update nypl core version to 2.21

README.md

@@ -10,7 +10,7 @@ For local development, it's easiest to just use local node binaries:
 
 ```
 nvm use; npm i
-nvm use; ENV=qa npm start
+nvm use; ENV=qa LOCAL=true npm start
 ```
 
 Note that when developing locally, if connecting to a IP ACL protected index (a practice we're currently deprecating), you may need to [add your IP to the access control policy of the relevant ES domain](https://github.com/NYPL/aws/blob/b5c0af0ec8357af9a645d8b47a5dbb0090966071/common/elasticsearch.md#2-make-the-domain-public-restrict-by-ip). If your IP has not been authorized, you will see errors such as the following in the application logs:

app.js

@@ -1,10 +1,11 @@
 const express = require('express')
 
-const esClient = require('./lib/es-client')
+const esClient = require('./lib/elasticsearch/client')
 const loadConfig = require('./lib/load-config')
 const { preflightCheck } = require('./lib/preflight_check')
 
 const swaggerDocs = require('./swagger.v1.1.x.json')
+
 const pjson = require('./package.json')
 
 const app = express()
@@ -52,6 +53,8 @@ app.init = async () => {
   app.get('/api/v0.1/discovery/swagger', function (req, res) {
     res.send(swaggerDocs)
   })
+
+  return app
 }
 
 app.start = async () => {

config/production.env

@@ -1,6 +1,6 @@
-# Currently, this is our ES 5.3 for Prod:
-ENCRYPTED_ELASTICSEARCH_URI=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAMMwgcAGCSqGSIb3DQEHBqCBsjCBrwIBADCBqQYJKoZIhvcNAQcBMB4GCWCGSAFlAwQBLjARBAyUXsOWkXt7ZqqD7uQCARCAfMIrLEGKd51lkX/hzWYlUms176JOee1HhdP/WIf5nd3xTwXJz1m4yOWwDsefoJEL4sez/2ZP43MVds+1lcM3PwkHMyN01sa2xpwVDrT68A/f22bDMtp6A9BmJqr7VBB5oq2j4FoXu2+0uaQWQRWJ0Ns+vQ6nbiR5V4Vv8zE=
-ENCRYPTED_RESOURCES_INDEX=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAHIwcAYJKoZIhvcNAQcGoGMwYQIBADBcBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDF+IpeiMrtx/UveMsgIBEIAvIolGVXJPBMseyyep3/fyf/qP+KdATwfxT6Gw4dnatbiL5609NBeyyLNIprEM90g=
+ENCRYPTED_ELASTICSEARCH_URI=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAJYwgZMGCSqGSIb3DQEHBqCBhTCBggIBADB9BgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDFWw8ECX9Pz81z0kvAIBEIBQGec9PCpwuvEgLH6imhqP6tx1fj8Vlf2ZipnUy06jzmpE262Qvk9LPAq7sIYPVkTCZctwilwcU9oC6yxasVoUlK87la77v03CeZsPIDwciFY=
+ENCRYPTED_RESOURCES_INDEX=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAHIwcAYJKoZIhvcNAQcGoGMwYQIBADBcBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDNsBsURm/qBI5aTUvwIBEIAvKqhQSnNEv4nla3EkKpcsMnl5IvCSgRTYFtTZYeo8WOZcK3CeHQSlaz3Ccuyat4E=
+ENCRYPTED_ELASTICSEARCH_API_KEY=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAJ4wgZsGCSqGSIb3DQEHBqCBjTCBigIBADCBhAYJKoZIhvcNAQcBMB4GCWCGSAFlAwQBLjARBAyPOPaQCBbvKQhJoPQCARCAV2TlWlRh+xKnCegpprEQgfldZGcVW48RND0LVd/pQpVTJnRTtbCpP7damT7k8ziJVdWZ3jsfs5fw5YnKc/EIQ1M//DRUzOJL98ir5LTTxE7QhflKDtUY+Q==
 
