check for duplicate keys in Java properties files, fix CodeMeta displ…

…ayName (#9176)

* ci: add workflow to check for duplicate keys in Java properties files #9169

* ci: add metadata blocks properties file check

As creating the properties files for metadata blocks is a
tedious manual process, this script ensures in CI everything
is present. It adds to checking for duplicates.

* feat,ci: move check scripts to separate files

- Easier to run and debug as separate files
- Can be used locally, too
- Use GraalVM to compile native binary for accents removal with same Java code as used in application (also uses JBang as build system)
- "Just" using JBang is not fast enough, JVM startup times are making it sluggish!

* fix,ci: remove shellspec tests on CentOS 7 as it is EOL

* test: add a stupid change to test checks run and find problems

* ci: fix JBang setup in properties check

* fix: make properties check script dir-safe

* fix: escape square brackets in pattern

* refactor: move CV mangling (lowercase, spaces) to Java

* fix(metadata): typo and missing property for CodeMeta

* Revert "test: add a stupid change to test checks run and find problems"

This reverts commit 33731ad.

* file for QA only. DO NOT CHECK IN

* file for QA only. DO NOT CHECK IN

* file for QA only. DO NOT CHECK IN

* Removing QA test files. Safe to merge

---------

Co-authored-by: Philip Durbin <philip_durbin@harvard.edu>
Co-authored-by: Steven Winship <39765413+stevenwinship@users.noreply.github.com>

.github/workflows/check_property_files.yml

@@ -0,0 +1,32 @@
+name: "Properties Check"
+on:
+    pull_request:
+        paths:
+            - "src/**/*.properties"
+            - "scripts/api/data/metadatablocks/*"
+jobs:
+    duplicate_keys:
+        name: Duplicate Keys
+        runs-on: ubuntu-latest
+        steps:
+            - uses: actions/checkout@v4
+            - name: Run duplicates detection script
+              shell: bash
+              run: tests/check_duplicate_properties.sh
+
+    metadata_blocks_properties:
+        name: Metadata Blocks Properties
+        runs-on: ubuntu-latest
+        steps:
+            - uses: actions/checkout@v4
+            - name: Setup GraalVM + Native Image
+              uses: graalvm/setup-graalvm@v1
+              with:
+                github-token: ${{ secrets.GITHUB_TOKEN }}
+                java-version: '21'
+                distribution: 'graalvm-community'
+            - name: Setup JBang
+              uses: jbangdev/setup-jbang@main
+            - name: Run metadata block properties verification script
+              shell: bash
+              run: tests/verify_mdb_properties.sh

.github/workflows/shellspec.yml

@@ -24,23 +24,6 @@ jobs:
               run: |
                   cd tests/shell
                   shellspec
-    shellspec-centos7:
-        name: "CentOS 7"
-        runs-on: ubuntu-latest
-        container:
-            image: centos:7
-        steps:
-            - uses: actions/checkout@v2
-            - name: Install shellspec
-              run: |
-                  curl -fsSL https://github.com/shellspec/shellspec/releases/download/${{ env.SHELLSPEC_VERSION }}/shellspec-dist.tar.gz | tar -xz -C /usr/share
-                  ln -s /usr/share/shellspec/shellspec /usr/bin/shellspec
-            - name: Install dependencies
-              run: yum install -y ed
-            - name: Run shellspec
-              run: |
-                  cd tests/shell
-                  shellspec
     shellspec-rocky8:
         name: "RockyLinux 8"
         runs-on: ubuntu-latest

src/main/java/propertyFiles/codeMeta20.properties

@@ -1,5 +1,6 @@
 metadatablock.name=codeMeta20
-metadatablock.displayName=Software Metadata (CodeMeta 2.0)
+metadatablock.displayName=Software Metadata (CodeMeta v2.0)
+metadatablock.displayFacet=Software
 datasetfieldtype.codeVersion.title=Software Version
 datasetfieldtype.codeVersion.description=Version of the software instance, usually following some convention like SemVer etc.
 datasetfieldtype.codeVersion.watermark=e.g. 0.2.1 or 1.3 or 2021.1 etc

tests/check_duplicate_properties.sh

@@ -0,0 +1,37 @@
+#!/bin/bash
+
+# This script will check Java *.properties files within the src dir for duplicates
+# and print logs with file annotations about it.
+
+set -euo pipefail
+
+FAIL=0
+
+while IFS= read -r -d '' FILE; do
+
+    # Scan the whole file for duplicates
+    FILTER=$(grep -a -v -E "^(#.*|\s*$)" "$FILE" | cut -d"=" -f1 | sort | uniq -c | tr -s " " | { grep -vs "^ 1 " || true; })
+
+    # If there are any duplicates present, analyse further to point people to the source
+    if [ -n "$FILTER" ]; then
+        FAIL=1
+
+        echo "::group::$FILE"
+        for KEY in $(echo "$FILTER" | cut -d" " -f3); do
+            # Find duplicate lines' numbers by grepping for the KEY and cutting the number from the output
+            DUPLICATE_LINES=$(grep -n -E -e "^$KEY=" "$FILE" | cut -d":" -f1)
+            # Join the found line numbers for better error log
+            DUPLICATE_NUMBERS=$(echo "$DUPLICATE_LINES" | paste -sd ',')
+
+            # This form will make Github annotate the lines in the PR that changes the properties file
+            for LINE_NUMBER in $DUPLICATE_LINES; do
+                echo "::error file=$FILE,line=$LINE_NUMBER::Found duplicate for key '$KEY' in lines $DUPLICATE_NUMBERS"
+            done
+        done
+        echo "::endgroup::"
+    fi
+done < <( find "$(git rev-parse --show-toplevel)" -wholename "*/src/*.properties" -print0 )
+
+if [ "$FAIL" -eq 1 ]; then
+    exit 1
+fi

