3 files changed, 90 insertions, 24 deletions
diff --git a/bh20sequploader/bh20seq-schema.yml b/bh20sequploader/bh20seq-schema.yml
index 6e0973a..5c962d1 100644
--- a/bh20sequploader/bh20seq-schema.yml
+++ b/bh20sequploader/bh20seq-schema.yml
@@ -1,36 +1,89 @@
+$base: http://biohackathon.org/bh20-seq-schema
+$namespaces:
+  sch: https://schema.org/
+  efo: http://www.ebi.ac.uk/efo/
+  obo: http://purl.obolibrary.org/obo/
 $graph:
 
-- name: sampleInformationSchema
+- name: hostSchema
   type: record
   fields:
-    location: string
-    host: string
-    sequenceTechnology: string
-    assemblyMethod: string
+    host_species:
+        type: string
+        jsonldPredicate:
+          _id: http://www.ebi.ac.uk/efo/EFO_0000532
+    host_id: string
+    host_common_name: string?
+    host_sex: string?
+    host_age: int?
+    host_age_unit: string?
+    host_health_status: string?
+    host_treatment:
+      type: string?
+      jsonldPredicate:
+          _id: http://www.ebi.ac.uk/efo/EFO_0000727
+    additional_host_information: string?
 
-- name: InstituteInformationSchema
+- name: sampleSchema
   type: record
   fields:
-    OriginatingLab: string
-    SubmittingLab: string
+    collector_name: string
+    collecting_institution: string
+    specimen_source: string?
+    collection_date: string?
+    collection_location:
+      type: string?
+      jsonldPredicate:
+        _id: https://schema.org/fromLocation
+    sample_storage_conditions: string?
+    additional_collection_information: string?
 
-- name: SubmitterInformationSchema
+- name: virusSchema
   type: record
   fields:
-    Submitter: string
-    submissionDate: string
+    virus_species: string?
+    virus_strain: string?
 
-- name: VirusDetailSchema
+- name: technologySchema
   type: record
   fields:
-    VirusName: string
-    AccessionId: string
+    sample_sequencing_technology:
+      type: string
+      jsonldPredicate:
+        _id: http://www.ebi.ac.uk/efo/EFO_0000532
+    sequence_assembly_method:
+      type: string?
+      jsonldPredicate:
+        _id: http://www.ebi.ac.uk/efo/EFO_0002699
+    sequencing_coverage:
+      type: string?
+      jsonldPredicate:
+        _id: http://purl.obolibrary.org/obo/FLU_0000848
+
+- name: submitterSchema
+  type: record
+  fields:
+    submitter_name: string
+    submitter_address: string?
+    originating_lab: string
+    lab_address: string?
+    provider_sample_id: string?
+    submitter_sample_id: string?
+    authors: string?
+    submitter_id: string?
 
 - name: MainSchema
   type: record
   documentRoot: true
   fields:
-    sampleInformation: sampleInformationSchema
-    InstituteInformation: InstituteInformationSchema
-    SubmitterInformation: SubmitterInformationSchema
-    VirusDetail: VirusDetailSchema
+    host: hostSchema
+    sample: sampleSchema
+    virus: virusSchema?
+    technology: technologySchema
+    submitter: submitterSchema
+    sequencefile:
+      doc: The subject (eg the fasta/fastq file) that this metadata describes
+      type: string?
+      jsonldPredicate:
+        _id: "@id"
+        _type: "@id"
diff --git a/bh20sequploader/main.py b/bh20sequploader/main.py
index d3ebc0c..bf74ea5 100644
--- a/bh20sequploader/main.py
+++ b/bh20sequploader/main.py
@@ -6,6 +6,7 @@ import json
 import urllib.request
 import socket
 import getpass
+import qc_metadata
 
 ARVADOS_API_HOST='lugli.arvadosapi.com'
 ARVADOS_API_TOKEN='2fbebpmbo3rw3x05ueu2i6nx70zhrsb1p22ycu3ry34m4x4462'
@@ -19,18 +20,26 @@ def main():
 
     api = arvados.api(host=ARVADOS_API_HOST, token=ARVADOS_API_TOKEN, insecure=True)
 
+    if not qc_metadata(args.metadata.name):
+        print("Failed metadata qc")
+        exit(1)
+
     col = arvados.collection.Collection(api_client=api)
 
-    print("Reading FASTA")
-    with col.open("sequence.fasta", "w") as f:
+    if args.sequence.name.endswith("fasta") or args.sequence.name.endswith("fa"):
+        target = "sequence.fasta"
+    elif args.sequence.name.endswith("fastq") or args.sequence.name.endswith("fq"):
+        target = "reads.fastq"
+
+    with col.open(target, "w") as f:
         r = args.sequence.read(65536)
         print(r[0:20])
         while r:
             f.write(r)
             r = args.sequence.read(65536)
 
-    print("Reading JSONLD")
-    with col.open("metadata.jsonld", "w") as f:
+    print("Reading metadata")
+    with col.open("metadata.yaml", "w") as f:
         r = args.metadata.read(65536)
         print(r[0:20])
         while r:
@@ -49,5 +58,7 @@ def main():
                  (properties['upload_user'], properties['upload_ip']),
                  properties=properties, ensure_unique_name=True)
 
+    print("Done")
+
 if __name__ == "__main__":
     main()
diff --git a/bh20sequploader/qc_metadata.py b/bh20sequploader/qc_metadata.py
index 78b31b2..ebe4dfc 100644
--- a/bh20sequploader/qc_metadata.py
+++ b/bh20sequploader/qc_metadata.py
@@ -1,6 +1,7 @@
 import schema_salad.schema
 import logging
 import pkg_resources
+import logging
 
 def qc_metadata(metadatafile):
     schema_resource = pkg_resources.resource_stream(__name__, "bh20seq-schema.yml")
@@ -17,5 +18,6 @@ def qc_metadata(metadatafile):
     try:
         doc, metadata = schema_salad.schema.load_and_validate(document_loader, avsc_names, metadatafile, True)
         return True
-    except:
-        return False
+    except Exception as e:
+        logging.warn(e)
+    return False