generated: '2026-08-06'
method: derived
source: openapi/belharra-therapeutics-content-openapi.yml + live anonymous responses and _links inspection on 2026-08-06
note: >-
  The entity-relationship graph of the Belharra Therapeutics content API, derived from the OpenAPI
  schemas, the id-reference fields on live responses, and the HAL-style `_links` relations
  WordPress attaches to every resource. Belharra publishes no object reference of its own. The
  notable shape here is that all three of Belharra's custom post types are FLAT — none declares a
  taxonomy, none carries an author or excerpt, and each relates outward only through
  featured_media. Editorial structure that would normally live in taxonomy terms lives instead in
  the untyped `acf` payload.
entities:
- name: Post
  rest_base: posts
  schema: '#/components/schemas/Post'
  count_at_harvest: 10
  identity: {field: id, type: integer}
  description: >-
    The blog/news archive. Categorised into Company News (7) and Press Releases (3). Spans
    2023-01-04 to 2024-06-18.
- name: Page
  rest_base: pages
  schema: '#/components/schemas/Page'
  count_at_harvest: 29
  identity: {field: id, type: integer}
  description: >-
    The corporate site. Self-referentially hierarchical via `parent` — /about/ parents Our Story,
    Our Values and Our Team, which in turn parents Leadership, Founders, Board of Directors, Senior
    Advisors and BTX Crew; /our-approach/ parents Platform and Publications; /science/ parents
    Program and Our Science.
- name: PressRelease
  rest_base: press-release
  schema: '#/components/schemas/CustomPost'
  count_at_harvest: 4
  identity: {field: id, type: integer}
  description: >-
    Belharra's own announcements — the January 2023 $130M debut, the January 2023 Genentech
    collaboration, the June 2024 Sanofi collaboration and the January 2025 CSO appointment.
- name: CompanyNewsItem
  rest_base: company-news
  schema: '#/components/schemas/CustomPost'
  count_at_harvest: 21
  identity: {field: id, type: integer}
  description: Third-party press coverage ("In the News"). Outbound article links live in `acf`.
- name: MultimediaFile
  rest_base: multimedia-file
  schema: '#/components/schemas/CustomPost'
  count_at_harvest: 10
  identity: {field: id, type: integer}
  description: >-
    Curated video and podcast records backing the Multimedia Library, including the "Birth of a
    Biotech" podcast series. Media references live in `acf`.
- name: MediaItem
  rest_base: media
  schema: '#/components/schemas/MediaItem'
  count_at_harvest: 430
  identity: {field: id, type: integer}
  description: >-
    The attachment library. Carries source_url, mime_type, filesize, alt_text and a media_details
    block holding the generated size variants.
- name: Term
  rest_base: categories | tags
  schema: '#/components/schemas/Term'
  count_at_harvest: 2
  identity: {field: id, type: integer}
  description: >-
    Taxonomy terms. Only two exist, both in the `category` taxonomy — Company News (id 1, count 7)
    and Press Releases (id 7, count 3). The `post_tag` taxonomy is registered but empty.
- name: Author
  rest_base: users
  schema: '#/components/schemas/Author'
  count_at_harvest: 3
  identity: {field: id, type: integer}
  description: >-
    The public author projection. Anonymously readable on this deployment (unlike many WordPress
    sites, which 401 this collection), but limited to id, name, url, description, link, slug and
    avatar_urls. No emails, roles or capabilities are exposed.
relationships:
- from: Post
  to: Author
  kind: belongs_to
  via: author
  link_rel: author
- from: Post
  to: MediaItem
  kind: has_one
  via: featured_media
  link_rel: 'wp:featuredmedia'
  note: 0 means no featured image is set.
- from: Post
  to: Term
  kind: has_many
  via: categories
  link_rel: 'wp:term'
  target_taxonomy: category
- from: Post
  to: Term
  kind: has_many
  via: tags
  link_rel: 'wp:term'
  target_taxonomy: post_tag
  note: Always empty — the post_tag taxonomy carries no terms on this deployment.
- from: Page
  to: Page
  kind: belongs_to
  via: parent
  link_rel: up
  note: Self-referential hierarchy. 0 means top level.
- from: Page
  to: Author
  kind: belongs_to
  via: author
  link_rel: author
- from: Page
  to: MediaItem
  kind: has_one
  via: featured_media
  link_rel: 'wp:featuredmedia'
- from: PressRelease
  to: MediaItem
  kind: has_one
  via: featured_media
  link_rel: 'wp:featuredmedia'
- from: CompanyNewsItem
  to: MediaItem
  kind: has_one
  via: featured_media
  link_rel: 'wp:featuredmedia'
- from: MultimediaFile
  to: MediaItem
  kind: has_one
  via: featured_media
  link_rel: 'wp:featuredmedia'
- from: MediaItem
  to: Post
  kind: belongs_to
  via: post
  note: The object the attachment was uploaded to. Null for library-only uploads.
- from: MediaItem
  to: Author
  kind: belongs_to
  via: author
- from: Term
  to: Term
  kind: belongs_to
  via: parent
  note: Both live categories are top level (parent 0).
- from: SearchResult
  to: [Post, Page, PressRelease, CompanyNewsItem, MultimediaFile]
  kind: polymorphic_reference
  via: subtype
  note: >-
    Resolve `subtype` to a rest_base via /wp/v2/types, then GET /wp/v2/{rest_base}/{id}. For all
    five subtypes on this deployment the subtype string IS the rest_base, but that is a property of
    this installation and should be confirmed from /wp/v2/types rather than assumed.
id_conventions:
  scheme: WordPress auto-increment integers
  shared_sequence: >-
    Posts, pages, media and all three custom post types draw from a single wp_posts ID sequence, so
    an ID is globally unique across those six collections but is NOT scoped per type — requesting a
    valid page ID on /wp/v2/posts/{id} returns 404 rest_post_invalid_id.
  taxonomy_sequence: >-
    Terms use a separate wp_terms sequence, so term id 1 and post id 1 are unrelated.
  prefixes: none
untyped_surface:
  field: acf
  present_on: [Post, Page, MediaItem, Term, Author, PressRelease, CompanyNewsItem, MultimediaFile]
  detail: >-
    An Advanced Custom Fields payload is inlined on every resource. It is schemaless from the
    consumer's point of view — the field keys are defined in the WordPress admin, which is
    401-gated — and it is where Belharra's real editorial structure lives: outbound coverage URLs
    on company-news, video and podcast references on multimedia-file, and the team/pipeline block
    content on pages. It is modelled as `additionalProperties: true` in the OpenAPI rather than
    guessed at, because the field definitions are not publicly readable.
render: null