datHere/ckan_geoconnex_bulk_runner
1
0
Fork 0
mirror of https://github.com/dathere/ckan_geoconnex_bulk_runner.git synced 2026-07-05 15:12:20 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: enhanced cargo workspace, NM usage, Dockerfile

Browse source
This commit is contained in:
rzmk 2026-06-15 11:20:51 -04:00
parent 71b08a53f0
commit 3a79fb2b0a
18 changed files with 362 additions and 2478 deletions
Download patch file Download diff file Expand all files Collapse all files

76
geoconnex_utils/src/schema.rs Normal file
View file

@ -0,0 +1,76 @@
use serde_json::json;
pub fn get_dataset_schema() -> serde_json::Value {
// Allow for "local" feature
#[allow(unused_mut)]
let mut dataset_schema = json!({
"type": "object",
"properties": {
"@context": {"type": ["string", "object"]},
"@type": {"type": ["string", "array"], "contains": {"anyOf": [{"const": "Dataset"}, {"const": "schema:Dataset"}]}},
"@id": {"type": "string"},
"name": {"type": "string"},
"schema:name": {"type": "string"},
"provider": {
"type": "object",
"properties": {"@type": {"type": "string"}, "name": {"type": "string"}},
},
"schema:provider": {
"type": "object",
"properties": {"@type": {"type": "string"}, "name": {"type": "string"}},
},
"gsp:hasGeometry": {"@type": "object"},
"about": {
"type": ["string", "array"],
"items": {
"type": "object",
"properties": {"@id": {"type": "string"}, "@type": {"const": "Place"}},
},
"minItems": 1
},
},
"anyOf": [
{ "required": ["@context", "@type", "@id", "name", "provider", "about"] },
{ "required": ["@context", "@type", "@id", "name", "provider", "gsp:hasGeometry"] },
{ "required": ["@context", "@type", "@id", "schema:name", "schema:provider", "about"] },
{ "required": ["@context", "@type", "@id", "schema:name", "schema:provider", "gsp:hasGeometry"] },
// { "required": ["@context", "@type", "@id", "name", "provider"] }
]
});
// Some JSON-LD for datasets (e.g. sciencebase) do not have about or gsp:hasGeometry yet are still valid as per SHACL shape
#[cfg(feature = "local")]
{
let required_array = dataset_schema
.get_mut("anyOf")
.unwrap()
.as_array_mut()
.unwrap();
required_array.insert(
required_array.len(),
json!({ "required": ["@context", "@type", "@id", "schema:name", "schema:provider"] }),
);
}
dataset_schema
}
pub fn get_location_schema() -> serde_json::Value {
json!({
"type": "object",
"properties": {
"@context": {"type": ["string", "object"]},
"@type": {"type": ["string", "array"], "contains": {"const": "Place"}},
"@id": {"type": "string"},
"name": {"type": "string"},
"provider": {
"type": "object",
"properties": {"@type": {"type": "string"}, "name": {"type": "string"}},
},
"geo": {"type": "object"},
"gsp:hasGeometry": {"type": "object"}
},
"anyOf": [
{ "required": ["@context", "@type", "@id", "name", "provider", "geo", "gsp:hasGeometry"] },
{ "required": ["@context", "@type", "@id", "schema:name", "schema:provider", "geo", "gsp:hasGeometry"] },
]
})
}