Refactor ORCID disambiguation to be more precise and comprehensive

README.md

@@ -69,6 +69,8 @@ Services:
 - `INDRA_DB_BASE_URL` : url for [INDRA (Integrated Network and Dynamical Reasoning Assembler)](https://indralab.github.io/)
 - `INDRA_ENGLISH_ASSEMBLER_URL` : url for service that assembles INDRA statements into models
 - `SEMANTIC_SEARCH_BASE_URL` : url for [semantic-search](https://github.com/PathwayCommons/semantic-search) web service
+- `ORCID_BASE_URL` : url for [ORCID](https://orcid.org/) website
+- `ORCID_PUBLIC_API_BASE_URL` : url for version of [ORCID](https://orcid.org/) public API
 - `NO_ABSTRACT_HANDLING` : labels directing how to sort documents missing query text. 'text' (default): autogenerate text from templates; 'date': sort by date and ignore text.
 
 Links:

src/client-env-vars.json

@@ -22,5 +22,6 @@
   "DEMO_CAN_BE_SHARED_MULTIPLE_TIMES",
   "SAMPLE_DOC_ID",
   "EMAIL_ADDRESS_INFO",
-  "MAX_WAIT_TWEET"
+  "MAX_WAIT_TWEET",
+  "ORCID_BASE_URL"
 ]

src/client/components/editor/credits.js

@@ -1,6 +1,8 @@
 import h from 'react-hyperscript';
 import { Component } from 'react';
 import _ from 'lodash';
+import { ORCID_BASE_URL } from '../../../config';
+import { findOrcidIdentifier } from '../../../util/pubmed';
 
 export class Credits extends Component {
   constructor(props){
@@ -17,16 +19,16 @@ export class Credits extends Component {
 
     const profiles = document.authorProfiles() || [];
     const isContributingAuthor = author => author.name === authorName;
-    const authorLink = _.get(profiles.find(isContributingAuthor), ['orcid']);
+    const orcid = findOrcidIdentifier( _.get(profiles.find(isContributingAuthor), ['orcid']) ); // Backwards compatible with URI
 
     return h('div.editor-credits', [
       `Article summary created by `,
-      (authorLink ? 
-        h('a.plain-link', { href: authorLink, target: '_blank' }, authorName) 
+      (orcid ?
+        h('a.plain-link', { href: `${ORCID_BASE_URL}${orcid}`, target: '_blank' }, authorName)
         :
-        h('span', authorName) 
+        h('span', authorName)
       ),
-      authorLink ? h('span', [
+      orcid ? h('span', [
         h('span', ' '),
         h('i.icon.icon-orcid')
       ]) : null

src/client/components/editor/info-panel.js

@@ -1,12 +1,13 @@
 import h from 'react-hyperscript';
 import { Component } from 'react';
-import { DOI_LINK_BASE_URL, PUBMED_LINK_BASE_URL, GOOGLE_SCHOLAR_BASE_URL } from '../../../config';
+import { DOI_LINK_BASE_URL, PUBMED_LINK_BASE_URL, GOOGLE_SCHOLAR_BASE_URL, ORCID_BASE_URL } from '../../../config';
 // import { Carousel, CAROUSEL_CONTENT } from '../carousel';
 import { makeClassList } from '../../dom';
 import ElementInfo from '../element-info/element-info';
 import RelatedPapers from '../related-papers';
 import _ from 'lodash';
 import Credits from './credits';
+import { findOrcidIdentifier } from '../../../util/pubmed';
 
 export class InfoPanel extends Component {
   constructor(props){
@@ -75,10 +76,10 @@ export class InfoPanel extends Component {
       h('div.editor-info-flags', [ retractFlag ]),
       h('div.editor-info-title', title),
       h('div.editor-info-authors', _.flatten(authorProfiles.map((a, i) => {
-        let orcidUri = a.orcid;
-        if ( orcidUri ) {
+        let orcid = findOrcidIdentifier( a.orcid ); // Backwards compatible with URI
+        if ( orcid ) {
           return [
-            h('a.editor-info-author.plain-link', { target: '_blank', href: orcidUri }, a.name),
+            h('a.editor-info-author.plain-link', { target: '_blank', href: `${ORCID_BASE_URL}${orcid}` }, a.name),
             h('i.icon.icon-orcid.editor-info-author-orcid'),
             i !== authorProfiles.length - 1 ? h('span.editor-info-author-spacer', ', ') : null
           ];

