solr-functions

#!/usr/bin/env bash

SOLR_HOST=${SOLR_HOST:-http://localhost:${PORT:-8983}}

check_core() {
  LOCAL_CORE=$1
  LOCAL_URL=$(printf "%s/solr/admin/cores?action=STATUS&core=%s" "$SOLR_HOST" "$LOCAL_CORE")
  curl -s "$LOCAL_URL" | jq .status | grep "\"$LOCAL_CORE\":" | grep -c -P '{$'
}

create_core() {
  LOCAL_CORE=$1
  echo "creating Solr index: $LOCAL_CORE at $SOLR_HOST"
  curl -s "$SOLR_HOST/solr/admin/cores?action=CREATE&name=$LOCAL_CORE&configSet=_default"
}

rename_core() {
  LOCAL_FROM=$1
  LOCAL_TO=$2
  echo "rename Solr index: $LOCAL_FROM to $LOCAL_TO"
  curl -s "$SOLR_HOST/solr/admin/cores?action=RENAME&core=$LOCAL_FROM&other=$LOCAL_TO"
}

swap_cores() {
  LOCAL_FROM=$1
  LOCAL_TO=$2
  echo "Swap Solr indexes $LOCAL_FROM and $LOCAL_TO"
  curl -s "$SOLR_HOST/solr/admin/cores?action=SWAP&core=$LOCAL_FROM&other=$LOCAL_TO"
}

prepare_schema() {
  LOCAL_CORE=$1
  SCHEMA_URL="$SOLR_HOST/api/cores/$LOCAL_CORE/schema"

  echo "prepare_schema $LOCAL_CORE"

  HAS_PROPER_SNI=$(curl -is "$SCHEMA_URL/dynamicfields/*_sni" | grep -c '"type":"string_big"')
  echo "Does $LOCAL_CORE have proper *_sni field definition? Answer: $HAS_PROPER_SNI"
  if [[ "$HAS_PROPER_SNI" -eq 0 ]]; then
    HAS_FIELD=$(curl -is "$SCHEMA_URL/dynamicfields/*_sni" | head -1 | grep -c "200 OK")
    echo "Does $LOCAL_CORE have  *_sni field definition? Answer: $HAS_FIELD"
    if [[ $HAS_FIELD -eq 1 ]]; then
      echo "Delete *_sni field definition from $LOCAL_CORE."
      curl -X POST -H 'Content-type:application/json' --data-binary '{
        "delete-dynamic-field":{"name":"*_sni",}
      }' "$SCHEMA_URL"
    fi

    HAS_TYPE=$(curl -is "$SCHEMA_URL/fieldtypes/string_big" | head -1 | grep -c "200 OK")
    if [[ "$HAS_TYPE" -eq 1 ]]; then
      echo "Delete string_big field type definition from $LOCAL_CORE."
      curl -X POST -H 'Content-type:application/json' --data-binary '{
        "delete-field-type":{"name":"string_big",}
      }' "$SCHEMA_URL"
    fi

    echo "Create string_big field type definition in $LOCAL_CORE."
    curl -X POST -H 'Content-type:application/json' --data-binary '{
      "add-field-type" : {
        "name":"string_big",
        "class":"solr.TextField",
        "positionIncrementGap":"100",
        "sortMissingLast":true,
        "omitNorms":true,
        "analyzer" : {
          "tokenizer":{
            "class":"solr.KeywordTokenizerFactory"
          },
        }
      }
    }' "$SCHEMA_URL"

    echo "Create *_sni dynamic field definition in $LOCAL_CORE."
    curl -X POST -H 'Content-type:application/json' --data-binary '{
      "add-dynamic-field":{
        "name":"*_sni",
        "type":"string_big",
        "stored":true,
        "indexed":false,
        "docValues":false,
        "multiValued":false,
    }}' "$SCHEMA_URL"
  fi

