Check for schema id to use schema endpoint (#415)

* Check for schema id to use schema endpoint * Add example in unit test * Add missing parameter inside _register_one. Add notebook example of schema handling * check if resource_id is given
BlueBrain · Jun 28, 2024 · 9ff022b · 9ff022b
1 parent a9c9398
commit 9ff022b
Show file tree

Hide file tree

Showing 5 changed files with 290 additions and 8 deletions.
diff --git a/examples/notebooks/use-cases/BBP KG Creating and Modifying Schemas.ipynb b/examples/notebooks/use-cases/BBP KG Creating and Modifying Schemas.ipynb
@@ -0,0 +1,254 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "b14744df-3567-40f3-bb0c-06c6767ab695",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import json\n",
+    "import uuid\n",
+    "\n",
+    "from kgforge.core import KnowledgeGraphForge\n",
+    "from kgforge.specializations.resources import Dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "72596414-c9ac-4312-a621-5bd4fdc3477b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import getpass\n",
+    "TOKEN = getpass.getpass()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "c124caae-b9ae-4eb4-8d8e-bc437d28d3e6",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "BUCKET = \"dke/kgforge\"\n",
+    "\n",
+    "forge = KnowledgeGraphForge(\"../use-cases/prod-forge-nexus.yml\",\n",
+    "    bucket=BUCKET,\n",
+    "    token=TOKEN\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "29bceff6-d7b4-4fda-9cc3-fe1923c0f865",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def make_id(i):\n",
+    "    return f\"{forge._store.endpoint}/schemas/{forge._store.bucket}/dummy_schema_{i}\"\n",
+    "\n",
+    "def create_schema(i):\n",
+    "    payload = {\n",
+    "    \"context\": \"https://incf.github.io/neuroshapes/contexts/schema.json\",\n",
+    "    \"id\": make_id(i),\n",
+    "    \"type\": \"Schema\",\n",
+    "    \"imports\": [\n",
+    "          \"https://neuroshapes.org/commons/entity\"\n",
+    "        ],\n",
+    "        \"shapes\": [\n",
+    "          {\n",
+    "            \"@id\": f\"{make_id(i)}/shapes/DummyShape\",\n",
+    "            \"@type\": \"NodeShape\",\n",
+    "            \"node\": \"https://neuroshapes.org/commons/entity/shapes/EntityShape\",\n",
+    "          }\n",
+    "        ]\n",
+    "    }\n",
+    "    resource = forge.from_json(payload)\n",
+    "    forge.register(resource, schema_id=forge._store.service.SHACL_SCHEMA)\n",
+    "    return resource"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e79f1807-dbfa-4cca-ab08-4809c9962517",
+   "metadata": {},
+   "source": [
+    "## 1. Create dummy schemas"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "9a022c3d",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<action> _register_one\n",
+      "<succeeded> True\n",
+      "<action> _register_one\n",
+      "<succeeded> True\n",
+      "<action> _register_one\n",
+      "<succeeded> True\n"
+     ]
+    }
+   ],
+   "source": [
+    "dummies = [create_schema(uuid.uuid4()) for j in range(3)]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d10b21d5-212b-48cc-b4a2-7ef2f739b2ed",
+   "metadata": {},
+   "source": [
+    "## 2. Retrieve the reciently created schemas using the ids and tag them"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "515b3e19",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "retrieved = [forge.retrieve(dummy.id) for dummy in dummies]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "2c5594ee",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{\n",
+      "    context: https://incf.github.io/neuroshapes/contexts/schema.json\n",
+      "    id: https://bbp.epfl.ch/nexus/v1/schemas/dke/kgforge/dummy_schema_7fab9433-af89-43eb-878a-607389240256\n",
+      "    type: Schema\n",
+      "    imports:\n",
+      "    [\n",
+      "        https://neuroshapes.org/commons/entity\n",
+      "    ]\n",
+      "    shapes:\n",
+      "    [\n",
+      "        {\n",
+      "            id: https://bbp.epfl.ch/nexus/v1/schemas/dke/kgforge/dummy_schema_7fab9433-af89-43eb-878a-607389240256/shapes/DummyShape\n",
+      "            type: NodeShape\n",
+      "            node: https://neuroshapes.org/commons/entity/shapes/EntityShape\n",
+      "        }\n",
+      "    ]\n",
+      "}\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(retrieved[0])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "efb9cb41",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "https://bluebrain.github.io/nexus/schemas/shacl-20170720.ttl\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(retrieved[0]._store_metadata['_constrainedBy'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "a4f95633",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<action> _tag_one\n",
+      "<succeeded> True\n",
+      "<action> _tag_one\n",
+      "<succeeded> True\n",
+      "<action> _tag_one\n",
+      "<succeeded> True\n"
+     ]
+    }
+   ],
+   "source": [
+    "for r in retrieved:\n",
+    "    forge.tag(r, 't01')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "551f3fdc-4760-4c6b-861b-6ac085f27e00",
+   "metadata": {},
+   "source": [
+    "## 3. Deprecate the schemas"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "e6859730",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<action> _deprecate_one\n",
+      "<succeeded> True\n",
+      "<action> _deprecate_one\n",
+      "<succeeded> True\n",
+      "<action> _deprecate_one\n",
+      "<succeeded> True\n"
+     ]
+    }
+   ],
+   "source": [
+    "for r in retrieved:\n",
+    "    forge.deprecate(r)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "dev-forge",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.8.18"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/kgforge/specializations/stores/bluebrain_nexus.py b/kgforge/specializations/stores/bluebrain_nexus.py
@@ -164,9 +164,8 @@ def register_callback(task: Task):
 
