feat(yasqe): always bundle the railML and RailTopoModel terms

.changeset/lov-rail-vocabularies.md

@@ -0,0 +1,5 @@
+---
+"@matdata/yasqe": patch
+---
+
+The bundled class and property suggestions always include the ERA ontology (`era:`), railML 3 (`railml:`) and RailTopoModel (`rtm:`), which are not in the LOV dump.

docs/developer-guide.md

@@ -1042,7 +1042,7 @@ Return the terms most used first; Yasqe filters them on what the user typed and
 
 #### Linked Open Vocabularies Terms
 
-The LOV suggestions come from `packages/yasqe/src/lov-terms.json`, which is bundled with Yasqe: no request is made to the LOV API at runtime. The file holds the classes and properties of the most used LOV vocabularies (ranked by the number of vocabularies that use them, up to about 600 KB) and is generated from the LOV N-Quads dump:
+The LOV suggestions come from `packages/yasqe/src/lov-terms.json`, which is bundled with Yasqe: no request is made to the LOV API at runtime. The file holds the classes and properties of the most used LOV vocabularies (ranked by the number of vocabularies that use them, up to about 600 KB) and is generated from the LOV N-Quads dump. The namespaces in `ALWAYS_INCLUDED` in `fetchLovData.js` (the ERA ontology, railML 3 and RailTopoModel) are added first, whatever their rank or the size budget, and their terms are read from the vocabulary's own source when the dump has none:
 
 ```bash
 npm run util:fetchLov                      # downloads https://lov.linkeddata.es/lov.nq.gz

docs/user-guide.md

@@ -858,7 +858,7 @@ Ctrl+clicking on `http://dbpedia.org/resource/European_Union` automatically quer
 Press `Ctrl+Space` after `a` (or `rdf:type`) to get class suggestions, or in the position of a property to get property suggestions. MatGUI suggests:
 
 1. **Used in this endpoint**: the classes and properties the endpoint of the tab actually uses, most used first, with the number of instances (classes) or triples (properties) next to them. This also covers vocabularies that are not registered anywhere, such as railML 3, the ERA ontology or your own ontologies.
-2. **Linked Open Vocabularies**: matching classes and properties of the most used vocabularies in [LOV](https://lov.linkeddata.es/dataset/lov/). MatGUI ships this list, so no request is made to LOV while you type. Terms that the endpoint already suggested are not repeated, and a term is inserted as a prefixed name with its `PREFIX` when LOV knows the vocabulary's preferred prefix.
+2. **Linked Open Vocabularies**: matching classes and properties of the most used vocabularies in [LOV](https://lov.linkeddata.es/dataset/lov/). MatGUI ships this list, so no request is made to LOV while you type. Terms that the endpoint already suggested are not repeated, and a term is inserted as a prefixed name with its `PREFIX` when LOV knows the vocabulary's preferred prefix. The ERA ontology (`era:`), railML 3 (`railml:`) and RailTopoModel (`rtm:`) are always included, even though they are not in LOV.
 
 Without typing anything, the endpoint's most used terms are listed. While you type, a suggestion matches on its full IRI, its prefixed name (`era:Op`) or, when you type a plain word (`Operat`), on any part of its local name.
 

fetchLovData.js

@@ -41,6 +41,18 @@ export const ALWAYS_INCLUDED = [
       "https://gitlab.com/era-europa-eu/public/interoperable-data-programme/era-ontology/era-ontology/-/raw/main/ontology.ttl",
     ],
   },
+  {
+    // railML 3 ontology (not in LOV)
+    namespace: "http://ontology.railml.org/railml3#",
+    prefix: "railml",
+    sources: ["https://ontology.railml.org/ontology.ttl"],
+  },
+  {
+    // RailTopoModel, as published in the railML ontology
+    namespace: "http://ontology.railml.org/railtopomodel#",
+    prefix: "rtm",
+    sources: ["https://ontology.railml.org/ontology.ttl"],
+  },
 ];
 
 const RDF_TYPE = "http://www.w3.org/1999/02/22-rdf-syntax-ns#type";