diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 4cce9d9..650a999 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -27,7 +27,7 @@
     - source `find .. -name section_helper.sh`
     - start_section install_linux_packages "Installing missing Linux packages"
     - apt-get -o dir::cache::archives="$APT_CACHE_DIR" update
-    - apt-get -o dir::cache::archives="$APT_CACHE_DIR" install -y libxml2-utils curl openjdk-17-jre-headless
+    - apt-get -o dir::cache::archives="$APT_CACHE_DIR" install -y libxml2-utils curl openjdk-17-jre-headless xmlstarlet
     - end_section install_linux_packages
     - start_section install_models "Installing models"
     - make models/dereko_domains_s.classifier
diff --git a/Makefile b/Makefile
index d55a415..7ec8d91 100644
--- a/Makefile
+++ b/Makefile
@@ -58,6 +58,7 @@
 	tail -n 1 xslt/idsCorpus-template.xml  >> $@
 
 test: models/dereko_domains_s.classifier i5valid
+	- sh test/test-xml.sh
 
 i5: $(foreach year,$(YEARS),$(TARGET_DIR)/dnb$(year).i5.xml)
 
diff --git a/test/assert.sh b/test/assert.sh
new file mode 100644
index 0000000..acceebf
--- /dev/null
+++ b/test/assert.sh
@@ -0,0 +1,248 @@
+#!/usr/bin/env bash
+
+#####################################################################
+##
+## title: Assert Extension
+##
+## description:
+## Assert extension of shell (bash, ...)
+##   with the common assert functions
+## Function list based on:
+##   http://junit.sourceforge.net/javadoc/org/junit/Assert.html
+## Log methods : inspired by
+##	- https://natelandau.com/bash-scripting-utilities/
+## author: Mark Torok
+##
+## date: 07. Dec. 2016
+##
+## license: MIT
+##
+#####################################################################
+
+if command -v tput &>/dev/null && tty -s; then
+  RED=$(tput setaf 1)
+  GREEN=$(tput setaf 2)
+  MAGENTA=$(tput setaf 5)
+  NORMAL=$(tput sgr0)
+  BOLD=$(tput bold)
+else
+  RED=$(echo -en "\e[31m")
+  GREEN=$(echo -en "\e[32m")
+  MAGENTA=$(echo -en "\e[35m")
+  NORMAL=$(echo -en "\e[00m")
+  BOLD=$(echo -en "\e[01m")
+fi
+
+log_header() {
+  printf "\n${BOLD}${MAGENTA}==========  %s  ==========${NORMAL}\n" "$@" >&2
+}
+
+log_success() {
+  printf "${GREEN}✔ %s${NORMAL}\n" "$@" >&2
+}
+
+log_failure() {
+  printf "${RED}✖ %s${NORMAL}\n" "$@" >&2
+}
+
+
+assert_eq() {
+  local expected="$1"
+  local actual="$2"
+  local msg="${3-}"
+
+  if [ "$expected" == "$actual" ]; then
+    return 0
+  else
+    [ "${#msg}" -gt 0 ] && log_failure "$expected == $actual :: $msg" || true
+    return 1
+  fi
+}
+
+assert_not_eq() {
+  local expected="$1"
+  local actual="$2"
+  local msg="${3-}"
+
+  if [ ! "$expected" == "$actual" ]; then
+    return 0
+  else
+    [ "${#msg}" -gt 0 ] && log_failure "$expected != $actual :: $msg" || true
+    return 1
+  fi
+}
+
+assert_true() {
+  local actual="$1"
+  local msg="${3-}"
+
+  assert_eq true "$actual" "$msg"
+  return "$?"
+}
+
+assert_false() {
+  local actual="$1"
+  local msg="${3-}"
+
+  assert_eq false "$actual" "$msg"
+  return "$?"
+}
+
+assert_array_eq() {
+
+  declare -a expected=("${!1-}")
+  # echo "AAE ${expected[@]}"
+
+  declare -a actual=("${!2}")
+  # echo "AAE ${actual[@]}"
+
+  local msg="${3-}"
+
+  local return_code=0
+  if [ ! "${#expected[@]}" == "${#actual[@]}" ]; then
+    return_code=1
+  fi
+
+  local i
+  for (( i=1; i < ${#expected[@]} + 1; i+=1 )); do
+    if [ ! "${expected[$i-1]}" == "${actual[$i-1]}" ]; then
+      return_code=1
+      break
+    fi
+  done
+
+  if [ "$return_code" == 1 ]; then
+    [ "${#msg}" -gt 0 ] && log_failure "(${expected[*]}) != (${actual[*]}) :: $msg" || true
+  fi
+
+  return "$return_code"
+}
+
+assert_array_not_eq() {
+
+  declare -a expected=("${!1-}")
