meta-rules-dat/.github/workflows/run.yml

name: Build rules dat files
on:
  workflow_dispatch:
  schedule:
    - cron: "0 22 * * *"
  push:
    branches:
      - master
    paths-ignore:
      - "**/README.md"
jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - name: Setup Go 1.19.6
        uses: actions/setup-go@v3
        with:
          go-version: 1.19.6

      - name: Set variables
        run: |
          echo "RELEASE_NAME=Released on $(date +%Y%m%d%H%M)" >> $GITHUB_ENV
          echo "TAG_NAME=$(date +%Y%m%d%H%M)" >> $GITHUB_ENV
          echo "CHINA_DOMAINS_URL=https://raw.githubusercontent.com/blackmatrix7/ios_rule_script/master/rule/Clash/ChinaMax/ChinaMax_Domain.txt" >> $GITHUB_ENV
          echo "GOOGLE_DOMAINS_URL=https://raw.githubusercontent.com/felixonmars/dnsmasq-china-list/master/google.china.conf" >> $GITHUB_ENV
          echo "APPLE_DOMAINS_URL=https://raw.githubusercontent.com/felixonmars/dnsmasq-china-list/master/apple.china.conf" >> $GITHUB_ENV
          # echo "EASYLISTCHINA_EASYLIST_REJECT_URL=https://easylist-downloads.adblockplus.org/easylistchina+easylist.txt" >> $GITHUB_ENV
          # echo "EASYPRIVACY_REJECT_URL=https://easylist-downloads.adblockplus.org/easyprivacy.txt" >> $GITHUB_ENV
          # echo "PETERLOWE_REJECT_URL=https://pgl.yoyo.org/adservers/serverlist.php?hostformat=hosts&showintro=1&mimetype=plaintext" >> $GITHUB_ENV
          # echo "ADGUARD_DNS_REJECT_URL=https://adguardteam.github.io/AdGuardSDNSFilter/Filters/filter.txt" >> $GITHUB_ENV
          # echo "DANPOLLOCK_REJECT_URL=https://someonewhocares.org/hosts/hosts" >> $GITHUB_ENV
          # echo "ANTIAD_DOMAINS_URL=https://raw.githubusercontent.com/privacy-protection-tools/anti-AD/master/anti-ad-domains.txt" >> $GITHUB_ENV
          echo "CUSTOM_DIRECT=https://raw.githubusercontent.com/Loyalsoldier/domain-list-custom/release/cn.txt" >> $GITHUB_ENV
          echo "CUSTOM_PROXY=https://raw.githubusercontent.com/Loyalsoldier/domain-list-custom/release/geolocation-!cn.txt" >> $GITHUB_ENV
          echo "WIN_SPY=https://raw.githubusercontent.com/crazy-max/WindowsSpyBlocker/master/data/hosts/spy.txt" >> $GITHUB_ENV
          echo "WIN_UPDATE=https://raw.githubusercontent.com/crazy-max/WindowsSpyBlocker/master/data/hosts/update.txt" >> $GITHUB_ENV
          echo "WIN_EXTRA=https://raw.githubusercontent.com/crazy-max/WindowsSpyBlocker/master/data/hosts/extra.txt" >> $GITHUB_ENV
        shell: bash

      - name: Checkout the "hidden" branch of this repo
        uses: actions/checkout@v3
        with:
          ref: hidden

      - name: Checkout Loyalsoldier/domain-list-custom
        uses: actions/checkout@v3
        with:
          repository: Loyalsoldier/domain-list-custom
          path: custom

      - name: Checkout v2fly/domain-list-community
        uses: actions/checkout@v3
        with:
          repository: v2fly/domain-list-community
          path: community

      - name: Checkout cokebar/gfwlist2dnsmasq
        uses: actions/checkout@v3
        with:
          repository: cokebar/gfwlist2dnsmasq
          path: gfwlist2dnsmasq