 ENCRYPTED_SCSB_URL=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAHwwegYJKoZIhvcNAQcGoG0wawIBADBmBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDKPFC8wFkVM5CyT6VQIBEIA5m4eLBkpChRA//ZNEWsRqIDGZmevb/thzI03a0NiAW6VfybSAYpFthh+bj/yAk1VEEBF6r1T4A2GP
 ENCRYPTED_SCSB_API_KEY=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAIMwgYAGCSqGSIb3DQEHBqBzMHECAQAwbAYJKoZIhvcNAQcBMB4GCWCGSAFlAwQBLjARBAw8tglwVzGKBduDD9wCARCAP4biSz13FvZVHyQ8LKCb0+uLcKUKmzWqC5abVJI0kTmQJvjr9ViHsuP9/qj94Y8E7K96sb+fn0+HZk8So6CssA==
@@ -10,7 +10,7 @@ NYPL_OAUTH_URL=https://isso.nypl.org/
 ENCRYPTED_NYPL_OAUTH_ID=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAGswaQYJKoZIhvcNAQcGoFwwWgIBADBVBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDMLKVUQA58B6vprNcAIBEIAoaz0lI9EL2M9NyTuEwT8JDmPBt6aXfMiFs027DEuwsCN0wS0qWeFL1g==
 ENCRYPTED_NYPL_OAUTH_SECRET=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAIcwgYQGCSqGSIb3DQEHBqB3MHUCAQAwcAYJKoZIhvcNAQcBMB4GCWCGSAFlAwQBLjARBAyWz91LOP2YP5fg0q0CARCAQ9inO9SV1M8R0Pkkx84r7UdwlU1FxfXvIjk/z6Qs81KBAVELhby2iD5LawQyDrR9tjhuMbotS6QnydwwMR/p8+qJXHI=
 
-NYPL_CORE_VERSION=v2.20
+NYPL_CORE_VERSION=v2.21
 
 LOG_LEVEL=info
 FEATURES=on-site-edd

config/qa.env

@@ -1,6 +1,6 @@
-# Currently, this is our ES 5.3 for QA:
-ENCRYPTED_ELASTICSEARCH_URI=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAALswgbgGCSqGSIb3DQEHBqCBqjCBpwIBADCBoQYJKoZIhvcNAQcBMB4GCWCGSAFlAwQBLjARBAyWLvUSzA/IAQCHl0MCARCAdNpF/Z1VJESwJ7hcwo/BqZz2mTDPA9NAPQ4zuPLsItz9A2lfHaP03bPuo9nq8VP5AKLOa4zPL0VoBmwEjj9qCCb+LSpQ3m+OoyM3BxG98/qYEcwXXOa8+0fH1x5asVrup/YICJdeD6jOewxttzzxCCGXEklL
-ENCRYPTED_RESOURCES_INDEX=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAHIwcAYJKoZIhvcNAQcGoGMwYQIBADBcBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDI88/9macimvmLyWCAIBEIAvMUOAtF2Miq+8u7/A9fzBz57LavqkeLJmv8dd7WQzdA9lhqPkjUK0pzYtxsPe6Nk=
+ENCRYPTED_ELASTICSEARCH_URI=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAJYwgZMGCSqGSIb3DQEHBqCBhTCBggIBADB9BgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDMIkDoQ9C/cCDCAq1wIBEIBQ+L3OgUGeOW9rs1CWkhpBjwM4LbbVRFIWedqew4UXIeSNMJ8cO9SNe4YGCUIoKwCDYt7W7ip3VtDRRRMVvz6QJw+Eg8ugTMVs2pbNFGNvaAQ=
+ENCRYPTED_RESOURCES_INDEX=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAHIwcAYJKoZIhvcNAQcGoGMwYQIBADBcBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDGjIeRRI3cN5LwfcmgIBEIAvkB3K4sxagpqspNa3b3BnJQgbw6Ic0jDhuXzk8gmrCiGNMR90YjTZmyLj9aVUN3M=
+ENCRYPTED_ELASTICSEARCH_API_KEY=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAJ4wgZsGCSqGSIb3DQEHBqCBjTCBigIBADCBhAYJKoZIhvcNAQcBMB4GCWCGSAFlAwQBLjARBAx+kryf2KUmGdBYD9sCARCAV3ygz3eXIdq8JX/wpG9JRWlTNMRcpNE1qT0zNlN4t+ZvXEoedLQa/3p1YjgHw06GIAdA9xtkMV4eH9a1K8uCvjP8XxxNKekcMj59TlResnu9QF3r7pGXuQ==
 
