From db36e517019fde50ab99bbbb0d48a568c215b91d Mon Sep 17 00:00:00 2001
From: Ian Young <ian@iay.org.uk>
Date: Tue, 7 Mar 2023 16:36:53 +0000
Subject: [PATCH] Implement basic continuous integration test

See ukf/ukf-meta#384.
---
 .gitlab-ci.yml               | 55 ++++++++++++++++++++++++++++++++++++
 ci-download.properties       | 15 ++++++++++
 ci-thin.properties           | 14 +++++++++
 utilities/thin_aggregate.xsl | 33 ++++++++++++++++++++++
 4 files changed, 117 insertions(+)
 create mode 100644 .gitlab-ci.yml
 create mode 100644 ci-download.properties
 create mode 100644 ci-thin.properties
 create mode 100644 utilities/thin_aggregate.xsl

diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
new file mode 100644
index 0000000..5ea7b85
--- /dev/null
+++ b/.gitlab-ci.yml
@@ -0,0 +1,55 @@
+#
+# Continuous integration tests for the ukf-meta repository.
+#
+
+#
+# Default image for all steps is Amazon Corretto 11, which is based
+# on Amazon Linux (a variety of yum-based Linux derived from RHEL 7).
+#
+image: "amazoncorretto:11"
+
+stages:
+  - test
+
+perform-test:
+  stage: test
+  script:
+    #
+    # Install the tools we need that are not provided by the base image.
+    #
+    - yum -y --quiet install ant git libxslt
+
+    #
+    # Create work directories under the current one.
+    #
+    - mkdir -p work/build work/temp
+
+    #
+    # Fetch full UK federation inventory.
+    #
+    # The clone is made with a depth of 1 because we don't need any history.
+    #
+    - git clone --depth=1 https://gitlab-ci-token:$CI_JOB_TOKEN@$CI_SERVER_HOST/ukf/ukf-data.git work/ukf-data
+
+    #
+    # Thin UK federation inventory: retain 100 random entities.
+    #
+    - find work/ukf-data/entities -type f -name uk0\*.xml | sort -R | tail -n +101 | xargs rm
+
+    #
+    # Acquire a copy of the eduGAIN aggregate.
+    #
+    - ant -Denv=ci-download flow.edugain.download
+    - ls -lh work/temp
+
+    #
+    # Thin the eduGAIN aggregate into another file containing just 1% of the original
+    # entities.
+    #
+    - xsltproc -o work/temp/edugain-thin.xml utilities/thin_aggregate.xsl work/temp/edugain-download.xml
+    - ls -lh work/temp
+
+    #
+    # Run the full generate pipeline on the thinned input data.
+    #
+    - ant -Denv=ci-thin flow.aggregates.generate
diff --git a/ci-download.properties b/ci-download.properties
new file mode 100644
index 0000000..287cb77
--- /dev/null
+++ b/ci-download.properties
@@ -0,0 +1,15 @@
+#
+# ci-download.properties
+#
+# Properties defined for the env=ci-download environment, the part of the
+# continuous integration job that downloads an eduGAIN aggregate.
+#
+
+# Shared workspace location is assumed to be under the current directory.
+shared.ws.dir=${basedir}/work
+
+#
+# Cached eduGAIN download goes into a different directory so that it
+# persists over comparison runs.
+#
+mda.output.edugain.download=${shared.ws.dir}/temp/edugain-download.xml
diff --git a/ci-thin.properties b/ci-thin.properties
new file mode 100644
index 0000000..67c9702
--- /dev/null
+++ b/ci-thin.properties
@@ -0,0 +1,14 @@
+#
+# ci-thin.properties
+#
+# Properties defined for the env=ci-thin environment, the part of the
+# continuous integration job that runs the test on the thinned data.
+#
+
+# Shared workspace location is assumed to be under the current directory.
+shared.ws.dir=${basedir}/work
+
+#
+# Location of the thinned version of the downloaded eduGAIN aggregate.
+#
+mda.output.edugain.download=${shared.ws.dir}/temp/edugain-thin.xml
diff --git a/utilities/thin_aggregate.xsl b/utilities/thin_aggregate.xsl
new file mode 100644
index 0000000..eb10c24
--- /dev/null
+++ b/utilities/thin_aggregate.xsl
@@ -0,0 +1,33 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!--
+    thin_aggregate.xsl
+
+    Thins the input aggregate so that only 1% of the entities remain.
+-->
+<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+    xmlns:md="urn:oasis:names:tc:SAML:2.0:metadata"
+    xmlns:math="http://exslt.org/math">
+
+    <!--
+        Force UTF-8 encoding for the output.
+    -->
+    <xsl:output omit-xml-declaration="no" method="xml" encoding="UTF-8"/>
+
+    <!-- Discard most entities. -->
+    <xsl:template match="md:EntityDescriptor[math:random()>0.01]">
+        <!-- discard -->
+    </xsl:template>
+
+    <!--By default, copy text blocks, comments and attributes unchanged.-->
+    <xsl:template match="text()|comment()|@*">
+        <xsl:copy/>
+    </xsl:template>
+
+    <!--By default, copy all elements from the input to the output, along with their attributes and contents.-->
+    <xsl:template match="*">
+        <xsl:copy>
+            <xsl:apply-templates select="node()|@*"/>
+        </xsl:copy>
+    </xsl:template>
+
+</xsl:stylesheet>