refactor(tools): format jq command for better readability in get_latest_gh_tag function

ct/dispatcharr.sh

@@ -54,7 +54,7 @@ function update_script() {
       cp /opt/dispatcharr/.env /tmp/dispatcharr.env.backup
     fi
     if [[ -f /opt/dispatcharr/start-gunicorn.sh ]]; then
-      rm -f /opt/dispatcharr/start-gunicorn.sh
+      cp /opt/dispatcharr/start-gunicorn.sh /tmp/start-gunicorn.sh.backup
     fi
     if [[ -f /opt/dispatcharr/start-celery.sh ]]; then
       cp /opt/dispatcharr/start-celery.sh /tmp/start-celery.sh.backup
@@ -83,6 +83,9 @@ function update_script() {
     if [[ -f /tmp/dispatcharr.env.backup ]]; then
       mv /tmp/dispatcharr.env.backup /opt/dispatcharr/.env
     fi
+    if [[ -f /tmp/start-gunicorn.sh.backup ]]; then
+      mv /tmp/start-gunicorn.sh.backup /opt/dispatcharr/start-gunicorn.sh
+    fi
     if [[ -f /tmp/start-celery.sh.backup ]]; then
       mv /tmp/start-celery.sh.backup /opt/dispatcharr/start-celery.sh
     fi
@@ -102,35 +105,7 @@ function update_script() {
     rm -rf .venv
     $STD uv venv --clear
     $STD uv sync
-    $STD uv pip install uwsgi gevent celery redis daphne
-    cat <<'UWSGI_EOF' >/opt/dispatcharr/start-uwsgi.sh
-#!/usr/bin/env bash
-cd /opt/dispatcharr
-set -a
-source .env
-set +a
-exec .venv/bin/uwsgi \
-    --chdir=/opt/dispatcharr \
-    --module=dispatcharr.wsgi:application \
-    --master \
-    --workers=4 \
-    --gevent=400 \
-    --http=0.0.0.0:5656 \
-    --http-keepalive=1 \
-    --http-timeout=600 \
-    --socket-timeout=600 \
-    --buffer-size=65536 \
-    --post-buffering=4096 \
-    --lazy-apps \
-    --thunder-lock \
-    --die-on-term \
-    --vacuum
-UWSGI_EOF
-    chmod +x /opt/dispatcharr/start-uwsgi.sh
-    if grep -q 'start-gunicorn.sh' /etc/systemd/system/dispatcharr.service; then
-      sed -i 's|start-gunicorn.sh|start-uwsgi.sh|g' /etc/systemd/system/dispatcharr.service
-      systemctl daemon-reload
-    fi
+    $STD uv pip install gunicorn gevent celery redis daphne
     msg_ok "Updated Dispatcharr Backend"
 
     msg_info "Building Frontend"

ct/mealie.sh

@@ -81,11 +81,7 @@ STARTEOF
     cp -r /opt/mealie/frontend/dist/* /opt/mealie/mealie/frontend/
     msg_ok "Copied Frontend"
 
-    msg_info "Updating NLTK Data"
-    mkdir -p /nltk_data/
-    cd /opt/mealie
-    $STD uv run python -m nltk.downloader -d /nltk_data averaged_perceptron_tagger_eng
-    msg_ok "Updated NLTK Data"
+    setup_nltk "averaged_perceptron_tagger_eng" "/nltk_data"
 
     msg_info "Starting Service"
     systemctl start mealie

ct/paperless-ngx.sh

@@ -164,13 +164,7 @@ function update_script() {
       fi
     fi
 
-    msg_info "Updating NLTK Data"
-    cd /opt/paperless
-    $STD uv run python -m nltk.downloader -d /usr/share/nltk_data snowball_data
-    $STD uv run python -m nltk.downloader -d /usr/share/nltk_data stopwords
-    $STD uv run python -m nltk.downloader -d /usr/share/nltk_data punkt_tab ||
-      $STD uv run python -m nltk.downloader -d /usr/share/nltk_data punkt
-    msg_ok "Updated NLTK Data"
+    setup_nltk "snowball_data stopwords punkt_tab" "/usr/share/nltk_data"
 
     msg_info "Starting all Paperless-ngx Services"
     systemctl start paperless-consumer paperless-webserver paperless-scheduler paperless-task-queue

install/dispatcharr-install.sh

@@ -38,7 +38,7 @@ msg_info "Installing Python Dependencies with uv"
 cd /opt/dispatcharr
 $STD uv venv --clear
 $STD uv sync
-$STD uv pip install uwsgi gevent celery redis daphne
+$STD uv pip install gunicorn gevent celery redis daphne
 msg_ok "Installed Python Dependencies"
 
 msg_info "Configuring Dispatcharr"
@@ -118,7 +118,7 @@ server {
         proxy_set_header X-Forwarded-Proto \$scheme;
     }
 
