docs: add clean up images job DOC-1232

.github/workflows/clean-up-unused-images.yaml

@@ -0,0 +1,71 @@
+name: Clean Up Unused Images
+
+on:
+  schedule:
+    # On the first of every month at 2 am
+    - cron: '0 2 1 * *'
+  workflow_dispatch:
+
+concurrency:
+  group: clean-up-images-${{ github.ref }}
+  cancel-in-progress: true
+
+env:
+  GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+jobs:
+  find_unused_images:
+    runs-on: ubuntu-latest
+
+    steps:
+      - id: checkout
+        name: Checkout Repository
+        uses: actions/checkout@v4
+
+      - name: Setup Nodejs
+        uses: actions/setup-node@v4
+        with:
+          node-version: "20"
+
+      - name: Find unused images
+        run: make find-unused-images
+
+      - name: Install GitHub CLI
+        run: sudo apt-get install gh
+
+      - name: Create PR with unused images
+        run: |
+          # Ensure that we are on master.
+          git checkout master
+
+          # Create a new branch.
+          branch_name="clean-up-unused-images-$(date +%Y%m%d%H%M%S)"
+          git checkout -b "$branch_name"
+
+          # Remove all the images identified as unused.
+          for img in $(cat unused_images.json); do
+            rm static/assets/docs/images/$img
+          done
+
+          # Clean up results file.
+          rm unused_images.json
+
+          # Commit and push branch
+          git add .
+          git commit -m "docs: clean up unused images"
+          git push origin $branch_name
+
+          # Create the pull request
+          gh pr create --base master --title "docs: clean up librarium unused images " --body "$(cat <<EOF
+            ## Describe the Change
+
+            This PR removes images identified as unused across all our branches. 
+
+            The images are identified using `scripts/find-unused-images.sh` script.
+
+            Please review this PR carefully before merging it.
+
+            EOF
+          )"
+    
+

.gitignore

@@ -77,3 +77,5 @@ static/img/packs
 vale/styles/spectrocloud/
 vale/styles/spectrocloud-docs-internal/
 vale/styles/config/vocabularies/spectrocloud-vocab
+
+unused_images.json

Makefile

@@ -65,7 +65,6 @@ clean-visuals:
 	@echo "Cleaning visual regression tests"
 
 	rm -rf test-results/  playwright-report/  screenshots/
-
 
 ##@ npm Targets
 
@@ -221,6 +220,12 @@ format-images: ## Format images
 	@echo "formatting images in /static/assets/docs/images/ folder"
 	./scripts/compress-convert-images.sh
 
+###@ Find unused images assets
+
+find-unused-images:
+	@echo "Find unused image assets"
+	./scripts/find-unused-images.sh	
+
 ###@ Generate _partials/index.ts required to automatic partials usage.
 
 generate-partials: ## Generate

scripts/find-unused-images.sh

@@ -0,0 +1,44 @@
+#!/bin/bash
+
+# Enable error handling
+set -e
+
+echo "Checking for unused image files in static/assets/docs/images folder..."
+
+# Create a list of all the images we have and save it to a json.
+# Trim the path static/assets/docs/images.
+find static/assets/docs/images -type f ! -name ".DS_STORE" ! -name ".DS_Store" | sed 's|static/assets/docs/images||g'  > all_images.json
+
+# List all the version branches
+branches="master"
+version_branches=$(git branch --list "version-[0-9]-[0-9]")
+branches+=" $version_branches"
+
+for current_branch in $branches; do
+    git checkout $current_branch
+    echo "Current branch: $current_branch"
+    git pull
+
+    find docs -type f -name "*.md" -exec grep -Hn "\.webp" {} + > docs_used_images.json
+    find _partials -type f -name "*.mdx" -exec grep -Hn "\.webp" {} + > partials_used_images.json
+    cat docs_used_images.json partials_used_images.json > used_images.json
+
+    line_number=1
+    for img in $(cat all_images.json); do
+        if grep -q $img used_images.json; then
+            sed -i .bak "${line_number}s|.*|${img},FOUND_USED|" all_images.json
+        fi
+        ((line_number++))
+    done
+
+done
+
+# Remove all marked used files to make up the list 
+sed '/FOUND_USED/d' all_images.json > unused_images.json
+
+# Clean up files
+rm all_images.json
+rm all_images.json.bak
+rm docs_used_images.json
+rm partials_used_images.json
+rm used_images.json