 ENCRYPTED_SCSB_URL=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAH8wfQYJKoZIhvcNAQcGoHAwbgIBADBpBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDBKllElmWYLxGOGopQIBEIA8JJyKde/8m8iCJGKR5D8HoTJhXHeyvw9eIDeuUNKiXLfJwoVz+PDAZSxkCQtM9O91zGhXbe3l6Bk1RlYJ
 ENCRYPTED_SCSB_API_KEY=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAGMwYQYJKoZIhvcNAQcGoFQwUgIBADBNBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDNw8KXkyN8HvtjAX0gIBEIAgX+XG2fxTj6kSchrd/dfHB05KU5pkT0LtPxUTuNCXoLc=
@@ -10,9 +10,9 @@ NYPL_OAUTH_URL=https://isso.nypl.org/
 ENCRYPTED_NYPL_OAUTH_ID=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAGswaQYJKoZIhvcNAQcGoFwwWgIBADBVBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDMLKVUQA58B6vprNcAIBEIAoaz0lI9EL2M9NyTuEwT8JDmPBt6aXfMiFs027DEuwsCN0wS0qWeFL1g==
 ENCRYPTED_NYPL_OAUTH_SECRET=AQECAHh7ea2tyZ6phZgT4B9BDKwguhlFtRC6hgt+7HbmeFsrsgAAAIcwgYQGCSqGSIb3DQEHBqB3MHUCAQAwcAYJKoZIhvcNAQcBMB4GCWCGSAFlAwQBLjARBAyWz91LOP2YP5fg0q0CARCAQ9inO9SV1M8R0Pkkx84r7UdwlU1FxfXvIjk/z6Qs81KBAVELhby2iD5LawQyDrR9tjhuMbotS6QnydwwMR/p8+qJXHI=
 
-NYPL_CORE_VERSION=v2.20
+NYPL_CORE_VERSION=v2.21
 
-LOG_LEVEL=debug
+LOG_LEVEL=info
 FEATURES=on-site-edd
 
 SEARCH_ITEMS_SIZE=3
@@ -22,3 +22,5 @@ HIDE_NYPL_SOURCE=
 
 BIB_HAS_VOLUMES_THRESHOLD=0.8
 BIB_HAS_DATES_THRESHOLD=0.8
+
+NAME_QUERIES=true

config/test.env

@@ -9,7 +9,7 @@ NYPL_OAUTH_URL=http://oauth.example.com
 ENCRYPTED_NYPL_OAUTH_ID=encrypted-nypl-oauth-id
 ENCRYPTED_NYPL_OAUTH_SECRET=encrypted-nypl-oauth-id
 
-NYPL_CORE_VERSION=v2.14
+NYPL_CORE_VERSION=v2.21
 
 LOG_LEVEL=error
 FEATURES=on-site-edd

lib/api-request.js

@@ -0,0 +1,75 @@
+/**
+* This class wraps request params to ease interpretting the query
+**/
+
+// Build regex pattern for matching a phrase fully enclosed in quotes (or smart quotes):
+const QUOTE_CHARS = '"\u201C\u201D\u201E\u201F\u2033\u2036'
+const IN_QUOTES_PATTERN = new RegExp(`^[${QUOTE_CHARS}][^${QUOTE_CHARS}]+[${QUOTE_CHARS}]$`)
+
+class ApiRequest {
+  static ADVANCED_SEARCH_PARAMS = ['title', 'subject', 'contributor']
+  static IDENTIFIER_NUMBER_PARAMS = ['isbn', 'issn', 'lccn', 'oclc']
+
+  constructor (params) {
+    this.params = params
+
+    // Make some substitutions for folks querying with "date:1997", etc
+    if (this.params.q) {
+      this.params.q = this.params.q
+        .replace(/date:/g, 'dateStartYear:')
+        .replace(/location:/g, 'locations:')
+        .replace(/subject:/g, 'subjectLiteral:')
+    }
+  }
+
+  /**
+  * Get array of params in the query that are also valid search-scopes:
+  */
+  advancedSearchParamsThatAreAlsoScopes () {
+    // Return search params that are also valid search_scope values
+    return ApiRequest.ADVANCED_SEARCH_PARAMS
+      .filter((key) => this.params[key])
+  }
+
+  hasKeyword () {
+    return !!this.params.q
+  }
+
+  hasSearch () {
+    return this.hasKeyword() || this.advancedSearchParamsThatAreAlsoScopes().length > 0
+  }
+
+  /**
+  * Returns true if search_scope matches one of the named scopes
+  **/
+  hasScope (scopes) {
+    return scopes.includes(this.params.search_scope)
+  }
+
+  /**
+  * Returns true if query is fully wrapped in quotes (or smart quotes)
+  **/
+  queryIsFullyQuoted () {
+    return IN_QUOTES_PATTERN.test(this.params.q)
+  }
+
+  /**
+  * Get keyword query without surrounding quotes (if fully quoted)
+  **/
+  querySansQuotes () {
+    return this.queryIsFullyQuoted()
+      ? this.params.q.substring(1, this.params.q.length - 1)
+      : this.params.q
+  }
+
+  hasIdentifierNumberParam () {
+    return Object.keys(this.params)
+      .some((userParam) => ApiRequest.IDENTIFIER_NUMBER_PARAMS.includes(userParam))
+  }
+
+  static fromParams (params) {
+    return new ApiRequest(params)
+  }
+}
+
+module.exports = ApiRequest

