Update NodeNorm and NameRes to Babel 2025jul22

helm/name-lookup/Chart.yaml

@@ -14,11 +14,11 @@ type: application
 
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
-version: 0.5.2
+version: 0.6.0
 
 # This is the version number of the application being deployed. This version number should be
 # incremented each time you make changes to the application.
-# 
+#
 # NameRes versions are based on the version of the NameRes Docker image concatenated with the
 # Babel release date.
-appVersion: 1.5.2_2025sep1
+appVersion: 1.7.0_2026jul22

helm/name-lookup/ncats-images-meta.yaml

@@ -1,10 +1,10 @@
 nameLookup:
   image: ghcr.io/ncatstranslator/nameresolution
-  version: v1.5.2
+  version: v1.7.0
 
 solr:
   image: solr
-  version: "9.1"
+  version: "9.10"
 
 renciPythonImage:
   image: ghcr.io/translatorsri/renci-python-image

helm/name-lookup/templates/restore-job.yaml

@@ -23,10 +23,8 @@ spec:
                   requests:
                     storage: {{ .Values.blocklist.storage }}
         {{ end }}
-      # Because of the way in which the Restore Job works -- by creating fields in
-      # the Solr database, including copy fields -- it can no longer be re-run on
-      # failure. Instead, it should fail so we can look at the error output and
-      # figure out what to do next.
+      # The restore job now only deletes blocklisted CURIEs (the schema and data
+      # come baked into the backup), so it is idempotent and safe to re-run.
       restartPolicy: Never
       containers:
         - name: restore-container

helm/name-lookup/templates/scripts-config-map.yaml

@@ -9,21 +9,52 @@ data:
     #!/bin/bash
     set -xa
     DATA_DIR="/var/solr/data"
-    BACKUP_NAME="snapshot.backup"
-    BACKUP_ZIP="${BACKUP_NAME}.tar.gz"
+    CORE_DIR="${DATA_DIR}/name_lookup"
+    BACKUP_ZIP="${DATA_DIR}/snapshot.backup.tar.gz"
     BACKUP_URL="{{ .Values.dataUrl }}"
+    STAMP="${CORE_DIR}/.dataUrl"
 
-    rm -rf $DATA_DIR/*
+    # An empty dataUrl means "leave the volume alone" -- keep whatever core is already
+    # there (or none). This is the switch for "do not (re)load", same as before.
+    if [ -z "${BACKUP_URL}" ]; then
+      echo "dataUrl is empty; leaving any existing core untouched."
+      exit 0
+    fi
+
+    # The backup is a self-contained Solr core (config + schema + index). Skip the
+    # download only if the extracted core came from THIS dataUrl: a pod restart with an
+    # unchanged dataUrl keeps the good core (translator-devops#609), but a new release
+    # (changed dataUrl) reloads instead of silently serving the old index on a PVC that
+    # outlived the version bump. A core with no stamp (old logic, or hand-restored)
+    # reloads once so the served data matches the configured dataUrl.
+    if [ -f "${CORE_DIR}/core.properties" ] && [ "$(cat "${STAMP}" 2>/dev/null)" = "${BACKUP_URL}" ]; then
+      echo "Core already present from ${BACKUP_URL}; skipping download."
+      exit 0
+    fi
+
+    # Otherwise start clean and download.
+    rm -rf ${DATA_DIR}/*
 
     # Download the file with retries if the download fails or if there is an existing file we can continue.
-    wget -c --progress=dot:giga --tries=10 --waitretry=10 --timeout=60 --read-timeout=30 -O $DATA_DIR/$BACKUP_ZIP $BACKUP_URL 
+    wget -c --progress=dot:giga --tries=10 --waitretry=10 --timeout=60 --read-timeout=30 -O ${BACKUP_ZIP} ${BACKUP_URL}
 
