Skip to main content
ExLibris
  • Subscribe by RSS
  • Ex Libris Knowledge Center

    Ex Libris Central Discovery Index (CDI) Content Submission Guide

    Ex Libris Central Discovery Index (CDI) Content Submission Guide

    The Central Discovery Index (CDI) from Ex Libris powers the Discovery products Primo and Summon.

    More information on CDI is available here: The Ex Libris Central Discovery Index (CDI) – An Overview

    In order to have content ingested and made discoverable in the index, please refer to this guide for metadata requirements and formats. Please note that ingestion in CDI means content will be available in both Summon and Primo.

    For questions and requests for more information, please contact Provider.Relations@exlibrisgroup.com

    Content / Resource Types

    CDI indexes these main content or resource types:

    • Electronic Serial Resources, such as journals and newspapers
      • Article-level content is required; more details below
    • Electronic Monographic Resources, such as ebooks
      • Chapter-level content is strongly requested; more details below
    • Video, music, images
    • Thumbnails and cover images
      • Provide URL links or syntax to display thumbnails/cover images

    For a more exhaustive list of types, please see:

    Metadata Elements

    Discovery of your hosted content is optimized by including as many of the metadata fields listed below as possible. In addition to item-level metadata, we also require expected record count for linking and quality assurance purposes.

    Below is a list of metadata elements for ebooks and serials (ejournals and newspapers). For all other content types, such as video, the basic required metadata elements are:

    • Unique record identifier
    • Content/resource type
    • Title
    • URL (as granular as possible)
    • Author/editor/creator
    • Date
    • Language
    • Subject terms
    • If relevant:
      • Open Access
      • Deletion indicator
      • Peer review indicator
      • Other identifiers (e.g. ORCID, DOI)

    Ebooks

    Required fields: Unique record identifier, ISBN, EISBN, Publication Title, Title-Level URL, Author(s) (can substitute Editor(s)), Date Published, Language

    Recommended fields: Parallel title (for languages other than English), Discipline, Editor(s), Publication Series Title, ORCID identifiers, Place of Publication, Full Text, LCCallNum, DEWEY, Series ISSN, Book DOI, Abstract, Type of work, Collection indicator, Subject Terms, Keywords, Deletion indicator, Open Access indicator, Peer Review indicator

    Additional requested fields: Author/Editor affiliation metadata such as email, Publication Subtitle, Publisher, Edition, Book Description, Page Count, Table of Contents (if not included in full text), Associated DOIs for each chapter, book cover thumbnail URL, Data Provider Logo (120 x 24 or 134 x 31), Grant metadata, Funding metadata, Creative Commons (CC) license metadata

    Ebook Chapters

    Required fields: Unique record identifier, ISBN, EISBN, Chapter Title, Publication Title, Chapter-level URL, Author(s) (can substitute Editor(s)), Date Published, Language, Start Page

    Recommended fields: Chapter Number, Parallel chapter title (for languages other than English), Parallel publication title (for languages other than English), ORCID identifiers, Discipline Editor(s), Full Text, LCCallNum, DEWEY, DOI, Publication Series Title, Subject Terms, Keywords, Series ISSN, Associated DOIs for the book and each additional chapter, Abstract, Type of work, Collection indicator, Deletion indicator, Open Access indicator, Peer Review indicator

    Additional requested fields: Author/Editor affiliation metadata such as email, Chapter Subtitle, Publication Subtitle, Publisher, Place of Publication, Edition, Book Description, End Page, Page Count, Table of Contents (if not included in full text), book cover thumbnail URL, Data Provider Logo (120 x 24 or 134 x 31), Grant metadata, Funding metadata, Creative Commons (CC) license metadata

    Journal/Newspaper Articles

    Required fields: Unique record identifier, ISSN, EISSN, Article Title, Publication Title, Publication Date, Volume, Issue, Start Page, Article-level URL, Author(s), Language

    Recommended fields: Parallel article title (for languages other than English), Parallel publication title (for languages other than English), ORCID identifiers, Author Affiliations, Discipline, Subject Terms, Keywords, Abstract, Full Text, DOI, End Page, Direct link to full article, Type of work, Collection indicator, Deletion indicator, Indicate whether a specified work is free-to-read (see NISO RP-22-2015), Open Access indicator, Reference to a URI carrying re-use rights granted to a reader of a work (see NISO RP-22-2015), Peer Review indicator

    Additional requested fields: Author/Editor affiliation metadata such as email, Article Subtitle, Publication Subtitle, Page Count, Dateline, Publisher, Total number of words, journal/newspaper cover thumbnail URL, Data Provider Logo (120 x 24 or 134 x 31), Grant metadata, Funding metadata, Creative Commons (CC) license metadata, (for Conference proceedings: name, location, dates).