      - name: Get geoip.dat relative files
        run: |
          wget -O geoip-lite.dat https://github.com/xishang0128/geoip/raw/release/geoip.dat
          wget https://github.com/Loyalsoldier/geoip/raw/release/geoip.dat

      - name: Generate GFWList domains
        run: |
          cd gfwlist2dnsmasq || exit 1
          chmod +x ./gfwlist2dnsmasq.sh
          ./gfwlist2dnsmasq.sh -l -o ./temp-gfwlist.txt

      - name: Get and add direct domains into temp-direct.txt file
        run: |
          curl -sSL ${CHINA_DOMAINS_URL} | sed '/^\s*#/d' | sed '/^[^\.]/ s/^/full: /' |  sed 's/^\.\([^.]*\)/\1/' > temp-direct.txt 
          curl -sSL ${CUSTOM_DIRECT} | perl -ne '/^(domain):([^:]+)(\n$|:@.+)/ && print "$2\n"' >> temp-direct.txt

      - name: Get and add proxy domains into temp-proxy.txt file
        run: |
          cat ./gfwlist2dnsmasq/temp-gfwlist.txt | 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"' > temp-proxy.txt
          curl -sSL $GOOGLE_DOMAINS_URL | perl -ne '/^server=\/([^\/]+)\// && print "$1\n"' >> temp-proxy.txt
          curl -sSL $APPLE_DOMAINS_URL | perl -ne '/^server=\/([^\/]+)\// && print "$1\n"' >> temp-proxy.txt
          curl -sSL ${CUSTOM_PROXY} | grep -Ev ":@cn" | perl -ne '/^(domain):([^:]+)(\n$|:@.+)/ && print "$2\n"' >> temp-proxy.txt

      # - name: Get and add reject domains into temp-reject.txt file
      #   run: |
      #     curl -sSL $EASYLISTCHINA_EASYLIST_REJECT_URL | perl -ne '/^\|\|([-_0-9a-zA-Z]+(\.[-_0-9a-zA-Z]+){1,64})\^$/ && print "$1\n"' | perl -ne 'print if not /^[0-9]{1,3}(\.[0-9]{1,3}){3}$/' > temp-reject.txt
      #     curl -sSL $EASYPRIVACY_REJECT_URL | perl -ne '/^\|\|([-_0-9a-zA-Z]+(\.[-_0-9a-zA-Z]+){1,64})\^$/ && print "$1\n"' | perl -ne 'print if not /^[0-9]{1,3}(\.[0-9]{1,3}){3}$/' >> temp-reject.txt
      #     curl -sSL $ADGUARD_DNS_REJECT_URL | perl -ne '/^\|\|([-_0-9a-zA-Z]+(\.[-_0-9a-zA-Z]+){1,64})\^$/ && print "$1\n"' | perl -ne 'print if not /^[0-9]{1,3}(\.[0-9]{1,3}){3}$/' >> temp-reject.txt
      #     curl -sSL $ANTIAD_DOMAINS_URL | 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"' >> temp-reject.txt
      #     curl -sSL $PETERLOWE_REJECT_URL | perl -ne '/^127\.0\.0\.1\s([-_0-9a-zA-Z]+(\.[-_0-9a-zA-Z]+){1,64})$/ && print "$1\n"' >> temp-reject.txt
      #     curl -sSL $DANPOLLOCK_REJECT_URL | perl -ne '/^127\.0\.0\.1\s([-_0-9a-zA-Z]+(\.[-_0-9a-zA-Z]+){1,64})/ && print "$1\n"' | sed '1d' >> temp-reject.txt

      - name: Reserve `full`, `regexp` and `keyword` type of rules from custom lists to "reserve" files
        run: |
          curl -sSL ${CUSTOM_DIRECT} | perl -ne '/^((full|regexp|keyword):[^:]+)(\n$|:@.+)/ && print "$1\n"' | sort --ignore-case -u > direct-reserve.txt
          curl -sSL ${CUSTOM_PROXY} | grep -Ev ":@cn" | perl -ne '/^((full|regexp|keyword):[^:]+)(\n$|:@.+)/ && print "$1\n"' | sort --ignore-case -u > proxy-reserve.txt

