Skip to content

DR-NTU: Repository search and metadata MCP tools

Service overview · CLI reference · All MCP tools · Data types

Repository search and metadata

Inputs are required only where marked. A required output field may still be null if its type allows it. See data types for nested objects.

Search DR-NTU publications by keywords or Solr query. Returns concise titles, authors, dates, identifiers and rights. Scope to a community/collection UUID, or apply facet filters using field/value/operator. Set entity_type=Person or all explicitly for broader searches.

Required: query, up to 500 characters. Optional: scope_id, entity_type, filters, page, page_size, default 10, and refresh.

Searches publications by default. General DR-NTU search also returns researcher profiles; select entity_type: "Person" or "all" explicitly for those results. Each match returns the item UUID, title, entity type, authors, dates, identifiers, published rights, and repository URL. Use drntu_get_item for full published metadata and abstracts.

Queries support the repository's Solr syntax. scope_id restricts results to a community or collection UUID from drntu_browse. Structured filters have field, value, and operator, one of equals, contains, or authority, default equals. Up to ten filters are accepted; repeated fields are retained. Facet operators and field support depend on repository configuration.

{"query": "machine learning", "page_size": 5}
{"query": "machine learning", "filters": [{"field": "dateIssued", "value": "[2020 TO 2026]", "operator": "equals"}]}

Inputs

Field Type Required Details
query string Yes maximum length: 500
scope_id string or null No default: null; pattern: ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$
entity_type "Publication", "Person", "all" No default: "Publication"
filters array of SearchFilter or null No default: null; maximum items: 10
page integer No minimum: 0; maximum: 100000; default: 0
page_size integer No minimum: 1; maximum: 50; default: 10
refresh boolean No default: false

Output

Field Type Required Details
retrieved_at string Yes format: date-time
source_url string Yes —
cached boolean Yes —
query string Yes —
items array of ItemSummary Yes —
pagination Page Yes —

Unknown fields are not accepted.

drntu_get_item

Read a repository item's published metadata, including abstracts, citations, DOI, subjects and rights where supplied. Metadata values retain language tags. Public metadata does not establish full-text access.

Required: item_id, a UUID from search results. Optional: refresh.

Returns title, entity type, repository URL, last-modified/withdrawn information when supplied, and a dictionary of published metadata fields. Metadata values retain their language tags. Abstracts are normally in dc.description.abstract; citations, DOI, subjects, degree information, and rights appear when the source provides them. Unknown field names are retained. A single title identical to the top-level title is omitted from metadata only when it has no language tag; alternative and multilingual titles remain.

Public metadata does not prove that a paper's full text is publicly downloadable.

Inputs

Field Type Required Details
item_id string Yes pattern: ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$
refresh boolean No default: false

Output

Field Type Required Details
retrieved_at string Yes format: date-time
source_url string Yes —
cached boolean Yes —
item_id string Yes —
title string Yes —
entity_type string or null Yes —
url string Yes —
last_modified string or null Yes —
withdrawn boolean or null Yes —
metadata object of array of MetadataValue Yes —

Unknown fields are not accepted.

drntu_browse

Browse repository communities or collections. Omit community_id for all; supply it for subcommunities or collections within that community. Use returned container_id as search scope_id. Negative/unavailable archived item counts become null.

Optional: kind, either communities, default, or collections; community_id; pagination; and refresh.

Without community_id, lists repository communities or collections. With a community UUID, lists its subcommunities or collections. Returns container IDs, names, published descriptions, and archived-item counts when known. DSpace's -1 count means unavailable and becomes null. Use a returned container_id as drntu_search.scope_id.

{"kind": "collections", "page_size": 10}

Inputs

Field Type Required Details
kind "communities", "collections" No default: "communities"
community_id string or null No default: null; pattern: ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$
page integer No minimum: 0; maximum: 100000; default: 0
page_size integer No minimum: 1; maximum: 50; default: 20
refresh boolean No default: false

Output

Field Type Required Details
retrieved_at string Yes format: date-time
source_url string Yes —
cached boolean Yes —
kind "communities", "collections" Yes —
containers array of Container Yes —
pagination Page Yes —

Unknown fields are not accepted.

drntu_get_facets

Discover search facet names, or supply a facet name to list its values and counts for the query/scope/filters. Facet definitions have no counts. Copy each value's filter object into search filters; date range labels can differ from their query values.

Optional: facet, query, scope_id, entity_type, filters, pagination, and refresh. The query/scope/filter options match search; entity type defaults to Publication.

Omit facet to discover filter definitions, such as author, subject, dateIssued, or entityType. Supply a name to get values and counts. Definitions are configuration metadata and have no result counts.

Each value includes a ready-to-use filter object extracted from the repository's own search link, when supplied. Copy it into drntu_search.filters. Display labels can differ from query values: a date label such as 2020 - 2026 may require [2020 TO 2026]. Authority-based values can also differ from their labels. Missing filter links remain null.

