From bd0e1a3d7b719c28a5a7e7727de912cb187240e4 Mon Sep 17 00:00:00 2001
From: loyalsoldier <10487845+Loyalsoldier@users.noreply.github.com>
Date: Mon, 2 Mar 2020 01:05:01 +0800
Subject: [PATCH] Remove redundant domains

---
 .github/workflows/build.yml | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)

diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 0e1f68f957..8aa36745a2 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -93,15 +93,28 @@ jobs:
           cat proxy.txt >> temp-proxy.txt
           cat direct.txt >> temp-direct.txt
 
-      - name: Sort and generate lists
+      - name: Sort and generate redundant lists
         run: |
-          cat temp-proxy.txt | sort --ignore-case -u | perl -ne '/^((?=^.{3,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})+)/ && print "$1\n"' > $GOPATH/src/$GEOSITE_REPO/data/proxy-list
+          cat temp-proxy.txt | sort --ignore-case -u | perl -ne '/^((?=^.{3,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})+)/ && print "$1\n"' > proxy-list-with-redundant
           cat temp-proxy.txt | sort --ignore-case -u | perl -ne 'print if not /^((?=^.{3,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})+)/' > proxy-excluse-list
-          cat temp-direct.txt | sort --ignore-case -u | perl -ne '/^((?=^.{3,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})+)/ && print "$1\n"' > $GOPATH/src/$GEOSITE_REPO/data/direct-list
+          cat temp-direct.txt | sort --ignore-case -u | perl -ne '/^((?=^.{3,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})+)/ && print "$1\n"' > direct-list-with-redundant
           cat temp-direct.txt | sort --ignore-case -u | perl -ne 'print if not /^((?=^.{3,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})+)/' > direct-excluse-list
-          cat temp-reject.txt | sort --ignore-case -u | perl -ne '/^((?=^.{3,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})+)/ && print "$1\n"' > $GOPATH/src/$GEOSITE_REPO/data/reject-list
+          cat temp-reject.txt | sort --ignore-case -u | perl -ne '/^((?=^.{3,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})+)/ && print "$1\n"' > reject-list-with-redundant
           cat temp-reject.txt | sort --ignore-case -u | perl -ne 'print if not /^((?=^.{3,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})+)/' > reject-excluse-list
 
+      - name: Remove redundant domains
+        run: |
+          chmod +x findRedundantDomain.py
+          ./findRedundantDomain.py ./direct-list-with-redundant ./direct-list-deleted-unsort
+          ./findRedundantDomain.py ./proxy-list-with-redundant ./proxy-list-deleted-unsort
+          ./findRedundantDomain.py ./reject-list-with-redundant ./reject-list-deleted-unsort
+          sort ./direct-list-deleted-unsort > ./direct-list-deleted-sort
+          sort ./proxy-list-deleted-unsort > ./proxy-list-deleted-sort
+          sort ./reject-list-deleted-unsort > ./reject-list-deleted-sort
+          diff ./direct-list-deleted-sort ./direct-list-with-redundant | awk '/^>/{print $2}' > $GOPATH/src/$GEOSITE_REPO/data/direct-list
+          diff ./proxy-list-deleted-sort ./proxy-list-with-redundant | awk '/^>/{print $2}' > $GOPATH/src/$GEOSITE_REPO/data/proxy-list
+          diff ./reject-list-deleted-sort ./reject-list-with-redundant | awk '/^>/{print $2}' > $GOPATH/src/$GEOSITE_REPO/data/reject-list
+
       - name: Add list into appropriate category file
         run: |
           cd $GOPATH/src/$GEOSITE_REPO/data