Library Stylesheets

The framework scaffolds a set of generic XSLT stylesheets into every new project under source/stylesheets/lib/. They handle the parts of the pipeline that don't depend on project-specific decisions: extracting metadata, aggregating indices and search data, generating Eleventy sidecar files, pruning multilingual content, and wrapping the EpiDoc/SigiDoc rendering stylesheets.

You typically don't edit these files. Project-specific logic lives in metadata-config.xsl (which these stylesheets invoke via hook templates), in pipeline.xml (which wires them together and passes parameters), and in overrides.xsl (imported by epidoc-to-html.xsl for rendering tweaks).

Because these files are part of the scaffold, they live inside your project and are yours to modify if you need to. See Designing Sustainable Projects for the rationale.

`extract-metadata.xsl`

Produces a per-document metadata XML file from one TEI/EpiDoc source document. The file it emits is the canonical intermediate artefact consumed by create-11ty-data.xsl, aggregate-indices.xsl, and aggregate-search-data.xsl.

For each configured language, it invokes three hook templates that live in your project's metadata-config.xsl, then stamps xml:lang onto their output and merges entities with the same @xml:id across languages into unified elements.

Parameters

Parameter	Default	Description
`source-file`	`base-uri()`	Path/URI of the source document (used to derive `documentId`)
`languages`	`'en'`	Space-separated list of language codes

Hooks (in `metadata-config.xsl`)

Mode	Purpose
`extract-metadata`	Page display fields (title, sortKey, dates, etc.)
`extract-all-entities`	`<entity indexType="...">` elements for indices
`extract-search`	Search facet data. Multi-valued fields emitted as `<item>` children are deduplicated automatically, per document and per language

All hooks receive a $language tunnel parameter.

Input / Output

Input: a TEI/EpiDoc XML document.

Output (abbreviated):

xml

<metadata>
  <documentId>Feind_Kr1</documentId>
  <sourceFile>Feind_Kr1.xml</sourceFile>
  <page>
    <title xml:lang="en">Kephalas, dignitary...</title>
    <title xml:lang="de">Kephalas, Würdenträger...</title>
    <sortKey xml:lang="en">kephalas...</sortKey>
  </page>
  <entities>
    <persons>
      <entity xml:id="p01" indexType="persons">
        <name xml:lang="en">Kephalas</name>
        <name xml:lang="de">Kephalas</name>
      </entity>
    </persons>
  </entities>
  <search>
    <centuries xml:lang="en"><item>8</item></centuries>
    <places xml:lang="en"><item>Kephalonia</item></places>
  </search>
</metadata>

`create-11ty-data.xsl`

Converts one metadata XML into a .11tydata.json sidecar file that Eleventy reads alongside the corresponding HTML fragment. The sidecar tells Eleventy which layout to use, what collection tags to assign, and what fields to expose to templates.

Parameters

Parameter	Default	Description
`layout`	`''`	Eleventy layout name (e.g. `document.njk`)
`tags`	`''`	Eleventy collection tags (e.g. `seals`)
`language`	`'en'`	Language to select from `/metadata/page/*[@xml:lang]`

Input / Output

Input: one metadata XML file (output of extract-metadata.xsl).

Output (abbreviated):

json

{
  "layout": "document.njk",
  "tags": "seals",
  "documentId": "Feind_Kr1",
  "title": "Kephalas, dignitary...",
  "sortKey": "kephalas..."
}

All /metadata/page/* elements matching the requested language are flattened into top-level keys. Elements with child elements (e.g. <textType><item>…</item><item>…</item></textType>) become JSON arrays.

`aggregate-indices.xsl`

Reads per-document metadata XML files from a whole collection and produces one JSON file per index type (e.g. persons.json, places.json) plus a _summary.json describing all indices. Driven by <idx:index> configuration in metadata-config.xsl.

Entities with the same @xml:id are merged into one index entry (the entry's references array records each document the entity appears in). Entities without @xml:id are unique per occurrence.

Parameters

Parameter	Default	Description
`metadata-files`	—	Space-separated absolute paths to metadata XML files
`metadata-config`	—	Absolute path to `metadata-config.xsl` (for reading `<idx:index>` declarations)

Multilingual handling

Fields with @xml:lang produce language-keyed JSON objects, {"en": "…", "de": "…"}. Fields without @xml:lang produce plain strings.

Input / Output

Input: multiple metadata XML files (via metadata-files) and the project's metadata-config.xsl (via metadata-config).