+  declare -a actual=("${!2}")
+
+  local msg="${3-}"
+
+  local return_code=1
+  if [ ! "${#expected[@]}" == "${#actual[@]}" ]; then
+    return_code=0
+  fi
+
+  local i
+  for (( i=1; i < ${#expected[@]} + 1; i+=1 )); do
+    if [ ! "${expected[$i-1]}" == "${actual[$i-1]}" ]; then
+      return_code=0
+      break
+    fi
+  done
+
+  if [ "$return_code" == 1 ]; then
+    [ "${#msg}" -gt 0 ] && log_failure "(${expected[*]}) == (${actual[*]}) :: $msg" || true
+  fi
+
+  return "$return_code"
+}
+
+assert_empty() {
+  local actual=$1
+  local msg="${2-}"
+
+  assert_eq "" "$actual" "$msg"
+  return "$?"
+}
+
+assert_not_empty() {
+  local actual=$1
+  local msg="${2-}"
+
+  assert_not_eq "" "$actual" "$msg"
+  return "$?"
+}
+
+assert_contain() {
+  local haystack="$1"
+  local needle="${2-}"
+  local msg="${3-}"
+
+  if [ -z "${needle:+x}" ]; then
+    return 0;
+  fi
+
+  if [ -z "${haystack##*$needle*}" ]; then
+    return 0
+  else
+    [ "${#msg}" -gt 0 ] && log_failure "$haystack doesn't contain $needle :: $msg" || true
+    return 1
+  fi
+}
+
+assert_not_contain() {
+  local haystack="$1"
+  local needle="${2-}"
+  local msg="${3-}"
+
+  if [ -z "${needle:+x}" ]; then
+    return 0;
+  fi
+
+  if [ "${haystack##*$needle*}" ]; then
+    return 0
+  else
+    [ "${#msg}" -gt 0 ] && log_failure "$haystack contains $needle :: $msg" || true
+    return 1
+  fi
+}
+
+assert_gt() {
+  local first="$1"
+  local second="$2"
+  local msg="${3-}"
+
+  if [[ "$first" -gt  "$second" ]]; then
+    return 0
+  else
+    [ "${#msg}" -gt 0 ] && log_failure "$first > $second :: $msg" || true
+    return 1
+  fi
+}
+
+assert_ge() {
+  local first="$1"
+  local second="$2"
+  local msg="${3-}"
+
+  if [[ "$first" -ge  "$second" ]]; then
+    return 0
+  else
+    [ "${#msg}" -gt 0 ] && log_failure "$first >= $second :: $msg" || true
+    return 1
+  fi
+}
+
+assert_lt() {
+  local first="$1"
+  local second="$2"
+  local msg="${3-}"
+
+  if [[ "$first" -lt  "$second" ]]; then
+    return 0
+  else
+    [ "${#msg}" -gt 0 ] && log_failure "$first < $second :: $msg" || true
+    return 1
+  fi
+}
+
+assert_le() {
+  local first="$1"
+  local second="$2"
+  local msg="${3-}"
+
+  if [[ "$first" -le  "$second" ]]; then
+    return 0
+  else
+    [ "${#msg}" -gt 0 ] && log_failure "$first <= $second :: $msg" || true
+    return 1
+  fi
+}
diff --git a/test/test-xml.sh b/test/test-xml.sh
new file mode 100755
index 0000000..65ade92
--- /dev/null
+++ b/test/test-xml.sh
@@ -0,0 +1,36 @@
+#!/usr/bin/env bash
+TESTDIR=$(dirname $0)
+ASSERTSH=${TESTDIR}/assert.sh
+set -e
+. ${ASSERTSH}
+
+TEXTS=6
+I5_FILE=target/dnb18.i5.xml
+if [ ! -f "$I5_FILE" ]; then
+  log_failure "File $I5_FILE does not exist"
+  exit 1
+fi
+
+
+observed=$(xmlstarlet  sel --net -t -v "count(//idsText)"  $I5_FILE)
+
+if $(assert_eq "$observed" "$TEXTS"); then
+  log_success "$I5_FILE contains $TEXTS idsText elements"
+else
+  log_failure "$I5_FILE does not contain $TEXTS idsText elements"
+fi
+
+
+observed=$(xmlstarlet sel --net -t -v "count(/idsCorpus/idsDoc/idsText/idsHeader/fileDesc/sourceDesc/biblStruct/monogr/h.author[normalize-space(.)])"  $I5_FILE)
+if $(assert_eq "$observed" "$TEXTS"); then
+  log_success "$I5_FILE contains $TEXTS non-empty h.author elements"
+else
+  log_failure "$I5_FILE does not contain $TEXTS non-empty h.author elements"
+fi
+
+observed=$(xmlstarlet sel --net -t -v "count(/idsCorpus/idsDoc/idsText/idsHeader/fileDesc/sourceDesc/biblStruct/monogr/h.author[contains(., '[')])"  $I5_FILE)
+if $(assert_eq "$observed" "0"); then
+  log_success "authors do not contain []"
+else
+  log_failure "authors contain []"
+fi
\ No newline at end of file