aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
-rw-r--r--bh20simplewebuploader/main.py5
-rw-r--r--workflows/pangenome-generate/collect-seqs.py5
2 files changed, 9 insertions, 1 deletions
diff --git a/bh20simplewebuploader/main.py b/bh20simplewebuploader/main.py
index 51048a4..bcdf8d8 100644
--- a/bh20simplewebuploader/main.py
+++ b/bh20simplewebuploader/main.py
@@ -695,6 +695,11 @@ union
<http://covid19.genenetwork.org/resource/{id}> pubseq:sample ?sample .
?sequenceuri pubseq:sample ?sample .
}}
+union
+{{
+ ?sequenceuri <http://biohackathon.org/bh20-seq-schema/collection_pdh> "{id}" .
+ ?sequenceuri pubseq:sample ?sample .
+}}
?sample <http://purl.obolibrary.org/obo/GAZ_00000448> ?geo .
?geo rdfs:label ?geoname .
diff --git a/workflows/pangenome-generate/collect-seqs.py b/workflows/pangenome-generate/collect-seqs.py
index 1a0807c..225a61f 100644
--- a/workflows/pangenome-generate/collect-seqs.py
+++ b/workflows/pangenome-generate/collect-seqs.py
@@ -36,11 +36,14 @@ if len(sys.argv) > 3:
for item in validated:
pdh = item["portable_data_hash"]
+ uuid = item["uuid"]
with arvados.collection.CollectionReader(pdh, api_client=api, keep_client=keepclient) as col:
with col.open("sequence.fasta", "rt") as fa:
- subject = "http://covid19.genenetwork.org/resource/%s" % pdh
+ subject = "http://covid19.genenetwork.org/resource/%s" % uuid
label = fa.readline().strip()
merged_metadata.write("<%s> <http://biohackathon.org/bh20-seq-schema/original_fasta_label> \"%s\" .\n" % (subject, label[1:].replace('"', '\\"')))
+ merged_metadata.write("<%s> <http://biohackathon.org/bh20-seq-schema/collection_pdh> \"%s\" .\n" % (subject, pdh))
+ merged_metadata.write("<%s> <http://biohackathon.org/bh20-seq-schema/collection_version> \"%s\" .\n" % (subject, item["version"]))
skip = (subject in blacklist or label[1:] in blacklist)
if skip:
merged_metadata.write("<%s> <http://biohackathon.org/bh20-seq-schema/excluded_from_graph> \"true\"^^<http://www.w3.org/2001/XMLSchema#boolean> .\n" % subject)