Merge pull request #35 from eScienceLab/development

AnnieZQH · web-flow · commit 86658a4445a7 · 2026-01-22T16:18:54.000Z
Merge dev changes into main
diff --git a/.docker/sampo-ui/deploy.dockerfile b/.docker/sampo-ui/deploy.dockerfile
@@ -1,6 +1,6 @@
-FROM node:16.13.0-alpine
+FROM node:22.21.1-alpine
 ARG API_URL
-ARG VERSION="3.0.0"
+ARG VERSION="3.0.0-1"
 
 # Based on https://nodejs.org/en/docs/guides/nodejs-docker-webapp/
 
@@ -11,7 +11,7 @@ WORKDIR /usr/src/app
 # mv commands: install app dependencies, Babel 7 presets and plugins, and bundle app source
 # Remove redundant files
 RUN <<EOF
-wget https://github.com/SemanticComputing/sampo-ui/archive/refs/tags/v$VERSION.zip
+wget https://github.com/UoMResearchIT/sampo-ui/archive/refs/tags/v$VERSION.zip
 unzip v$VERSION.zip
 mv ./sampo-ui-$VERSION/package*.json ./
 mv ./sampo-ui-$VERSION/webpack*.js ./
diff --git a/.docker/sampo-ui/dev.dockerfile b/.docker/sampo-ui/dev.dockerfile
@@ -1,7 +1,7 @@
-FROM node:16.13.0-alpine
+FROM node:22.21.1-alpine
 
 # Specify the Sampo UI version
-ARG VERSION="3.0.0"
+ARG VERSION="3.0.0-1"
 
 RUN apk update && apk add bash
 
@@ -12,7 +12,7 @@ WORKDIR /usr/src/app
 # mv commands: install app dependencies, Babel 7 presets and plugins, and bundle app source
 # Remove redundant files
 RUN <<EOF
-wget https://github.com/SemanticComputing/sampo-ui/archive/refs/tags/v$VERSION.zip
+wget https://github.com/UoMResearchIT/sampo-ui/archive/refs/tags/v$VERSION.zip
 unzip v$VERSION.zip
 mv ./sampo-ui-$VERSION/package*.json ./
 mv ./sampo-ui-$VERSION/webpack*.js ./
diff --git a/.github/workflows/check-profile-pr.yml b/.github/workflows/check-profile-pr.yml
@@ -0,0 +1,30 @@
+name: Check added profile link (if applicable)
+
+on:
+  pull_request:
+    types: [opened, synchronize, reopened]
+  workflow_dispatch:
+jobs:
+  check:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+      - name: Check if profile URLs changed
+        id: changes
+        run: |
+          git fetch origin ${{ github.event.pull_request.base.ref }}
+          if git diff --name-only origin/${{ github.event.pull_request.base.ref }}...HEAD | grep -qx "scripts/profile_urls.txt"; then 
+            echo "changed=true" >> "$GITHUB_OUTPUT"
+          else
+            echo "changed=false" >> "$GITHUB_OUTPUT"
+          fi
+      - name: Install dependencies
+        if: steps.changes.outputs.changed == 'true'
+        run: python3 -m pip install -r requirements.workflow.txt
+      - name: Run script
+        if: steps.changes.outputs.changed == 'true'
+        run: python upload_profiles.py --dry-run
+        working-directory: ./scripts
diff --git a/.github/workflows/upload-profiles.yml b/.github/workflows/upload-profiles.yml
@@ -0,0 +1,24 @@
+name: Upload to Jena Fuseki
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - 'scripts/profile_urls.txt'
+  workflow_dispatch:
+jobs:
+  upload:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+      - name: Install dependencies
+        run: python3 -m pip install -r requirements.workflow.txt
+      - name: Run script
+        run: python upload_profiles.py
+        working-directory: ./scripts
+        env:
+          FUSEKI_UPLOAD_ENDPOINT: ${{ vars.ENV_FUSEKI_UPLOAD_ENDPOINT }}
+          FUSEKI_PASSWORD: ${{ secrets.ENV_FUSEKI_PASSWORD }}
+
diff --git a/.gitignore b/.gitignore
@@ -1,2 +1,3 @@
 # Environments
 .env
+venv/
diff --git a/README.md b/README.md
@@ -9,7 +9,7 @@ RO-Crates are a method for packaging research data with their metadata. RO-Crate
 The profile portal is accepting contributions! 
 
