farside/.github/workflows/update-instances.yml

on:
  workflow_dispatch:
  schedule:
    - cron: '0 0 * * *'

jobs:
  update-instances:
    runs-on: ubuntu-latest

    steps:
    - uses: actions/checkout@v2
    - name: Install dependencies
      run: sudo apt-get install -y jq dnsutils

    - uses: webfactory/ssh-agent@v0.5.3
      with:
        ssh-private-key: |
          ${{ secrets.PRIMARY_REPO_KEY }}          

    - name: Fetch instances
      run: |
        function apply_update() {
          mv services-tmp.json services-full.json
          rm -f *-tmp.json

          # Ensure no trailing slashes for any instance
          sed -i 's/\/"/"/g' services-full.json
        }

        # ==============================================================
        # Git config
        # ==============================================================
        git config --global user.name github-actions
        git config --global user.email 41898282+github-actions[bot]@users.noreply.github.com
        git remote set-url origin git@github.com:benbusby/farside.git
        git checkout main

        # ==============================================================
        # Bibliogram update
        # ==============================================================
        curl -s https://bibliogram.art/api/instances | \
          jq '[
            .data |
            .[] |
            select(.onion_site | not) |
            .address
          ] | sort' > bibliogram-tmp.json
        jq --slurpfile bibliogram bibliogram-tmp.json \
          '( .[] | select(.type == "bibliogram") )
          .instances |= $bibliogram[0]' services-full.json > services-tmp.json

        apply_update

        # ==============================================================
        # searx update
        # ==============================================================
        curl -s https://searx.space/data/instances.json | \
          jq '[
            .instances |
            to_entries[] |
            select(.value.network_type == "normal") |
            select(.value.version | . != null) |
            select(.value.version | startswith("1.0.0")) |
            select(.value.network.asn_privacy == 0) |
            select(.value.http.error == null) |
            select(.value.tls.grade == "A+" or .value.tls.grade == "A") |
            select(.value.http.grade == "A+" or .value.http.grade == "A") |
            select(.value.html.grade == "V" or .value.html.grade == "F") |
            select(.key | contains(".i2p") | not) |
            .key
          ] | sort' > searx-tmp.json

        cat searx-tmp.json | jq .

        jq --slurpfile searx searx-tmp.json \
          '( .[] | select(.type == "searx") )
          .instances |= $searx[0]' services-full.json > services-tmp.json

        apply_update

        # ==============================================================
        # SearXNG update
        # ==============================================================
        curl -s https://searx.space/data/instances.json | \
          jq '[
            .instances |
            to_entries[] |
            select(.value.network_type == "normal") |
            select(.value.generator | . != null) |
            select(.value.generator | contains("searxng")) |
            select(.value.version | . != null) |
            select(.value.network.asn_privacy == 0) |
            select(.value.http.error == null) |
            select(.value.tls.grade == "A+" or .value.tls.grade == "A") |
            select(.value.http.grade == "A+" or .value.http.grade == "A") |
            select(.value.html.grade == "V" or .value.html.grade == "F") |
            select(.key | contains(".i2p") | not) |
            .key
          ] | sort' > searxng-tmp.json

        cat searxng-tmp.json | jq .

        jq --slurpfile searxng searxng-tmp.json \
          '( .[] | select(.type == "searxng") )
          .instances |= $searxng[0]' services-full.json > services-tmp.json

        apply_update

        # ==============================================================
        # SimplyTranslate update
        # ==============================================================
        curl -s https://simple-web.org/instances/simplytranslate > st-out
        sed -i -e 's/^/https:\/\//' st-out
        jq -ncR '[inputs]' <<< "$(cat st-out | sort)" > st-tmp.json

        jq --slurpfile simplytranslate st-tmp.json \
          '( .[] | select(.type == "simplytranslate") )
          .instances |= $simplytranslate[0]' services-full.json > services-tmp.json

        apply_update

        # ==============================================================
        # Whoogle update
        # ==============================================================
        curl -s https://raw.githubusercontent.com/benbusby/whoogle-search/main/misc/instances.txt > whoogle-out
        jq -ncR '[inputs]' <<< "$(cat whoogle-out | sort)" > whoogle-tmp.json

        jq --slurpfile whoogle whoogle-tmp.json \
          '( .[] | select(.type == "whoogle") )
          .instances |= $whoogle[0]' services-full.json > services-tmp.json

        apply_update

        # ==============================================================
        # Invidious update
        # ==============================================================
        curl -s https://api.invidious.io/instances.json | \
          jq '[
            .[] |
            select(.[1].type | contains("https")) ] |
            [.[][1].uri] |
            sort' > invidious-tmp.json
        jq --slurpfile invidious invidious-tmp.json \
          '( .[] | select(.type == "invidious") )
          .instances |= $invidious[0]' services-full.json > services-tmp.json

        apply_update

        # ==============================================================
        # Scribe update
        # ==============================================================
        curl -s https://git.sr.ht/~edwardloveall/scribe/blob/main/docs/instances.json > scribe-tmp.json
        jq --slurpfile scribe scribe-tmp.json \
          '( .[] | select(.type == "scribe") )
          .instances |= $scribe[0]' services-full.json > services-tmp.json

        apply_update

        # ==============================================================
        # teddit update
        # ==============================================================
        curl -s https://codeberg.org/teddit/teddit/raw/branch/main/instances.json | \
          jq '[
            .[] |
            select(.url | contains("https")) |
            .url ] |
            sort' > teddit-tmp.json
        jq --slurpfile teddit teddit-tmp.json \
          '( .[] | select(.type == "teddit") )
          .instances |= $teddit[0]' services-full.json > services-tmp.json

        apply_update

        # ==============================================================
        # Wikiless update
        # ==============================================================
        curl -s https://codeberg.org/orenom/Wikiless/raw/branch/main/instances.json | \
          jq '[
            .[] |
            select((.url | contains(".onion") | not) and (.url | contains("https"))).url ] |
            sort' > wikiless-tmp.json
        jq --slurpfile wikiless wikiless-tmp.json \
          '( .[] | select(.type == "wikiless") )
          .instances |= $wikiless[0]' services-full.json > services-tmp.json

        apply_update

        # ==============================================================
        # TODO: Update instances for other services
        # ==============================================================

        # ==============================================================
        # Push changes
        # ==============================================================
        git add services.json services-full.json
        git commit -m '[CI] Auto update instances' || exit 0
        git push