Research Library

Research Library is a three-module Drupal sub-suite, classification_x, library_x and research_library, that catalogues external content such as lectures, books, research papers and maps into a trust-scored registry, classifies each item against bibliographic and subject-classification schemes, and gives visitors a faceted page to browse the result.

It builds on MADDev's External Entities suite: rather than fetching content itself, it records where an item lives and which processor module already knows how to fetch it.

Modules

classification_x, the classification subsystem; library_x, the catalogue hub; research_library, the public browse layer

Drupal core

11

Requires

Core Field, Options, Taxonomy, Datetime, Text and Media; the key module; the contributed external_entities module and MADDev's own entities_ext

Catalogue types

lecture, book, research_paper and map, each a code-defined library_item bundle

Field type

classification, a generic field type: a scheme, code and label tuple, with more than one scheme able to sit on the same field

Works with

Any registered External Entities processor module, for example species_ext, wikipedia_ext, publications_ext or entity_ext

Status

In development. Built and verified in June 2026 with a first collection of 20 lectures and 11 books for The Neville Goddard Library; the deduplication and merge engine followed. Not yet on a public site.

The /library page facets the catalogue by item type, collection and classification scheme, with a search box for titles.

Opening an item shows its detail page: title, date, sources, classification chips and, where a copy has been mirrored, the item's own body text.

Entity types

library_item: the catalogue record, in four code-defined bundles: lecture, book, research_paper and map. library_source: one row per trusted upstream repository, holding that source's own trust sub-scores.

Fields and taxonomy

• classification: a generic field type holding a scheme, code and label; more than one scheme can sit on the same field
• library_collection: a taxonomy vocabulary for grouping items into a project or collection
• Four base fields on library_item carry the suite's reconciliation pattern: field_classification; field_local_copy, a pointer to a corrected local copy that later harvesting will not overwrite; field_duplicate_of, a soft link to the canonical item; and field_merge_provenance, a JSON audit trail of how a merge decision was reached

Plugin types

• @ClassificationScheme: one class per classification scheme; a new one appears in the settings form and the field widget's scheme dropdown without any schema or configuration change
• @LibraryItemType: a catalogue bundle such as lecture or map
• @LibraryMergeStrategy: how two catalogue items are reconciled; the shipped strategies are soft_merge (the default), confidence and provenance

Pages

• /library: the public faceted browse page, by item type, collection and classification scheme, with a title search
• /library/item/{id}: an item's detail page, with its title, date, sources, classification chips and, where a copy has been mirrored, its body text
• /admin/config/content/classification: classification scheme settings
• /admin/content/library/duplicates: dedup curation
• /admin/config/library/merge-policy: merge-policy configuration

Services

library_x.dedup finds likely duplicate items by normalised title within the same bundle and collection. library_x.classifier tags items against schemes such as LCC, Dewey and Wikipedia categories using Claude, run from Drush rather than automatically. classification_x.iconclass_motif_bridge softly bridges an Iconclass notation to a folklore motif entity, and does nothing when that other module is absent.

Drush commands

lxin bulk-registers a first slice of catalogue entries from CSV. lxmir mirrors trusted items: it fetches the content, creates a managed file and a document media item, and attaches it to the catalogue item.

Permission

The /library route requires View published library items, a permission defined by library_x.

Only classification_x reaches an external source, and only softly. Labels for Getty's Art & Architecture Thesaurus resolve against Getty's Linked Open Data vocabulary service when MADDev's data_ext module is installed, and fall back to the bare code when it is not.

Dewey Decimal Classification captions are OCLC-licensed, so classification_x stores and shows the bare code by default and only shows the caption text once a licence key is configured in the key module. MADDev applies the same discipline at the level of whole works, not just catalogue metadata, through Copyrights Guard.

• Strict one-way layering. classification_x depends on nothing above it; library_x depends down on classification_x; research_library depends down on library_x and defines no entity of its own. Neither layer feeds back up.
• New classification schemes need no schema change. Dropping in one new @ClassificationScheme class is enough: it appears in the settings form and the field widget's scheme dropdown by itself.
• Catalogue bundle types are code, not config. A new library_item type is a @LibraryItemType plugin, and the browse page in research_library picks it up in its facets automatically, with no UI code of its own.
• Reconciliation keeps the record, not just the merge. A catalogue item can point at a locally corrected copy that later harvesting will not overwrite, be marked a duplicate of another item, and carry a JSON audit trail of how that decision was reached, under one of three pluggable merge strategies.
• A named upgrade path, not a rewrite. The browse page filters in PHP for now, a choice suited to a first slice of the collection; a Search API-backed version is the planned replacement as it grows.