      - name: Add proxy, direct and reject domains from "hidden" branch to appropriate temp files
        run: |
          cat proxy.txt >> temp-proxy.txt
          cat direct.txt >> temp-direct.txt
          # cat reject.txt >> temp-reject.txt

      - name: Sort and generate redundant lists
        run: |
          cat temp-proxy.txt | sort --ignore-case -u > proxy-list-with-redundant
          cat temp-direct.txt | sort --ignore-case -u > direct-list-with-redundant
          # cat temp-reject.txt | sort --ignore-case -u > reject-list-with-redundant

      - 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
          [ ! -f "direct-list-deleted-unsort" ] && touch direct-list-deleted-unsort
          [ ! -f "proxy-list-deleted-unsort" ] && touch proxy-list-deleted-unsort
          # [ ! -f "reject-list-deleted-unsort" ] && touch 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}' > ./direct-list-without-redundant
          diff ./proxy-list-deleted-sort ./proxy-list-with-redundant | awk '/^>/{print $2}' > ./proxy-list-without-redundant
          # diff ./reject-list-deleted-sort ./reject-list-with-redundant | awk '/^>/{print $2}' > ./reject-list-without-redundant

      - name: Remove domains from "need-to-remove" lists in "hidden" branch
        run: |
          diff ./direct-need-to-remove.txt ./direct-list-without-redundant | awk '/^>/{print $2}' > temp-cn.txt
          diff ./proxy-need-to-remove.txt ./proxy-list-without-redundant | awk '/^>/{print $2}' > temp-geolocation-\!cn.txt
          # diff ./reject-need-to-remove.txt ./reject-list-without-redundant | awk '/^>/{print $2}' > temp-category-ads-all.txt

      - name: Remove domains end with ".cn" in "temp-geolocation-!cn.txt" and write lists to data directory
        run: |
          cat temp-cn.txt | sort --ignore-case -u | perl -ne '/^((?=^.{1,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})*)/ && print "$1\n"' > ./community/data/cn
          cat temp-cn.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-tld-list.txt
          cat temp-geolocation-\!cn.txt | sort --ignore-case -u | perl -ne '/^((?=^.{1,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})*)/ && print "$1\n"' | perl -ne 'print if not /\.cn$/' > ./community/data/geolocation-\!cn
          cat temp-geolocation-\!cn.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-tld-list.txt
          # cat temp-category-ads-all.txt | sort --ignore-case -u | perl -ne '/^((?=^.{1,255})[a-zA-Z0-9][-_a-zA-Z0-9]{0,62}(\.[a-zA-Z0-9][-_a-zA-Z0-9]{0,62})*)/ && print "$1\n"' > ./community/data/category-ads-all
          # cat temp-category-ads-all.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-tld-list.txt

      - name: Add `full`, `regexp` and `keyword` type of rules back into "cn", "geolocation-!cn" and "category-ads-all" list
        run: |
          [ -f "direct-reserve.txt" ] && cat direct-reserve.txt >> ./community/data/cn
          [ -f "proxy-reserve.txt" ] && cat proxy-reserve.txt >> ./community/data/geolocation-\!cn
          # [ -f "reject-reserve.txt" ] && cat reject-reserve.txt >> ./community/data/category-ads-all
          cp ./community/data/cn direct-list.txt
          cp ./community/data/geolocation-\!cn proxy-list.txt
          # cp ./community/data/category-ads-all reject-list.txt
          # awk '{print "0.0.0.0 " $0}' reject-list.txt > hosts