 > [!IMPORTANT]
-> To be accepted, the profile must be a [Profile Crate](url) accessible on the public internet.
+> To be accepted, the profile must be a [Profile Crate](https://www.researchobject.org/ro-crate/specification/1.2/profiles.html#profile-crate) accessible on the public internet.
 
 To add your profile (or a profile you feel is missing):
 - Open `scripts/profile_urls.txt` in this repo for editing (click file, then click pencil icon)
diff --git a/env.template b/env.template
@@ -1,6 +1,7 @@
 # Template file for populating .env
 
 FUSEKI_PASSWORD=${ENV_FUSEKI_PASSWORD}
+FUSEKI_UPLOAD_ENDPOINT=
 
 # Local http://localhost:3006/api/v1
 SAMPO_API_URL=${ENV_SAMPO_API_URL}
diff --git a/requirements.workflow.in b/requirements.workflow.in
@@ -0,0 +1,4 @@
+rdflib==7.4.0
+validators==0.35.0
+requests==2.32.5
+python-dotenv==1.2.1
diff --git a/requirements.workflow.txt b/requirements.workflow.txt
@@ -0,0 +1,24 @@
+#
+# This file is autogenerated by pip-compile with Python 3.12
+# by the following command:
+#
+#    pip-compile --output-file=requirements.workflow.txt requirements.workflow.in
+#
+certifi==2025.11.12
+    # via requests
+charset-normalizer==3.4.4
+    # via requests
+idna==3.11
+    # via requests
+pyparsing==3.2.5
+    # via rdflib
+python-dotenv==1.2.1
+    # via -r requirements.workflow.in
+rdflib==7.4.0
+    # via -r requirements.workflow.in
+requests==2.32.5
+    # via -r requirements.workflow.in
+urllib3==2.6.2
+    # via requests
+validators==0.35.0
+    # via -r requirements.workflow.in
diff --git a/scripts/gaps.ttl b/scripts/gaps.ttl
@@ -0,0 +1,17 @@
+@prefix schema: <http://schema.org/> .
+
+<https://w3id.org/workflowhub/workflow-ro-crate/1.0> 
+    a schema:Profile ;
+    schema:keywords "workflow" .
+<https://w3id.org/ro/wfrun/process/0.5> 
+    schema:datePublished "2024-06-19";
+    schema:keywords "provenance", "process", "informatics" .
+<https://w3id.org/ro/wfrun/workflow/0.5> 
+    schema:datePublished "2024-06-19";
+    schema:keywords "workflow", "provenance", "workflow execution", "process", "informatics" .
+<https://w3id.org/ro/wfrun/provenance/0.5> 
+    schema:datePublished "2024-06-19";
+    schema:keywords "workflow", "provenance", "workflow execution", "process", "informatics" .
+<https://w3id.org/5s-crate/0.4> 
+    schema:identifier "https://w3id.org/5s-crate/0.4" ;
+    schema:keywords "workflow execution", "five safes", "transparency", "sensitive data", "trusted research environment", "secure data environment", "data safe haven" .
diff --git a/scripts/profile_urls.txt b/scripts/profile_urls.txt
@@ -0,0 +1,5 @@
+https://trefx.uk/5s-crate/0.4/ro-crate-metadata.json
+https://www.researchobject.org/workflow-run-crate/profiles/0.5/process_run_crate/ro-crate-metadata.json
+https://www.researchobject.org/workflow-run-crate/profiles/0.5/workflow_run_crate/ro-crate-metadata.json
+https://www.researchobject.org/workflow-run-crate/profiles/0.5/provenance_run_crate/ro-crate-metadata.json
+https://www.researchobject.org/ro-crate/specification/1.2/ro-crate-metadata.json
diff --git a/scripts/upload_profiles.py b/scripts/upload_profiles.py
@@ -0,0 +1,109 @@
+import os
+import re
+import validators
+import requests
+import argparse
+
+from datetime import datetime, date
+from dotenv import load_dotenv
+from hashlib import md5
+from rdflib import Graph, URIRef, Literal, XSD, RDF, OWL, Namespace
+from urllib.parse import urljoin
+
+
+SCHEMA = Namespace("http://schema.org/")
+
+def map_sameas_uri(uri):
+        id = md5(uri.encode('utf-8')).hexdigest()
+        new_uri = URIRef(f"https://profiles.ro-crate.org/data/profile/{id}")
+        return new_uri
+
+def main(dry_run):
+    g = Graph()
+    profile_class = URIRef("http://www.w3.org/ns/dx/prof/Profile")
+
+    with open("profile_urls.txt", "r") as file:
+        profile_urls = [line.strip() for line in file if line.strip()]
+
+    for url in profile_urls:
