This document maps between the three vocabularies used in BADGERS and the external academic standards that consumers are most likely to target.
BADGERS uses two complementary vocabularies at different stages of the pipeline:
| Vocabulary | Schema File | Stage | Purpose |
|---|---|---|---|
| Elements | badgers-elements.xsd |
Analysis (per-specialist) | Semantic content classification — "what is this?" |
| Content Tree (Spine) | badgers-content-tree.xsd |
Correlation (unified output) | PDF/UA structure tagging — "how should this be tagged?" |
The elements vocabulary is what individual specialists produce. The content tree vocabulary is what the correlation specialist produces after merging all specialist outputs into a single hierarchical document spine. The content tree uses PDF/UA (ISO 14289-1) tag names directly.
| Elements Specialist Type | Elements Sub-types | Content Tree (PDF/UA) Tag(s) | Notes |
|---|---|---|---|
document_title |
— | H1 |
Mapped to top-level heading |
section |
Chapter, Main section, Subsection, Sub-subsection | Sect |
Sections become structural grouping containers |
heading |
Title (H1), Section heading (H2), Subsection heading (H3), Sub-subsection heading (H4), Lower-level heading (H5, H6), Running head | H1, H2, H3, H4, H5, H6 |
Direct mapping by level. Running head → Artifact |
paragraph |
Body, Introduction, Conclusion | P |
All paragraph sub-types map to P |
citation |
In-text, Footnote, Endnote | Reference (inline marker) + Note (body) |
In-text citations become Reference elements; footnote/endnote text becomes Note |
reference |
Book, Journal article, Web page, Conference paper | BibEntry |
All reference sub-types map to BibEntry |
caption |
Figure caption, Table caption, Image caption | Caption |
Direct mapping; parent_ref links to parent Figure or Table |
quote |
Block quote, Inline quote, Pull quote | Quote or BlockQuote (block) / Span (inline) |
Block quotes get their own Sect with tag="BlockQuote" |
equation |
Inline equation, Display equation | Formula |
placement="Inline" or placement="Block" |
list |
Bulleted, Numbered, Definition | L → LI → Lbl + LBody |
Full substructure in content tree |
footnote |
— | Note |
Paired with Reference via ref attribute |
table_of_contents |
— | TOC → TOCI |
Table of contents with item sub-elements |
abstract |
— | P (within a Sect) |
No dedicated PDF/UA tag; rendered as paragraph in a labeled section |
keyword |
— | (enrichment only) | Captured in <enrichments><keyword_topic> — no PDF/UA structure tag |
| (no equivalent) | — | Table, THead, TBody, TR, TH, TD |
Table substructure only exists in the content tree |
| (no equivalent) | — | Link |
Inline hyperlinks — only in content tree |
| (no equivalent) | — | Code |
Code blocks — only in content tree |
| (no equivalent) | — | Artifact |
Non-content elements (page numbers, headers, footers, watermarks) |
| (no equivalent) | — | Span |
Inline formatting — only in content tree |
For scholarly publishing workflows. JATS is the standard for NLM/PubMed/PMC journal articles.
| Content Tree Tag | JATS Element | Notes |
|---|---|---|
H1 |
<article-title> |
Document title |
H2 |
<sec><title> |
Section heading (section wraps in <sec>) |
H3–H6 |
<sec><title> (nested) |
Nested section levels |
P |
<p> |
Direct mapping |
Figure |
<fig> |
Alt text → <alt-text> |
Caption |
<caption><p> |
Inside parent <fig> or <table-wrap> |
Table |
<table-wrap><table> |
THead/TBody/TR/TH/TD map to HTML table model |
L |
<list> |
list-type="bullet" or list-type="order" |
LI |
<list-item> |
Direct mapping |
Formula |
<disp-formula> / <inline-formula> |
By placement attribute |
BibEntry |
<ref><mixed-citation> |
Inside <ref-list> |
Reference |
<xref ref-type="bibr"> |
Inline citation marker |
Note |
<fn> |
Inside <fn-group> |
Link |
<ext-link> |
ext-link-type="uri" |
Code |
<code> |
JATS v1.2+ |
Quote/BlockQuote |
<disp-quote> |
Direct mapping |
Artifact |
(omit) | Not applicable in JATS |
TOC |
(omit) | Not applicable in JATS |
For digital library layout-aware OCR workflows. ALTO preserves spatial coordinates.
| Content Tree Tag | ALTO Element | Notes |
|---|---|---|
P |
<TextBlock> |
Each paragraph → a TextBlock |
H1–H6 |
<TextBlock STYLEREFS="heading"> |
Headings are TextBlocks with style refs |
Figure |
<Illustration> |
Alt text → <Description> (ALTO v4) |
Table |
<ComposedBlock TYPE="table"> |
Table → ComposedBlock containing TextBlocks |
L |
<ComposedBlock TYPE="list"> |
List → ComposedBlock |
Artifact |
(omit or mark as <GraphicalElement> ) |
Page decorations |
| Grid coordinates | HPOS, VPOS, WIDTH, HEIGHT |
BADGERS grid → pixel coordinates via page dimensions |
Note: ALTO is layout-centric, not semantic. The content tree's semantic tags (Formula, BibEntry, Note) don't have ALTO equivalents — they would be preserved as metadata annotations or lost in a pure ALTO export.
For general metadata interoperability. See also badgers-elements.jsonld.
| Content Tree Tag | Dublin Core | Schema.org | Notes |
|---|---|---|---|
H1 |
dc:title |
schema:name |
Document title |
P |
dc:description (if abstract) |
schema:text |
General text |
Figure |
— | schema:image |
With schema:caption from Caption |
BibEntry |
dc:references |
schema:citation |
Bibliographic reference |
Note |
— | schema:comment |
Footnotes/endnotes |
Formula |
— | schema:mathExpression |
Mathematical content |
Table |
— | schema:Table |
Tabular data |
TOC |
dcterms:tableOfContents |
— | Table of contents |
For digital preservation workflows. PREMIS tracks provenance events — pairs well with W3C PROV.
| BADGERS Concept | PREMIS Element | Notes |
|---|---|---|
| Analysis run | <premis:event> with eventType="analysis" |
One event per specialist execution |
| Source document | <premis:object> (intellectual entity) |
The input PDF |
| Content tree output | <premis:object> (representation) |
The correlation XML |
| Specialist Lambda | <premis:agent> with agentType="software" |
Agent that performed the event |
| Foundation model | <premis:agent> with agentType="software" |
Model version as agent identifier |
| Session provenance | <premis:event> chain |
Ordered sequence of analysis events |
| If you need to... | Use this schema | Map to this external standard |
|---|---|---|
| Classify document elements semantically | Elements (badgers-elements.xsd) |
Dublin Core / Schema.org via JSON-LD |
| Build an accessible PDF structure tree | Content Tree (badgers-content-tree.xsd) |
PDF/UA (ISO 14289-1) — already native |
| Export to a scholarly publishing pipeline | Content Tree | JATS XML |
| Integrate with a digital library system | Content Tree + Elements | ALTO XML (layout) + Dublin Core (metadata) |
| Track provenance of analysis results | Content Tree metadata + PROV | PREMIS or W3C PROV-O/PROV-XML |
| Serve linked-data consumers | Elements JSON-LD context | Schema.org / Dublin Core |
This mapping tracks BADGERS v5.0 (September 2026). The content tree schema version is 2.0 (introduced with the hierarchical content_tree replacing the flat content_spine). The elements vocabulary has no independent version number — it tracks the BADGERS release.