads, [&](size_t nidx_in_set, common::Range1d r) {
         auto tidx = omp_get_thread_num();

src/tree/hist/histogram.h

@@ -194,13 +194,27 @@ class HistogramBuilder {
         collective::Allreduce<collective::Operation::kSum>(
         reinterpret_cast<double *>(this->hist_[first_nidx].data()), n);
       } else {
-        // Secure mode, we need to call interface to perform AllReduce
+        // Secure mode, we need to call interface to perform encryption and decryption
         // for simplicity, use AllGather instead of AllReduce
-        // note that aggregation will be performed at server side
+        // note that the actual aggregation will be performed at server side
         auto first_nidx = nodes_to_build.front();
         std::size_t n = n_total_bins * nodes_to_build.size() * 2;
-        collective::Allreduce<collective::Operation::kSum>(
-        reinterpret_cast<double *>(this->hist_[first_nidx].data()), n);
+        auto hist_to_aggr = std::vector<double>();
+        for (int hist_idx = 0; hist_idx < n; hist_idx++) {
+          double hist_item = reinterpret_cast<double *>(this->hist_[first_nidx].data())[hist_idx];
+          hist_to_aggr.push_back(hist_item);
+        }
+
+        // For FL with secure horizontal
+        //auto hist_buf_ckks = processor_instance->EncodeHist(hist_to_aggr);
+        //auto hist_entries = collective::AllgatherV(hist_buf_ckks);
+        //std::vector<double> hist_aggr = processor_instance->DecodeHist(hist_entries.data(), n);
+
+        // Assign the aggregated histogram back to the local histogram
+        collective::Allreduce<collective::Operation::kSum>(hist_to_aggr.data(), n);
+        for (int hist_idx = 0; hist_idx < n; hist_idx++) {
+          reinterpret_cast<double *>(this->hist_[first_nidx].data())[hist_idx] = hist_to_aggr[hist_idx];
+        }
       }
     }