+        if not validators.url(url):
+            raise ValueError(f"{url} is not an URL")
+
+        headers = {
+            "Accept": "application/ld+json, application/json"
+        }
+
+        try:
+            response = requests.head(url, headers=headers, allow_redirects=True)
+        except requests.RequestException as e:
+            raise ValueError(f"Unable to reach {url} (Error: {e})")
+        
+        # Fallback
+        if response.status_code >= 400:
+            response = requests.get(url, headers=headers)
+
+        content_type = response.headers.get("Content-Type", "").lower()
+        if "json" not in content_type:
+            raise ValueError(f"{url} does not return JSON-LD (Content type: {content_type})")
+
+        temp_g = Graph()
+        base_iri = urljoin(url, '.') if url.endswith("ro-crate-metadata.json") else f"{url.rstrip('/')}/"
+        temp_g.parse(url, format="json-ld", publicID=base_iri)
+        if any(temp_g.subjects(RDF.type, profile_class)):
+            g += temp_g
+        else:
+            raise ValueError(f"No profile entity found in {url}")
+
+    g.parse("gaps.ttl", format="turtle")
+
+    datetime_pattern = re.compile(r"^-?\d{4}-\d{2}-\d{2}T\d{2}:\d{2}:\d{2}(Z|[+-]\d{2}:\d{2})?$")
+    date_pattern = re.compile(r"^-?\d{4}-\d{2}-\d{2}$")
+    time_pattern = re.compile(r"^\d{2}:\d{2}:\d{2}(Z|[+-]\d{2}:\d{2})?$")
+
+    for s, p, o in g.triples((None, None, None)):
+        typed_o = None
+
+        if datetime_pattern.match(o):
+            typed_o = Literal(o, datatype=XSD.dateTime)
+        elif date_pattern.match(o):
+            typed_o = Literal(o, datatype=XSD.date)
+        elif time_pattern.match(o):
+            typed_o = Literal(o, datatype=XSD.time)
+
+        # Add type to datetime, date and time data
+        if typed_o is not None:
+            g.add((s, p, typed_o))
+            g.remove((s, p, o))
+
+        # Create profile URI for use in the profile portal
+        if p == RDF.type and o == profile_class:
+            new_s = map_sameas_uri(s)
+            g.add((new_s, RDF.type, o))
+            g.add((new_s, OWL.sameAs, s))
+
+        # Add triple <new profile URI> schema:datePublished "yyyy-mm-dd"^^xsd:date
+        if p == SCHEMA.datePublished and typed_o is not None:
+            value = typed_o.toPython()
+            if isinstance(value, datetime):
+                new_s = map_sameas_uri(s)
+                g.add((new_s, p, Literal(value.date(), datatype=XSD.date)))
+            if isinstance(value, date):
+                new_s = map_sameas_uri(s)
+                g.add((new_s, p, Literal(value.isoformat(), datatype=XSD.date)))
+
+    ttl_data = g.serialize(format="turtle")
+    if not dry_run:
+        FUSEKI_ENDPOINT = os.getenv('FUSEKI_UPLOAD_ENDPOINT')
+        FUSEKI_PASSWORD = os.getenv('FUSEKI_PASSWORD')
+
+        response = requests.post(FUSEKI_ENDPOINT, data=ttl_data, 
+                                headers={"Content-Type": "text/turtle"}, 
+                                auth=("admin", FUSEKI_PASSWORD))
+        response.raise_for_status()
+
+if __name__ == "__main__":
+    load_dotenv()
+    parser = argparse.ArgumentParser(description="Parse profile Ro Crate and upload graph to Jena Fuseki instance")
+    parser.add_argument("--dry-run", action="store_true")
+    args = parser.parse_args()
+    main(args.dry_run)
diff --git a/src/client/components/App.js b/src/client/components/App.js
@@ -1,7 +1,7 @@
 import React from 'react'
 import { ThemeProvider, createTheme } from '@mui/material/styles'
-import AdapterMoment from '@mui/lab/AdapterMoment'
-import LocalizationProvider from '@mui/lab/LocalizationProvider'
+import { AdapterMoment } from '@mui/x-date-pickers/AdapterMoment'
+import { LocalizationProvider } from '@mui/x-date-pickers/LocalizationProvider'
 import SemanticPortal from '../containers/SemanticPortal'
 import portalConfig from '../../configs/portalConfig.json'
 