  NUMBER_OF_COPY_FIELD=$(curl -is "$SCHEMA_URL/copyfields?source.fl=*_ss" | grep -c '"source":"\*_ss"')
  echo "NUMBER_OF_COPY_FIELD: $NUMBER_OF_COPY_FIELD"
  if [[ "$NUMBER_OF_COPY_FIELD" -eq 0 ]]; then
    echo "add *_ss copyfield definition to $LOCAL_CORE"
    # add _ss
    # <copyField source="*_ss" dest="_text_"/>
    curl -X POST -H 'Content-type:application/json' --data-binary '{
      "add-copy-field":{
        "source":"*_ss",
        "dest":"_text_"
    }}' "$SCHEMA_URL"
  elif [[ "$NUMBER_OF_COPY_FIELD" -gt 1 ]]; then
    while [[ "$NUMBER_OF_COPY_FIELD" -gt 1 ]]; do
      echo "delete extra copy fields"
      curl -s -X POST -H 'Content-type:application/json' --data-binary '{
        "delete-copy-field":{ "source":"*_ss", "dest":"_text_" }
      }' "$SCHEMA_URL"
      NUMBER_OF_COPY_FIELD=$(curl -is "$SCHEMA_URL/copyfields?source.fl=*_ss" | grep -c '"source":"\*_ss"')
    done
  fi

  HAS_PROPER_TT=$(curl -is "$SCHEMA_URL/dynamicfields/*_tt" | grep -c '"type":"text_general"')
  echo "Does $LOCAL_CORE have proper *_tt field definition? Answer: $HAS_PROPER_TT"
  if [[ "$HAS_PROPER_TT" -eq 0 ]]; then
    # add _tt
    # <dynamicField name="*_tt" type="text_general" indexed="true" stored="false"/>
    echo "Create *_tt dynamic field definition in $LOCAL_CORE."
    curl -X POST -H 'Content-type:application/json' --data-binary '{
      "add-dynamic-field":{
        "name":"*_tt",
        "type":"text_general",
        "stored":false,
        "indexed":true,
    }}' "$SCHEMA_URL"
  fi
}

status() {
  echo "Solr index status at $SOLR_HOST"
  echo 
  curl -s "$SOLR_HOST/solr/admin/cores?action=STATUS" \
    | jq . \
    | grep instanceDir \
    | head -1 \
    | awk '{print $2}' \
    | sed -E 's,^"(.*solr).+$,\1,' \
    | xargs -I '@@' echo "Solr directory: @@"

  FORMAT="%-20s | %-15s | %10s | %10s | %19s\n"
  printf "$FORMAT" "core" "location" "nr of docs" "size" "last modified"
  printf "$FORMAT" "...................." "..............." ".........." ".........." "..................."
  curl -s "$SOLR_HOST/solr/admin/cores?action=STATUS" \
    | jq '.status[] | [.name, .instanceDir, .index.numDocs, .index.size, .index.lastModified]' \
    | sed -r 's,".*/server/solr/(.+)",\1,' \
    | paste - - - - - - - \
    | sed -r 's,(\[|]|\t|\,),,g' \
    | sed -r 's,([0-9])T([0-9]),\1 \2,g' \
    | sed -r 's,\.[0-9]{3}Z,,g' \
    | xargs printf "$FORMAT"
  exit 1
}

purge_core() {
  LOCAL_CORE=$1
  SOLR_DB_URL="$SOLR_HOST/solr/$LOCAL_CORE"
  echo "Delete records in $SOLR_DB_URL"
  curl -s "$SOLR_DB_URL/update" \
       -H "Content-type: text/xml" \
       --data-binary '<delete><query>*:*</query></delete>' \
       > /dev/null
}

optimize_core() {
  LOCAL_CORE=$1
  SOLR_DB_URL="$SOLR_HOST/solr/$LOCAL_CORE"
  echo "Optimizing $SOLR_DB_URL"
  curl -s "$SOLR_DB_URL/update?optimize=true" \
       -H 'Content-type: text/xml' \
       --data-binary '<commit/>' \
       > /dev/null
}

purge_and_exit() {
  LOCAL_CORE=$1
  if [[ -n "$LOCAL_CORE" ]]; then
    purge_core "$LOCAL_CORE"
    optimize_core "$LOCAL_CORE"
  else
    echo "You should give the name with --core parameter"
  fi
  exit 1
}

store_fields() {
  SOLR_URL=$1
  OUTPUT_DIR=$2
  echo "Store fields into $OUTPUT_DIR/solr-fields.json"
  echo "curl -s \"$SOLR_URL/admin/luke?wt=json\" | jq '.fields | keys_unsorted'"
  curl -s "$SOLR_URL/admin/luke?wt=json" \
    | jq '.fields | keys_unsorted' \
    > "$OUTPUT_DIR/solr-fields.json"
}

extract_host() {
  URL=$1
  HOST=$(echo "$URL" | grep -oP "^https?://[^/]+" || true)
  echo "$HOST"
}

extract_core() {
  URL=$1
  CORE=$(echo "$URL" | grep -oP "[^/]+$" || true)
  echo "$CORE"
}