src/config.js

@@ -72,6 +72,8 @@ export const NCBI_EUTILS_API_KEY = env('NCBI_EUTILS_API_KEY', 'b99e10ebe0f90d815
 export const INDRA_DB_BASE_URL = env('INDRA_DB_BASE_URL', 'https://db.indra.bio/');
 export const INDRA_ENGLISH_ASSEMBLER_URL = env('INDRA_ENGLISH_ASSEMBLER_URL', 'http://api.indra.bio:8000/assemblers/english');
 export const SEMANTIC_SEARCH_BASE_URL = env('SEMANTIC_SEARCH_BASE_URL', 'https://main.semanticsearch.baderlab.org/');
+export const ORCID_BASE_URL = env('ORCID_BASE_URL', 'https://orcid.org/');
+export const ORCID_PUBLIC_API_BASE_URL = env('ORCID_PUBLIC_API_BASE_URL', 'https://pub.orcid.org/v3.0/');
 
 // Links
 export const UNIPROT_LINK_BASE_URL = env('UNIPROT_LINK_BASE_URL', 'http://www.uniprot.org/uniprot/');

src/server/routes/api/document/index.js

@@ -52,7 +52,8 @@ import { BASE_URL,
   EMAIL_ADDRESS_ADMIN,
   BULK_DOWNLOADS_PATH,
   BIOPAX_DOWNLOADS_PATH,
-  BIOPAX_IDMAP_DOWNLOADS_PATH
+  BIOPAX_IDMAP_DOWNLOADS_PATH,
+  ORCID_PUBLIC_API_BASE_URL
  } from '../../../../config';
 
 import { ENTITY_TYPE } from '../../../../model/element/entity-type';
@@ -344,60 +345,83 @@ const fillDocArticle = async doc => {
   }
 };
 
-const makeOrcidKey = ( s1, s2 ) => s1 + '_' + s2;
-
-const getOrcidUriMap = async doi => {
-  let url = `https://pub.orcid.org/v3.0/csv-search?q=`
-      + `doi%3D${doi}`;
-
-  const textToMap = text => {
-    let lines = text.split(/\r\n|\r|\n/);
-    let obj = {};
-    lines.forEach( ( line, i ) => {
-      if ( i == 0 ) {
-        return;
-      }
-      let vals = line.split(',');
-      let orcidUri = 'https://orcid.org/' + vals[ 0 ];
-      let givenNames = vals[ 2 ];
-      let familyName = vals[ 3 ];
+/**
+ * searchOrcid
+ * Search the ORCID API using an article DOI
+ *
+ * @param {string} doi article DOI
+ * @returns A unique set of search {@link https://info.orcid.org/documentation/integration-guide/orcid-record/ records}
+ */
+const searchOrcid = async doi => {
+  const getDoiSearchUrl = doi => `${ORCID_PUBLIC_API_BASE_URL}expanded-search?q=doi-self:${doi}`;
+  const fetchJson = url => fetch( url, { headers: { 'accept': 'application/json' } } );
+  const toJson = res => res.json();
+  const fetchAllJson = async urls => {
+    const responses = await Promise.all( urls.map( fetchJson ) );
+    return await Promise.all( responses.map( toJson ) );
+  };
+  const mergeExpandedResults = searchResults => {
+    const getSearchResult = response => response['expanded-result']; //possibly null
+    let records = _.compact( searchResults.map( getSearchResult ) );
+    return _.unionBy( ...records, 'orcid-id' );
+  };
 
-      if ( !givenNames || !familyName ) {
-        return;
-      }
+  try {
+    // DOIs can be set by clients, with varying case (e.g. eLife vs elife)
+    let rawUrl = getDoiSearchUrl( doi );
+    let normalizedUrl = getDoiSearchUrl( doi.toLowerCase() );
+    let urls = [ rawUrl, normalizedUrl ];
 
-      let key = makeOrcidKey( givenNames, familyName );
-      obj[ key ] = orcidUri;
-    } );
+    let expandedSearchResults = await fetchAllJson( urls );
+    let searchRecords = mergeExpandedResults( expandedSearchResults );
+    return searchRecords;
 
-    return obj;
-  };
+  } catch ( err ) {
+    logger.error( `Error finding Orcid URI: ${err.message}`);
+    return null;
+  }
+};
 