lib/availability_resolver.js

@@ -2,7 +2,7 @@ const scsbClient = require('./scsb-client')
 const logger = require('./logger')
 const ResourceSerializer = require('./jsonld_serializers').ResourceSerializer
 const NyplSourceMapper = require('discovery-store-models/lib/nypl-source-mapper')
-const { nonRecapItemStatusAggregation } = require('./es-client')
+const { nonRecapItemStatusAggregation } = require('./elasticsearch/client')
 const { deepValue, isInRecap } = require('./util')
 
 const ITEM_STATUSES = {
@@ -94,7 +94,9 @@ class AvailabilityResolver {
       logger.debug('_fixItemStatusAggregation: Skipping because no SCSB statuses provided')
       return Promise.resolve()
     }
-    const bnum = resp.hits.hits[0]._id
+    const bnum = resp.hits.hits[0]?._id
+    if (!bnum) return Promise.resolve()
+
     const { nyplSource } = NyplSourceMapper.instance().splitIdentifier(bnum)
 
     // Get total number of items:
@@ -135,7 +137,10 @@ class AvailabilityResolver {
             nonRecapStatusAggregation
           )
           logger.debug(`_fixItemStatusAggregation: Final aggregation: ${JSON.stringify(resp.aggregations.item_status._nested.buckets)}`)
-        }).catch((e) => console.error('Got error: ', e))
+        }).catch((e) => {
+          console.error('Got error: ', e)
+          return Promise.resolve()
+        })
     }
   }
 

lib/es-client.js → lib/elasticsearch/client.js

@@ -1,8 +1,8 @@
 const elasticsearch = require('@elastic/elasticsearch')
 const url = require('node:url')
-const { deepValue } = require('./util')
-const logger = require('./logger')
-const { IndexConnectionError } = require('./errors')
+const { deepValue } = require('../util')
+const logger = require('../logger')
+const { IndexConnectionError } = require('../errors')
 
 const clientWrapper = {}
 
@@ -11,16 +11,38 @@ const clientWrapper = {}
  */
 clientWrapper.esClient = function () {
   if (!this._esClient) {
-    // Parse ES connection string:
-    const { protocol, auth, host, port } = url.parse(process.env.ELASTICSEARCH_URI)
-    const [username, password] = auth ? auth.split(':') : []
-    const options = {
-      node: `${protocol}//${host}`,
-      port,
-      auth: { username, password }
+    // Parse ES connection string, which is likely multiple http base URIs
+    // separated by a comma:
+    const elasticUris = process.env.ELASTICSEARCH_URI.split(',')
+    const urisParsed = elasticUris.map((uri) => {
+      // Extract parts of the URI:
+      const { protocol, auth, host } = url.parse(uri)
+      const [username, password] = auth ? auth.split(':') : []
+      return {
+        protocol,
+        host,
+        username,
+        password
+      }
+    })
+    // Build ES client connection config:
+    const config = {}
+    config.nodes = urisParsed.map((uri) => `${uri.protocol}//${uri.host}`)
+
+    // Configure auth:
+    if (process.env.ELASTICSEARCH_API_KEY) {
+      // Auth with `apiKey`:
+      config.auth = { apiKey: process.env.ELASTICSEARCH_API_KEY }
+    } else if (urisParsed[0].username) {
+      // Auth with username, password:
+      config.auth = { username: urisParsed[0].username, password: urisParsed[0].password }
     }
-    logger.info(`Connecting to ES at ${host}:${port} ${username && password ? 'with creds' : 'w/out creds'}`)
-    this._esClient = new elasticsearch.Client(options)
+
+    // Log out some of the connection details for debugging purposes:
+    const authMethod = urisParsed[0].username ? 'with creds' : (process.env.ELASTICSEARCH_API_KEY ? 'with apiKey' : 'w/out creds')
+    logger.info(`Connecting to ES at ${urisParsed.map((u) => u.host).join(',')}/${process.env.RESOURCES_INDEX} ${authMethod}`)
+
+    this._esClient = new elasticsearch.Client(config)
   }
   return this._esClient
 }.bind(clientWrapper)
@@ -30,7 +52,7 @@ clientWrapper.esClient = function () {
  *  the configured ES
  */
 clientWrapper.search = function (body) {
-  logger.debug(`Performing ES search: ${JSON.stringify(body)}`)
+  logger.debug(`Performing ES search: ${JSON.stringify(body, null, 2)}`)
   return this.esClient().search({
     index: process.env.RESOURCES_INDEX,
     body