tests/verify_mdb_properties.sh

@@ -0,0 +1,99 @@
+#!/bin/bash
+
+# This script will check our metadata block files and scan if the properties files contain all the matching keys.
+
+set -euo pipefail
+
+if ! which jbang > /dev/null 2>&1; then
+  echo "Cannot find jbang on path. Did you install it?" >&2
+  exit 1
+fi
+if ! which native-image > /dev/null 2>&1; then
+  echo "Cannot find GraalVM native-image on path. Did you install it?" >&2
+  exit 1
+fi
+
+FAIL=0
+
+# We need a small Java app here, replacing spaces, converting to lower case but especially to replace UTF-8 chars with nearest ascii / strip accents because of
+# https://github.com/IQSS/dataverse/blob/dddcf29188a5c35174f3c94ffc1c4cb1d7fc0552/src/main/java/edu/harvard/iq/dataverse/ControlledVocabularyValue.java#L139-L140
+# This cannot be replaced by another tool, as it behaves rather individually.
+DIR=$(mktemp -d)
+SOURCE="$DIR/stripaccents.java"
+STRIP_BIN="$(dirname "$0")/stripaccents"
+cat > "$SOURCE" << EOF
+///usr/bin/env jbang "\$0" "\$@" ; exit \$?
+//JAVA 11+
+//DEPS org.apache.commons:commons-lang3:3.12.0
+import org.apache.commons.lang3.StringUtils;
+import java.nio.charset.StandardCharsets;
+import java.io.IOException;
+class stripaccents {
+    public static void main(String[] args) throws IOException {
+        String input = new String(System.in.readAllBytes(), StandardCharsets.UTF_8).toLowerCase().replace(" ", "_");
+        System.out.println(StringUtils.stripAccents(input));
+    }
+}
+EOF
+jbang export native --force --fresh -O "$STRIP_BIN" "$SOURCE"
+
+while IFS= read -r -d '' MDB; do
+
+    echo "::group::$MDB"
+    BLOCK_NAME=$(sed -n "2p" "$MDB" | cut -f2)
+    BLOCK_DISPLAY_NAME=$(sed -n "2p" "$MDB" | cut -f4)
+    PROPERTIES_FILE="$(git rev-parse --show-toplevel)/src/main/java/propertyFiles/$BLOCK_NAME.properties"
+
+    # Check correct file exists
+    if [ ! -r "$PROPERTIES_FILE" ]; then
+        echo "::error::Missing properties file for metadata block '$BLOCK_NAME', expected at '$PROPERTIES_FILE'"
+        FAIL=1
+        continue
+    fi
+
+    # Check metadata block properties exist and are equal to TSV source
+    if ! grep -a -q -e "^metadatablock.name=$BLOCK_NAME$" "$PROPERTIES_FILE"; then
+        echo "::error::Missing 'metadatablock.name=$BLOCK_NAME' or different from TSV source in $PROPERTIES_FILE"
+        FAIL=1
+    fi
+    if ! grep -a -q -e "^metadatablock.displayName=$BLOCK_DISPLAY_NAME$" "$PROPERTIES_FILE"; then
+        echo "::error::Missing 'metadatablock.displayName=$BLOCK_DISPLAY_NAME' or different from TSV source in $PROPERTIES_FILE"
+        FAIL=1
+    fi
+    if ! grep -a -q -e "^metadatablock.displayFacet=" "$PROPERTIES_FILE"; then
+        echo "::error::Missing 'metadatablock.displayFacet=...' in $PROPERTIES_FILE"
+        FAIL=1
+    fi
+
+    # Check dataset fields
+    for FIELD in $(grep -a -A1000 "^#datasetField" "$MDB" | tail -n+2 | grep -a -B1000 "^#controlledVocabulary" | head -n-1 | cut -f2); do
+        for ENTRY in title description watermark; do
+            if ! grep -a -q -e "^datasetfieldtype.$FIELD.$ENTRY=" "$PROPERTIES_FILE"; then
+                echo "::error::Missing key 'datasetfieldtype.$FIELD.$ENTRY=...' in $PROPERTIES_FILE"
+                FAIL=1
+            fi
+        done
+    done
+
+    # Check CV entries
+    while read -r LINE; do
+        FIELD_NAME=$(echo "$LINE" | cut -f1)
+        # See https://github.com/IQSS/dataverse/blob/dddcf29188a5c35174f3c94ffc1c4cb1d7fc0552/src/main/java/edu/harvard/iq/dataverse/ControlledVocabularyValue.java#L139-L140
+        # Square brackets are special in grep with expressions activated, so escape them if present!
+        FIELD_VALUE=$(echo "$LINE" | cut -f2 | "$STRIP_BIN" | sed -e 's/\([][]\)/\\\1/g' )
+
+        if ! grep -q -a -e "^controlledvocabulary.$FIELD_NAME.$FIELD_VALUE=" "$PROPERTIES_FILE"; then
+            echo "::error::Missing key 'controlledvocabulary.$FIELD_NAME.$FIELD_VALUE=...' in $PROPERTIES_FILE"
+            FAIL=1
+        fi
+    done < <(grep -a -A1000 "^#controlledVocabulary" "$MDB" | tail -n+2)
+
+    echo "::endgroup::"
+
+done < <( find "$(git rev-parse --show-toplevel)/scripts/api/data/metadatablocks" -name '*.tsv' -print0 )
+
+rm "$SOURCE" "$STRIP_BIN"
+
+if [ "$FAIL" -eq 1 ]; then
+    exit 1
+fi