-  return fetch( url, { method: 'GET', headers: { 'accept': 'text/csv' } } )
-    .then( res => res.text() )
-    .then( textToMap )
-    .catch( error => {
-      logger.error( `Error finding Orcid URI: ${error.message}`);
-      return null;
-    } );
+const findAllIndexes = ( collection, key, val ) => {
+  var indexes = [], i = -1;
+  while ( ( i = _.findIndex( collection, [ key, val ], i + 1 ) ) != -1 ){
+    indexes.push( i );
+  }
+  return indexes;
 };
 
+/**
+ * fillDocAuthorProfiles
+ * Supplement PubMed author ORCIDs from ORCID itself.
+ * Match by last name when unique, else by first and last names.
+ *
+ * @param {Object} doc Document
+ * @returns An array of author profile information
+ */
 const fillDocAuthorProfiles = async doc => {
   const citation = doc.citation();
-  const { authors, doi } = citation;
-
-  let orcidUrisMap = await getOrcidUriMap( doi );
-
-  orcidUrisMap = orcidUrisMap || null;
-
-  let authorProfiles = authors.authorList.map( a => {
-    let key = makeOrcidKey( a.ForeName, a.LastName );
-    let orcid = orcidUrisMap[ key ] || null;
-    let authorProfile =  _.assignIn({}, a, { orcid });
-    return authorProfile;
-  });
-
-
+  const { authors: { authorList }, doi } = citation;
+
+  let orcidRecords = await searchOrcid( doi );
+  let authorProfiles = authorList
+    .map( author => {
+      let authorProfile = _.assign( {}, author );
+      const { orcid, ForeName, LastName } = author;
+      if( orcid == null ){
+        let match;
+        const byLastName =  [ 'family-names', LastName ];
+        const byFirstLastNames = o => LastName == o['family-names'] && ForeName == o['given-names'];
+        let authorListIndices = findAllIndexes( authorList, 'LastName', LastName );
+        let lastNameIsUnique = authorListIndices.length == 1;
+
+        // Find by last name alone, if it is unique
+        match = lastNameIsUnique ?
+          _.find( orcidRecords, byLastName ) :
+          _.find( orcidRecords, byFirstLastNames );
+        if( match ) _.set( authorProfile, ['orcid'], match['orcid-id'] );
+      }
+      return authorProfile;
+    });
   await doc.authorProfiles( authorProfiles );
 };
 

src/util/pubmed.js

@@ -94,6 +94,31 @@ const getAuthorAbbrev = AuthorList => {
 
 const getContacts = AuthorList => AuthorList.map( getContact ).filter( contact => !_.isEmpty( _.get( contact, 'email' ) ) );
 
+/**
+ * findOrcidIdentifier
+ * Helper to extract an ORCID accession from a string
+ *
+ * @param {string} raw the string to examine
+ * @returns ORCID accession, possibly null
+ */
+const findOrcidIdentifier = raw => {
+  let orcidId = null;
+  if( raw ){
+  const ORCID_REGEX = /\d{4}-\d{4}-\d{4}-\d{3}(\d|X)/g;
+    const matches = raw.match( ORCID_REGEX );
+    const hasMatch = !_.isEmpty( matches );
+    if( hasMatch ) orcidId = _.head( matches );
+  }
+  return orcidId;
+};
+
+const getAuthorOrcid = Author => {
+  const isOrcId = id => id['Source'] === 'ORCID';
+  const Identifier = _.get( Author, ['Identifier'] );
+  let raw =  _.get( _.find( Identifier, isOrcId ), ['id'] );
+  return findOrcidIdentifier( raw );
+};
+
 const getAuthorList = AuthorList => {
   return AuthorList.map( Author => {
     const { ForeName, LastName } = getAuthorNameParts( Author );
@@ -103,7 +128,8 @@ const getAuthorList = AuthorList => {
       LastName,
       email: _.head( getEmail( Author ) ) || null,
       abbrevName: getAbbrevAuthorName( Author ),
-      isCollectiveName: !_.isNull( _.get( Author, 'CollectiveName' ) )
+      isCollectiveName: !_.isNull( _.get( Author, 'CollectiveName' ) ),
+      orcid: getAuthorOrcid( Author )
     };
   });
 };
@@ -276,4 +302,4 @@ class ArticleIDError extends Error {
   }
 }
 
-export { getPubmedCitation, createPubmedArticle, ArticleIDError };
+export { getPubmedCitation, createPubmedArticle, ArticleIDError, findOrcidIdentifier };