    If your content does not meet the above critieria, please contact us to discuss the matter further and explore what options might be available.

    Supported File Delivery Methods

    (in order of preference)

    • Delivery Method Description Ex Libris-hosted FTP Site (FTP push)
      • Upon request, Ex Libris will provide a secure login and site
    • Provider-hosted FTP Site (FTP pull)
      • A secure FTP site hosted on the content provider’s domain
    • OAI-PMH
    • HTTP link to download files (see Metadata Standards Supported below)

    At this time we do not support web crawls.

    Other delivery methods may be possible, but will require evaluation to determine feasibility and will likely require additional processing time.

    Metadata Standards Supported

    • XML
      • including XML in Dublin Core/XML with NLM DTD/XML with ONIX DTD/XML with custom DTD/Custom XML without DTD
    • Delimited Text Format (Tab Delimited or CSV)
    • ONIX and BITS for book/chapter-level metadata; JATS for journal article-level metadata
      • Download samples here:
    • JSON
    • MARC21/MARC UNIMARC (binary – not ASCII)/MARCXML
    • Excel spreadsheets

    XML_examples.zip

    Other standards may be possible, but will require evaluation to determine feasibility and will likely require additional processing time.

    Full Text

    If applicable, full text (used for indexing purposes) can be accepted in the following formats:

    • XML
      • Preferably < body > tags within article/book/chapter records
    • PDF
      • File name must be included in XML record

    Other formats may be possible, but will require evaluation to determine feasibility and will likely require additional processing time.

    Direct Linking

    To ensure access to your digital content, we require direct links to each record. This allows our mutual customers to directly access the content hosted on your platform if their library has subscribed and the user has been authenticated per the terms of your website. In addition, include DOI’s if available.

    The users expect to click on a link from the Primo or Summon results page and be directed to the specific record. Links to a generic home page or search form provide an unsatisfactory experience. Our mutual clients do not expect to rerun their search in your platform to locate the desired item.

    An alternative to providing us with direct links is to provide us with the linking syntax and the necessary field names. We can create the links independent of your content.

    Client Specific URL

    Provide a detailed description of requirements for substituting a client ID into the outgoing URL.

    See examples listed here: Summon and 360 Link: Index-Enhanced Direct Linking Special Providers

    Username/Password

    Providing a username and password for authentication (even temporary) is helpful to troubleshoot linking and verify metadata mapping.

    Subscription Rights

    We will need to know how content is sold to our mutual clients, so that they are able to accurately discover content. Our mutual clients will activate either at the collection level or the title level in the knowledgebases (KBs): Alma CZ, SFX, and 360 KB.

    Additionally we prefer to receive a regular list of collections (manifest) with their names, update frequencies (or indication that it is static/closed), and title count. This allows us to regularly audit the KBs for accuracy and completeness, and update the mappings for subscription rights.

    Title level rights

    The ISSN or ISBN in the KB collection is used as the match point between the KB and Discovery feed.

    Title level rights is the most common scenario. It cannot be used if the content does not contain ISSNs or ISBNs.

    Collection level rights

    An identifier in the Discovery feed will need to be available to map the entire collection in the KB to the applicable records in CDI.

    Collection level rights are typically used when the collection is sold as-is (no single-title purchasing or “cherry-picking”) or for content that does not contain ISSNs or ISBNs.

    Deletes

    We support record deletes using any one of the following methods:

    • The record to be deleted included in the update feed with a delete marker
    • The record to be deleted included in the delete feed
    • A file listing the IDs of the records to be deleted included in the delete feed. Must match the ID used in the metadata file
    • Periodic new full load (NFL)

    Further Inquiry

    If you have any questions about the described CDI inclusion criteria for Primo and Summon, please contact Provider.Relations@exlibrisgroup.com. We would be happy to answer any questions you may have and/or arrange a demonstration of the Primo and Summon products.


    • Article last edited: 14-June-2020