lib/elasticsearch/config.js

@@ -0,0 +1,103 @@
+// Configure search scopes:
+const SEARCH_SCOPES = {
+  all: {
+    fields: [
+      'title^5',
+      'title.folded^2',
+      'description.folded',
+      'subjectLiteral^2',
+      'subjectLiteral.folded',
+      'creatorLiteral^2',
+      'creatorLiteral.folded',
+      'contributorLiteral.folded',
+      'note.label.folded',
+      'publisherLiteral.folded',
+      'seriesStatement.folded',
+      'titleAlt.folded',
+      'titleDisplay.folded',
+      'contentsTitle.folded',
+      'donor.folded',
+      'parallelTitle.folded^5',
+      'parallelTitleDisplay.folded',
+      'parallelTitleAlt.folded',
+      'parallelSeriesStatement.folded',
+      'parallelCreatorLiteral.folded',
+      'parallelPublisher',
+      'uniformTitle.folded',
+      'parallelUniformTitle',
+      'formerTitle',
+      'addedAuthorTitle',
+      'placeOfPublication',
+      { field: 'items.idBarcode', on: (q) => /\d{6,}/.test(q) },
+      // Try to detect shelfmark searches (e.g. JFD 16-5143)
+      { field: 'items.shelfMark', on: (q) => /^[A-Z]{1,3} \d{2,}/.test(q) }
+    ]
+  },
+  title: {
+    fields: [
+      'title^5',
+      'title.folded^2',
+      'titleAlt.folded',
+      'uniformTitle.folded',
+      'titleDisplay.folded',
+      'seriesStatement.folded',
+      'contentsTitle.folded',
+      'donor.folded',
+      'parallelTitle.folded^5',
+      'parallelTitleDisplay.folded',
+      'parallelSeriesStatement.folded',
+      'parallelTitleAlt.folded',
+      'parallelCreatorLiteral.folded',
+      'parallelUniformTitle',
+      'formerTitle',
+      'addedAuthorTitle'
+    ]
+  },
+  contributor: {
+    fields: ['creatorLiteral^4', 'creatorLiteral.folded^2', 'contributorLiteral.folded', 'parallelCreatorLiteral.folded', 'parallelContributorLiteral.folded']
+  },
+  subject: {
+    fields: ['subjectLiteral^2', 'subjectLiteral.folded', 'parallelSubjectLiteral.folded']
+  },
+  series: {
+    fields: ['seriesStatement.folded']
+  },
+  journal_title: {
+    fields: null
+  },
+  callnumber: {
+    // We do custom field matching for this search-scope
+  },
+  standard_number: {
+    // We do custom field matching for this search-scope
+  }
+}
+
+const FILTER_CONFIG = {
+  owner: { operator: 'match', field: 'items.owner_packed', repeatable: true, path: 'items' },
+  subjectLiteral: { operator: 'match', field: 'subjectLiteral_exploded', repeatable: true },
+  holdingLocation: { operator: 'match', field: 'items.holdingLocation_packed', repeatable: true, path: 'items' },
+  buildingLocation: { operator: 'match', field: 'buildingLocationIds', repeatable: true },
+  language: { operator: 'match', field: 'language_packed', repeatable: true },
+  materialType: { operator: 'match', field: 'materialType_packed', repeatable: true },
+  mediaType: { operator: 'match', field: 'mediaType_packed', repeatable: true },
+  carrierType: { operator: 'match', field: 'carrierType_packed', repeatable: true },
+  publisher: { operator: 'match', field: 'publisherLiteral.raw', repeatable: true },
+  contributorLiteral: { operator: 'match', field: 'contributorLiteral.raw', repeatable: true },
+  creatorLiteral: { operator: 'match', field: 'creatorLiteral.raw', repeatable: true },
+  issuance: { operator: 'match', field: 'issuance_packed', repeatable: true },
+  createdYear: { operator: 'match', field: 'createdYear', repeatable: true },
+  dateAfter: {
+    operator: 'custom',
+    type: 'int'
+  },
+  dateBefore: {
+    operator: 'custom',
+    type: 'int'
+  }
+}
+
+module.exports = {
+  SEARCH_SCOPES,
+  FILTER_CONFIG
+}