Better handling of duplicate sequences

Also save original fasta label in metadata

1 files changed, 10 insertions, 1 deletions

diff --git a/bh20sequploader/bh20seq-schema.yml b/bh20sequploader/bh20seq-schema.yml
index 64008f2..982447c 100644
--- a/bh20sequploader/bh20seq-schema.yml
+++ b/bh20sequploader/bh20seq-schema.yml
@@ -18,6 +18,7 @@ $graph:
         jsonldPredicate:
           _id: http://www.ebi.ac.uk/efo/EFO_0000532
           _type: "@id"
+          identity: true
     host_id:
         doc: Identifer for the host. If you submit multiple samples from the same host, use the same host_id for those samples
         type: string
@@ -29,6 +30,7 @@ $graph:
         jsonldPredicate:
           _id: http://purl.obolibrary.org/obo/PATO_0000047
           _type: "@id"
+          identity: true
     host_age:
         doc: Age of the host as number (e.g. 50)
         type: int?
@@ -40,6 +42,7 @@ $graph:
         jsonldPredicate:
           _id: http://purl.obolibrary.org/obo/NCIT_C42574
           _type: "@id"
+          identity: true
     host_health_status:
         doc: A condition or state at a particular time
         type: string?
@@ -79,12 +82,14 @@ $graph:
       jsonldPredicate:
           _id: http://purl.obolibrary.org/obo/OBI_0001479
           _type: "@id"
+          identity: true
     specimen_source2:
       doc: Method how the specimen was derived as NCIT IRI, e.g. http://purl.obolibrary.org/obo/NCIT_C155835 (=throat swabb)
       type: string?
       jsonldPredicate:
           _id: http://purl.obolibrary.org/obo/OBI_0001479
           _type: "@id"
+          identity: true
     collection_date:
       doc: Date when the sample was taken
       type: string
@@ -96,6 +101,7 @@ $graph:
       jsonldPredicate:
         _id: http://purl.obolibrary.org/obo/GAZ_00000448
         _type: "@id"
+        identity: true
     sample_storage_conditions:
       doc: Information about storage of a specified type, e.g.  frozen specimen, paraffin, fresh ....
       type: string?
@@ -126,6 +132,7 @@ $graph:
       jsonldPredicate:
           _id: http://edamontology.org/data_1875
           _type: "@id"
+          identity: true
     virus_strain:
       doc: Name of the virus strain
       type: string?
@@ -141,12 +148,14 @@ $graph:
       jsonldPredicate:
         _id: http://purl.obolibrary.org/obo/OBI_0600047
         _type: "@id"
+        identity: true
     sample_sequencing_technology2:
       doc: Technology that was used to sequence this sample (e.g Sanger, Nanopor MiniION)
       type: string?
       jsonldPredicate:
         _id: http://purl.obolibrary.org/obo/OBI_0600047
         _type: "@id"
+        identity: true
     sequence_assembly_method:
       doc: Protocol which provides instructions on the alignment of sequencing reads to reference genome
       type: string?
@@ -215,7 +224,7 @@ $graph:
       jsonldPredicate:
           _id: http://semanticscience.org/resource/SIO_000115
           _type: "@id"
-          noLinkCheck: true
+          identity: true
 
 - name: MainSchema
   type: record