      - name: Create `google-cn`、`apple-cn`、`gfw` lists
        run: |
          curl -sSL $GOOGLE_DOMAINS_URL | perl -ne '/^server=\/([^\/]+)\// && print "full:$1\n"' > ./community/data/google-cn
          curl -sSL $APPLE_DOMAINS_URL | perl -ne '/^server=\/([^\/]+)\// && print "full:$1\n"' > ./community/data/apple-cn
          cat ./gfwlist2dnsmasq/temp-gfwlist.txt | 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"' >> ./community/data/gfw
          curl -sSL $WIN_SPY | grep "0.0.0.0" | awk '{print $2}' > ./community/data/win-spy
          curl -sSL $WIN_UPDATE | grep "0.0.0.0" | awk '{print $2}' > ./community/data/win-update
          curl -sSL $WIN_EXTRA | grep "0.0.0.0" | awk '{print $2}' > ./community/data/win-extra

      - name: merge 'ios_rule_script/SteamCN'
        env:
          SED: sed '/^\s*#/d' | sed 's/DOMAIN,//g' | sed 's/DOMAIN-SUFFIX,//g' | sed 's/DOMAIN-KEYWORD,/keyword:/g'
        run: |
          curl -sSL https://raw.githubusercontent.com/blackmatrix7/ios_rule_script/master/rule/Clash/SteamCN/SteamCN.list | ${{ env.SED }} > steamcn.txt
          while read line; do grep -q "$line @cn" ./community/data/steam || sed -i "/$line/ s/$/ @cn/" ./community/data/steam; done < steamcn.txt
          cat ./community/data/steam

      - name: Add `meta` data
        env:
          SED: sed '/^\s*#/d' | sed 's/^PROCESS-NAME,//' | sed 's/DOMAIN,/full:/g' | sed 's/DOMAIN-SUFFIX,//g' | sed 's/DOMAIN-KEYWORD,/keyword:/g'
        run: |
          curl -sSL https://raw.githubusercontent.com/xishang0128/rules/main/biliintl.list > ./community/data/biliintl
          curl -sSL https://raw.githubusercontent.com/blackmatrix7/ios_rule_script/master/rule/Clash/OneDrive/OneDrive.list | ${{ env.SED }} > ./community/data/onedrive
          echo "sharepoint.cn" >> ./community/data/onedrive && sort ./community/data/onedrive | uniq -i > tmp.txt && mv tmp.txt ./community/data/onedrive
          curl -sSL https://raw.githubusercontent.com/xishang0128/rules/main/sharepoint.list > ./community/data/sharepoint
          curl -sSL https://raw.githubusercontent.com/blackmatrix7/ios_rule_script/master/rule/Clash/PrivateTracker/PrivateTracker.list | ${{ env.SED }} > ./community/data/tracker1
          curl -sSL https://gitea.com/XIU2/TrackersListCollection/raw/branch/master/all.txt | grep -i "\.[A-Z]" | grep -v tracker | sed 's/^.*\/\///g' | sed 's/:.*\/.*//g' >> ./community/data/tracker1
          sort ./community/data/tracker1 | uniq -i > ./community/data/tracker
          rm ./community/data/tracker1
          
      - name: Build geosite.dat file
        run: |
          cd custom || exit 1
          go run ./ --datapath=../community/data



