Add ShEx validation

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

1 files changed, 12 insertions, 1 deletions

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)