Add ShEx validation

Arvados-DCO-1.1-Signed-off-by: Peter Amstutz <peter.amstutz@curii.com>

3 files changed, 23 insertions, 3 deletions

diff --git a/bh20sequploader/validation_shape.rdf b/bh20sequploader/bh20seq-shex.rdf
index dd34fa9..8d3f5fc 100644
--- a/bh20sequploader/validation_shape.rdf
+++ b/bh20sequploader/bh20seq-shex.rdf
@@ -1,4 +1,4 @@
-PREFIX : <http://whatever/>
+PREFIX : <https://raw.githubusercontent.com/arvados/bh20-seq-resource/master/bh20sequploader/bh20seq-shex.rdf#>
 PREFIX MainSchema: <http://biohackathon.org/bh20-seq-schema#MainSchema/>
 PREFIX hostSchema: <http://biohackathon.org/bh20-seq-schema#hostSchema/>
 PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
diff --git a/bh20sequploader/main.py b/bh20sequploader/main.py
index e0a6a9a..49d012d 100644
--- a/bh20sequploader/main.py
+++ b/bh20sequploader/main.py
@@ -21,16 +21,25 @@ def main():
     parser = argparse.ArgumentParser(description='Upload SARS-CoV-19 sequences for analysis')
     parser.add_argument('sequence', type=argparse.FileType('r'), help='sequence FASTA/FASTQ')
     parser.add_argument('metadata', type=argparse.FileType('r'), help='sequence metadata json')
+    parser.add_argument("--validate", action="store_true", help="Dry run, validate only")
     args = parser.parse_args()
 
     api = arvados.api(host=ARVADOS_API_HOST, token=ARVADOS_API_TOKEN, insecure=True)
 
-    target = qc_fasta(args.sequence)
+    try:
+        target = qc_fasta(args.sequence)
+    except ValueError as e:
+        print(e)
+        exit(1)
 
     if not qc_metadata(args.metadata.name):
         print("Failed metadata qc")
         exit(1)
 
+    if args.validate:
+        print("Valid")
+        exit(0)
+
     col = arvados.collection.Collection(api_client=api)
 
     with col.open(target, "w") as f:
diff --git a/bh20sequploader/qc_metadata.py b/bh20sequploader/qc_metadata.py
index e477f21..fbfd286 100644
--- a/bh20sequploader/qc_metadata.py
+++ b/bh20sequploader/qc_metadata.py
@@ -1,9 +1,12 @@
 import schema_salad.schema
 import schema_salad.ref_resolver
+import schema_salad.jsonld_context
 import logging
 import pkg_resources
 import logging
 import traceback
+from rdflib import Graph, Namespace
+from pyshex.evaluate import evaluate
 
 def qc_metadata(metadatafile):
     schema_resource = pkg_resources.resource_stream(__name__, "bh20seq-schema.yml")
@@ -13,13 +16,21 @@ def qc_metadata(metadatafile):
      schema_metadata,
      metaschema_loader) = schema_salad.schema.load_schema("https://raw.githubusercontent.com/arvados/bh20-seq-resource/master/bh20sequploader/bh20seq-schema.yml", cache=cache)
 
+    shex = pkg_resources.resource_stream(__name__, "bh20seq-shex.rdf").read().decode("utf-8")
+
     if not isinstance(avsc_names, schema_salad.avro.schema.Names):
         print(avsc_names)
         return False
 
     try:
         doc, metadata = schema_salad.schema.load_and_validate(document_loader, avsc_names, metadatafile, True)
-        return True
+        g = schema_salad.jsonld_context.makerdf("workflow", doc, document_loader.ctx)
+        rslt, reason = evaluate(g, shex, doc["id"], "https://raw.githubusercontent.com/arvados/bh20-seq-resource/master/bh20sequploader/bh20seq-shex.rdf#submissionShape")
+
+        if not rslt:
+            print(reason)
+
+        return rslt
     except Exception as e:
         traceback.print_exc()
         logging.warn(e)