From 8f5c8b18788de79f6ce20e08e9ea3291ba801249 Mon Sep 17 00:00:00 2001
From: tooomm <tooomm@users.noreply.github.com>
Date: Wed, 23 Apr 2025 21:13:32 +0200
Subject: [PATCH] CI: Show useful data about token URLs (#192)

---
 .github/workflows/check_links.yml | 45 +++++++++++++++++++++++++++----
 1 file changed, 40 insertions(+), 5 deletions(-)

diff --git a/.github/workflows/check_links.yml b/.github/workflows/check_links.yml
index 5d2e1f1..83dbf5f 100644
--- a/.github/workflows/check_links.yml
+++ b/.github/workflows/check_links.yml
@@ -23,8 +23,6 @@ jobs:
 
     name: Check image links
     runs-on: ubuntu-latest
-    env:
-      lychee_args: '--no-progress --require-https --cache --max-cache-age 8h --exclude http://www.w3.org'
 
     steps:
       - name: Checkout
@@ -39,13 +37,50 @@ jobs:
           key: cache-lychee-${{ github.sha }}
           restore-keys: cache-lychee-
 
-      # Extract and check URLs directly from token files
-      - name: Check token art URLs
+      # Extract picture URLs from tokens.xml
+      - name: Extract URLs
+        id: tokens_pic_urls
         uses: lycheeverse/lychee-action@v2
         with:
-          args: '${{env.lychee_args}} -- tokens.xml'
+          args: --dump --exclude www.w3.org -- tokens.xml
+          output: url_list.md
+          jobSummary: false
+
+      # Check dumped URLs
+      - name: Check token art URLs
+        if: steps.tokens_pic_urls.outcome == 'success'
+        uses: lycheeverse/lychee-action@v2
+        with:
+          args: --no-progress --require-https --cache --max-cache-age 8h -- url_list.md
           jobSummary: true
 
+      # List some helpful stats on missing pictures
+      - name: List missing image link counts
+        shell: bash
+        run: |
+          echo "❌ **Missing Image Links (<kbd>master</kbd> branch)**" >> $GITHUB_STEP_SUMMARY
+          echo "Token with missing \`set\` element: [![](https://img.shields.io/badge/dynamic/xml?label=&colorB=white&query=count%28%2F%2Fcard%5Bnot%28set%29%5D%29&url=https%3A%2F%2Fraw.githubusercontent.com%2FCockatrice%2FMagic-Token%2Fmaster%2Ftokens.xml)](https://raw.githubusercontent.com/Cockatrice/Magic-Token/master/tokens.xml)" >> $GITHUB_STEP_SUMMARY
+          echo "Missing picURL attribute (\`<set>\`): [![](https://img.shields.io/badge/dynamic/xml?label=&colorB=white&query=count%28%2F%2Fset%29-count%28%2F%2Fset%5B%40picURL%5D%29&url=https%3A%2F%2Fraw.githubusercontent.com%2FCockatrice%2FMagic-Token%2Fmaster%2Ftokens.xml)](https://raw.githubusercontent.com/Cockatrice/Magic-Token/master/tokens.xml)" >> $GITHUB_STEP_SUMMARY
+          echo "Empty picURL value (\`<set picURL=\"\">\`): [![](https://img.shields.io/badge/dynamic/xml?label=&colorB=white&query=count%28%2F%2Fset%5B%40picURL%3D%22%22%5D%29&url=https%3A%2F%2Fraw.githubusercontent.com%2FCockatrice%2FMagic-Token%2Fmaster%2Ftokens.xml)](https://raw.githubusercontent.com/Cockatrice/Magic-Token/master/tokens.xml)" >> $GITHUB_STEP_SUMMARY
+
+      # Analyse extracted links (1/2)
+      - name: List duplicated image links
+        if: steps.tokens_pic_urls.outcome == 'success'
+        shell: bash
+        # Remove empty lines | trim trailing integers (Scryfall) | sort | count and list duplicates
+        run: |
+          echo "🪞 **Duplicated Image Links**" >> $GITHUB_STEP_SUMMARY
+          grep . url_list.md | sed 's/\.jpg?.*/.jpg/' | sort | uniq -cd >> $GITHUB_STEP_SUMMARY
+
+      # Analyse extracted links (2/2)
+      - name: List image hosting sources
+        if: steps.tokens_pic_urls.outcome == 'success'
+        shell: bash
+        # Extract 3rd field (domain name) | remove empty lines | sort | count duplicates and list with numbers | sort descending
+        run: |
+          echo "📶 **Image Hosting Statistics**" >> $GITHUB_STEP_SUMMARY
+          awk -F/ '{print $3}' url_list.md | grep . | sort | uniq -c | sort -nr >> $GITHUB_STEP_SUMMARY
+
       # Always save cache
       - name: Save lychee cache
         uses: actions/cache/save@v4