Back to Reference
Contents
Reference / GLAM Sectors
487 words
Jun 28, 2026

KBART e-resource feeds (vendor harvesting)

KBART (Knowledge Bases And Related Tools, a NISO recommended practice) is the standard tab-separated format vendors use to publish the title lists and date coverage of the e-resource packages you subscribe to. The library module can register one or more vendor KBART feeds, download and parse them, and keep an audit log of every harvest — so your catalogue knows which electronic titles a package covers and for which date ranges.

Where it lives

Library → KBART vendors, or directly: /library/kbart/vendors

The vendors page lists every registered feed with its name, source URL, active flag, and the time of its last fetch.

Adding a vendor feed

/library/kbart/vendor/add

Provide:

  • Name — a label for the package/provider (e.g. "ProQuest Ebook Central", "JSTOR Arts & Sciences").
  • URL — the HTTPS address of the vendor's KBART .txt/.tsv file.
  • Active — marks the vendor as one you intend to keep harvesting (see Scheduling below).

Edit an existing vendor at /library/kbart/vendor/edit/:id, enable/disable it at /library/kbart/vendor/toggle/:id, and remove it at /library/kbart/vendor/delete/:id.

Fetching a feed

/library/kbart/vendor/fetch/:id — the Fetch now action.

Fetching downloads the vendor's KBART file (30-second timeout) and parses it as tab-separated values keyed by the KBART column names, including the standard fields:

  • publication_title
  • print_identifier / online_identifier (ISSN/ISBN)
  • date_first_issue_online, num_first_vol_online, num_first_issue_online
  • date_last_issue_online, num_last_vol_online, num_last_issue_online
  • title_url, first_author, title_id
  • embargo_info, coverage_depth, coverage_notes

Each fetch records how many rows were read and any parse/validation problems.

Import log

/library/kbart/vendor/log/:id

Every fetch writes an entry you can review here: when it ran, how many rows the feed contained, and any errors (unreachable URL, malformed TSV, missing required columns). Use it to confirm a harvest succeeded and to diagnose a feed a vendor has changed or broken.

Scheduling

Harvesting is currently manual — open a vendor and click Fetch now to pull the latest feed. The Active flag marks vendors you want kept up to date; automatic, scheduled harvesting of all active vendors is not yet enabled, so re-fetch feeds by hand (or after a vendor notifies you of a holdings change) until scheduled harvesting ships.

Data

KBART harvesting uses two tables:

  • library_kbart_vendor — one row per registered feed (name, URL, active flag, last-fetch timestamp).
  • library_kbart_import_log — one row per fetch (timestamp, row count, error detail).

Prerequisites

  • The server must be able to reach the vendor URL over outbound HTTPS.
  • The feed must be a valid KBART tab-separated file with the standard header row; non-KBART or HTML responses are rejected and recorded in the import log.
  • Z39.50 / SRU Integration — for harvesting bibliographic (MARC) records from remote catalogues.
  • MARC Import/Export — for loading and exporting catalogue records directly.

KBART export, OpenURL & ODI scorecard (v3.46.0)

  • KBART export — publish the library's holdings as a NISO KBART title list (TSV) from Library → KBART → Export (/library/kbart/export); serials_only limits it to titles with an ISSN.
  • OpenURL resolver — an inbound NISO OpenURL 1.0 endpoint at /openurl (accepts rft.isbn / rft.issn / rft.title). A single match redirects to the catalogue record; multiple matches list the candidates.
  • ODI quality scorecard — Admin → Library → ODI (/admin/library/odi) scores discovery-metadata completeness (title, creator, subject, publisher, date, identifier, call number) across published items, with per-field fill rates and an overall score.
  • MARCXML — available via the metadata-export module (marc21 format).