Updating schema and examples. This is still work in progress but we get there

3 files changed, 91 insertions, 32 deletions

diff --git a/bh20sequploader/bh20seq-schema.yml b/bh20sequploader/bh20seq-schema.yml
index 6e0973a..38cfb48 100644
--- a/bh20sequploader/bh20seq-schema.yml
+++ b/bh20sequploader/bh20seq-schema.yml
@@ -1,36 +1,60 @@
 $graph:
 
-- name: sampleInformationSchema
+- name: hostSchema
   type: record
   fields:
-    location: string
-    host: string
-    sequenceTechnology: string
-    assemblyMethod: string
+    host_id: string
+    host_species: string
+    host_common_name: string?
+    host_sex: string?
+    host_age: int?
+    host_age_unit: string?
+    host_health_status: string?
+    host_treatment: string?
+    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: string?
+    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: string
+    sequence_assembly_method: string?
+    sequencing_coverage: string?
+
+- 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
\ No newline at end of file
diff --git a/example/metadata.yaml b/example/metadata.yaml
index 587d0be..8a93379 100644
--- a/example/metadata.yaml
+++ b/example/metadata.yaml
@@ -1,17 +1,38 @@
-sampleInformation:
-  location: "USA"
-  host : "Homo Sapiens"
-  sequenceTechnology: "Sanger"
-  assemblyMethod: "CLC Genomics"
+host:
+    host_id: XX1
+    host_species: string
+    host_common_name: string
+    host_sex: string
+    host_age: 20
+    host_age_unit: string
+    host_health_status: string
+    host_treatment: string
+    additional_host_information: string
 
-InstituteInformation:
-  OriginatingLab: "Erik's kitchen"
-  SubmittingLab: "National Institute for Viral Disease Control and Prevention, China CDC"
+sample:
+    collector_name: XXX
+    collecting_institution: XXX
+    specimen_source: XXX
+    collection_date: XXX
+    collection_location: XXX
+    sample_storage_conditions: XXX
+    additional_collection_information: XXX
 
-SubmitterInformation:
-  Submitter: "National Institute for Viral Disease Control and Prevention, China CDC"
-  submissionDate: "04-04-2020"
+virus:
+    virus_species: XX
+    virus_strain: XX
 
-VirusDetail:
-  VirusName: "hCoV-19/USA/identifer/2020"
-  AccessionId: "EPI_ISL_Random"
+technology:
+    sample_sequencing_technology: XX
+    sequence_assembly_method: XX
+    sequencing_coverage: 70x
+
+submitter:
+    submitter_name: tester
+    submitter_address: testerAdd
+    originating_lab: testLab
+    lab_address: labAdd
+    provider_sample_id: string
+    submitter_sample_id: string
+    authors: testAuthor
+    submitter_id: X12
\ No newline at end of file
diff --git a/example/minimal_example.yaml b/example/minimal_example.yaml
new file mode 100644
index 0000000..201b080
--- /dev/null
+++ b/example/minimal_example.yaml
@@ -0,0 +1,14 @@
+host:
+    host_id: XX
+    host_species: string
+
+sample:
+    collector_name: XXX
+    collecting_institution: XXX
+
+technology:
+    sample_sequencing_technology: XX
+
+submitter:
+    submitter_name: tester
+    originating_lab: testLab
\ No newline at end of file