{"facet": "dateIssued", "query": "machine learning", "page_size": 5}

Inputs

Field Type Required Details
facet string or null No default: null; pattern: ^[A-Za-z][A-Za-z0-9_.-]{0,79}$
query string No maximum length: 500; default: ""
scope_id string or null No default: null; pattern: ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$
entity_type "Publication", "Person", "all" No default: "Publication"
filters array of SearchFilter or null No default: null; maximum items: 10
page integer No minimum: 0; maximum: 100000; default: 0
page_size integer No minimum: 1; maximum: 50; default: 10
refresh boolean No default: false

Output

Field Type Required Details
retrieved_at string Yes format: date-time
source_url string Yes —
cached boolean Yes —
facet string or null Yes —
definitions array of FacetDefinition Yes —
values array of FacetValue Yes —
pagination Page or null Yes —

Unknown fields are not accepted.

drntu_list_bundles

List an item's file groups. ORIGINAL usually holds deposited papers; TEXT, THUMBNAIL and LICENSE are separate bundles. Pass a returned bundle_id to drntu_list_files.

Required: item_id. Optional: pagination and refresh.

Returns named file groups and their UUIDs. ORIGINAL usually contains deposited papers; TEXT, LICENSE, and THUMBNAIL are different groups. Pass a returned bundle_id to drntu_list_files. Follow bundle pagination before assuming an item has no originals.

Inputs

Field Type Required Details
item_id string Yes pattern: ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$
page integer No minimum: 0; maximum: 100000; default: 0
page_size integer No minimum: 1; maximum: 50; default: 20
refresh boolean No default: false

Output

Field Type Required Details
retrieved_at string Yes format: date-time
source_url string Yes —
cached boolean Yes —
item_id string Yes —
bundles array of Bundle Yes —
pagination Page Yes —

Unknown fields are not accepted.

drntu_list_files

List a bundle's file metadata: names, byte sizes, checksums, published rights and content URLs. Does not fetch file bytes or verify download permission. Restricted content remains restricted even when its metadata is public.

Required: bundle_id. Optional: pagination and refresh.

Returns published file UUIDs, names, byte sizes, checksums, rights, and content URLs. No file bodies are fetched, and download permission is not tested. The example supplied for development has public metadata with restricted rights; the integration preserves those rights. Missing rights remain an empty list of published statements, which does not establish open access.

Inputs

Field Type Required Details
bundle_id string Yes pattern: ^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$
page integer No minimum: 0; maximum: 100000; default: 0
page_size integer No minimum: 1; maximum: 50; default: 20
refresh boolean No default: false

Output

Field Type Required Details
retrieved_at string Yes format: date-time
source_url string Yes —
cached boolean Yes —
bundle_id string Yes —
files array of File Yes —
pagination Page Yes —

Unknown fields are not accepted.

drntu_get_dataset

Read a published DR-NTU Data dataset by doi: identifier: title, authors, descriptions, subjects, license and paginated file inventory. Omit version for latest, or use a release such as 1.0. CAPTCHA/access denials are returned as actionable errors without bypass or login.

Required: persistent_id in doi: form. Optional: version such as 1.0, limit from 1–50, default 50, offset, default 0, and refresh.

Reads a published DR-NTU (Data) dataset's citation metadata and file inventory: title, authors, descriptions, subjects, release version/state, publication date, license, filenames, sizes, media types, directory labels, restriction flags, and checksums where provided. Omit version for the latest published version. Draft versions are unsupported. Unknown restriction flags remain null.

The dataset API returns the complete version metadata; limit and offset bound the files returned through MCP. Follow next_offset until null. Very large source responses may exceed the HTTP size limit. Additional discipline-specific dataset metadata is not exposed by this concise citation view.

{"persistent_id": "doi:10.21979/N9/UD1IM9", "version": "1.0", "limit": 10}

HTML responses, including CAPTCHA pages with HTTP 200, and access denials stop retrieval with an actionable error. The integration does not start a browser, collect credentials, solve CAPTCHA, or reuse personal NTU sessions.

Inputs

Field Type Required Details
persistent_id string Yes maximum length: 250; pattern: ^doi:10\.[0-9]{4,9}/[^\s?#]+$
version string or null No default: null; maximum length: 20; pattern: ^[0-9]+\.[0-9]+$
limit integer No minimum: 1; maximum: 50; default: 50
offset integer No minimum: 0; default: 0
refresh boolean No default: false

Output

Field Type Required Details
retrieved_at string Yes format: date-time
source_url string Yes —
cached boolean Yes —
persistent_id string Yes —
dataset_id integer Yes greater than: 0
version string Yes —
version_state string Yes —
title string or null Yes —
authors array of string Yes —
descriptions array of string Yes —
subjects array of string Yes —
publication_date string or null Yes —
license_name string or null Yes —
license_url string or null Yes —
url string Yes —
files array of DatasetFile Yes —
total_files integer Yes minimum: 0
next_offset integer or null Yes minimum: 0

Unknown fields are not accepted.