Principal output (_summary.json):

json

{
  "indices": [
    { "id": "persons", "title": {"en": "Persons", "de": "Personen"},
      "description": {"en": "All persons in the corpus."},
      "order": 1, "nav": "indices", "entryCount": 145 }
  ]
}

Per-index result documents (e.g. persons.json):

json

{
  "id": "persons",
  "title": {"en": "Persons"},
  "columns": [
    { "key": "name", "header": {"en": "Name"} }
  ],
  "entries": [
    {
      "name": {"en": "Kephalas"},
      "sortKey": "kephalas",
      "references": [
        { "documentId": "Feind_Kr1", "name": {"en": "Kephalas"} }
      ]
    }
  ]
}

`aggregate-search-data.xsl`

Reads per-document metadata XML files and emits a JSON array of search documents for one language, consumed by the client-side efes-search component (which builds a FlexSearch index and computes facet counts from it at page load).

For a multi-language project, run one pipeline node per language with a different $language parameter and a different output filename (e.g. documents_en.json, documents_de.json).

Parameters

Parameter	Default	Description
`metadata-files`	—	Space-separated absolute paths to metadata XML files
`language`	first available	Language code to extract

Input / Output

Input: multiple metadata XML files.

Output (abbreviated):

json

[
  { "documentId": "Feind_Kr1",
    "centuries": ["8"],
    "places": ["Kephalonia"],
    "title": "Kephalas, dignitary..." },
  { "documentId": "Feind_Kr2", "... ": "..." }
]

Multi-valued fields (those emitted as <item> children by the extract-search hook) become JSON arrays. Scalar fields become strings.

`prune-to-language.xsl`

Strips elements with an xml:lang attribute that doesn't match the target language, leaving a mono-lingual TEI document. Used as a preprocessing step when the source XML encodes multiple language variants and the downstream rendering or aggregation only expects one.

Identity-transform at heart; the only rules are:

Remove *[@xml:lang != $language]
Preserve tei:foreign (language-tagged but shouldn't be pruned)
Preserve tei:div[@type='edition'], its textpart and ab descendants, and tei:div[@type='translation'] (they carry their own language semantics handled by the EpiDoc stylesheets)

Parameters

Parameter	Default	Description
`language`	—	Target language code (e.g. `en`, `de`, `el`)

Input / Output

Input: a multilingual TEI/EpiDoc document.

Output: the same document with non-matching-language material removed.

`epidoc-to-html.xsl`

Thin wrapper around the upstream EpiDoc (or SigiDoc) stylesheets' start-edition.xsl. Gives the pipeline a single entry point for rendering a source document to an HTML fragment, with project-specific parameter wiring and an optional i18n substitution pass.

This is the one library stylesheet most likely to need editing, because it scaffolds the import path for the upstream stylesheets and the name of the body-structure template (sigidoc-body-structure or inslib-body-structure) based on the stylesheet flavour chosen at project-creation time. If you switch flavours, or add custom rendering params, adjust this file (and overrides.xsl, which it also imports).

Parameters

Pipeline wires values through for these tunnel parameters:

Group	Parameters
EpiDoc rendering	`edn-structure`, `edition-type`, `external-app-style`, `internal-app-style`, `leiden-style`, `line-inc`, `verse-lines`, `hgv-gloss`, `css-loc`, `image-loc`
Bibliography	`bib`, `bibloc`, `bib-link-template`
Authority files	`authority-dir`, `symbols-file`, `places-file`, `institutions-file`
Glyphs	`glyph-variant`
i18n	`language`, `messages-file`

When language (or messages-file) is set, the stylesheet loads ../../translations/messages_{language}.xml and replaces every <i18n:text i18n:key="…"/> in the rendered output with the corresponding translation. Unknown keys render as [key] so missing translations are visible, not silent.

Input / Output

Input: a TEI/EpiDoc or SigiDoc XML document.

Output: an HTML fragment (rendered transcription, metadata, commentary), ready for assembly into a full page by Eleventy.

Library Stylesheets ​

extract-metadata.xsl ​

Parameters ​

Hooks (in metadata-config.xsl) ​

Input / Output ​

create-11ty-data.xsl ​

Parameters ​

Input / Output ​

aggregate-indices.xsl ​

Parameters ​

Multilingual handling ​

Input / Output ​

aggregate-search-data.xsl ​

Parameters ​

Input / Output ​

prune-to-language.xsl ​

Parameters ​

Input / Output ​

epidoc-to-html.xsl ​

Parameters ​

Input / Output ​

Library Stylesheets

`extract-metadata.xsl`

Parameters

Hooks (in `metadata-config.xsl`)

Input / Output

`create-11ty-data.xsl`

Parameters

Input / Output

`aggregate-indices.xsl`

Parameters

Multilingual handling

Input / Output

`aggregate-search-data.xsl`

Parameters

Input / Output

`prune-to-language.xsl`

Parameters

Input / Output

`epidoc-to-html.xsl`

Parameters

Input / Output