      - name: Build geosite-lite.dat file
        env:
          NO_SKIP: true
          SED: sed '/^\s*#/d' | sed 's/^PROCESS-NAME,//' | sed 's/DOMAIN,/full:/g' | sed 's/DOMAIN-SUFFIX,//g' | sed 's/DOMAIN-KEYWORD,/keyword:/g'
        run: |
          cd community || exit 1
          mkdir -p data-lite
          curl -sSL https://raw.githubusercontent.com/xishang0128/rules/main/biliintl.list > ./data-lite/biliintl
          curl -sSL https://raw.githubusercontent.com/blackmatrix7/ios_rule_script/master/rule/Clash/ChinaMax/ChinaMax_Domain.txt  | sed '/^\s*#/d' | sed '/^[^\.]/ s/^/full:/' |  sed 's/^\.\([^.]*\)/\1/' > cn
          curl -sSL https://raw.githubusercontent.com/Loyalsoldier/domain-list-custom/release/cn.txt | perl -ne '/^(domain):([^:]+)(\n$|:@.+)/ && print "$2\n"' >> cn
          curl -sSL https://raw.githubusercontent.com/Loyalsoldier/domain-list-custom/release/cn.txt | perl -ne '/^((full|regexp|keyword):[^:]+)(\n$|:@.+)/ && print "$1\n"' | sort --ignore-case -u >> cn
          sort ./cn | uniq -i > data-lite/cn
          go run ./ --datapath=./data-lite --outputname geosite-lite.dat




      - name: Build db file
        env:
          NO_SKIP: true
        run: |
          mkdir -p v2box
          cd v2box || exit 1
          wget https://github.com/xishang0128/meta-rules-dat/releases/download/v2box/v2box
          chmod 755 v2box
          ./v2box migrate geosite -i ../custom/publish/geosite.dat -o ./geosite.db
          ./v2box migrate geosite -i ../community/geosite-lite.dat -o ./geosite-lite.db
          ./v2box migrate geoip -i ../geoip.dat -o ./geoip.db
          ./v2box migrate geoip -i ../geoip-lite.dat -o ./geoip-lite.db



      - name: Move and zip files
        run: |
          mkdir -p ./publish/
          wget https://raw.githubusercontent.com/xishang0128/geoip/release/Country.mmdb -O ./publish/country-lite.mmdb
          wget https://raw.githubusercontent.com/Loyalsoldier/geoip/release/Country.mmdb -O ./publish/country.mmdb
          install -Dp ./geoip-lite.dat ./publish/
          install -Dp ./geoip.dat ./publish/
          install -Dp ./custom/publish/geosite.dat ./publish/
          install -Dp ./community/geosite-lite.dat ./publish/

          install -Dp ./v2box/geosite.db ./publish/
          install -Dp ./v2box/geosite-lite.db ./publish/
          install -Dp ./v2box/geoip.db ./publish/
          install -Dp ./v2box/geoip-lite.db ./publish/
          curl -sSL https://raw.githubusercontent.com/blackmatrix7/ios_rule_script/master/rule/Clash/ChinaMax/ChinaMax_Domain.yaml | sed '/^\s*#/d' > ./publish/cn_domain.yaml
          curl -sSL https://raw.githubusercontent.com/blackmatrix7/ios_rule_script/master/rule/Clash/Proxy/Proxy_Domain.yaml | sed '/^\s*#/d' > ./publish/proxy.yaml
          cd ./publish || exit 1


      - name: Delete current release assets
        uses: andreaswilli/delete-release-assets-action@v2.0.0
        with:
          github_token: ${{ secrets.GITHUB_TOKEN }}
          tag: latest
          deleteOnlyFromDrafts: false


      - name: Release and upload assets
        uses: softprops/action-gh-release@v0.1.6
        with:
          name: latest
          tag_name: latest
          draft: false
          prerelease: false
          files: |
            ./publish/*
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

      - name: Git push assets to "release" branch
        run: |
          cd publish || exit 1
          git init
          git config --local user.name "github-actions[bot]"
          git config --local user.email "41898282+github-actions[bot]@users.noreply.github.com"
          git checkout -b release
          git add .
          git commit -m "${{ env.RELEASE_NAME }}"
          git remote add origin "https://${{ github.actor }}:${{ secrets.GITHUB_TOKEN }}@github.com/${{ github.repository }}"
          git push -f -u origin release

      - name: Purge jsdelivr CDN
        run: |
          cd publish || exit 1
          for file in $(ls); do
            curl -i "https://purge.jsdelivr.net/gh/${{ github.repository }}@release/${file}"
          done