diff --git a/src/client/translations/rocrate/localeEN.json b/src/client/translations/rocrate/localeEN.json
@@ -159,7 +159,7 @@
       "label": "Profiles",
       "facetResultsType": "Profiles",
       "shortDescription": "Click here to enter the profile portal",
-      "longDescription": "<p class=\"MuiTypography-root MuiTypography-body1 MuiTypography-paragraph\"><a href='https://www.researchobject.org/ro-crate/' target='_blank' rel='noopener noreferrer'>RO-Crates</a> are a method for packaging research data with their metadata. RO-Crate Profiles define the conventions, types and properties to do this for specific communities and domains.<br><br>Welcome to the RO-Crate Profile Portal. [Visit the webpage] to learn about RO-Crates, or [read this guide] to learn about profiles. Click below to enter the RO-Crate Profile Portal, or read on [how to submit your profile]</p>",
+      "longDescription": "<p class=\"MuiTypography-root MuiTypography-body1 MuiTypography-paragraph\"><a href='https://www.researchobject.org/ro-crate/' target='_blank' rel='noopener noreferrer'>RO-Crates</a> are a method for packaging research data with their metadata. RO-Crate Profiles define the conventions, types and properties to do this for specific communities and domains.<br><br>Welcome to the RO-Crate Profile Portal. <a href='https://www.researchobject.org/ro-crate/' target='_blank' rel='noopener noreferrer'>Visit the webpage</a> to learn about RO-Crates, or <a href='https://www.researchobject.org/ro-crate/profiles' target='_blank' rel='noopener noreferrer'>read this guide</a> to learn about profiles. Click below to enter the RO-Crate Profile Portal, or read on <a href='https://github.com/eScienceLab/sampo-dashboard?tab=readme-ov-file#submitting-a-profile-to-the-portal' target='_blank' rel='noopener noreferrer'>how to submit your profile</a></p>",
       "instancePage": {
         "label": "Profile",
         "description": "TODO: Instance page description"
diff --git a/src/configs/rocrate/search_perspectives/profiles.json b/src/configs/rocrate/search_perspectives/profiles.json
@@ -6,7 +6,7 @@
         "prefixesFile": "SparqlQueriesPrefixes.js"
     },
     "sparqlQueriesFile": "SparqlQueriesProfiles.js",
-    "baseURI": "http://example.org/data",
+    "baseURI": "https://profiles.ro-crate.org/data",
     "URITemplate": "<BASE_URI>/profile/<LOCAL_ID>",
     "facetClass": "prof:Profile",
     "defaultConstraint": "FILTER EXISTS { <SUBJECT> owl:sameAs [] }",
@@ -139,7 +139,7 @@
         "publicationDate": {
             "containerClass": "five",
             "filterType": "dateNoTimespanFilter",
-            "predicate": "owl:sameAs/ns11:datePublished",
+            "predicate": "ns11:datePublished",
             "sortByPredicate": "owl:sameAs/ns11:datePublished",
             "max": "2118-01-01",
             "min": "2018-01-01"
diff --git a/src/server/sparql/rocrate/sparql_queries/SparqlQueriesPrefixes.js b/src/server/sparql/rocrate/sparql_queries/SparqlQueriesPrefixes.js
@@ -14,6 +14,6 @@ export const prefixes = `
   PREFIX ns2: <https://w3id.org/ro/terms/test#>
   PREFIX rel: <https://www.w3.org/ns/iana/link-relations/relation#>
   PREFIX wfrun: <https://w3id.org/ro/terms/workflow-run#>
-  PREFIX rocrate: <http://example.org/data/> # TODO: Change URI
+  PREFIX rocrate: <https://profiles.ro-crate.org/data/>
   PREFIX prof: <http://www.w3.org/ns/dx/prof/>
 `

Original file line number	Diff line number	Diff line change
`@@ -1,2 +1,3 @@`
`1`	`1`	`# Environments`
`2`	`2`	`.env`
	`3`	`+venv/`