-    # All other requests proxy to uWSGI
+    # All other requests proxy to Gunicorn
     location / {
         include proxy_params;
         proxy_pass http://127.0.0.1:5656;
@@ -131,30 +131,20 @@ systemctl restart nginx
 msg_ok "Configured Nginx"
 
 msg_info "Creating Services"
-cat <<EOF >/opt/dispatcharr/start-uwsgi.sh
+cat <<EOF >/opt/dispatcharr/start-gunicorn.sh
 #!/usr/bin/env bash
 cd /opt/dispatcharr
 set -a
 source .env
 set +a
-exec .venv/bin/uwsgi \\
-    --chdir=/opt/dispatcharr \\
-    --module=dispatcharr.wsgi:application \\
-    --master \\
+exec uv run gunicorn \\
     --workers=4 \\
-    --gevent=400 \\
-    --http=0.0.0.0:5656 \\
-    --http-keepalive=1 \\
-    --http-timeout=600 \\
-    --socket-timeout=600 \\
-    --buffer-size=65536 \\
-    --post-buffering=4096 \\
-    --lazy-apps \\
-    --thunder-lock \\
-    --die-on-term \\
-    --vacuum
+    --worker-class=gevent \\
+    --timeout=300 \\
+    --bind 0.0.0.0:5656 \\
+    dispatcharr.wsgi:application
 EOF
-chmod +x /opt/dispatcharr/start-uwsgi.sh
+chmod +x /opt/dispatcharr/start-gunicorn.sh
 
 cat <<EOF >/opt/dispatcharr/start-celery.sh
 #!/usr/bin/env bash
@@ -194,7 +184,7 @@ After=network.target postgresql.service redis-server.service
 [Service]
 Type=simple
 WorkingDirectory=/opt/dispatcharr
-ExecStart=/opt/dispatcharr/start-uwsgi.sh
+ExecStart=/opt/dispatcharr/start-gunicorn.sh
 Restart=on-failure
 RestartSec=10
 User=root

install/kitchenowl-install.sh

@@ -47,8 +47,7 @@ msg_info "Setting up KitchenOwl"
 cd /opt/kitchenowl/backend
 $STD uv sync --no-dev
 sed -i 's/default=True/default=False/' /opt/kitchenowl/backend/wsgi.py
-mkdir -p /nltk_data
-$STD uv run python -m nltk.downloader -d /nltk_data averaged_perceptron_tagger_eng
+setup_nltk "averaged_perceptron_tagger_eng" "/nltk_data"
 JWT_SECRET=$(openssl rand -hex 32)
 mkdir -p /opt/kitchenowl/data
 cat <<EOF >/opt/kitchenowl/kitchenowl.env

install/mealie-install.sh

@@ -55,11 +55,7 @@ mkdir -p /opt/mealie/mealie/frontend
 cp -r /opt/mealie/frontend/dist/* /opt/mealie/mealie/frontend/
 msg_ok "Copied Frontend"
 
-msg_info "Downloading NLTK Data"
-mkdir -p /nltk_data/
-cd /opt/mealie
-$STD uv run python -m nltk.downloader -d /nltk_data averaged_perceptron_tagger_eng
-msg_ok "Downloaded NLTK Data"
+setup_nltk "averaged_perceptron_tagger_eng" "/nltk_data"
 
 msg_info "Writing Environment File"
 SECRET=$(openssl rand -hex 32)

install/paperless-ngx-install.sh

@@ -94,18 +94,12 @@ user.save()
 EOF
 msg_ok "Set up admin Paperless-ngx User & Password"
 
