From 2e3fe6e00e0009136fb8d933e2fd61caad681ade Mon Sep 17 00:00:00 2001
From: Marlon Saglia <marlon@vespa.ai>
Date: Tue, 13 Aug 2024 15:34:51 +0200
Subject: [PATCH] feat: migrate "link-checker-sample-apps" workflow from
 Screwdriver to Github

temporary test with separate branch
---
 .github/workflows/link-checker.yml | 38 ++++++++++++++++++++++++++++++
 screwdriver.yaml                   | 11 +++++----
 2 files changed, 45 insertions(+), 4 deletions(-)
 create mode 100644 .github/workflows/link-checker.yml

diff --git a/.github/workflows/link-checker.yml b/.github/workflows/link-checker.yml
new file mode 100644
index 000000000..3fab681d8
--- /dev/null
+++ b/.github/workflows/link-checker.yml
@@ -0,0 +1,38 @@
+name: Link checker
+
+on:
+  workflow_dispatch:
+  push:
+    branches:
+      - master
+  pull_request:
+    branches:
+      - master
+  schedule:
+    - cron: "00 3 * * 1-5"
+
+jobs:
+  test:
+    uses: vespa-engine/gh-actions/.github/workflows/jekyll-link-checker.yml@remove-redirect-files-before-link-check
+    with:
+      additional-args: "--directory-index-file README.html"
+      add-no-render-with-liquid: "true"
+      build-plugins: |
+        _plugins-linkcheck
+      pre-check-script: |
+        mv links-to-check.html _site
+        echo "Broken links in links-to-check.html are extracted from services/hosts/deployment.xml in the sample apps, as well as links in javadoc."
+        echo "Search for broken links in these files."
+      ignore-files: |-
+        /tensor-playground/src/main/resources/playground/
+        /src/main/resources/site/index.html/
+        /dense-passage-retrieval-with-ann/src/main/resources/frontend/index.html/
+      ignore-urls: |-
+        src/main/application/schemas/wiki.sd#L80
+        /localhost:8080/
+        src/main/application/search/query-profiles
+        /hub.docker.com/r/vespaengine/vespa/tags/
+      swap-urls: |-
+        (https\://github.com.*/master/.*)#.*:\1
+        (https\://github.com.*/main/.*)#.*:\1
+        ^(?!https)(.*)\.md:\1.html
diff --git a/screwdriver.yaml b/screwdriver.yaml
index 6685fae1e..221d0b1c0 100644
--- a/screwdriver.yaml
+++ b/screwdriver.yaml
@@ -48,13 +48,16 @@ jobs:
           echo "Search for broken links in these files."
       - check-links: |
           bundle exec htmlproofer \
-            --assume-extension --check-html --no-check-external-hash --no-enforce-http \
-            --typhoeus '{"connecttimeout": 10, "timeout": 30, "accept_encoding": "zstd,br,gzip,deflate"}' \
-            --hydra '{"max_concurrency": 1}' \
+            --assume-extension \
+            --check-html \
             --directory-index-file README.html \
-            --ignore-urls 'src/main/application/schemas/wiki.sd#L80,/localhost:8080/,src/main/application/search/query-profiles,/hub.docker.com/r/vespaengine/vespa/tags/' \
+            --hydra '{"max_concurrency": 1}' \
             --ignore-files '/tensor-playground/src/main/resources/playground/,/src/main/resources/site/index.html/,/dense-passage-retrieval-with-ann/src/main/resources/frontend/index.html/' \
+            --ignore-urls 'src/main/application/schemas/wiki.sd#L80,/localhost:8080/,src/main/application/search/query-profiles,/hub.docker.com/r/vespaengine/vespa/tags/' \
+            --no-check-external-hash \
+            --no-enforce-http \
             --swap-urls '(https\://github.com.*/master/.*)#.*:\1,(https\://github.com.*/main/.*)#.*:\1,^(?!https)(.*)\.md:\1.html' \
+            --typhoeus '{"connecttimeout": 10, "timeout": 30, "accept_encoding": "zstd,br,gzip,deflate"}' \
             _site
 
   verify-billion-scale-vector-search: