diff options
Diffstat (limited to 'gn3/api/metadata.py')
| -rw-r--r-- | gn3/api/metadata.py | 349 |
1 files changed, 10 insertions, 339 deletions
diff --git a/gn3/api/metadata.py b/gn3/api/metadata.py index 91dc115..3f28f5d 100644 --- a/gn3/api/metadata.py +++ b/gn3/api/metadata.py @@ -5,7 +5,6 @@ from string import Template from pathlib import Path from authlib.jose import jwt - from flask import Blueprint from flask import request from flask import current_app @@ -14,135 +13,20 @@ from gn3.auth.authorisation.errors import AuthorisationError from gn3.db.datasets import (retrieve_metadata, save_metadata, get_history) -from gn3.db.rdf import RDF_PREFIXES from gn3.db.rdf import (query_frame_and_compact, - query_and_compact, - query_and_frame) - - -BASE_CONTEXT = { - "data": "@graph", - "id": "@id", - "type": "@type", - "gnc": "http://genenetwork.org/category/", - "gnt": "http://genenetwork.org/term/", - "rdfs": "http://www.w3.org/2000/01/rdf-schema#", - "rdf": "http://www.w3.org/1999/02/22-rdf-syntax-ns#>", -} - -DATASET_CONTEXT = { - "accessRights": "dct:accessRights", - "accessionId": "dct:identifier", - "acknowledgement": "gnt:hasAcknowledgement", - "altLabel": "skos:altLabel", - "caseInfo": "gnt:hasCaseInfo", - "classifiedUnder": "xkos:classifiedUnder", - "contributors": "dct:creator", - "contactPoint": "dcat:contactPoint", - "created": "dct:created", - "dcat": "http://www.w3.org/ns/dcat#", - "dct": "http://purl.org/dc/terms/", - "description": "dct:description", - "ex": "http://example.org/stuff/1.0/", - "experimentDesignInfo": "gnt:hasExperimentDesignInfo", - "experimentType": "gnt:hasExperimentType", - "foaf": "http://xmlns.com/foaf/0.1/", - "geoSeriesId": "gnt:hasGeoSeriesId", - "gnt": "http://genenetwork.org/term/", - "inbredSet": "gnt:belongsToGroup", - "label": "rdfs:label", - "normalization": "gnt:usesNormalization", - "platformInfo": "gnt:hasPlatformInfo", - "notes": "gnt:hasNotes", - "organization": "foaf:Organization", - "prefLabel": "skos:prefLabel", - "citation": "dct:isReferencedBy", - "GoTree": "gnt:hasGOTreeValue", - "platform": "gnt:usesPlatform", - "processingInfo": "gnt:hasDataProcessingInfo", - "rdfs": "http://www.w3.org/2000/01/rdf-schema#", - "skos": "http://www.w3.org/2004/02/skos/core#", - "specifics": "gnt:hasContentInfo", - "title": "dct:title", - "xkos": "http://rdf-vocabulary.ddialliance.org/xkos#", - "tissueInfo": "gnt:hasTissueInfo", - "tissue": "gnt:hasTissue", - "contactWebUrl": "foaf:homepage", - "contactName": "foaf:name", -} - -SEARCH_CONTEXT = { - "pages": "ex:pages", - "hits": "ex:hits", - "result": "ex:result", - "results": "ex:items", - "resultItem": "ex:resultType", - "currentPage": "ex:currentPage", -} - -DATASET_SEARCH_CONTEXT = SEARCH_CONTEXT | { - "classifiedUnder": "xkos:classifiedUnder", - "created": "dct:created", - "dct": "http://purl.org/dc/terms/", - "ex": "http://example.org/stuff/1.0/", - "inbredSet": "ex:belongsToInbredSet", - "title": "dct:title", - "name": "rdfs:label", - "rdfs": "http://www.w3.org/2000/01/rdf-schema#", - "type": "@type", - "xkos": "http://rdf-vocabulary.ddialliance.org/xkos#", -} + query_and_compact) +from gn3.db.constants import ( + RDF_PREFIXES, BASE_CONTEXT, + DATASET_CONTEXT, + DATASET_SEARCH_CONTEXT, PUBLICATION_CONTEXT, + PHENOTYPE_CONTEXT +) -PUBLICATION_CONTEXT = { - "dct": "http://purl.org/dc/terms/", - "fabio": "http://purl.org/spar/fabio/", - "prism": "http://prismstandard.org/namespaces/basic/2.0/", - "xsd": "http://www.w3.org/2001/XMLSchema#", - "title": "dct:title", - "journal": "fabio:Journal", - "volume": "prism:volume", - "page": "fabio:page", - "creator": "dct:creator", - "abstract": "dct:abstract", - "year": { - "@id": "fabio:hasPublicationYear", - "@type": "xsd:gYear", - }, - "month": { - "@id": "prism:publicationDate", - "@type": "xsd:gMonth" - }, -} +from gn3.api.metadata_api import wiki -PHENOTYPE_CONTEXT = BASE_CONTEXT | PUBLICATION_CONTEXT | { - "skos": "http://www.w3.org/2004/02/skos/core#", - "dcat": "http://www.w3.org/ns/dcat#", - "prism": "http://prismstandard.org/namespaces/basic/2.0/", - "traitName": "skos:altLabel", - "trait": "rdfs:label", - "altName": "rdfs:altLabel", - "description": "dct:description", - "abbreviation": "gnt:abbreviation", - "labCode": "gnt:labCode", - "submitter": "gnt:submitter", - "dataset": "dcat:Distribution", - "contributor": "dct:contributor", - "mean": "gnt:mean", - "locus": "gnt:locus", - "lodScore": "gnt:lodScore", - "references": "dct:isReferencedBy", - "additive": "gnt:additive", - "sequence": "gnt:sequence", - "prefLabel": "skos:prefLabel", - "identifier": "dct:identifier", - "chromosome": "gnt:chr", - "mb": "gnt:mb", - "peakLocation": "gnt:locus", - "species": "gnt:belongsToSpecies", - "group": "gnt:belongsToGroup", -} metadata = Blueprint("metadata", __name__) +metadata.register_blueprint(wiki.wiki_blueprint) @metadata.route("/datasets/<name>", methods=["GET"]) @@ -208,7 +92,7 @@ CONSTRUCT { (Path( current_app.config.get("DATA_DIR") ) / "gn-docs/general/datasets" / - Path(__result.get("id", "")).stem).as_posix() + Path(__result.get("id", "")).stem).as_posix() ) @@ -348,69 +232,6 @@ def edit_dataset(): lambda x: ("Edit successfull", 201) ) -@metadata.route("/datasets/search/<term>", methods=["GET"]) -def search_datasets(term): - """Search datasets""" - args = request.args - page = args.get("page", 0) - page_size = args.get("per-page", 10) - _query = Template(""" -$prefix - -CONSTRUCT { - ex:result rdf:type ex:resultType ; - ex:pages ?pages ; - ex:hits ?hits ; - ex:currentPage $offset ; - ex:items [ - rdfs:label ?label ; - dct:title ?title ; - ex:belongsToInbredSet ?inbredSetName ; - xkos:classifiedUnder ?datasetType ; - ] -} WHERE { -{ - SELECT DISTINCT ?dataset ?label ?inbredSetName ?datasetType ?title - WHERE { - ?dataset rdf:type dcat:Dataset ; - rdfs:label ?label ; - ?datasetPredicate ?datasetObject ; - xkos:classifiedUnder ?inbredSet . - ?inbredSet ^skos:member gnc:Set ; - rdfs:label ?inbredSetName . - ?datasetObject bif:contains "'$term'" . - OPTIONAL { - ?dataset dct:title ?title . - } . - OPTIONAL { - ?classification ^xkos:classifiedUnder ?dataset ; - ^skos:member gnc:DatasetType ; - ?typePredicate ?typeName ; - skos:prefLabel ?datasetType . - } - } ORDER BY ?dataset LIMIT $limit OFFSET $offset -} - -{ - SELECT (COUNT(DISTINCT ?dataset)/$limit+1 AS ?pages) - (COUNT(DISTINCT ?dataset) AS ?hits) WHERE { - ?dataset rdf:type dcat:Dataset ; - ?p ?o . - ?o bif:contains "'$term'" . - } -} - -} -""").substitute(prefix=RDF_PREFIXES, term=term, limit=page_size, offset=page) - _context = { - "@context": BASE_CONTEXT | DATASET_SEARCH_CONTEXT, - "type": "resultItem", - } - return query_frame_and_compact( - _query, _context, - current_app.config.get("SPARQL_ENDPOINT") - ) - @metadata.route("/publications/<name>", methods=["GET"]) def publications(name): @@ -436,65 +257,6 @@ CONSTRUCT { ) -@metadata.route("/publications/search/<term>", methods=["GET"]) -def search_publications(term): - """Search publications""" - args = request.args - page = args.get("page", 0) - page_size = args.get("per-page", 10) - _query = Template(""" -$prefix - -CONSTRUCT { - ex:result rdf:type ex:resultType ; - ex:totalCount ?totalCount ; - ex:currentPage $offset ; - ex:items [ - rdfs:label ?publication ; - dct:title ?title ; - ] -} WHERE { -{ - SELECT ?publication ?title ?pmid WHERE { - ?pub rdf:type fabio:ResearchPaper ; - ?predicate ?object ; - dct:title ?title . - ?object bif:contains "'$term'" . - BIND( STR(?pub) AS ?publication ) . - } ORDER BY ?title LIMIT $limit OFFSET $offset - } -{ - SELECT (COUNT(*)/$limit+1 AS ?totalCount) WHERE { - ?publication rdf:type fabio:ResearchPaper ; - ?predicate ?object . - ?object bif:contains "'$term'" . - } -} -} -""").substitute(prefix=RDF_PREFIXES, term=term, limit=page_size, offset=page) - _context = { - "@context": BASE_CONTEXT | SEARCH_CONTEXT | { - "dct": "http://purl.org/dc/terms/", - "ex": "http://example.org/stuff/1.0/", - "rdfs": "http://www.w3.org/2000/01/rdf-schema#", - "fabio": "http://purl.org/spar/fabio/", - "title": "dct:title", - "pubmed": "fabio:hasPubMedId", - "currentPage": "ex:currentPage", - "url": "rdfs:label", - }, - "type": "resultItem", - "paper": { - "@type": "fabio:ResearchPaper", - "@container": "@index" - } - } - return query_and_frame( - _query, _context, - current_app.config.get("SPARQL_ENDPOINT") - ) - - @metadata.route("/phenotypes/<name>", methods=["GET"]) @metadata.route("/phenotypes/<group>/<name>", methods=["GET"]) def phenotypes(name, group=None): @@ -630,97 +392,6 @@ CONSTRUCT { ) -@metadata.route("/genewikis/gn/<symbol>", methods=["GET"]) -def get_gn_genewiki_entries(symbol): - """Fetch the GN and NCBI GeneRIF entries""" - args = request.args - page = args.get("page", 0) - page_size = args.get("per-page", 10) - _query = Template(""" -$prefix - -CONSTRUCT { - ?symbol ex:entries [ - rdfs:comment ?comment ; - ex:species ?species_ ; - dct:created ?createTime ; - dct:references ?pmids ; - dct:creator ?creator ; - gnt:belongsToCategory ?categories ; - ] . - ?symbol rdf:type gnc:GNWikiEntry ; - ex:totalCount ?totalCount ; - ex:currentPage $offset . -} WHERE { -{ - SELECT ?symbol ?comment - (GROUP_CONCAT(DISTINCT ?speciesName; SEPARATOR='; ') AS ?species_) - ?createTime ?creator - (GROUP_CONCAT(DISTINCT ?pubmed; SEPARATOR='; ') AS ?pmids) - (GROUP_CONCAT(DISTINCT ?category; SEPARATOR='; ') AS ?categories) - WHERE { - ?symbol rdfs:label ?label ; - rdfs:comment _:entry . - ?label bif:contains "'$symbol'" . - _:entry rdf:type gnc:GNWikiEntry ; - rdfs:comment ?comment . - OPTIONAL { - ?species ^xkos:classifiedUnder _:entry ; - ^skos:member gnc:Species ; - skos:prefLabel ?speciesName . - } . - OPTIONAL { _:entry dct:created ?createTime . } . - OPTIONAL { _:entry dct:references ?pubmed . } . - OPTIONAL { - ?investigator foaf:name ?creator ; - ^dct:creator _:entry . - } . - OPTIONAL { _:entry gnt:belongsToCategory ?category . } . - } GROUP BY ?comment ?symbol ?createTime - ?creator ORDER BY ?createTime LIMIT $limit OFFSET $offset -} - -{ - SELECT (COUNT(DISTINCT ?comment)/$limit+1 AS ?totalCount) WHERE { - ?symbol rdfs:comment _:entry ; - rdfs:label ?label . - _:entry rdfs:comment ?comment ; - rdf:type gnc:GNWikiEntry . - ?label bif:contains "'$symbol'" . - } -} -} -""").substitute(prefix=RDF_PREFIXES, symbol=symbol, - limit=page_size, offset=page) - _context = { - "@context": BASE_CONTEXT | { - "ex": "http://example.org/stuff/1.0/", - "rdfs": "http://www.w3.org/2000/01/rdf-schema#", - "gnt": "http://genenetwork.org/term/", - "gnc": "http://genenetwork.org/category/", - "dct": "http://purl.org/dc/terms/", - "xsd": "http://www.w3.org/2001/XMLSchema#", - "entries": "ex:entries", - "comment": "rdfs:comment", - "species": "ex:species", - "category": 'gnt:belongsToCategory', - "author": "dct:creator", - "pubmed": "dct:references", - "currentPage": "ex:currentPage", - "pages": "ex:totalCount", - "created": { - "@id": "dct:created", - "@type": "xsd:datetime" - }, - }, - "type": "gnc:GNWikiEntry" - } - return query_frame_and_compact( - _query, _context, - current_app.config.get("SPARQL_ENDPOINT") - ) - - @metadata.route("/genewikis/ncbi/<symbol>", methods=["GET"]) def get_ncbi_genewiki_entries(symbol): """Fetch the NCBI GeneRIF entries""" |