-msg_info "Installing Natural Language Toolkit (Patience)"
-cd /opt/paperless
-$STD uv run python -m nltk.downloader -d /usr/share/nltk_data snowball_data
-$STD uv run python -m nltk.downloader -d /usr/share/nltk_data stopwords
-$STD uv run python -m nltk.downloader -d /usr/share/nltk_data punkt_tab ||
-  $STD uv run python -m nltk.downloader -d /usr/share/nltk_data punkt
+setup_nltk "snowball_data stopwords punkt_tab" "/usr/share/nltk_data"
 for policy_file in /etc/ImageMagick-6/policy.xml /etc/ImageMagick-7/policy.xml; do
   if [[ -f "$policy_file" ]]; then
     sed -i -e 's/rights="none" pattern="PDF"/rights="read|write" pattern="PDF"/' "$policy_file"
   fi
 done
-msg_ok "Installed Natural Language Toolkit"
 
 msg_info "Creating Services"
 cat <<EOF >/etc/systemd/system/paperless-scheduler.service

misc/tools.func

@@ -2095,10 +2095,10 @@ get_latest_gh_tag() {
     local count
     count=$(jq 'length' "$temp_file" 2>/dev/null || echo 0)
     if [[ "$count" -gt 0 ]]; then
-      tag=$(jq -r '.[].ref' "$temp_file" \
-        | sed 's|^refs/tags/||' \
-        | sort -V \
-        | tail -n1)
+      tag=$(jq -r '.[].ref' "$temp_file" |
+        sed 's|^refs/tags/||' |
+        sort -V |
+        tail -n1)
     fi
   else
     # No prefix: just take the first (newest) tag from /tags
@@ -9439,3 +9439,73 @@ function fetch_and_deploy_gl_release() {
   msg_ok "Deployed: $app ($version)"
   rm -rf "$tmpdir"
 }
+
+# ------------------------------------------------------------------------------
+# Download NLTK data packages directly from GitHub, bypassing Python.
+# Avoids CPU-instruction failures (SIGILL) on older hardware lacking AVX.
+#
+# Usage:
+#   setup_nltk "averaged_perceptron_tagger_eng" "/nltk_data"
+#   setup_nltk "snowball_data stopwords punkt_tab" "/usr/share/nltk_data"
+#
+# Parameters:
+#   $1 - Space-separated list of NLTK package IDs
+#   $2 - Target directory (default: /usr/share/nltk_data)
+#
+# Returns: 0 on success, non-zero if any package failed
+# ------------------------------------------------------------------------------
+function setup_nltk() {
+  local packages="${1:?setup_nltk requires at least one package name}"
+  local target_dir="${2:-/usr/share/nltk_data}"
+  local NLTK_INDEX_URL="https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/index.xml"
+  local index_xml rc=0
+
+  ensure_dependencies unzip
+
+  index_xml=$(curl_with_retry "$NLTK_INDEX_URL" "-") || {
+    msg_error "Failed to fetch NLTK package index"
+    return 1
+  }
+
+  local pkg
+  for pkg in $packages; do
+    msg_info "Downloading NLTK: $pkg"
+    local pkg_line subdir pkg_url do_unzip tmp_zip
+
+    pkg_line=$(echo "$index_xml" | grep "id=\"${pkg}\"" | head -1)
+    if [[ -z "$pkg_line" ]]; then
+      msg_error "NLTK package not found in index: $pkg"
+      rc=1
+      continue
+    fi
+
+    subdir=$(echo "$pkg_line" | grep -oP 'subdir="\K[^"]+')
+    pkg_url=$(echo "$pkg_line" | grep -oP 'url="\K[^"]+')
+    do_unzip=$(echo "$pkg_line" | grep -oP 'unzip="\K[^"]+')
+
+    if [[ -z "$subdir" || -z "$pkg_url" ]]; then
+      msg_error "Could not parse NLTK index entry for: $pkg"
+      rc=1
+      continue
+    fi
+
+    mkdir -p "${target_dir}/${subdir}"
+    tmp_zip=$(mktemp --suffix=.zip)
+
+    if CURL_TIMEOUT=120 curl_with_retry "$pkg_url" "$tmp_zip"; then
+      if [[ "$do_unzip" == "1" ]]; then
+        $STD unzip -q -o "$tmp_zip" -d "${target_dir}/${subdir}/"
+        rm -f "$tmp_zip"
+      else
+        mv "$tmp_zip" "${target_dir}/${subdir}/${pkg}.zip"
+      fi
+      msg_ok "Downloaded NLTK: $pkg"
+    else
+      msg_error "Failed to download NLTK package: $pkg"
+      rm -f "$tmp_zip"
+      rc=1
+    fi
+  done
+
+  return $rc
+}