Implement dataset metadata API endpoint.

* guix.scm: Import (gnu packages rdf).
(genenetwork3)[propagated-inputs]: Add python-sparqlwrapper.
* gn3/settings.py (SPARQL_ENDPOINT): New variable.
* gn3/api/general.py: Import datasets from gn3.db.
(dataset_metadata): New API endpoint.
* gn3/db/datasets.py: Import re, Template from string, Dict and Optional from
typing, JSON and SPARQLWrapper from SPARQLWrapper, SPARQL_ENDPOINT from
gn3.settings.
(sparql_query, dataset_metadata): New functions.

4 files changed, 113 insertions, 2 deletions

diff --git a/gn3/api/general.py b/gn3/api/general.py
index 69ec343..e0bfc81 100644
--- a/gn3/api/general.py
+++ b/gn3/api/general.py
@@ -7,7 +7,7 @@ from flask import request
 
 from gn3.fs_helpers import extract_uploaded_file
 from gn3.commands import run_cmd
-
+from gn3.db import datasets
 
 general = Blueprint("general", __name__)
 
@@ -68,3 +68,8 @@ def run_r_qtl(geno_filestr, pheno_filestr):
     cmd = (f"Rscript {rqtl_wrapper} "
            f"{geno_filestr} {pheno_filestr}")
     return jsonify(run_cmd(cmd)), 201
+
+@general.route("/dataset/<accession_id>")
+def dataset_metadata(accession_id):
+    """Return info as JSON for dataset with ACCESSION_ID."""
+    return jsonify(datasets.dataset_metadata(accession_id))
diff --git a/gn3/db/datasets.py b/gn3/db/datasets.py
index 6c328f5..e4c779a 100644
--- a/gn3/db/datasets.py
+++ b/gn3/db/datasets.py
@@ -1,7 +1,11 @@
 """
 This module contains functions relating to specific trait dataset manipulation
 """
-from typing import Any
+import re
+from string import Template
+from typing import Any, Dict, Optional
+from SPARQLWrapper import JSON, SPARQLWrapper
+from gn3.settings import SPARQL_ENDPOINT
 