-    cd $DATA_DIR
-    tar -xf $DATA_DIR/$BACKUP_ZIP -C $DATA_DIR
-    rm $DATA_DIR/$BACKUP_ZIP
+    # Extract the core into the Solr home. Solr auto-discovers name_lookup on startup,
+    # with its schema and config baked in -- nothing else to set up.
+    tar -xf ${BACKUP_ZIP} -C ${DATA_DIR}
+    rm ${BACKUP_ZIP}
+
+    # Record which backup this core came from, so the next pod start can tell a restart
+    # (skip) from a version bump (reload). Written after a successful extract so a
+    # failed download never leaves a stamp that would suppress the retry.
+    printf '%s' "${BACKUP_URL}" > "${STAMP}"
 
   restore.sh: |-
     #!/bin/sh
+    #
+    # The backup already contains the loaded index AND the schema/config, so there
+    # is no collection to create, no replication restore, and no fields to add. The
+    # only post-restore work is deleting blocklisted CURIEs, if a blocklist is set.
+    # (This makes the job idempotent and safe to re-run -- see NameResolution#185.)
 
     BLOCKLIST_DIR="/var/blocklist"
     BLOCKLIST_CHUNK_SIZE=500
@@ -46,169 +77,19 @@ data:
     COLLECTION_NAME="name_lookup"
     SOLR_SERVER=http://{{ include "name-lookup.fullname" . }}-solr-svc:{{ .Values.solr.service.port }}
 
