Documentation
¶
Overview ¶
Package catalogindex builds a read-only, periodically-rebuildable analytics view of the whole catalog over an in-memory gochickpeas graph. It answers cross-work questions the per-grain scans and the workindex identity layer cannot: which works share a controlled subject authority, and how heavily each authority is used across the corpus.
The view is loaded from a catalog.nq dataset (the `lcat serialize` output), where each bf:Work is a "Work"-labelled node, each bf:subject link is an outgoing "subject" rel to the authority node, and every IRI node carries its own IRI under the "uri" property. Blob grains remain the source of truth; this Snapshot is derived and disposable -- rebuild a fresh one when the catalog changes rather than mutating an existing one.
Subject authorities are the cross-work dimension modelled here because they are IRI-identified and therefore shared across works. Agents are blank nodes scoped to their grain (a MARC feed mints a fresh blank per contribution), so two works by the same author do not share an agent node; author co-occurrence stays a string-keyed concern of the similar package, not a graph query.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type AuthorityUse ¶
type AuthorityUse struct {
URI string `json:"uri"`
Label string `json:"label,omitempty"`
Scheme string `json:"scheme,omitempty"`
Works int `json:"works"`
}
AuthorityUse is one controlled subject authority and how many distinct works in the corpus are subject-linked to it.
type Snapshot ¶
type Snapshot struct {
// contains filtered or unexported fields
}
Snapshot is an immutable analytics view of one catalog dataset. Build it with Open or FromNQuads; its query methods are safe for concurrent use.
func FromNQuads ¶
FromNQuads builds a Snapshot from an in-memory catalog.nq dataset.
func (*Snapshot) AuthorityUsage ¶
func (s *Snapshot) AuthorityUsage() []AuthorityUse
AuthorityUsage reports every controlled subject authority referenced by at least one work, with the count of distinct works referencing it, sorted by descending use then URI. A subject a single work links more than once (e.g. asserted in both a feed graph and the editorial graph) counts once for that work.
Only controlled authorities are reported -- subjects whose URI resolves to a recognized authority scheme (project.SchemeForURI). Bare-label topics (blank nodes with no URI) and per-grain local subjects (document-relative "#..." fragment IRIs) are omitted: they carry no shared cross-work identity to aggregate, which is the whole point of the tally.
func (*Snapshot) WorksUsingAuthority ¶
WorksUsingAuthority returns the grain ids of the works subject-linked to the given authority URI, deduplicated and sorted. An unknown or unreferenced URI yields nil. The id is the local part of a work's "#<id>Work" node -- the same id the rest of libcat keys works by.
Source Files