 def retrieve_probeset_trait_dataset_name(
         threshold: int, name: str, connection: Any):
@@ -289,3 +293,100 @@ def retrieve_trait_dataset(trait_type, trait, threshold, conn):
         **dataset_fns[trait_type](),
         **group
     }
+
+def sparql_query(query: str) -> Dict[str, Any]:
+    """Run a SPARQL query and return the bound variables."""
+    sparql = SPARQLWrapper(SPARQL_ENDPOINT)
+    sparql.setQuery(query)
+    sparql.setReturnFormat(JSON)
+    return sparql.queryAndConvert()['results']['bindings']
+
+def dataset_metadata(accession_id: str) -> Optional[Dict[str, Any]]:
+    """Return info about dataset with ACCESSION_ID."""
+    # Check accession_id to protect against query injection.
+    # TODO: This function doesn't yet return the names of the actual dataset files.
+    pattern = re.compile(r'GN\d+', re.ASCII)
+    if not pattern.fullmatch(accession_id):
+        return None
+    # KLUDGE: We split the SPARQL query because virtuoso is very slow on a
+    # single large query.
+    queries = ["""
+PREFIX gn: <http://genenetwork.org/>
+SELECT ?name ?dataset_group ?status ?title ?geo_series
+WHERE {
+  ?dataset gn:accessionId "$accession_id" ;
+           rdf:type gn:dataset ;
+           gn:name ?name .
+  OPTIONAL { ?dataset gn:datasetGroup ?dataset_group } .
+  # FIXME: gn:datasetStatus should not be optional. But, some records don't
+  # have it.
+  OPTIONAL { ?dataset gn:datasetStatus ?status } .
+  OPTIONAL { ?dataset gn:title ?title } .
+  OPTIONAL { ?dataset gn:geoSeries ?geo_series } .
+}
+""",
+             """
+PREFIX gn: <http://genenetwork.org/>
+SELECT ?platform_name ?normalization_name ?species_name ?inbred_set_name ?tissue_name
+WHERE {
+  ?dataset gn:accessionId "$accession_id" ;
+           rdf:type gn:dataset ;
+           gn:normalization / gn:name ?normalization_name ;
+           gn:datasetOfSpecies / gn:menuName ?species_name ;
+           gn:datasetOfInbredSet / gn:name ?inbred_set_name .
+  OPTIONAL { ?dataset gn:datasetOfTissue / gn:name ?tissue_name } .
+  OPTIONAL { ?dataset gn:datasetOfPlatform / gn:name ?platform_name } .
+}
+""",
+             """
+PREFIX gn: <http://genenetwork.org/>
+SELECT ?specifics ?summary ?about_cases ?about_tissue ?about_platform
+       ?about_data_processing ?notes ?experiment_design ?contributors
+       ?citation ?acknowledgment
+WHERE {
+  ?dataset gn:accessionId "$accession_id" ;
+           rdf:type gn:dataset .
+  OPTIONAL { ?dataset gn:specifics ?specifics . }
+  OPTIONAL { ?dataset gn:summary ?summary . }
+  OPTIONAL { ?dataset gn:aboutCases ?about_cases . }
+  OPTIONAL { ?dataset gn:aboutTissue ?about_tissue . }
+  OPTIONAL { ?dataset gn:aboutPlatform ?about_platform . }
+  OPTIONAL { ?dataset gn:aboutDataProcessing ?about_data_processing . }
+  OPTIONAL { ?dataset gn:notes ?notes . }
+  OPTIONAL { ?dataset gn:experimentDesign ?experiment_design . }
+  OPTIONAL { ?dataset gn:contributors ?contributors . }
+  OPTIONAL { ?dataset gn:citation ?citation . }
+  OPTIONAL { ?dataset gn:acknowledgment ?acknowledgment . }
+}
+"""]
+    result = {'accession_id': accession_id,
+              'investigator': {}}
+    query_result = {}
+    for query in queries:
+        if sparql_result := sparql_query(Template(query).substitute(accession_id=accession_id)):
+            query_result.update(sparql_result[0])
+        else:
+            return None
+    for key, value in query_result.items():
+        result[key] = value['value']
+    investigator_query_result = sparql_query(Template("""
+PREFIX gn: <http://genenetwork.org/>
+SELECT ?name ?address ?city ?state ?zip ?phone ?email ?country ?homepage
+WHERE {
+  ?dataset gn:accessionId "$accession_id" ;
+           rdf:type gn:dataset ;
+           gn:datasetOfInvestigator ?investigator .
+  OPTIONAL { ?investigator foaf:name ?name . }
+  OPTIONAL { ?investigator gn:address ?address . }
+  OPTIONAL { ?investigator gn:city ?city . }
+  OPTIONAL { ?investigator gn:state ?state . }
+  OPTIONAL { ?investigator gn:zipCode ?zip . }
+  OPTIONAL { ?investigator foaf:phone ?phone . }
+  OPTIONAL { ?investigator foaf:mbox ?email . }
+  OPTIONAL { ?investigator gn:country ?country . }
+  OPTIONAL { ?investigator foaf:homepage ?homepage . }
+}
+""").substitute(accession_id=accession_id))[0]
+    for key, value in investigator_query_result.items():
+        result['investigator'][key] = value['value']
+    return result
diff --git a/gn3/settings.py b/gn3/settings.py
index 0ac6698..c945fbf 100644
--- a/gn3/settings.py
+++ b/gn3/settings.py
@@ -13,6 +13,9 @@ REDIS_JOB_QUEUE = "GN3::job-queue"
 TMPDIR = os.environ.get("TMPDIR", tempfile.gettempdir())
 RQTL_WRAPPER = "rqtl_wrapper.R"
 
+# SPARQL endpoint
+SPARQL_ENDPOINT = "http://localhost:8891/sparql"
+
 # SQL confs
 SQL_URI = os.environ.get(
     "SQL_URI", "mysql://webqtlout:webqtlout@localhost/db_webqtl")
diff --git a/guix.scm b/guix.scm
index 9bf23c8..e2e49ab 100644
--- a/guix.scm
+++ b/guix.scm
@@ -42,6 +42,7 @@
              (gnu packages python-web)
              (gnu packages python-xyz)
              (gnu packages python-science)
+             (gnu packages rdf)
              ((guix build utils) #:select (with-directory-excursion))
              (guix build-system python)
              (guix gexp)
@@ -78,6 +79,7 @@
                        ("python-redis" ,python-redis)
                        ("python-requests" ,python-requests)
                        ("python-scipy" ,python-scipy)
+                       ("python-sparqlwrapper" ,python-sparqlwrapper)
                        ("r-optparse" ,r-optparse)
                        ("r-qtl" ,r-qtl)
                        ("r-rjson" ,r-rjson)