-    # liveliness check
-
-    HEALTH_ENDPOINT=http://{{ include "name-lookup.fullname" . }}-solr-svc:{{ .Values.solr.service.port }}/solr/admin/cores?action=STATUS
+    # Wait for Solr and the auto-discovered core to be ready.
+    HEALTH_ENDPOINT=${SOLR_SERVER}/solr/${COLLECTION_NAME}/admin/ping
     response=$(wget --spider --server-response ${HEALTH_ENDPOINT} 2>&1 | grep "HTTP/" | awk '{ print $2 }') >&2
     until [ "$response" = "200" ]; do
       response=$(wget --spider --server-response ${HEALTH_ENDPOINT} 2>&1 | grep "HTTP/" | awk '{ print $2 }') >&2
-      echo "  -- SOLR  is unavailable - sleeping"
+      echo "  -- SOLR is unavailable - sleeping"
       sleep 3
     done
-
-    # solr is ready Now we create collection if it doesn't exist
-
-    EXISTS=$(wget -O - ${SOLR_SERVER}/solr/admin/collections?action=LIST | grep name_lookup)
-
-    # create collection / shard
-    if [ -z "$EXISTS" ]
-    then
-      wget -O- ${SOLR_SERVER}/solr/admin/collections?action=CREATE'&'name=${COLLECTION_NAME}'&'numShards=1'&'replicationFactor=1
-      sleep 3
-    fi
-
-    # Setup fields for search
-    wget --post-data '{"set-user-property": {"update.autoCreateFields": "false"}}' \
-        --header='Content-Type:application/json' \
-        -O- ${SOLR_SERVER}/solr/${COLLECTION_NAME}/config
-    sleep 1
-
-    # Restore data
-    BACKUP_NAME="backup"
-    CORE_NAME=${COLLECTION_NAME}_shard1_replica_n1
-    RESTORE_URL=${SOLR_SERVER}/solr/${CORE_NAME}/replication?command=restore'&'location=/var/solr/data/var/solr/data/'&'name=${BACKUP_NAME}
-    wget -O - $RESTORE_URL
-    sleep 10
-    RESTORE_STATUS=$(wget -q -O - ${SOLR_SERVER}/solr/${CORE_NAME}/replication?command=restorestatus 2>&1 | grep "success") >&2
-    echo "Restore status: " $RESTORE_STATUS
-    until [ ! -z $RESTORE_STATUS ] ; do
-      echo "restore not done , probably still loading. Note: if this takes too long please check solr health"
-      RESTORE_STATUS=$(wget -O - ${SOLR_SERVER}/solr/${CORE_NAME}/replication?command=restorestatus 2>&1 | grep "success") >&2
-      sleep 10
-    done
-    echo "restore done"
-    wget --post-data '{
-              "add-field-type" : {
-                  "name": "LowerTextField",
-                  "class": "solr.TextField",
-                  "positionIncrementGap": "100",
-                  "analyzer": {
-                      "tokenizer": {
-                          "class": "solr.StandardTokenizerFactory"
-                      },
-                      "filters": [{
-                          "class": "solr.LowerCaseFilterFactory"
-                      }]
-                  }
-              }}' \
-        --header='Content-Type:application/json' \
-        -O- ${SOLR_SERVER}/solr/${COLLECTION_NAME}/schema
-    sleep 1
-    # exactish type taken from https://stackoverflow.com/a/29105025/27310
-    wget --post-data '{
-              "add-field-type" : {
-                  "name": "exactish",
-                  "class": "solr.TextField",
-                  "analyzer": {
-                      "tokenizer": {
-                          "class": "solr.KeywordTokenizerFactory"
-                      },
-                      "filters": [{
-                          "class": "solr.LowerCaseFilterFactory"
-                      }]
-                  }
-              }}' \
-        --header='Content-Type:application/json' \
-        -O- ${SOLR_SERVER}/solr/${COLLECTION_NAME}/schema
-    sleep 1
-    wget --post-data '{
-                "add-field": [
-                    {
-                        "name":"names",
-                        "type":"LowerTextField",
-                        "stored": true,
-                        "multiValued": true
-                    },
-                    {
-                        "name":"names_exactish",
-                        "type":"exactish",
-                        "indexed":true,
-                        "stored":true,
-                        "multiValued":true
-                    },
-                    {
-                        "name":"curie",
-                        "type":"string",
-                        "stored":true
-                    },
-                    {
-                        "name": "preferred_name",
-                        "type": "LowerTextField",
-                        "stored": true
-                    },
-                    {
-                        "name": "preferred_name_exactish",
-                        "type": "exactish",
-                        "indexed": true,
-                        "stored": false,
-                        "multiValued": false
-                    },
-                    {
-                        "name": "types",
-                        "type": "string",
-                        "stored": true,
-                        "multiValued": true
-                    },
-                    {
-                        "name": "shortest_name_length",
-                        "type": "pint",
-                        "stored": true
-                    },
-                    {
-                        "name": "curie_suffix",
-                        "type": "plong",
-                        "docValues": true,
-                        "stored": true,
-                        "required": false,
-                        "sortMissingLast": true
-                    },
-                    {
-                        "name": "taxa",
-                        "type": "string",
-                        "stored": true,
-                        "multiValued": true
-                    },
-                    {
-                        "name": "clique_identifier_count",
-                        "type": "pint",
-                        "stored": true
-                    }
-                ]
-            }' \
-          --header='Content-Type:application/json' \
-          -O- ${SOLR_SERVER}/solr/${COLLECTION_NAME}/schema
-    sleep 1
-    wget --post-data '{
-              "add-copy-field" : {
-                "source": "names",
-                "dest": "names_exactish"
-              }}' \
-        --header='Content-Type:application/json' \
-        -O- ${SOLR_SERVER}/solr/${COLLECTION_NAME}/schema
-    wget --post-data '{
-              "add-copy-field" : {
-                "source": "preferred_name",
-                "dest": "preferred_name_exactish"
-              }}' \
-        --header='Content-Type:application/json' \
-        -O- ${SOLR_SERVER}/solr/${COLLECTION_NAME}/schema
-    sleep 1
+    echo "Solr core ${COLLECTION_NAME} is up."
 
