#!
Module for handling SPARQL primitives.
Note that GN queries should go into gn/db/sources - this is currently not
the case.
!#
(define-module (gn db sparql)
#:use-module (json)
#:use-module (ice-9 match)
#:use-module (ice-9 format)
#:use-module (ice-9 iconv)
#:use-module (ice-9 receive)
#:use-module (ice-9 string-fun)
#:use-module (web client)
#:use-module (web request)
#:use-module (web uri)
#:use-module (gn cache memoize)
#:use-module (gn db sources wikidata)
#:use-module (web gn-uri)
#:export (memo-sparql-species
memo-sparql-species-meta
sparql-species-meta
sparql-groups-meta
sparql-group-info
memo-sparql-wd-species-info
compile-species
compile-groups-meta
get-rows
tsv->scm
strip-lang
make-table
make-pairs
)
)
(define (strip-lang s)
"Strip quotes and language tag (@en) from RDF entries"
(list->string (match (string->list s)
[(#\"rest ... #\") rest]
[(#\"rest ... #\" #\@ #\e #\n) rest]
[rest rest]))
)
(define (gn-sparql-endpoint-url)
"https://sparql.genenetwork.org/sparql")
(define (wd-sparql-endpoint-url)
"https://query.wikidata.org/sparql")
(define (gn-sparql-prefix query)
(string-append
"
PREFIX gn: <http://genenetwork.org/id/>
PREFIX gnt: <http://genenetwork.org/term/>
PREFIX gnc: <http://genenetwork.org/category/>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
" query))
(define (wdt-taxon-name) "wdt:P225")
(define (sparql-exec endpoint-url query)
"Execute raw SPARQL query returning response as a UTF8 string"
(bytevector->string (receive (response-status response-body)
(http-request (string-append endpoint-url "?default-graph-uri=&query=" (uri-encode query) "&format=application%2Fsparql-results%2Bjson"))
response-body) "UTF-8"))
(define (sparql-tsv endpoint-url query)
"Execute raw SPARQL query returning response as a UTF8 string, e.g.
(tsv->scm (sparql-tsv (wd-sparql-endpoint-url) \"wd:Q158695\"))
"
; GET /sparql?query=SELECT%20DISTINCT%20%2A%20where%20%7B%0A%20%20wd%3AQ158695%20wdt%3AP225%20%3Fo%20.%0A%7D%20limit%205 HTTP/2
(receive (response-status response-body)
(http-get (pk (string-append endpoint-url "?query=" (uri-encode query))) #:headers '((Accept . "text/tab-separated-values")(user-agent . "curl/7.74.0")))
response-body))
(define (unpack field response)
"Helper to get nested JSON field from SPARQL response"
(cdr (assoc field response)))
(define (sparql-names response)
"Helper to get the names part of a SPARQL query"
(unpack "vars" (unpack "head" response)))
(define (sparql-results response)
"Helper to get the results part of a SPARQL query"
(unpack "bindings" (unpack "results" response)))
(define (sparql-scm endpoint-url query)
"Return dual S-exp 'resultset' of varnames and results"
(let ((response (json-string->scm
(sparql-exec endpoint-url (gn-sparql-prefix query)))))
(values (sparql-names response) (sparql-results response))))
(define (tsv->scm text)
"Split a TSV string into a list of fields. Returns list of names header) and rows"
(let ([lst (map (lambda (f) (string-split f #\tab) ) (delete "" (string-split text #\newline)))])
(values (car lst) (cdr lst))
))
#!
(define-values (names res) (sparql-species-meta))
(define table (get-rows names res))
(define recs '())
(define h (compile-species recs table))
(assoc "http://genenetwork.org/species_drosophila_melanogaster" h)
(assoc-ref h "http://genenetwork.org/id/Drosophila_melanogaster")
(define d (car h))
(assoc-ref (list d) "http://genenetwork.org/species_drosophila_melanogaster")
(scm->json #(1 (("2" . 3))))
;; [1,{"2":3}]
(scm->json #("http://genenetwork.org/species_drosophila_melanogaster" (("http://genenetwork.org/menuName" . "Drosophila") ("http://genenetwork.org/binomialName" . "Drosophila melanogaster") )))
;; ["http://genenetwork.org/species_drosophila_melanogaster",{"http://genenetwork.org/menuName":"Drosophila","http://genenetwork.org/binomialName":"Drosophila melanogaster"}]
l
;; (("http://genenetwork.org/menuName" "Drosophila") ("http://genenetwork.org/name" "Drosophila") ("http://genenetwork.org/binomialName" "Drosophila melanogaster"))
(scm->json (map (lambda (i) (cons (car i) (car (cdr i)))) l))
;; {"http://genenetwork.org/menuName":"Drosophila","http://genenetwork.org/name":"Drosophila","http://genenetwork.org/binomialName":"Drosophila melanogaster"}
curl -G https://query.wikidata.org/sparql -H "Accept: application/json; charset=utf-8" --data-urlencode query="SELECT DISTINCT * where {
wd:Q158695 wdt:P225 ?o .