     def _register_one(self, resource: Resource, schema_id: str) -> None:
         method, url, resource, exception_, headers, params, payload = (
-            prepare_methods.prepare_create(service=self.service, resource=resource)
+            prepare_methods.prepare_create(service=self.service, resource=resource, schema_id=schema_id)
         )
-
         response = requests.request(
             method=method,
             url=url,

diff --git a/kgforge/specializations/stores/nexus/prepare_methods.py b/kgforge/specializations/stores/nexus/prepare_methods.py
@@ -18,6 +18,20 @@
 #  but since the service is available, the param can be retrieved here
 
 
+def _make_url(service: Service,
+              schema_id: str,
+              identifier: str):
+    if schema_id == service.SHACL_SCHEMA:
+        url = Service.add_resource_id_to_endpoint(
+            service.url_schemas, resource_id=identifier
+        )
+    else:
+        url = Service.add_schema_and_id_to_endpoint(
+            service.url_resources, schema_id=schema_id, resource_id=identifier
+        )
+    return url
+
+
 def prepare_create(
         service: Service,
         resource: Resource,
@@ -30,9 +44,7 @@ def prepare_create(
 
     identifier = resource.get_identifier()
 
-    url = Service.add_schema_and_id_to_endpoint(
-        service.url_resources, schema_id=schema_id, resource_id=identifier
-    )
+    url = _make_url(service, schema_id, identifier)
 
     context = service.model_context or service.context
 
@@ -86,9 +98,7 @@ def _prepare_uri(
     if schema_id == service.UNCONSTRAINED_SCHEMA and not keep_unconstrained:
         schema_id = None
 
-    url = Service.add_schema_and_id_to_endpoint(
-        service.url_resources, schema_id, resource_id=resource.id
-    )
+    url = _make_url(service, schema_id, resource.id)
 
     rev = resource._store_metadata._rev
     params = {"rev": rev}

diff --git a/kgforge/specializations/stores/nexus/service.py b/kgforge/specializations/stores/nexus/service.py
@@ -63,6 +63,7 @@ class Service:
     DEPRECATED_PROPERTY_FALLBACK = f"{NEXUS_NAMESPACE_FALLBACK}deprecated"
     PROJECT_PROPERTY_FALLBACK = f"{NEXUS_NAMESPACE_FALLBACK}project"
     UNCONSTRAINED_SCHEMA = "https://bluebrain.github.io/nexus/schemas/unconstrained.json"
+    SHACL_SCHEMA = "https://bluebrain.github.io/nexus/schemas/shacl-20170720.ttl"
 
     SPARQL_ENDPOINT_TYPE = "sparql"
     ELASTIC_ENDPOINT_TYPE = "elastic"
@@ -150,6 +151,7 @@ def __init__(
         self.url_files = Service.make_endpoint(self.endpoint, "files", org, prj)
         self.url_resources = Service.make_endpoint(self.endpoint, "resources", org, prj)
         self.url_resolver = Service.make_endpoint(self.endpoint, "resolvers", org, prj)
+        self.url_schemas = Service.make_endpoint(self.endpoint, "schemas", org, prj)
 
         self.metadata_context = Context(
             recursive_resolve(
@@ -210,6 +212,15 @@ def make_endpoint(endpoint: str, endpoint_type: str, organisation: str, project:
             (endpoint, endpoint_type, quote_plus(organisation), quote_plus(project))
         )
 
+    @staticmethod
+    def add_resource_id_to_endpoint(
+            endpoint: str, resource_id: Optional[str]
+    ):
+        if resource_id:
+            return "/".join([endpoint, quote_plus(resource_id)])
+        else:
+            return endpoint
+
     @staticmethod
     def add_schema_and_id_to_endpoint(
             endpoint: str, schema_id: Optional[str], resource_id: Optional[str]

diff --git a/tests/specializations/stores/test_bluebrain_nexus.py b/tests/specializations/stores/test_bluebrain_nexus.py
@@ -304,6 +304,14 @@ def test_to_resource(nexus_store, registered_building, building_jsonld, store_co
             ("/".join((NEXUS, "resources", BUCKET, quote_plus("_"), "{}", "tags"))),
             id="tag-unconstrained",
         ),
+        pytest.param(
+            (Service.SHACL_SCHEMA),
+            (Service.SHACL_SCHEMA),
+            ({"rev": 1}),
+            ("/".join((NEXUS, "schemas", BUCKET, "{}"))),
+            ("/".join((NEXUS, "schemas", BUCKET, "{}", "tags"))),
+            id="tag-schema",
+        ),
     ],
 )
 def test_prepare_tag_uri(