     # Delete the blocklist terms from the Solr database.
     {{ if .Values.blocklist.url }}
 
-    BLOCKLIST_DIR=/var/blocklist
-
     # Split the blocklist into files of 500 CURIEs each.
     rm -rf ${BLOCKLIST_DIR}/blocklist_*
     split -l ${BLOCKLIST_CHUNK_SIZE} ${BLOCKLIST_DIR}/blocklist.txt ${BLOCKLIST_DIR}/blocklist_

helm/name-lookup/templates/secrets.yaml

@@ -1,3 +1,12 @@
+{{- /*
+  Fail fast rather than quietly deploying a blocklist that cannot authenticate: if the
+  blocklist is switched on (blocklist.url set) there must be a token to go with it. The
+  token is looked up nil-safely here so that a values file which drops the `secrets` map
+  entirely still gets this message instead of a raw nil-pointer error.
+*/ -}}
+{{- if and .Values.blocklist.url (not (.Values.blocklist.secrets).github_personal_access_token) }}
+{{- fail "blocklist.url is set but blocklist.secrets.github_personal_access_token is empty. Set the token, or clear blocklist.url to turn the blocklist off. (If the blocklist URL needs no auth, set any non-empty placeholder.)" }}
+{{- end }}
 {{ if .Values.blocklist.secrets.github_personal_access_token }}
 apiVersion: v1
 kind: Secret

helm/name-lookup/templates/solr-deployment.yaml

@@ -44,9 +44,10 @@ spec:
       containers:
         - name: {{ .Chart.Name }}
           image: "{{ .Values.solr.image.repository }}:{{ .Values.solr.image.tag }}"
+          # Standalone mode: no -DzkRun/ZooKeeper. The backup is a self-contained
+          # core that the download init-container extracts into /var/solr/data, so
+          # Solr auto-discovers name_lookup on startup with its schema baked in.
           args:
-          - '-DzkRun'
-          - '-DzkClientTimeout=5000'
           - '-q'
           - '-Dlog4j2.disable.jmx=true'
           - '-Dlog4j2.formatMsgNoLookups=true'
@@ -57,6 +58,11 @@ spec:
               value: {{ include "name-lookup.fullname" . }}-solr-svc
             - name: GC_TUNE
               value: {{ .Values.solr.gc | quote }}
+            {{- if .Values.solr.gc_log }}
+            # Overrides Solr's default file-based GC logging (see solr.gc_log in values.yaml).
+            - name: GC_LOG_OPTS
+              value: {{ .Values.solr.gc_log | quote }}
+            {{- end }}
               # "-XX:-UseLargePages -XX:+UseG1GC -XX:MaxGCPauseMillis=500 -XX:+UnlockExperimentalVMOptions -XX:G1MaxNewSizePercent=30 -XX:G1NewSizePercent=5 -XX:G1HeapRegionSize=32M -XX:InitiatingHeapOccupancyPercent=70"
               # "-XX:-UseLargePages -XX:+UseG1GC -XX:MaxGCPauseMillis=100 -XX:+UnlockExperimentalVMOptions -XX:G1MaxNewSizePercent=20 -XX:G1NewSizePercent=5 -XX:G1HeapRegionSize=32M -XX:InitiatingHeapOccupancyPercent=20"
               # "-XX:+UseG1GC -XX:MaxGCPauseMillis=1000 -XX:+UnlockExperimentalVMOptions -XX:G1MaxNewSizePercent=40 -XX:G1NewSizePercent=5 -XX:G1HeapRegionSize=32M -XX:InitiatingHeapOccupancyPercent=90"

helm/name-lookup/templates/web-deployment.yaml

@@ -40,11 +40,17 @@ spec:
             value: "{{ .Values.app.otel.jaegerHost }}"
           - name: "JAEGER_PORT"
             value: "{{ .Values.app.otel.jaegerPort }}"
+          # Logging level.
+          - name: LOGLEVEL
+            value: "{{ .Values.app.logLevel }}"
           # Babel version information.
           - name: BABEL_VERSION
             value: "{{ .Values.data.babelVersion }}"
           - name: BABEL_VERSION_URL
             value: "{{ .Values.data.babelVersionURL }}"
+          # Biolink Model tag.
+          - name: BIOLINK_MODEL_TAG
+            value: "{{ .Values.data.biolinkModelTag }}"
           ports:
             - name: http
               containerPort: {{ .Values.webServer.port }}