} limit 5"
{
"head" : {
"vars" : [ "o" ]
},
"results" : {
"bindings" : [ {
"o" : {
"type" : "literal",
"value" : "Arabidopsis thaliana"
}
} ]
}
}
!#
(define (sparql-wd-species-info species)
"Returns wikidata entry for species, e.g.:
(sparql-wd-species-info \"Q158695\") generates something like
SELECT DISTINCT * where { wd:Q158695 wdt:P225 ?o . } limit 10
"
(sparql-tsv (wd-sparql-endpoint-url) (string-append "
SELECT DISTINCT ?taxon ?ncbi ?descr where {
wd:" species " " (wdt-taxon-name) " ?taxon ;
wdt:P685 ?ncbi ;
schema:description ?descr .
?species wdt:P685 ?ncbi .
FILTER (lang(?descr)='en')
} limit 5
")))
(define memo-sparql-wd-species-info
(memoize sparql-wd-species-info))
#!
gn:Mus_musculus rdf:type gnc:species .
gn:Mus_musculus gnt:name "Mouse" .
gn:Mus_musculus rdfs:label "Mouse (Mus musculus, mm10)" .
gn:Mus_musculus gnt:binomialName "Mus musculus" .
gn:Mus_musculus gnt:family "Vertebrates" .
gn:Mus_musculus gnt:organism taxon:10090 .
!#
(define (sparql-species)
(sparql-scm (gn-sparql-endpoint-url) "
SELECT DISTINCT ?species WHERE {
?species rdf:type gnc:species .
}"))
(define memo-sparql-species
(memoize2 sparql-species))
(define (sparql-species-meta)
"Return values names recs"
(sparql-scm (gn-sparql-endpoint-url) "
SELECT ?species ?p ?o WHERE {
MINUS { ?species rdf:type ?o . }
{
SELECT DISTINCT ?species ?p ?o WHERE {
?species rdf:type gnc:species .
?species ?p ?o .
}}}"))
(define memo-sparql-species-meta
(memoize2 sparql-species-meta))
#!
dump-species-metadata.ttl:gn:Axbxa rdf:type gnc:inbredSet .
dump-species-metadata.ttl:gn:Axbxa rdfs:label "AXB/BXA Family" .
dump-species-metadata.ttl:gn:Axbxa gnt:family "Reference Populations (replicate average, SE, N)" .
dump-species-metadata.ttl:gn:Axbxa gnt:mappingMethod "AXBXA" .
dump-species-metadata.ttl:gn:Axbxa gnt:code "AXB" .
dump-species-metadata.ttl:gn:Axbxa gnt:belongsToSpecies gn:Mus_musculus .
!#
(define (get-values names row)
"Get values by name from a resultset row"
(map (lambda (n) (unpack "value" (unpack n row))) (array->list names)))
(define (get-rows names results)
"Format results as a list of values ordered by names"
(map (lambda (row) (get-values names row)) (array->list results)))
(define (get-pairs names results)
"Format results as a list of key-values ordered by names"
(map (lambda (row)
(let ([tuple (get-values names row)])
(cons (car tuple) (car (cdr tuple)))
))
(array->list results)))
;; from the triples first harvest the species URIs, followed by creating records of information
(define (compile-species recs rows)
"Compile a matrix of species triples into records"
(for-each (lambda (r)
(let* ([species (car r)]
[v (cdr (cdr r))]
[p (car (cdr r))]
[nrec '()]
[kvs (assoc species recs)]) ; find record to fill based on subject
;; for each gnid add value pair
(if (not kvs)
(set! nrec `(("gnid" ,species)))
(set! nrec (cdr kvs))
)
(set! nrec (assoc-set! nrec p v))
(set! recs (assoc-set! recs species nrec))
))
rows)
recs)
;; ------------------------------------------------------------------------------
(define (sparql-groups-meta)
"Return values names recs - (set-id, species, descr)"
(sparql-scm (gn-sparql-endpoint-url) "
SELECT DISTINCT ?set ?species ?descr WHERE {
?set rdf:type gnc:inbredSet ;
gnt:belongsToSpecies ?species .
OPTIONAL {?set rdfs:label ?descr } .
}"))
(define memo-sparql-groups-meta
(memoize2 sparql-groups-meta))
(define (make-table sparql-thunk)
"Make a tuple of column names and rows"
(receive (names res) (sparql-thunk)
(let ([rows (get-rows names res)])
(list names rows))))
(define (make-pairs sparql-thunk)
"Make a tuple of column names and rows"
(receive (names res) (sparql-thunk)
(let ([rows (get-pairs names res)])
(list names rows))))
(define (compile-groups-meta)
"Return tuple of names and rows containing #(set species descr)"
(receive (names res) (memo-sparql-groups-meta)
(let ([rows (get-rows names res)])
(list names rows))))
(define (sparql-group-info gnid)
"Return set/group info - used by meta and data output"
(sparql-scm (gn-sparql-endpoint-url) (string-append "
SELECT DISTINCT ?key ?value WHERE {
" gnid " ?key ?value .
# FILTER ( !EXISTS{ " gnid " gnt:hasTissue ?value })
}")))
